AEA Poster Session
Friday, Jan. 6, 2023 7:00 AM - 6:00 PM (CST)
Saturday, Jan. 7, 2023 7:00 AM - 6:00 PM (CST)
Sunday, Jan. 8, 2023 7:00 AM - 3:00 PM (CST)
A Corporate Finance Perspective on Environmental Regulation (Q5, G3)
AbstractThis paper examines optimal environmental policy when external financing is costly for firms. We introduce emission externalities and industry equilibrium in the Holmstrom and Tirole (1997) workhorse model of corporate finance. While a (Pigouvian) cap-and-trading system optimally governs both firms' abatement activities (internal emission margin) and industry size (external emission margin) when firms have sufficient funds on their own, external financing constraints introduce a wedge between these two objectives. When a sector is financially constrained in the aggregate,
the optimal cap is strictly above the Pigouvian benchmark and emission allowances should be allocated below market prices. When a sector is not financially constrained in the aggregate, a cap that is below the Pigiouvian benchmark optimally shifts market share to less polluting firms and,
moreover, there should be no "grandfathering" of emission allowances. With financial constraints and heterogeneity across firms or sectors, a uniform policy, such as a single cap-and-trade system, is typically not optimal.
A Quantitative Analysis of Relaxing UI Eligibility Requirements: Evidence from the Pandemic Unemployment Assistance (J6, E6)
AbstractETA reports show that 10% of unemployment insurance (UI) applications were rejected before the pandemic due to applicants’ failure to meet the past earning eligibility requirements. During the pandemic recession, UI relaxed its past earning requirement for the first time under the pandemic unemployment assistance (PUA). By using an equilibrium search model, I quantify the effect of relaxing the past earning requirement on heterogeneous workers during a pandemic recession. I find that PUA helped recover consumption by 20%. It made UI more complete by supporting workers who needed UI the most. I use the model to predict the optimal duration of the whole suite of programs under the CARES Act, given the start date. I find that ending benefits at 52 weeks provides the highest increase in welfare per dollar of spending.
A Thousand Words Tell More Than Just Numbers: Financial Crises and Historical Headlines (G0, G2)
AbstractWe show that financial crises are preceded by changes in specific types of narrative information contained in newspaper article titles. Our novel international dataset and the resulting empirical evidence are gathered by integrating information from a large panel of economic news articles in global newspapers between the years 1870 and 2016 with conventional macroeconomic and financial indicators. We find that the predictive information of newspaper article titles that signals coming crisis episodes is substantial over and above the macroeconomic and financial indicators. The new indicators capture common features that have often been discussed as potential causes of specific crises but not incorporated yet into empirical models.
Accounting for Variety (O4)
AbstractThis paper proposes a decomposition of output growth between price and volume indices, plus the cross-entropy of GDP and weights assumed in constructing aggregate indices. Over time, the cross-entropy of GDP with respect to a benchmark captures the change in its distribution, analogous to `information loss' -- the degree of representation -- in price and volume indices across time. The correction addresses a long-standing issue concerning estimates for historical real GDP growth, since the comparability of output in the past to output today is a function of their cross-entropy. Using US World Klems data, this decomposition demonstrates the volume of a representative 1947 product, weighted by nominal output shares, grew 1pp slower per year than the average 2014 product, despite a lower inflation rate for the 1947 product. The changing distribution of GDP is thus a more plausible explanation for small estimates of historical living standards than downward biases in CPI.
Adjustable CBDC Interest Rate: Easing the Macroeconomic Trilemma (E5, F3)
AbstractCentral banks in open economies are constrained by the macroeconomic trilemma where the objectives of exchange rate stability, monetary independence and capital mobility cannot be achieved together. This paper shows that the design choice of an adjustable CBDC interest rate allows the central bank to relax the trilemma. A domestic CBDC monetary policy gives monetary autonomy to achieve price stability while allowing the central bank to stabilize exchange rates via policy interest rate.
We develop a two-agent dynamic stochastic general equilibrium (DSGE) model with an adjustable CBDC interest rate as a secondary instrument and a policy interest rate as a primary instrument. In our model, bank deposits and CBDC are competing medium of exchange and are both competing assets of domestic and foreign bonds. We derive parity conditions between CBDC vis-à-vis its competing assets, revealing the transactional efficiency of CBDC as crucial in the implications of the CBDC monetary policy on exchange rate movements. Calibrating the model with New Zealand data, our simulation exercises show that a) A counter-cyclical CBDC monetary policy improves social welfare and economic stability; b) CBDC monetary policy insulates exchange rate movements more in the context of foreign shocks than domestic shocks; c)The adjustable CBDC interest rate allows the central bank to achieve both objectives of monetary autonomy and exchange rate stability, analogous to a sterilized foreign exchange intervention.
AI Training for Online Entrepreneurs: An Experiment with Two Million New Sellers on an E-commerce Platform (L1, M2)
AbstractWe conduct a year-long experiment among new sellers on a large e-commerce platform. The treatment group receives access to a free and customized entrepreneur training program, in which an AI algorithm assigns training materials to sellers based on their real-time operations data.
With a 24% take-up rate, new sellers that are eligible for training see 1.7% higher revenue on average (6.6% ATT), driven largely by higher traffic and enhanced marketing activities. To dissect the mechanisms behind this effect and the economic surplus generated, we develop a model of platform matching as well as consumers' visit and purchase decisions. We then use search records to construct a panel dataset of consumer-seller pairs across consideration sets. With exhaustive fixed effects, we show that training raises new sellers' product value, boosting consumers' visit and purchase rates simultaneously. But the effect is small compared to an entry barrier driven by search friction, which is unchanged for treated sellers. Lastly, model estimation shows that although only 0.25% of products in consumers' consideration sets are from treated new sellers, removing the program would reduce consumer surplus by 0.07%.
Air Pollution and School Absences in New York City (I1, Q5)
AbstractIn this paper, we analyze the effects of daily air pollution levels on daily absences for New York City schools from 2006 to 2018. We combine EPA air quality data from monitoring stations with climate modeling to get daily pollution maps for all of New York City for several air pollutants. With a fixed effects model controlling for individual schools and seasonal variation, we find significant effects on increased daily absences from exposure to high concentrations of PM2.5, NO2, or Ozone. Using wind direction and strength as an instrument, or adding lagged pollution from previous days, we find similar results. Our work shows the improvements over time in air quality in New York City but also highlights the disparate impacts of air pollution based on location.
An Ounce of Prevention or a Pound of Cure? The Value of Health Risk Information (I1, D8)
AbstractIndividuals infer their health risk from observing the health experiences of people around them, particularly family members. I assess whether people correctly interpret new information from household health events and analyze resulting welfare implications. When an individual is diagnosed with a new chronic condition, unaffected family members increase their healthcare spending by over 10 percent. Informational spillovers are associated with increased use of both high- and low-return care, including takeup of new services and renewed adherence to extant ones. I show these responses are consistent with individual reevaluations of health risk and reject other mechanisms. To assess welfare implications, I estimate a structural model of health choices in which individuals learn about risk after health events reveal information. I find that consumers over-respond to recent, salient health events by over-weighting their risks ex-post. This leads to annual welfare losses of $2,788 per family on average; suppressing responsiveness results in net gains for 86 percent of households. Revealing health risk information can be optimally targeted on household demographics to improve social welfare gains.
Announcement and Implementation Effects of Central Bank Asset Purchases (E5, C3)
AbstractWe provide a unified empirical dynamic framework for studying the joint impact of announcement and implementation (or flow) effects of central bank asset purchases, an issue so far neglected by the literature. Announcement and implementation shocks are identified exploiting a novel combination of survey based external instruments and zero-sign restrictions within a standard VAR model estimated at daily frequency. We document three main results. First, both purchase announcements and actual purchases tend to exert meaningful effects on yields and stock prices, although the former are found to be more persistent. Second, and in line with the evidence on conventional monetary policy tools, we show that their variation is largely driven by a systematic reaction of the central bank to macro and financial shocks and that asset purchase shocks only account for a small share of the variation in non-policy variables. Third, for a given stock of announced purchases, the way the latter are actually implemented can play a non-negligible role in stabilizing financial conditions. By means of structural policy counterfactuals we document that the substantial frontloading of ECB purchases during the pandemic crisis accounted for about 1/4 of the overall impact of the “pandemic” QE on long-term yields.
Are Heated Tobacco Products Substitutes or Complement to Cigarettes? Empirical Evidence across Countries (H2, I1)
AbstractHeated Tobacco Products (HTPs) are a new form of tobacco products that heat raw tobacco sticks to generate an aerosol containing tobacco flavor, nicotine, and other chemicals. Essentially because the tobacco is not combusted for HTPs, in contrast to traditional cigarettes, companies have aggressively promoted them as ``reduced-risk'' products that could help consumers to quit cigarette use. HTPs have rapidly gained market shares over traditional cigarettes in several countries. Despite the lack of evidence on their impact on health of demand, most countries that sell HTPs have heavily subsidized these products though lower excise taxes than cigarettes. Data and Methods: We construct a unique database on HTPs and cigarettes quarterly demand and prices from Q4 of 2014 to Q1 of 2022, obtained from Philip Morris International. We combine this database with a unique database on HTP and cigarettes excise taxes developed by the Campaign for Tobacco Free Kids across countries over the same period, and estimate own- and cross- price elasticity of cigarette and HTP demand using a seemingly unrelated regressions model, using a variety of models. Conclusions and Policy Implications: We find the own tax elasticity of HTP demand to be TBD, and the own price elasticity of HTP demand to be TBD. The cross-price and tax elasticities suggest that HTPs and cigarettes are substitutes: A TBD% increase in HTP prices(taxes) increase the demand for cigarette consumption by TBD%; conversely, A TBD % increase in cigarette prices(taxes) increase the demand for HTP consumption by TBD% ??
Are Sanctuary Cities Attracting Illegal Immigrants? (J6)
AbstractSanctuary cities, enacted by local governments, are jurisdictions that do not require cooperation with the U.S. federal government in the enforcement of immigration policies and, as a result, often carry the connotation of “a paradise for illegal immigrants”. The sanctuary city policy and its social and economic implications have been much debated. A fundamental question about sanctuary cities that underlies much of the controversy is whether they attract undocumented immigrants. The impact of the policy on major socioeconomic variables such as labor markets hinges on the answer to this critical question. If a city or county becomes a sanctuary but does not attract more undocumented immigrants, it would have implications for any presumptions about its effect beyond its primary goal of encouraging the communities of undocumented immigrants to be more integrated and less afraid to contribute and cooperate when it is needed. Using individual-level data by county between 2010 and 2019, this study constructs the inflow and outflow rates of likely undocumented immigrants (LUI) as defined in the literature, and uses these inter-county mobility measurements to test the effect of sanctuary city status on the net inflow of LUIs in a difference-in-difference framework. The study finds no significant effects; that is, undocumented immigrants do not gravitate toward sanctuary cities. Further examination shows that this coincides with the lack of significant effects on deportations. These interesting results are contrary to a priori expectations and have important policy implications.
Are Solar Incentive Programs Effective at Reducing CO2 Emissions? Evidence from Massachusetts (Q4, L9)
AbstractHow effective are demand-side incentive programs at encouraging households to adopt solar panels? I use data on residential solar panel installations in Massachusetts to estimate a dynamic model of solar panel adoption (or demand) that accounts for both current and future savings. The model allows me to evaluate several solar incentive programs implemented in Massachusetts in terms of their impacts on adoption rates, consumer welfare, and contribution to the reduction of CO2 emissions. In addition, I analyze each program’s cost effectiveness by comparing the social benefit generated due to displaced CO2 emissions to the government’s expenditure on each program. My estimates suggest that the social benefits generated are modest relative to the magnitude of public spending.
Assessing the Importance of an Attribute in a Demand System (C5, D4)
AbstractFirms can prioritize product features based on consumer valuations using market-level data. A number of approaches exist to measure variable importance, including regression coefficients. However, most regression models are not designed for deriving relative feature importance, because they predict with a combination of features that are vulnerable to multicollinearity, which produces biased coefficients and negative inputs. Structural estimation of product importance is particularly challenging when there is a large number of attributes, e.g., in high-dimensional data, hence some of them are more likely to be related, or the model is highly nonlinear.
We propose a machine-learning (ML) algorithm based on Random Forests (RF) and Shapley values that can detect consumers’ sensitivity to product attributes—measured by their relative contribution. We show the theoretical connections, and use simulations to rank the attributes when prices are experimental or endogenous, coefficients are homogenous or random, and instrumental and demographic variables are unavailable. A link is numerically derived between the true parameters and the ML importance by optimizing the hyperparameters. The ML importance consistently correlates with true parameters in ranking and magnitude even when product attributes are large in numbers, and the rates increase with the sample size. We empirically test the model using scanner data for yogurt and find that the ML importance closely follows the structural estimates. Finally, we discuss the managerial and policy implications of the model.
Asset Prices in a Labor Search Model with Confidence Shocks (E3, G1)
AbstractThis paper studies a labor search model with agents who are averse to ambiguity (Knightian uncertainty). Shocks to confidence about future productivity are modeled as changes in ambiguity. Using the Survey of Professional Forecasters data, I find that confidence shocks help explain the equity premium and the stock market volatility, including their term structure. Ambiguity amplifies the response of the economy to productivity shocks, helping the model produce more realistic dynamics of unemployment, vacancies, labor market tightness, stock prices, and returns. Returns in the model are predictable with price-dividend ratios; both returns and dividends are predictable with unemployment, like in the data. Two extensions consider shocks to bargaining power and to separation rate and find similar implications of ambiguity about these shocks.
Back to Edgeworth? Estimating the Value of Time Using Hedonic Experiences (R4, D6)
AbstractFollowing early economist Francis Y. Edgeworth's proposal to measure people's hedonic experiences as they go about their daily lives, we use a smartphone app that randomly asked a panel of 30,936 UK residents (N=2,235,733) over the course of eight years about their momentary feelings and activities, to derive a causal parameter for the value of time (VOT), a key input into cost-benefit analyses. Exploiting the randomisation of messages for identification, we arrive at a VOT of GBP 8.3 (USD 11.2) per hour, considerably smaller than estimates currently used by UK Government. While our parameter is more generalisable than previous ones, applying across a broad range of activities and their contexts, our unique data and method also allow us to derive activity-specific estimates. These closely resemble those from studies using revealed preferences, suggesting that using hedonic experiences to value intangibles may lead to very similar results as observed behaviour. We are the first to estimate the VOT (or indeed any intangible) using hedonic experiences in real-time, which has the potential to value other intangibles too.
Big Techs, QR Code Payments and Financial Inclusion (G3, D2)
AbstractUsing a unique dataset of 500,000 Chinese firms that use a QR-code-based mobile
payment system, we find that (i) the creation of a digital payment footprint allows firms
to access credit provided by the same big tech company; (ii) transaction data generated
via QR code generate spillover effects on access to bank credit; and (iii) there are
positive effects of access to big tech credit on sales, including during the Covid-19
shock. The findings suggest that access to innovative payment methods help micro
firms build up credit history and the use of big tech credit can ease access to bank credit.
Black Farmers’ Experience with USDA Financing: A Qualitative Study (J1, Q1)
AbstractBlack farmers are on a historical decline. The share of Black or mixed-race farmers has decreased from 14% to only about 1.4% in the last century. One of the many determinants of the decline in the number of farms can be credit constraints. Among financing options, grant and loan options offered by the US Department of Agriculture (USDA) are more convenient for farmers because they are characterized by low or no down payments, competitive interest rates, broader eligibility, and overall greater flexibility compared to most other commercial sources.
However, the relationship between black farmers and USDA has not been perfect. There have been lawsuits like the Pigford v. Glickman case alleging racial discrimination against African-American farmers in USDA’s allocation of farm loans and assistance. This event may have long-term effects on the relationship between black farmers and USDA.
We asked farmers to rate their experience with USDA offices and loan application processes, using in-depth phone interviews with 31 farmers across the country. We find that, compared to white farmers, black farmers are 80% more likely to rate their experience with local offices “negative”, and 48% more likely to rate their loan application experience “negative”.
The thematic analysis generates three themes (1) black farmers experience a lack of inclusiveness in local offices, which may improve if officers could understand and connect with farmers, (2) lack of assistance in the preparation of complicated paperwork, which can be addressed by offering training and dedicated technical personnel, and (3) imperfect communication between USDA and black farmers, which could have been improved by timely communication and internet access. The discourse analysis indicates that black farmers feel unwelcomed, but improving the condition is difficult because USDA is just another microcosm of the entire USA where racial bias still exists.
Carbon Taxes and Tariffs, Financial Frictions, and International Spillovers (F3, Q5)
AbstractClimate policy, coupled with financial frictions, has the potential to create macrofinancial stability risk through asset stranding. Such stability risk may expand beyond the economy implementing climate policy, because of the use of carbon tariffs or simply because countries are connected through trade and financial channels. Hence, we study the effects and design of climate policies in the world economy with international trade and financial flows and frictions. We develop an original multi-sector, multi-country, dynamic stochastic general equilibrium (DSGE) model with pollution externalities, financial frictions, and climate and macroprudential policies. Using the calibrated model, we evaluate spillovers from domestic carbon pricing to foreign economies through international trade and financial linkages. We find that an ambitious domestic climate policy affects foreign economies through international capital flows, besides more studied trade channels. On the one hand, domestic financial companies' positions are impaired due to their carbon exposures forcing them to pull back from investing in foreign economies. On the other hand, a high domestic carbon tax makes foreign polluting industries more attractive destination for capital flows creating a risk of “finance-driven carbon leakage.” Further, we study a novel trade-off faced by a country when imposing a carbon border tax on goods imported from the foreign country, where the latter is assumed not to have any domestic carbon pricing in place. While the carbon border tax may bring several key benefits - leveling the playing field and to some extent preventing carbon leakage - it may also trigger ``asset stranding'' in the carbon-intensive foreign economy. Asset stranding in the foreign economy, in turn, may have destabilizing effects on the domestic economy through cross-border exposure of domestic financial sector to foreign assets. Our model serves as a laboratory to quantify this unexplored trade-off and to assess the effectiveness of combining climate with macroprudential policies.
Cash Transfers and Adolescent Nutrition Outcomes: Evidence from the Child Support Grant in South Africa (I3, I1)
AbstractCash transfers are a popular intervention to improve nutrition outcomes in children. The bulk of the literature focuses on the effects of safety nets during the critical and sensitive period of the first 1000 days of life. However, in addition to early childhood, adolescence is a period characterized by rapid biological, physiological and psychosocial growth, which increases nutritional needs.
We examine the effect of the Child Support Grant (CSG), an unconditional means-tested grant provided to age-eligible poor children in South Africa, on Body Mass Index (BMI)-for-age z-scores and the likelihood of being underweight, normal, overweight, and obese.
We use rich longitudinal data from the 2008, 2010, and 2012 waves the National Income Dynamics Study (NIDS) South Africa. To estimate the causal effect of the CSG on adolescent nutritional outcomes, we employ a fuzzy regression discontinuity approach that exploits the changes in the age-eligibility criterion as the source of exogenous birth cohort variation in grant receipt. Since the CSG program started in 1998, the age eligibility criterion has extended over time. From 1 January 2010, the age eligibility was extended from 14 to 16 years, followed by an extension to less than 18 years in January 2012. These changes resulted in extended receipt for the same cohorts.
We find that CSG receipt increases the BMI z-scores of females. While CSG receipt decreases the likelihood of being underweight and obese, it increases the likelihood of being overweight among females beneficiaries. This may have negative long-term health implications considering the co-morbidities and other economic costs associated with being overweight, although most of this evidence is from developed countries. We do not find significant effects on the nutritional outcomes of males, except an increase in BMI at the lowest end of the BMI z-score distribution.
Unearned cash transfers are likely to prevent females from being in a state of malnutrition, i.e., at extreme ends of the BMI z-scores distribution. Therefore, poverty-alleviation interventions may have additional benefits of improving nutrition and reproductive health of adolescent girls. Such policies may contribute to reducing the health care costs associated with extreme undernutrition and overnutrition, especially in the context of a double burden of malnutrition.
Causal Inference in Big Health Data: What Can Health Economists and Machine Learning Scientists Learn from Each Other? (I1, C8)
Health economists usually build a theory model to analyze causal inference in healthcare, while machine learning scientists analyze patterns in the data to build an interpretable model. Causal inference methods in economics are quite established, while interpretable machine learning models are involving as the initial tasks of machine learning models are prediction (of an event or outcome).
I aimed to explore whether causal inference methods in economics are suitable for analyzing big healthcare data in its current forms or in combination with machine learning techniques. I also aimed to give recommendations whether machine learning scientists can use any econometric tool box for their interpretable machine learning models in health.
I reviewed economic and machine learning studies using causal inference methods in health and compared them in various aspects, such as research topics, research questions, data, methods, models, tools, outcomes, countries, software, ease of technical, instrument variables, reporting results, and quality of the study. All studies published between 2000 and 2021 in English was included. The search was done on Scopus and Econlit databases in January 2022.
I identified over 1200 papers and further examining the abstracts limited to over 20 studies. I found that causal inference in econometrics more advance in term of problem formulations and tools. However, machine learning models may be more real-time data focus, more practical, and still evolving fast. Econometric models focus more on policy interventions at a macro-level, while machine learning more on healthcare at an individual-level. I made some specific recommendations for health economists and machine learning scientists in learning from each other.
Discussion and conclusions:
Given the time of big data, more collaboration between health economists and machine learning scientists would be fruitful. Machine learning can lay a foundation or suggesting new risk factors for economic models in healthcare.
Certificate of Need and the Labor Market (I1, J4)
Abstract34 US states currently have Certificate of Need (CON) laws, which require health care providers to prove their “economic necessity” to a state board before they can legally open or expand. While dozens of articles have evaluated the effect of CON on hospitals and consumers, no published article has evaluated its effect on health care workers. Economists expect entry barriers such as CON to reduce the number of firms and so to reduce employment. The effect of this reduced competition on wages, however, is theoretically ambiguous: the lower number of required workers and the increased bargaining power of firms relative to workers pushes wages down, but the increased bargaining power of firms relative to consumers increases rents, which may be shared with workers to push wages up. The American Health Planning Association provides data on CON laws themselves, noting which states had CON in a given year and which types of health providers are covered by the law in each state. Labor market data come primarily from the Current Population Survey, supplemented by administrative data. The main empirical strategy is a difference-in-difference analysis of CON repeals. I find that CON reduces overall health sector employment by 4.2%, but does not significantly affect wages.
Coherence without Rationality at the Zero Lower Bound (E4, C6)
AbstractStandard rational expectations models with an occasionally binding zero lower bound (ZLB) constraint either admit no solutions (incoherency) or multiple solutions (incompleteness). This paper shows that deviations from full information rational expectations mitigate concerns about incoherency and incompleteness. Models with no rational expectations equilibrium admit self-confirming equilibria involving the use of simple mis-specified forecasting models. Completeness and coherency is restored if expectations are adaptive or if agents are less forward-looking due to some information or behavioral friction. In the case of incompleteness, the E-stability criterion selects an equilibrium.
Collateral Choice (E4, G0)
AbstractThis is the first paper studying collateral choices in one of the main short-term funding instruments, the repurchase agreement (repo). In general collateral repos, the borrower can choose which bond he posts as collateral out of a predefined list. I show that collateral availability and opportunity cost explain the collateral choice. In the aggregate, on-the-run bonds are more likely to be delivered than cheapest-to-post securities which is surprising given that the former is more expensive. I rationalize those findings in a theoretical framework that links the repo to the bond market and show that bond market spreads depend on repo collateral choices.
Commuting Behavior and Gender Norms by Sexual Orientation (J1, R2)
AbstractWe assess the role of gender-conforming social norms in household decision-making and gender inequalities in the labor market with a parsimonious household model that endogenizes commuting time. Using the American Community Survey 2008-2019, we test the model predictions and find that women in same-sex couples have a longer commute to work than working women in different-sex couples, whereas the commute to work of men in same-sex couples is shorter than the one of working men in different-sex couples, even after controlling for demographic characteristics, partner’s characteristics, location, fertility, and marital status. These differences among men and women amount to 50%, and 100%, respectively, of the gender commuting gap estimated in the literature, and are particularly stark among married couples with children. Within-couple gaps in commuting time are also significantly smaller in same-sex couples, and labor supply disparities mimic the commuting ones. According to our model, these differences are interpreted as gender-conforming social norms leading women in different-sex couples into jobs with a shorter commute and fewer hours worked while their male partners/spouses hold jobs with a longer commute and more hours worked, thus reinforcing gender inequalities.
Comparative Effectiveness of Machine Learning Methods for Causal Inference (C5, Q1)
AbstractRecent developments in machine learning (ML) literature have helped researchers refine causal analysis models. Most ML models do not require matrix inversion, and generate importance factors for variables corresponding to their contribution to the task of prediction. This feature enables dimension reduction in big data where confoundedness is more likely. A double machine learning method, for instance, utilizes the feature to learn the average treatment effect and the average treatment effect on the treated under unconfoundedness. Contrarily, the causal forest utilizes ML’s classification power to maximize the difference across recursively generated data partitions in the relationship between an outcome variable and a treatment variable, thus uncovers heterogeneity in a causal effect. Many studies have adopted these methods for more flexible and informative causal analysis. However, no study so far has compared the effectiveness of these models.
The current study compares three approaches (1) causal forest (CF), (2) double machine learning (DML), and (3) propensity score matching with ML (PSM-ML) for causal analysis. We find that the using PSM-ML is only advantageous in smaller data set, especially under linear specifications, followed by CF and DML. However, when the functional form is nonlinear, CF performs relatively better. DML has a relatively lower rate of error and exhibits better estimation accuracy as the data dimension grows.
For empirical evidence, we use a randomized controlled trial (Dhurandhar et al. 2014) to obtain the impact of skipping breakfast on body weight. Interestingly, a double ML model generates a positive significant result—that was found null earlier by linear models. This indicates that dimension reduction and a more flexible functional form with ML might help find the causal estimate. Thus, ML models can offer a viable alternative to traditional ATE models for their flexible, data-driven nature, better predictive accuracy, and dimension reduction capabilities.
Comparative Immigration Policies and the Effect of International Students in U.S. Higher Education (J6, I2)
AbstractThe international student population has been steadily growing globally in recent decades, yet there have been ongoing debates, especially in the U.S., over immigration policies regarding international students and their impact on domestic students. Proponents argue that international students contribute to U.S. economy and help smooth budget fluctuations in U.S. public universities; while opponents seek to restrict the size of and post-graduation work visas for international students claiming that they displace U.S. domestic students and workers. In this paper, I examine the causal effect of international student enrollment on college completion of U.S. domestic students by leveraging a restrictive immigration policy change in the U.K. that induced more international students from former British colonies to enroll in U.S. universities. Using newly obtained administrative data on all international students in the U.S. between 2003 and 2015, I find that an additional international student per program leads to 0.7 more domestic students to obtain a college degree four years later. The effect is concentrated in public four-year institutions. Additionally, I find positive cross-degree-level effect of international students in master’s degree programs on U.S. domestic students in bachelor’s degree programs. The positive impact is most likely through cross subsidization of tuition, serving as evidence of resource effects. The findings show that international students respond to comparative immigration policies. The restrictive immigration policies and anti-immigrant sentiment in the U.S. during the Trump administration could have contributed to the recent decline in international enrollment that indirectly harms domestic students and U.S. higher education.
Compensation and Fatigue Related Risk Factors for Occupational Safety: A Structural Equation Modeling Approach to Long-Haul Truck Driving (J3, J3)
AbstractThis paper explores how compensation and occupational safety are related among employee
long-haul truck drivers in the U.S. We hypothesize that drivers’ compensation is related to safety
performance in two ways. First, compensation is directly related to safety because greater
compensation may cause higher safety incentives among drivers. Second, compensation is
indirectly related to safety by the intermediary of work fatigue because lowly compensated
drivers are likely to suffer from greater work-stress. Using a structural equation model, we
analyze the direct and indirect relationships between compensation and safety independently.
Our results suggest that compensation is indirectly related to safety by the intermediary of sleep and work fatigue whereas the direct relationship between compensation and safety is statistically
insignificant. The result implies that greater compensation seems to mitigate drivers’ sleep and work fatigue and enhance occupational safety.
Contraceptive Usage and Fertility: What Happens When Doorstep Access Comes at a Price? (I1, J1)
AbstractContraceptive usage usually increases with easier access but evidently decreases as prices rise. We study a unique policy from India where home delivery of minimally priced contraceptives replaced the practice of acquiring contraceptives free of cost from village centers. Using a quasi-experimental estimation framework, we find that this intervention led to higher usage of contraceptives and lower fertility, potentially attributable to easier access. However, households substitute away from the priced modern contraception methods towards traditional or permanent forms of contraception, for which prices remained unchanged, reflecting a revealed preference towards costless contraception or high fixed-cost but low variable-cost based methods.
Corruption and Human Capital: Evidence from Public Jobs in China (D7, I2)
AbstractCorruption may disincentivize productive activities and undermine human capital. This paper examines how reducing corruption reshapes household talent allocation and accumulation. I do so by exploiting China's staggered anti-corruption inspections that dampen perceived returns to rent-seeking sharply. Using a representative child-parent panel and personnel data from a state-owned enterprise, I find that reducing corruption induces positive selection for integrity and public mindedness into the state sector, which remains present even when conditioning on ability and family background. The results hold for both unrealized occupational preferences and realized labor market outcomes. I also show suggestive evidence that parents put greater emphasis on human capital accumulation following the crackdown, increasing inputs in their children's education. Taken together, my results point to the short-run and long-run human capital costs of rent-seeking.
Covid19 and Consumer Animus Towards Chinese Products: Evidence from Amazon Data (J7, F1)
AbstractIn this paper, we study consumers’ discriminative behaviors towards Chinese products post Covid-19 pandemic. China is the first country to report Covid-19 and consumers might discriminate Chinese products based on health concerns or purely out of prejudice. To study consumer discrimination, we collect data from Amazon where consumers can leave long reviews. For all face masks sold on Amazon from Sep. 2019 to Sep. 2020, we collect all information that is available to a real consumer, including rating, sales rank, price, product description, customer questions and answers, and every consumer review. The great richness of data allows us to identify discriminative reviews towards China/Chinese products, as well as identify the origin countries of face masks. The existence of discriminative reviews lowers the average rating for Chinese face masks sold on Amazon. This impact peaks at one week after review with -0.25 point, and gradually fades out in 5 weeks. Under the DID design, we also find consumers give lower ratings and switch away from a product once its Chinese identity is revealed. The results are even stronger right after some key public health/political events related to Covid-19. The impact, however, is short-lived. All results are robust to control of product quality, both directly and indirectly measured.
Cracks in the Glass Ceiling and Gender Equality: An Examination using Administrative Data (F1, J7)
AbstractUsing Brazilian administrative employer-employee matched data of worker demographics, industry of affiliation, occupation, and wages, we examine whether females in managerial and executive positions (cracks in the glass ceiling) lead to more gender equal workplace outcomes. The evidence is mixed. In response to the large and unanticipated 1999 Brazilian Real exchange rate devaluation, male-led export-oriented firms significantly accelerated the overall hiring of female employees. Over the same time
period, female-led firms implemented incremental but sometimes offsetting changes in hiring practices. The male-female wage gap widened in all firms, but less so in female-led firms. Overall, we do find significant differences between female and male-led firms, some of which seem counterintuitive.
Credible Interval Estimates of the Size and Legal Composition of the US Foreign-Born Population (J1, C8)
AbstractGovernment agencies and academic researchers typically report the size and legal composition of the foreign-born population as point estimates. As these estimates are generally produced using survey data, they are impacted by both sampling and nonsampling error. This paper considers nonsampling error due to item nonresponse in the estimates of the size and legal composition of the foreign-born population produced using the American Community Survey. The standard practice to deal with item nonresponse is to impute values under the assumption that nonresponse is conditionally random. We follow a procedure that allows us to form credible interval estimates that make no assumptions about the values of missing data by taking into account all uncertainty due to item nonresponse. Without any assumptions on the distribution of citizenship status among non-respondents, the size of the foreign-born population in the US falls somewhere between 40.4 and 59.4 million as of 2019 compared to the Census estimate of 44.9 million. When taking into account item nonresponse from all questions used in the imputation procedure to assign legal status, the size of the undocumented population fall between 7.3 and 23.3 million compared to the widely accepted estimate of 11 million undocumented immigrants.
Cronicle of a Death Foretold: Does Higher Volatility Anticipate Corporate Default? (C5, G3)
AbstractWe test whether a simple measure of corporate insolvency based on equity return
volatility and denoted as Distance to Insolvency (DI) delivers better prediction
of corporate defaults than the widely-used Expected Default Frequency (EDF)
measure computed by Moody’s. We look at the relationship and predictive power
between future defaults and current DI and EDF both at a firm-level and aggregate
level. At the granular level, the DI and EDF both have predictive power for subsequent
corporate defaults, but the DI dominates the EDF when both variables are
pooled together. Moreover, the DI has a superior information content at longer forecasting
horizons. At the aggregate level, the DI shows superior forecasting power
compared to the EDF for horizons between 3 and 12 months. Lastly, we use the DI
measure to assess the evolution of corporate defaults during the COVID-19 crisis if
the ECB had not implemented the Public Emergency Purchase Program (PEPP).
Culture and Privatization: The Agricultural Legacy and Politicians’ Decisions in China (P3, Z1)
AbstractThis paper studies the role of local politicians’ cultural attributes, in particular individualism vs. collectivism, in their policy decisions on the massive privatization of China’s state-owned enterprises (SOEs). Drawing on the well-documented link between wheat cultivation and individualistic culture, we employ the soil suitability for wheat relative to rice in their birth places to measure Municipal Party Secretaries’ (MPS) individualism. We find that individualistic MPS introduced more private ownership and were more likely to transfer control rights to private owners in privatization. Furthermore, the MPS individualism only affected SOEs supervised by governments at the city level or lower but not those managed by central or provincial governments. The results still hold when we implement a fuzzy regression discontinuity (RD) design by exploiting the variations in traditional farming practices generated by a natural geographic boundary in China. For robustness, we check and confirm that the assignment of MPS workplace was not influenced by the farming practices in their birth places, thus ruling out reverse causality. In addition, using household surveys, we show that wheat cultivation could influence one’s attitude toward the role of government in the economy. Finally, we find that MPS individualism had little effect on the efficiency gains of privatization. Overall, our findings suggest an important factor—politicians’ cultural attributes—in shaping China’s privatization.
Curiosity in News Consumption (D8, D9)
AbstractWe analyze how curiosity drives news consumption. We apply natural language processing (NLP) methods and test predictions of the information-gap theory of curiosity in the context of news consumption on WeChat. As predicted, we find that people are more likely to consume news when: 1) the headline raises a salient question; 2) it appears more important, e.g., because the headline was displayed in a higher position on the webpage; 3) the topics mentioned in the headline are more surprising, as measured by the KL-divergence of the distribution over topics referenced in the headline; and 4) the headline has lower valence.
Data Science for Justice: Evidence from a Randomized Judicial Reform in the Kenyan Judiciary (K0, O1)
AbstractCan Data Science be used to improve the functioning of courts, and unlock the positive effects of institutions on economic development? In a nationwide randomized experiment in Kenyan courts, we develop and implement an algorithm that uses data regularly captured by administrative systems, identifies for each court their main sources of delay, and provides court-specific actionable recommendations on how to increase performance. We find that this intervention reduces delays, especially when the information is also shared with court user committees that include representatives from civil society, lawyers, and police. We find downstream economic effects of court speed, especially on contract-intensive industries.
Deciphering Monetary Policy Shocks (E5, G1)
AbstractA key question in economics and finance is how monetary policy impacts financial markets and the real economy. In addressing this question, a standard approach is to assess monetary policy shocks based on high-frequency market price reactions around central bank announcements. To give these shocks an economic interpretation, researchers use economic theory to (indirectly) infer what type of news may have plausibly moved asset prices in the direction observed. In this paper, we propose to identify the sources of such monetary policy shocks directly form the contents of central banks' verbal communication.
We decipher monetary policy shocks by connecting them to the stance a central bank expresses in its communication about different topics. To measure topic-specific stances, we apply textual analysis techniques to press conference statements of the European Central Bank (ECB). Using three sets of shocks established in the literature, based on market reactions in interest rates, the entire term structure, or the joint response in interest rates and stock prices, we find that markets react to news on the topics 'rate guidance', 'economic activity' or 'financial and monetary conditions'.
Our results provide a text-based validation for shock measures used in numerous studies in monetary economics and asset pricing. For example, implementing the approach of Cieslak and Schrimpf (2019), we show that their monetary shocks relate to the ECB changing its stance towards 'rate guidance', growth shocks to 'economic activity', and risk premium shocks to 'financial and monetary conditions'. We present similar results for the shocks proposed by Altavilla et al. (2019) and Jarocinski and Karadi (2020), and we show that ECB stance changes matter for price movements in sovereign bond and currency markets as well.
At a more general level, our findings how different topics move different asset prices should be useful for the design of policy communication.
Delays in Public Good Provision and Green Transition Risks (E2, H4)
AbstractThe European Green Deal, the US Green New Deal and other similar initiatives around the world call for the public sector's strong involvement by mobilizing mass investments in public infrastructure to fasten the transition to a low-carbon economy. In this paper, we develop a general equilibrium model with two productive sectors (a dirty sector and a low-carbon sector) with capital accumulation and convex adjustment costs. Each productive sector employ both private and public capital, with varying elasticity of substitution. Given this setting, we analyze the consequences of delays associated with the provision of public infrastructure on both sectors, in the context of a transition to low-carbon economy. Uncertainty about the arrival of public capital can more than offset its positive spillovers for private-sector productivity. In a decentralized economy, unanticipated delays in the provision of public capital generate too much consumption and very little private investment relative to first-best optimum. With this model, we also study capital asset pricing and reallocation of private capital. For instance, we look into the sectoral risk measures, by deriving the beta of each sector as a function of its share of aggregate private capital.
Demographics, Deep Habits, and Markups (E3, J1)
AbstractI document a robust correlation between markups and demographic factors across countries and over time. I then develop an Overlapping Generations New Keynesian model with age-dependent Deep Habits to rationalize these findings. The model implies a link between the demographic transition to an ageing society and increasing market power by firms.
Depressive Disorders Among Breast Cancer Survivors of Racial and Ethnic Minorites in the US: New Evidence from Longitudinal MEPS Analysis (I1, D6)
AbstractMental health disorders, including major and minor depressions, among breast cancer survivors have been associated with poor cancer treatment outcomes, lower quality of life, risk of suicidal thoughts and lengthened hospitalizations. Existing studies have also shown that black breast cancer survivors are more likely to suffer from other chronic health conditions, including chronic mental health issues. However, there is limited evidence on the differential mental health impact of breast cancer across racial and ethnic groups. We use the 2009-2019 data from Medical Expenditure Panel Survey (MEPS) which includes detailed information on demographic characteristics, breast cancer survivorship, mental health status, and healthcare utilization. Our main outcome is whether the person has mental health disorders or not, and we utilize several measures available in MEPS Self-Administered Questionnaire (SAQ): the Mental Component Summary (MCS), the Veterans RAND 12 Item Health Survey the Kessler Index (K6) of non-specific psychological distress, and the Patient Health Questionnaire (PHQ-2). We further use restricted-use geographic identifiers to supplement the MEPS-HC data with the information on community characteristics and local healthcare resources, such as racial/ethnic makeup of the neighborhood (at the Census block group level), Social Vulnerability Index (at the Census tract level), Health Professional Shortage Area (HPSA) (at the county level), and other variables available in AHRQ’s Social Determinants of Health (SDOH) database. The empirical results show significant differences in mental health outcomes among racial and ethnic minorities. We further implement the Blinder-Oaxaca decomposition method to understand how much of the racial/ethnic disparities in mental health outcomes are explained by the differences in observable characteristics. Finally, we use a Difference-in-Differences model to explore whether the ACA Medicaid expansion decreased the racial/ethnic disparities in mental health among cancer survivors. This work enhances our understanding of the disparities in access to care faced by racial and ethnic minority groups.
Desk Rejection of Submissions to Academic Journals: An Efficient Screening Process? (D8, J0)
AbstractWith the advent of the Internet, word processing software, and personal computers, it has become easier for people to access existing research and submit new research to journals. However, the flood of content may exceed the time it takes for editors and referees to evaluate all these submissions. Therefore, more and more journals make increasing reliance on desk reject decisions to allow referees to focus their precious time on papers that editors believe are most likely to be worthy of publication.
This paper studies the desk reject decisions using the submissions in one leading journal and matches the citations and publication records via Microsoft Academic. The result shows that the desk reject decisions are not random. First, this paper finds that desk-reject submissions are less likely to be published in other journals compared with papers that pass the desk review round but end up being rejected. Second, this paper develops a control function model and shows that non-desk-reject papers tend to have significantly higher citation counts than desk-reject papers when controlling the predicted probability of avoiding desk rejections.
Moreover, by using natural language processing (NLP) and machine learning (ML) tools, this paper manages to construct and select important features from the abstract of the submissions, which are overlooked by previous literature. The machine learning models attempt to predict the desk reject decisions and achieve an accuracy of 72%, which is 10% higher than the baseline model which assumes rejecting every paper. The NLP models show that papers with more concrete and descriptive words are more likely to be published. The way a paper is written has proven to be very important in editorial decisions.
Determinants of the Presence, Density, and Popularity of U.S. Food Retailers (I1, R1)
AbstractThis paper uses census tract-level information to determine the presence, density and popularity of U.S. food retailers. We merge census tract-level demographic data, neighborhood amenities, and relative location efficiency information with food retailer location and foot traffic information obtained from cellphone location data.
We first utilize machine learning models to predict the location of four types of food retailers, namely, large retailers, small retailers, fast food restaurants, and full service restaurants. Our models can detect these food stores with a prediction accuracy of 71.62%, 91.87%, 77.62%, and 73.78% respectively. We then use Poisson-based regression techniques to determine the factors that affect the density and popularity of these food stores.
We find that the density and popularity of every type of store is consistently and significantly related to the presence, density, and popularity of other store types, indicating that nearby businesses are often complementary and helpful in driving customer traffic to other stores. This is particularly true for small stores such as small retailers, fast food, and full service restaurants, which are shown to benefit from traffic generated by large retailers.
Walkability, transit density and auto network density, which are measures of ease of accessibility by walking, public transport, and motor vehicles respectively, are all positively related to the density and popularity of small food stores. Similarly, reliance on public transportation decreases foot traffic to food stores of all types. All these point to the importance of accessibility and visibility in determining store density and popularity.
For policymakers, the findings of the study have the potential to tease out proactive approaches to alter food store environments by encouraging establishment of healthy food store options in underserved areas. As an example, our results suggest that policies targeted at improving access and transportation options can affect food retailer choice, and consequently, diet quality.
Determinants of unhealthy BMI among child-bearing age women: new evidence from Bangladesh Demographic and Health Surveys (I3, I1)
AbstractThe medical care and other economic costs of unhealthy body mass index (BMI), as measured by increasing health expenditure, lost economic growth and labor productivity, are immense. Limited empirical evidence exists in the economics literature on the determinants of unhealthy BMI for developing countries. This study specifies and estimates a model of the determinants of unhealthy BMI categories separately for child-bearing urban and rural women aged 15-49 years old, employing data from multiple waves (2004-18) of the nationally representative Demographic and Health Surveys (DHS) of Bangladesh. Using multilevel nested logit empirical models, we detect robust heterogeneous associations of unhealthy BMI categories with individual-level socioeconomic variables and cluster (community and region) factors. After removing cluster effects to obtain more precise estimates, we document that underweight status risk is inversely related to years of education for urban and rural women. In contrast, a direct relationship exists between the risk of being overweight and years of education for urban and rural women. The risk of being obese is positively associated with years of education only for urban women. Women’s employment status at the extensive margin lowers the risk of being overweight and obese in both cohorts. Women are less likely to be underweight and more likely to be overweight and obese as their household's socioeconomic status improves. Study findings could inform both policy and practice aimed at reducing economic burdens of the unhealthy BMI to achieve important targets such as sustainable development goals.
Difference Is Good: Evidence from the Effect of Ethnicity Diversity on Corporate Interstate Investments (J1, G3)
AbstractWe examine the effect of local ethnicity diversity on attracting interstate capital investments. Using a novel dataset of interstate investments at the project level, we show that a county’s ethnicity diversity is positively associated with the likelihood of the county being chosen as the location of interstate investment projects made by firms from other states. This result is stronger when the investing firm’s home county has higher ethnicity diversity or when the investing firm ranks higher on diversity policies, has higher R&D intensity, or operates in a high-technology industry. After aggregating the investment data at the county level, we find that a county with greater ethnicity diversity also secures greater amounts of project investments made by firms from other states. Overall, our results suggest that an ethnically diversified local environment helps attract more outside investments, especially investments from high-technology, innovative firms, or firms with a similar diversity orientation. The evidence speaks to the debate over the effects of ethnicity diversity on economic development.
Disrupting Science (O3, M0)
AbstractIn this paper, we explore how the rise of remote collaboration has shaped the pursuit
of new ideas in scientific discovery. To systematically distinguish between disruptive
breakthroughs and incremental discoveries, we use citations data for over ten
million research teams—publishing in 11 fields of research from 1961 to 2020. On average,
we document a robust and significant negative impact of remote collaboration
on breakthrough discovery. However, beginning in the 2010s, the negative impact tapers
off and even becomes positive. We provide evidence suggesting that the reversal is
driven by improvements in key technologies needed for effective remote collaboration.
Do Board Connections between Product Market Peers Impede Competition? (G3, L4)
AbstractUsing a treated-control matched sample, we find that after a new direct board connection is formed to a product market peer under the Hoberg-Phillips industry classification, a firm's gross margin significantly increases by 0.8%. Gross margin also rises after a new indirect board connection is formed to a product market peer through a third intermediate firm. We see consistent results when the new connections are caused by changes on the board of an intermediate firm. Such third-party initiated changes are unlikely to be related to the economic prospects of the focal firm. Consistent with the anti-competitive mechanism, the effects are stronger when the newly connected peers are located closer to each other or have more similar businesses and when the firms are in industries with greater potential benefits of collusion.
This paper has several important regulatory implications. First, our results indicate the role of directors in anti-competitive practices and provide support for the current ban of interlocking directorates between competing firms. Second, the results suggest that text-based analyses are powerful in identifying competitors in the market place and can have the potential to aid the execution of antitrust regulations. In addition, we find that indirect connections via an intermediate firm also have positive effects on profitability even though their economic magnitudes are smaller than those of direct connections. This argues for going beyond interlocking directorates and putting restraints on indirect board connections between competitors as well, especially in cases where the detrimental effects of anti-competitive practices on consumer welfare are substantial.
Do Fair Housing Policies Help or Hinder? (J1, R3)
AbstractSeveral U.S. cities have imposed strict regulations on landlords to combat discrimination and reduce racial disparities in housing. This paper asks whether these policies help minority citizens or inadvertently exacerbate racial disparities in housing. It does so in the context of 1. Washington state, which imposed fines on landlords for using blanket bans on applicants with a criminal record in 2017, and 2. Seattle, which has imposed multiple fair housing policies, such as a ban on background checks and a “First in Line” policy. To identify effects, I use a difference-in-differences approach to compare outcomes of black and white residents in Washington state and Seattle over time. Results indicate that the policies did not affect the likelihood of renting or moving but that housing spending increased for black WA citizens by approximately $100.
Do Women Need to Provide More Skill Signals to Advance Their Careers? (J1, G3)
AbstractOn average, additional qualification signals increase male directors' probability to become CEO by 10.1%, while this probability increases by more than 30% for female directors. With respect to compensation, additional qualification signals increase male directors' compensation by 10.4% ($444,145). Female directors' compensation benefits more than twice as much and increases by 21.1% ($849,500). Additionally, the composition of professional networks matters for women's career advancement. We find that connections to male executives and executives who are members of an old boys' club are more valuable for women than having access to all-female networks.
Does Civic Education Foster Democratic Attitudes? (I2, H4)
AbstractWe study the effect of civic education in school on students’ democratic attitudes and civic engagement later in life. We exploit variation in weekly hours of civic education in secondary school from curricular reforms across states, school tracks, and over time in a generalized differences-in-differences framework. We combine a novel database documenting seven decades of state curricular reforms in Germany with the German Socio-Economic Panel. Our findings suggest that an increase in weekly hours of civic education significantly raises civic engagement later in life. We rule out that effects are driven by state-specific trends or by the political party enacting the reform.
Does Compulsory Schooling Exacerbate Son-Preference Practice? Evidence from China (J1, O1)
AbstractGovernment policies may alter cultural practices that persist in part due to economic incentives. In this paper, I develop a conceptual framework that shows how compulsory schooling affects parental incentives and exacerbates pre-existing son-preferences in China. I make six testable predictions and test these predictions with a difference-in-differences method using China’s 1986 compulsory schooling law (CSL) and 2008 free compulsory schooling law. My empirical results show that the CSL reduces the gender gap in educational achievement, skews the sex ratio toward males, and reduces the fertility rate. The change in the sex ratio is mainly driven by prenatal sex selection and is larger in locations where accessing ultrasound technology is less costly. In addition, where prenatal sex selection is impossible, the CSL leads to relatively higher rates of malnutrition in girls. On the other hand, when schooling is cheaper, the sex ratio increase is reversed and son-preference practice is mitigated. As policy implications, I further explore the pre-law characteristics and show that compulsory schooling with more school resources and higher education returns will reduce the unintended consequences it generates.
Does Demography Determine Democratic Attitudes? (D7, P4)
AbstractThis paper presents new evidence on how demography affects attitudes toward democracy and policy preferences. The empirical analysis disentangles age effects from cohort effects and separates their role from economic and political factors that shape political preferences in a given period, using survey responses for more than 200,000 individual observations from 90 countries. The results show that the support for democracy increases with age and, at the same time, depends on cohort-specific factors that are related to past experiences with democracy and socioeconomic status. The findings shed new light on the role of demography in terms of life-cycle and cohort effects for
Does Hiring Credit Increase Jobs for Vulnerable Workers? Evaluating the First Job Act in Colombia (J2, J4)
AbstractIn this paper, we assess the impact of a hiring credit policy in Colombia for the program-defined vulnerable workers: youth, career-disrupted women, and low-income workers. The First Job Act of 2010 incentivized formal job creation for vulnerable workers by offering a corporate income tax rebate for firms' payroll tax accruing to new vulnerable workers. We estimate the impact of such cash windfall for firms on the employment and wages of the vulnerable. A distinct feature of the First Job Act is that the policy precludes reductions in the headcount of workers or wage bill reductions for eligibility for the tax rebate. As such, our estimated employment effects show job creation on the extensive margin within firms. The difference-in-differences (DID) and propensity score matching (PSM) results for all populations show substantial formal job creation effects for vulnerable workers, in particular, the youth and career-disrupted females. The employment increase accompanied slight wage increase for the vulnerable workers. The evidence is consistent with the results from Kugler and Kugler (2009) and Kugler et al. (2017). The matched sample DID proves that the formal job growth was not the result of substitution for similar non-beneficiary workers.
Does Intelligence Shield Children from the Effects of Parental Unemployment? (I2, J6)
AbstractThe negative effect of parental job loss on various outcomes of children is well-documented. In this paper, I provide new evidence on how these effects change with the intelligence of children. I find that higher intelligence mitigates some of the impacts, but not all. Parental unemployment is more harmful to the education of children at the top of the distribution. This forces them to start their careers at lower-paying jobs and continues to weigh down on their wages even later in life. Nevertheless, higher intelligence alleviates the effect on labour supply, job ranking and monthly earnings over time. I also provide evidence that the results may be driven by the loss of income and/or psychological distress following the unemployment of parents.
Does It Matter Where You Grow Up? Childhood Exposure Effects in Latin America and the Caribbean (J6, R2)
AbstractI study whether the observed differences in intergenerational educational mobility across regions in Latin America and the Caribbean are due to the sorting of families or the effect of growing up in these different places. I exploit differences in the age of children at the time their families move across locations to isolate regional childhood exposure effects from sorting. I find a convergence rate of 3.5% per year of exposure between age 1 to 11, implying that children who move at the age of 1 would pick up 35% of the observed differences in mobility between origin and destination. These results are robust to using a specification that identifies the effect of place within households, the use of only anomalously high migration outflows, instrumenting the choice of destination with historical migration, and a combination of both approaches.
Does Media Coverage of Firms’ Environment, Social, and Governance (ESG) Incidents Affect Analyst Coverage and Forecasts? (M4, M0)
AbstractMedia coverage of environment, social, and governance (ESG) issues provides useful information for analysts as corporate social irresponsibility events potentially influence corporate performance and risks. Our study explores whether and how analysts respond to media coverage of corporate social irresponsibility by examining its relationship with analyst coverage and forecasts. We find evidence that the level of analyst coverage is negatively associated with a firm’s ESG incidents covered by the media. This association is more pronounced for firms with high business risk, high information risk, more intense industrial product market competition, and more severe ESG scandals. We also find a positive association between media-covered ESG incidents and analyst forecast error and dispersion, suggesting that analysts might fail to incorporate the ESG risk exposures into their forecasts in an appropriate manner. The business risk and information risk tend to be higher for firms that are covered by the media for having got involved in ESG incidents, thereby explaining why the analyst coverage and forecasts for these firms are adversely affected. Overall, our results suggest that corporate social irresponsibility undermines the role analysts play as information intermediaries for investors in the stock market.
Driving, Dropouts, and Drive-throughs: Mobility Restrictions and Teen Human Capital (I2, J2)
AbstractWe provide evidence that graduated driver licensing (GDL) laws, originally intended to improve public safety, impact both high school completion and teen employment. Many teens use automobiles to commute to both school and to employment. Because school and work decisions are interrelated, the effects of automobile-specific mobility restrictions are ex ante ambiguous. Combining variation in the timing of both GDL law adoption and changes in compulsory school laws into a triple-difference research design shows that restricting teen mobility significantly reduces high school dropout rates and teen employment. These findings are consistent with a model in which teens use automobiles to access educational distractions (employment or even risky behaviors). We develop a discrete choice model that reflects reduced access to school, work, and other activities, which reveals that limiting access to work alone cannot explain the reduction in high school dropout rates.
Early Child Care and Maternal Labor Supply: A Field Experiment (J2, D1)
AbstractWe present experimental evidence that enabling access to universal early child care for families with
lower socioeconomic status (SES) increases maternal labor supply. Our intervention provides families with
customized help for child care applications, resulting in a large increase in enrollment among lower-SES
families. The treatment increases lower-SES mothers’ full-time employment rates by 9 percentage points
(+160%), household income by 10%, and mothers’ earnings by 22%. The effect on full-time employment
is largely driven by increased care hours provided by child care centers and fathers. Overall, the treatment substantially improves intra-household gender equality in terms of child care duties and earnings.
Economic Shocks and Social Preferences (E7, Z1)
AbstractAltruistic and pro-social behaviors form the building blocks of modern economic life and individual well-being. Yet, little is known about how such behaviors and associated preferences arise. Which factors can explain the differences in pro-social behaviors across and within countries?
We build upon recent insights from psychology and economics which suggest that economic shocks during early adulthood, also known as the “formative years” (ages 18-25), can affect economic preferences in the long run. Our study is the first to consider whether such shocks affect altruism and pro-social behavior. A priori, the direction of the effect (if any) is unclear: people who were exposed to a recession might be more selfish since they feel more deserving; alternatively, they might empathize more with the ones who are less fortunate and therefore be more altruistic.
Our study exploits within-country, across-cohort variation in exposure to economic shocks (recessions) during a person’s formative years. We draw upon data on individual preferences (altruism, trust, positive and negative reciprocity) from the Global Preferences Survey and combine it with information on country-level economic shocks using data from the Maddison Project Database. The data spans 75 countries from six continents, and 78 birth cohorts (1914-1991).
We find that individuals who were exposed to a recession during early adulthood are less pro-social than those who were not exposed. Specifically, being exposed is associated with a decrease in positive reciprocity by 3.7 percent of a standard deviation. This effect is robust to accounting for changes in country-level institutions over time and holds for various definitions of recessions. Placebo tests using exposure during other age brackets confirm that the effects occur only through exposure during the formative years.
Economic Shocks, Food Insufficiency and Mental Health during the COVID-19 Pandemic (I1)
AbstractMillions of Americans experienced a sudden loss of income along with hunger early in the COVID-19 outbreak. These shocks had crucial implications for the economy, society, and individuals. In particular, a mental health crisis was associated with these food and economic hardships during this outbreak. In this paper, we employ weekly household-level survey data from April 23, 2020 to March 29, 2021 to investigate the effect of food insufficiency and economic shocks, defined as income loss and unemployment, on the mental health of adults during the COVID-19 pandemic. Using neighboring states' food insufficiency and unemployment as instrument variables and excluding people who were voluntary to quit their jobs to control for endogeneity, we find a significant impact of the pandemic on both food sufficiency and mental health, with heterogeneous effects of food insufficiency and income loss on mental health across different groups. Subsample analyses show heterogeneous effects of food insufficiency and economic shocks across groups during the pandemic and reveal disadvantaged groups. We find that, in households, males were more likely to be more anxious if their households experienced food security problems. No significant effects of economic shocks on mental health were found in the top 15 Metropolitan areas, but they did exist in non-metro areas. Even when we include Supplemental Nutrition Assistance Program (SNAP) and unemployment insurance application as control variables, the negative effect of food insufficiency on mental health persists, suggesting a limited impact of public benefits to reduce mental health through decreasing food insufficiency. These findings provide policy implications for identifying the most vulnerable groups and for providing prompt mental health assistance during shocks such as those created by economic or public health crises.
Effect of Welfare Reform on Intergenerational Adult Outcome - Comparing Immigrant and Native Children (I3, J6)
AbstractThis paper studies the effect of welfare reform on intergenerational adult outcomes in both immigrant and native households receiving public assistance. While an implicit aim of the 1996 welfare reform was to disrupt the intergenerational transmission of disadvantage across all U.S. households receiving welfare, little is known about the varying effects of welfare reform on any particular next-generation adulthood development indicator across nativity status. Using the Panel Study of Income Dynamics and cross-state variation in reform policy to address potential endogenous selection, we find that childhood exposure to welfare reform increased immigrant children's educational attainment more than the children of U.S. natives. Reform has also increased immigrant children's labor market participation as adults, while estimated effects for children of natives are negligible, indicating the importance of children's nativity status in understanding the impacts of welfare reform.
Electricity Market Design and Market-Based Environmental Policies in India (Q4, L9)
AbstractHow do market-based environmental policy instruments perform in non-market-based electricity systems? I explore this question in the context of the electric power sector in India, where electricity is transacted primarily via long-term bilateral contracts and state-owned distribution utilities self-schedule contracted power plants to meet their demand. The absence of a centralized and dynamic market-based economic dispatch mechanism generates short-run misallocation in electricity dispatch and distorts long-run investment decisions, such as the incentive to invest in flexible generation capacity and energy storage to complement renewable-based capacity. Using detailed data on power plant operations and the network of long-term bilateral contracts, I find that the demand for electricity from coal-fired power plants with a higher share of capacity allocated under long-term contract(s) is less sensitive to changes in coal prices, implying that the existing market design could erode some of the environmental benefits of carbon pricing.
Employment Effect of Means-Tested Program: Evidence from a Pension Reform in Chile (J2, D8)
AbstractI study the employment effects of a 2008 policy reform in the Chilean pension system. The reform i) increased pension benefits, ii) changed the accrual rates, iii) relaxed eligibility criteria for low-savings retirees. I estimate the labor supply responses using a unique database that combines monthly administrative records with a representative panel survey. Using a Difference in Differences (DID) approach with multiple time periods, I find that, on average, the reform increased labor force participation and hours worked for men between 60-64 years old by 15% and 4%, respectively. I find heterogeneous effects, depending on how their accrual rate changes, which depends on where they are in the distribution of pension wealth and years of contribution. I find that lack of pension knowledge is crucial in explaining heterogeneity within different groups of workers. People with similar levels of pension wealth and monthly contributions behave differently depending on what they believe their pension assets and number of contributions are. I then develop a three-period model where workers decide consumption, non-pension assets and labor supply. I conclude that the net impacts on labor decisions depend on pension wealth, and the knowledge about the pension system. These findings are consistent with the empirical results and suggest that more access to relevant information regarding pension system may facilitate people to make well-educated decisions when facing a massive pension reform, especially in the case of a country with a high level of informality.
Environmental Stress, Genetic Engineering, and Agricultural Productivity (Q1, Q5)
AbstractExploring the synergies of biotechnology, environmental stresses, and farm management is a new opportunity to enhance agricultural productivity. However, literature on the nexus between biotechnology and environmental stresses remains scant despite a growing body of literature on the interplay of the other combinations among the three components. Ground-level ozone pollution, a major source of environmental stresses other than climate change, has been identified as a hidden threat to U.S. agriculture; meanwhile, genetically modified (GM) crops swiftly spread out the U.S. (e.g., GM corn soared up from zero in 1995 to 92% in 2020). This research is the first one that investigates whether biotechnology damages or improves crops’ ability to deal with environmental stresses using observation data. Based on some datasets from USDA, TROPOMI, and ECMWF, combined with a survey dataset on GM adoption, we constructed county-level panel data from 2003 to 2020 for corn and soybeans. Utilizing the lightning and upwind ozone as instruments for focal ozone concentrations, we adopt a two-way fixed effects model to investigate this research question. Our results suggest a hidden role of biotechnology: it significantly damages the crops’ ability to cope with ozone pollution despite biotechnology itself improving agricultural productivity, and the positive effects of biotechnology itself disappear when ozone pollution exceeds approximately 95 ppb. More importantly, our results are robust for both corn and soybeans in different econometric specifications and subsamples. Our results highlight the importance of breeding GM seeds that are resilient to environmental stress in addition to existing traits.
Equity of COVID-19-Induced Job Loss Duration and Effect of Unemployment Insurance Generosity on Labor Supply (J2, D0)
AbstractI study COVID-19-induced job losses using administrative data on the universe of Unemployment Insurance claims for the state of Indiana. I show that during the pandemic, women and Blacks not only lost more jobs but also stayed unemployed longer. Using a difference-in-difference strategy exploiting the eligibility for Lost Wages Assistance (LWA) benefits, I find that a change in weekly benefit amount by $300 is associated with a change in the exit rate out of unemployment by 2.1 percentage points (p.p.) in the opposite direction. This is because of both a decline in the reemployment rate (1.7 p. p.) which represents the disincentive or the moral hazard effect of UI and a decline in labor force exits (0.4 p. p.) which suggests the ability of UI to keep workers in the labor force or the labor force participation effect. Access to six weeks of increased benefit amounts did not change workers’ probability of switching employers, nor significantly affect their earnings after reemployment, suggesting worker-firm match quality was unaffected.
Estimation of a Latent Reference Point: Method and Application to NYC Taxi Drivers (C1, J2)
AbstractI use a dynamic discrete choice model with a latent variable to flexibly estimate reference- dependent utility models. The structure and evolution of the reference point are estimated directly from observational data. I apply the model to the daily labor-supply choices of NYC taxi drivers and use a Bayesian estimation approach. I find that rational expectations are an important determinant of the reference point but do not fully explain its evolution. The refer- ence point adjusts asymmetrically, responding more to positive income shocks than to negative ones. The reference point also has an important transitory component: a shock to the reference point dissipates within hours.
Evaluating the Role of Policy Interactions and Behavioral Mistakes in Understanding Consumer Demand for Energy Efficiency (Q5, D9)
AbstractThis research investigates the effect of policy interactions and behavioral mistakes on consumer adoption of energy-efficient technology. Information provision and incentives are among the most common tools used by environmental policy-makers and -implementers to address energy efficiency. This work aims to identify potential interactions between incentives and information, as well as interactions between the provision of different types of information — energy cost savings information and emissions impact information. Potential behavioral explanations for these policy interactions are also explored, with a focus on behavioral mistakes, or errors in economic reasoning, caused by not only imperfect information and inattention but also overreactions to price changes framed as discounts, a lack of deliberative thinking and a lack of a framework to evaluate choice optimality. The primary outcome of interest is demand for an energy-efficient good, a smart power strip. Data for this research will be collected through a theoretically motivated online survey-experiment in which respondents are randomized into treatments representing different policy mixes and varying in the information and incentives provided. Respondents will also be prompted to deliberate over the consistency of their preferences with each other and a random utility model of behavior.
Results are expected to offer insights on how to effectively design interventions aimed at increasing residential energy efficiency. The evidence generated by the portion of the experiment designed to study interaction effects can be used to make rich comparisons between policy treatments that represent different mixes of policy tools. For example, interactions between information and incentives would imply that a mix of information and incentives is either ineffective, effective but less effective than anticipated, or more effective than anticipated in increasing demand relative to information or incentives alone. The treatment effects associated with guided deliberation can be used to determine whether or not deliberation based on economic principles can improve policy efficacy, particularly by counteracting negative interaction effects and correcting behavioral mistakes where they exist. Effects of deliberation that depend on the information or incentives provided would suggest that policy choice affects the process consumers use to make purchase decisions about energy efficient products. The effects of information, incentives and deliberation are expected to be heterogeneous, which would suggest that energy efficiency policy be carefully targeted and blanket policy recommendations be avoided.
Excess Volatility of British Pound: Jumps or Regime Switches? (F3, C0)
AbstractThis paper estimates the dynamics of British pound exchange rate using continuous-time models which combine both Lévy process and regime-switching. It also compares the diffusion model, the regime-switching diffusion model, and the regime-switching Lévy model, and discusses the appropriate model selection for the nominal exchange rate.
This paper finds that jumps and regime switches play different roles in explaining the excess volatility of the nominal exchange rate. In Regime 1, the pound has higher value and is more stable and more persistent. However, it faces a much larger sudden depreciation or appreciation when a jump event happens. Conversely, in Regime 2, the pound has lower value and larger variance, but is less subjected to extreme value variations. The probability of staying within Regime 2 is smaller compared with that of Regime 1. These dynamic patterns shown by the estimated Markov regimes with jumps have a great match for the features of the business cycles and the exchange rate regimes in the U.K. history.
This paper helps us better decompose the driving forces behind the spikes and slumps of the exchange rate and establishes a foundation for future studies to endogenously explain the excess volatility of the nominal exchange rate using macroeconomic variables.
Expert Recommender and Reputation Failure (D8, C7)
AbstractThis paper studies the role of a recommender’s career concerns in his long-run relationship with a consumer when the recommender has a private type in his expertise. An informed type’s expertise is valuable for ongoing purchasing decisions of the consumer, whereas an uninformed type lacks the expertise. The uninformed type cannot mimic the informed type, suggesting that the informed type can build a reputation for competence and that an equilibrium should exhibit information transmission. Nonetheless, I find that the relationship completely breaks down if the recommender is sufficiently patient and thus has stronger career concerns. Moreover, this occurs even when the uncertainty of expertise is arbitrarily small.
Farm To Firm: Clustering and Returns to Scale in Agricultural Value Chains (Q1, F1)
AbstractMovement away from low productivity subsistence agriculture and towards high value, export oriented, agricultural chains is often thought of as a key driver of structural change. However, these agricultural value chains are unequally distributed across space, often times located in one or two productive clusters within a country. I argue that this relationship is driven by crop-specific farm-to-firm linkages, which generate regions of high specialization in exporting and specific crop cultivation, but leave many regions without these linkages unable to specialize. Empirically, I show casual evidence that farms located closer to exporters of a given crop are more likely to cultivate it. Using farm and firm microdata from Mexico, I estimate the fixed costs of entry for firms in the agricultural value chain as well as input costs of production for exporting farms, both of which are partially driven by the costs of compliance with phytosanitary barriers. I then develop a model which can accommodate these patterns of clustering, and can explain aggregate patterns of land use. Finally, using model based counterfactual analysis, I examine which policies can best facilitate the goal of high value agricultural export promotion as well as the ability of agricultural value chains to shift under climate change.
Financial Statement Errors and Analysts: Obstacle or Opportunity? (M4, M0)
AbstractFinancial analysts need to decide whether to follow firms with financial statement errors and, if they do, commit to additional effort and resources to develop accurate forecasts of earnings, relative to firms with fewer or no errors. Further, inaccurate forecasts could impair analysts’ reputation. On the other hand, firms with financial statement errors present an opportunity for analysts to add value to investors by providing informative forecasts. We examine which of these two explanations prevail in the post-Regulation Fair Disclosure (Reg FD) era where firms are prohibited from disclosing private information to analysts. Using the financial statement divergence (FSD) scores (Amiram et al. 2015) as the measure of financial statement errors, we find that FSD scores are positively associated with the number of analysts following and forecasting for firms. This association is more pronounced for firms with lower institutional stock ownership. We also find that financial statement errors strengthen the stock market reaction to analyst earnings forecasts, and this strengthening effect is stronger for firms with lower institutional ownership. Lastly, we find that financial statement errors increase the errors in the earnings forecasts by analysts with varying degrees of sophistication. In a nutshell, although financial statement errors lead to an increase in investors’ demand for analyst forecasts and in analyst coverage, analysts tend to commit more errors in their earnings forecasts. These findings contribute to understanding the relation of financial statement errors with analyst coverage and forecast accuracy.
Finding out who you are: a self-exploration view of education (I2, D8)
AbstractI study the role of education as self-exploration. Students in my model have different priors about their talents and update their beliefs after receiving noisy signals about themselves. I characterize a socially optimal design of the signal structure. An optimal structure encourages a career in which participating students are on average more confident. I apply the model to students in the United States and estimate its parameters from data. Advanced science classes in high school tend to encourage science majors in college. Their estimated self-exploration value is a 5-percent average increase in future earnings.
Fiscal Monetary Services and Inflation (H6, E3)
AbstractIn this paper I use a Fisher ideal index to track the monetary services provided by marketable US government debt. To do so, I first develop the theory necessary to consider using such a statistical index number and show how the value of these fiscal monetary services expand the fiscal capacity to borrow. I then use Jorda (2005) projections to estimate the impact of such monetary services on inflation. I find that a one-percent increase in fiscal monetary services produces a positive and statistically-significant inflationary response that peaks between four and five basis points and persists for eight months. Given that the average growth rate of the fiscal monetary services in my sample is 2.5 percent, the impact is also economically significant. Together, these results suggest that there is a monetary services channel to the fiscal theory of the price level.
Forecasting Agricultural Commodity Prices Using CNN-GRU Networks with Likelihood Loss Functions (C1, Q1)
AbstractThis paper exploits a CNN-GRU neural network model in forecasting commodity prices. This paper's contribution is twofold. First, the commodity market is commonly considered seasonal due to the nature of the product. Traditional models such as VAR and ARIMA can perform poorly in keeping track of long-term lagged information. This model is able to capture the long-term nonlinear dependency and detect anomalies by leveraging a flexible structure consisting of GRU and CNN layers. One of the significant problems in modeling commodity price returns is their non-normal distribution. This paper's second contribution is a likelihood-based SGD approach that captures the asymmetrically distributed returns with a fat-tail. Instead of using a distance-based loss function, this paper investigates the disturbance distribution and minimizes a likelihood-related loss function to achieve convergence. Out-of-sample evaluation results suggest that this model can alleviate the above mentioned problems and improve forecasting accuracy.
Foreign Mining, Labor Welfare and Local Trust: Evidence from Kyrgyzstan Gold Mine (Q3, O0)
AbstractThere has been controversy over the impact of foreign natural resource investment on worker welfare and host country politics. This article explores this question by analyzing one of Kyrgyzstan’s dominant natural resource FDIs, which accounted for 12.5\% of Kyrgyzstan’s GDP in 2020. Using geolocated data from household panel surveys in Kyrgyzstan, the study shows Kumtor, the largest foreign mine, offers workers better wages and social benefits. Foreign mining investment has caused national discontent, not because it worsens the local economy, but because it creates greater inequality and social division. This study also believes that the emergence of Kumtor is associated to the high level of trust in the central government by local residents, especially high-income miners. It also fosters mistrust among its employees, local communities and the foreign investor.
Foreign Student Share and Enrollment Changes in U.S. Higher Education during the COVID-19 Pandemic (I2)
AbstractForeign student enrollment in U.S. postsecondary education dropped by 15 percent in the fall 2020 semester. I examine whether the COVID-19 pandemic had heterogeneous effects on postsecondary institutions with different pre-COVID shares of international students. I use a sample of 261 public four-year universities from 2016 to 2020 and exploit an exogenous shock to enrollment in the fall 2020. Using the pre-COVID share of foreign students as a treatment, I find a positive differential effect on foreign student enrollment. Difference-in-differences model result suggests that a 5 percentage point difference in foreign student share is associated with a 3.5 percentage point increase in foreign student enrollment in the fall 2020. This positive differential effect is partially due to the reopening status and policies to improve foreign student retention. I do not find a statistically significant result for domestic students.
Frequent vs. Lumpy: Impact of Remittance Patterns on Investment, Consumption, and Poverty (F2, D1)
AbstractThis paper examines the relationship between remittance patterns and household-level economic outcomes. Remittance patterns vary with some households receiving remittance in bulky less-frequent sums while others receiving smaller amounts in a more frequent manner. High frequency transfers in the presence of high transaction costs mean that both the senders and receivers value the importance of remittances on a more regular basis. Simple economic intuition derived from the permanent income hypothesis suggests that a higher likelihood of receiving frequent remittances over their lifetime leads households to increase present consumption. In other words, receiving remittances in a regular manner signals to households that migrants will continue to send remittances in the future, therefore, households increases their consumption accordingly. Using a rich dataset from a nationally representative household survey from Nepal, we document the effect of remittance frequency on household decisions. Controlling for the total value of remittances received by a household, we find that more frequent remittances indeed increase consumption. We also find that households who receive more frequent remittances are less likely to be below the poverty line. To address potential endogeneity issues in our estimation, we use women migrants as an exogenous instrument to isolate the causal influence of frequency of the remitted funds on households’ consumption outcomes. Women tend to remit more frequently due to familial and social constraints and they tend to be more altruistic and attached to their families back home. Therefore, women’s remittances tend to be regular and driven by factors exogenous to the economic outcomes, but also our instrument passes statistical tests for validity. Regardless of the motive of migrants for frequent remittances, in our work, we document the impact of this pattern on stayers’ consumption. To our knowledge, we are the first to examine the effect of remittance regularity, rather than its volume.
Friends and Drugs: The Role of Social Networks in the Opioid Epidemic (I0, I1)
AbstractMany infectious diseases, such as SARS-CoV-2, are easily transmissible through personal contact and occur in waves that show exponential growth in case numbers. Although not physically transmissible, non-infectious diseases like substance use disorder, obesity, or depression are similar in this regard because they can also spread rapidly through the population. A natural question is: What role do friendship networks play in the spread of non-infectious diseases? We study this question in the context of the US opioid epidemic.
The US is currently experiencing a severe opioid epidemic that has claimed half a million lives over the past two decades. Opioid consumption is influenced by many socio-economic factors like personal characteristics or the availability of opioids. Our main contribution is to document that friendship networks are an important driver of the opioid epidemic. We proceed in two main steps:
First, we use panel regressions to analyze the spatial spread of the opioid epidemic and show that having many friendship links to counties with high exposure to the opioid epidemic positively correlates with future overdose death rates. This correlation is not driven by physical proximity and socio-economic characteristics.
Second, to establish causality, we exploit the 2010 OxyContin reformulation and the staggered introduction of must-access PDMPs. Both interventions had the unintended consequence that users switched to illegal opioids, thereby constituting shocks to illegal drug consumption that are exogenous to pre-existing friendship networks. We show that these shocks propagate through friendship networks to socially proximate counties. The friendship network effect is driven by illegal opioids (e.g., heroin and fentanyl) and economically sizable even when considering geographically distant counties. Consistent with a causal mechanism, we do not find evidence for differential pre-trends between treatment and control counties.
At a general level, our results provide first evidence on the social transmission of non-infectious diseases.
From Employee to Entrepreneur: The Role of Unemployment Risk (J6, J4)
AbstractEconomists have long presumed that there is too little entrepreneurial activity and an emerging literature considers various labor market frictions that impede entrepreneurial activities. The “last-in-first-out” (LIFO) seniority rule is a major cornerstone in the Swedish Employment Protection Act (EPA). This seniority principle stipulates that the employee who was employed last has to go first when a firm decides to downsize. A reform of the LIFO principle took place in January 2001. After the reform, firms with ten or fewer employees are allowed to exempt two workers from layoff according to the LIFO principle. We exploit this quasi-experimental shock to shed light on how employment protection affects entrepreneurship using rich firm- and individual-level data from the full population of Swedish firms.
We use a standard difference-in-difference approach and compare the post-reform outcomes for individuals who work for firms which employ no more than 10 employees pre-reform (treatments) with individuals who work for firms which employs greater than 10 employees pre-reform (controls). To sharpen the identification, we will adopt an approach similar to the regression discontinuity design to refine the treatment groups to firms which employ 9 and 10 employees and the control groups to firms which employ 11 and 12 employees. Moreover, we will examine the decisions by the individuals whose spouses work for the treatment and control firms.
We find that the entrepreneurial activities increase significantly after the relaxation of employment protection policies. Employees whose job security are negatively affected after this reform are substantially more likely to start new business and their business start-ups exhibit greater growth than controls. Our results provide some of the first empirical evidence on how employment protection policy can impede entrepreneurship.
Gender Biases in Police Decisions (K4, J1)
AbstractThis paper studies the gender gap in officer decisions to frisk male and female subjects in nearly 40,000 stop and frisk interactions in Seattle from 2015-2019. Female subjects are 14 percentage points less likely to be frisked after being stopped by the police, but this gap closes by 46% when the officer is female. This cross-gender pattern in frisks is not driven by the systematic assignment of female officers to female subjects, and is robust to conditioning on officer fixed effects and a large set of observable subject, stop, and officer characteristics. The cross-patterns are consistent with preference-based discrimination, in which male and female officers face a differential cost to frisking female subjects. Taking advantage of the start of the #MeToo movement and the staggered introduction of body cameras, we do not find strong support for these costs being driven by inappropriateness norms or concerns about accountability. Though only suggestive, the findings point towards the importance of paternalism. Finally, we demonstrate that increasing female representation in policing could reduce the gender frisk gap both directly but also indirectly by impacting the behavior of male officers assigned to mixed gender crews or squads with female supervisors.
Gender Differences in Remote Learning amid COVID-19: Disruptive Peers and Self-Control (I2)
AbstractA shift to remote and blended learning following pandemic-induced school closures changed the nature of the learning environment for students. It also led to changes in the relative importance of educational inputs and their impacts on student outcomes. Further, concerns have been raised that the pandemic may have exacerbated pre-existent achievement gaps by race/ethnicity, gender, and socio-economic status. In this paper, I explore another dimension of achievement growth differences during the pandemic, student gender. There is emerging evidence of differences in achievement growth for male and female students, yet there has been little to no work to understand the casual mechanisms of such gender differences and the implications for education policies. I focus on two potential mechanisms associated with the switch from in-person to remote learning – changes in the nature of peer interactions and differences in the level of self-control required to learn in a remote environment. I employ the Blinder-Oaxaca decomposition and Two-Stage Least Squares methods to examine how changes in exposure to disruptive peers and gender-based differences in self-control due to the pandemic-induced remote learning impacted student learning trajectories. Analyses during pre-pandemic in-person instruction indicate that being exposed to historically disruptive peers and a lack of personal self-control negatively affect student achievement and that boys are disproportionately negatively impacted by disruptive peers, compared to girls. The two mechanisms continue to be significant determinants of student achievement over the course of the pandemic, where differences in students’ pre-pandemic self-control explain a moderate share of the observed gender achievement gaps. Also, results indicate that during blended remote learning, where students received a mix of in-person and virtual instruction, math achievement gaps widened for those who remained remote whereas there were no significant gaps for those returned to school.
Geographic Variation in Mental Health Care: Evidence from Migration (I1)
AbstractMental illnesses, including depression and anxiety, are widespread among elderly adults in the U.S. Although treatments are known to be effective for many mental health conditions, care utilization rates vary substantially across geographic areas. Using administrative data from Medicare, this paper isolates the patient- and place-specific drivers of the geographic variation in mental health care use among elderly adults. Specifically, I use an event-study framework with individual fixed effects to study changes in mental health care utilization for patients who move across areas with differing rates of average utilization. My results show that 60 percent of the geographic variation is attributed to place-specific factors. I then explore components of the "place effect", finding that mental health care provider capacity explains only one fifth of it. Beyond that, local attitudes toward mental health play an important role, as shown by asymmetric responses for people who move from low-to-high and high-to-low care utilization areas, especially among those who were never diagnosed with any mental illness before moving. Lastly, I find a strong negative correlation between area-level mental health care utilization and suicide rates, and suggestive evidence that moving to high utilization areas is associated with a lower risk of self-harm-related Emergency Department visits. These findings suggest that promoting mental health care could benefit the elderly population, and that there is substantial scope for achieving this goal with interventions targeting place-specific factors.
Gini in Taylor Rule: Should Fed Care about Inequality? (E5)
AbstractThis study investigates whether the Federal Reserve (Fed) should care about inequality. We develop a Heterogeneous Agent New Keynesian (HANK) model, which generates empirically realistic inequalities and business cycle properties observed in the U.S. data. Households in the model economy are subject to the aggregate productivity shock and to the idiosyncratic labor efficiency and preference shocks. In addition, the model distinguishes the extensive and intensive margins of labor supply. We consider the income Gini coefficient in a monetary policy rule to see how an inequality-targeting monetary policy might affect aggregate and disaggregate outcomes, as well as economic welfare. First, we find that a monetary policy rule with an explicit inequality target can be welfare improving, even if inequality becomes volatile. In particular, the policy reform can improve the welfare of the poorest the most. Second, there is an efficiency-equity trade-off: an economy should sacrifice a more volatile output in order to have smaller cyclical variations in its inequality measures. Lastly, when the Fed targets the employment of a specific group of households, it can improve economic welfare and stabilize inequality at the same time.
Governance Risk and the Cross-Section of Stock Returns: Do Business Cycles Help to Solve the Puzzle? (G3, G1)
AbstractThis paper investigates the asset-pricing implications of corporate governance decisions. Since Gompers, Ishii, and Metrick (2003), the question of whether governance indices are priced into stocks has been debated. From an asset pricing perspective, it remains a puzzle that firms with low indices or good governance also deliver higher risk premiums. We propose a framework with governance risk and business cycles to explain this puzzle and to provide theoretical and empirical evidence of a connection between corporate governance decisions and asset-pricing theory. Our finding is made possible by introducing business cycles and Epstein-Zin type of investors into corporate finance models with agency conflicts.
Government Credit and Monetary Policy Transmission: Evidence from Brazil (E5, G2)
AbstractWe examine the role of government credit in monetary policy transmission, using a comprehensive dataset from Brazil and an empirical specification that allows the identification of the credit supply channel. We find that (i) the interest rate for government bank loans responds more to the policy rate than that for private bank loans; (ii) the interest rate for non-earmarked private loans respond more to the policy rate than that for earmarked private loans; (iii) having a past earmarked-loan relationship between a firm and a bank makes the interest rate for private banks’ non-earmarked loans less responsive to the policy rate; (iv) the higher the firm’s exposure to the earmarked market, the interest rate for private banks’ non-earmarked loans respond more to the policy rate. These suggest that the transmission of monetary policy through private bank lending may be weaker in the presence of significant government involvement in credit markets.
Grab a Bite? Prices in the Food Away From Home Industry During the Covid19 Pandemic (E3, L1)
AbstractDue to social distancing measures, among other health related policies, the Food Away From Home (FAFH) industry has been one of the most affected by the Covid19 pandemic. As FAFH prices account for a non-negligible share of CPI baskets, this study examines FAFH price-setting behaviour in Mexico City from two complementary angles. First, using web scraped data from an online food ordering and delivery platform, classified through machine learning techniques, this study shows that in 2020 and 2021 (i) independent and multi-outlet restaurants report similar price trends; (ii) prices of soups and beverages without alcohol, potentially substituted by home-production, exhibit low price growth rates; (iii) in contrast, prices of beverages with alcohol, desserts and starters have been on the rise; (iv) the heterogeneous growth rates across dish categories seem to be explained by the extensive margin; and (v) the escalation and deescalation in the Covid infections increased and decreased price-stickiness, respectively. Second, employing shrinkage and non-linear machine learning frameworks able to deal with large number of explanatory variables, this study analyses potential price determinants. Contrary to what is expected, it seems that costs stemming from electricity, permanent workers and prices of beans and rice are more relevant than those from gas, temporal workers and meat prices as inflation determinants for this industry in Mexico City. The results from this paper adds on understanding price-setting during the fast-changing Covid19 pandemic.
Growing by Mentoring: The Effect of Mentoring Program on Mentors (I2, C9)
AbstractWe organized an online mentoring program, in which college students from an elite university provided voluntary assistance to middle-school students from a less developed area one hour a week for eight weeks. Eligible volunteer applicants were randomly divided into two groups, and an evaluation is conducted after the early group completed the service and before the late group started it. Results show that one-semester mentoring service can (1) improve students’ perceived skill of caretaking; (2) increase their appreciation of government anti-poverty effort; (3) generate more egalitarianism in the redistribution; (4) increase compromise between fairness and efficiency; and (5) form generalized in-group favoritism.
Hate in the Time of Covid: Racial Crimes against East Asians (I1, K0)
AbstractWe provide evidence of the impact of the coronavirus pandemic on racial hate crime in England and Wales. Using various data sources, including unique data collected from UK Police Forces by Freedom of Information (FOI) requests, and an event study method and regression discontinuity design we find that racial hate crime against East Asians increased by an average 50% beginning in early February through the end of 2020. This effect was greatest in weeks leading up to the first national lockdown in the U.K. with evidence that the shock was lower during the lockdown before increasing again in the summer 2020. We present evidence that hate crime increased as COVID-19 cases in China increased and following announcements from the government which may signal that China or Chinese individuals pose a public health risk to the U.K., indicating that protectionism played an important role in the observed hate crime spike.
Health Insurance Coverage Changes under Affordable Care Act among High Housing Cost Households (I1)
AbstractUnaffordable housing is a particular economic burden which forces individuals to make trade-offs between either paying the rent or spending money on fresh food, health care services and other basic needs. Because of their potential high healthcare needs which are caused by the high housing costs, obtaining insurance coverage is crucial. This paper examines the degree of improvement of coverage since the ACA went into effect among individuals living in different rent-to-income ratio households.
To estimate the impact of the ACA regarding the portion of income spent solely on housing, we use difference-in-differences and triple differences approaches to compare insurance coverage in 2010−13 with 2015−19 to determine how individuals with various levels of households’ rent to income ratios gained health insurance coverage in expansion and non-expansion states.
For individuals with incomes between 100 and 138% of FPL, adjusted uninsured rates decreased by 4.6 percentage points (pp) more in expansion states than in non-expansion states after the ACA went into effect among those whose gross rent is between 30% and 100% of their income. Similarly, relative to non-expansion states, uninsured rates decreased by 3.3 pp more in expansion states for individuals with incomes up to 400% of FPL. We also find the effects on Medicaid coverage improved by 6.4 and 7.3 pp more in expansion states for those whose gross rent is below and above 30% of their income among individuals with incomes up to 100% of FPL. There are relatively larger pre-post improvements in Medicaid between expansion and non-expansion states for individuals with income between 100 and 138% of FPL. Relative to those whose gross rent is below 30% of their income, the percentage of Medicaid coverage change in expansion states compared to non-expansion states shows to be 1.8 pp larger for individuals with gross
Healthy Homeowners, Healthy Housing Market? (R2, I1)
AbstractMedical conditions, both acute and chronic, can pose a significant financial burden for many US households. Furthermore, unpaid medical debt can further impair household financial positions by impairing personal credit scores and access to future credit. We investigate the relationship between health and housing market outcomes. Using data from the CDC Places: Local Data for Better Health, which contains information at a zip code level about the incidence of a variety of health issues at the zip code level. We match this data to zip code level housing market outcomes. Our initial results show that areas with higher levels of disease, particularly chronic disease, experience lower house price growth and have lower access to mortgage credit. This research has important policy implications, given that major US credit bureaus announced in early 2022 their intent to remove most medical debt from credit reports.
High Stakes Evaluation Ratings and Principal Transitions (I2, J4)
AbstractWith the expansion of school accountability, performance-based incentives and high stakes evaluations, there is growing interest in understanding how these affect school performance and leadership composition. This paper studies a reform in a large urban public school district that ties principal compensation to their performance. Using a regression discontinuity design, I investigate the effect of performance rating on principal transitions and future school leadership performance for equally productive principals. Results indicate that principals with evaluation scores that place them just below the higher rating cutoff are more likely to turnover, exit the district and are less likely to be promoted. Most of the differential turnover across the rating cutoffs is driven by Middle and High school principals. I also find some evidence of future changes in leadership performance at the higher rating thresholds. Principal transitions across the rating cutoffs may be driven by the label effect (motivation and district decisions) or the salary effect. The next step is to isolate the effect of salary on transitions using a difference-in-discontinuity design.
Hiring with Algorithmic Fairness Constraints: Theory and Empirics (J7, C6)
AbstractWe present, solve, and empirically test a 2-stage hiring model to understand the interplay between algorithmic fairness constraints and human decisions in the hiring context.
In the first stage, a screening algorithm screens n_a candidates, estimates a noisy signal of the quality of each candidate, and shortlists the top n_s candidates.
Because the data used to train the screening algorithm may be biased, the algorithm has a gender-parity constraint such that it shortlists an equal number of men and women.
In the second stage, an unbiased human hiring manager interviews the shortlisted candidates, estimates the quality based on her own assessment, and hires the best candidate.
We solve this model analytically and identify 2 key parameters that affect the effectiveness of the fairness constraint (i.e. gender proportion of the hires): (1) size of the candidate pool, (2) correlation between the algorithm's and the hiring manager's assessment scores.
We show that parity constraint in the shortlist does not necessarily lead to gender parity in hires.
The effectiveness of the fairness constraint increases with the size of the candidate pool, and decreases with the correlation between the algorithm's and the hiring manager's assessment scores.
To test the empirical validity of this model, we estimate the correlation parameter using real-world hiring data from IT firms.
First, we use a deep-learning model to train a resume screening algorithm on 530k resumes to predict whether a candidate passed screening (AUC=0.91).
Similarly, we train a hiring manager model to predict whether a candidate passed the interview (AUC=0.75).
We use the predictions from these models on a hold-out test set to get estimates of screening and hiring manager scores for each candidate.
We find: (1) high correlation between screening and interview scores (between 0.4-0.7) for most jobs, (2) a high level of heterogeneity in correlation across jobs.
Household Debt Overhang and Human Capital Investment (G5, D1)
AbstractUnlike labor income, human capital is inseparable from individuals and does not accrue to creditors at default. As a consequence, human capital investment should be more resilient to ``debt overhang'' than labor supply. We develop a dynamic model displaying this important difference.
We find that while both labor supply and human capital investment are hump-shaped in leverage, human capital investment tails off more slowly as leverage builds up. This is especially the case when human capital depreciation rates are lower. We provide empirical support for the model using individually identifiable data. To strengthen the causal link between leverage and human capital investment, we perform an instrumental variable analysis based on plausibly exogenous variation on individuals' mortgage-loan-to-value ratio due to the dynamics of housing market conditions.
How Do Physicians Respond to New Medical Research? Evidence from an Influential, but Subsequently Overturned Clinical Trial (I1, I0)
AbstractWhat happens when the findings of a prominent medical study are overturned? Using a medical trial on breech births, we show that the reversal of such a medical study had an arguably large immediate and lasting impact on physician choices that might lead to longer-run health issues. In particular, we find that the reversal of a multi-site high profile randomized control trial on the appropriate delivery of term breech births, the Term Breech Trial (TBT), led to a twenty percent decline in C-sections for such births at a time when the overall trend in C-sections was rising. However, we do not find that such a change in practice had significant impacts on infant health. Interestingly, we find that the initial publication of the TBT didn’t have as strong of an impact as its reversal, potentially because it recommended the use of a procedure that was already broadly implemented by medical professionals. To our knowledge, this is the first paper to examine how the reversal of the findings of a prominent medical study impacted procedural choice and related outcomes.
How Effective Are Work Search Requirements? Evidence from a UK Reform Targeting Single Parents (H3, J2)
AbstractIn the last few decades, an increasing number of OECD countries have strengthened the work search or 'conditionality' requirements to be eligible for out-of-work benefits. While a large literature examines the aggregate employment and benefit caseload impacts of such policies, much less is known about the wider effects which determine the consequences for welfare. In this paper, we exploit the staggered roll-out of a policy that introduced such requirements for UK single parents. We study the impact on the kinds of jobs that the reform induces in terms of hours, pay, and tax liabilities. By also estimating the effect on substitution to alternative benefits, we are able to recover the full fiscal impact of the policy. In addition, we analyse conditionality's effects on mental health and reported life satisfaction. We find that the introduction of conditionality induced a significant decrease in the share of single parents claiming the benefit (7.5ppts, from a baseline of 34%), around 4ppts of which is explained by higher employment, and 3.5ppts by substitution to (more costly) incapacity and disability benefits. But the new jobs single parents took were entirely part-time and largely low paid, meaning that their tax liabilities were low and benefit entitlement remained high. The result is that the fiscal impact of the policy was approximately zero. While this would appear to make the reform a worsening in welfare terms (by removing a choice for lone parents with no positive fiscal externalities), we find no evidence of a decline in mental health or reported life satisfaction. These results suggest that strong effects of conditionality requirements on employment and caseload may mask a considerably more mixed picture when a more comprehensive set of outcomes is considered.
How Increased Labor Demand at the Start of Your Career Can Improve Long Run Outcomes (J2, E3)
AbstractThe literature has traditionally focused on the local unemployment rate one faces at the beginning of their career to measure how initial economic conditions affect long-run outcomes. However, the unemployment rate moves in response to changes in labor supply or labor demand. Using JOLTS State Estimates for job openings, hires, and separations along with Local Area Unemployment Statistics, I test how changes in more direct measures of demand at labor market entry affect long run outcomes. I find that for every one point increase in the local unemployed-to-job-opening ratio, annual earnings are reduced by 4.53% and remain depressed for 13 years. Conversely, I find that a one percentage point increase in the local job openings rate or the local quits rate, increases initial annual earnings by 8.15% and 14.23%, respectively.
How to Make People Work Without Direct Supervision: A Network Bargaining Game of Collusion (D7, D8)
AbstractIs it possible to make the workers exert costly efforts when they can collude, and the managers cannot observe their efforts or the outcomes? The answer is yes if the principal can limit the initial communication network among the agents. This paper provides a method to stop workers from deviating to corruption and maintain a stable effort level using peer supervision. I model two methods for joint deviation: deviation by voting and deviation by commitment. I find that when all agents are fully connected in a communication network, deviation by voting can be stopped if and only if the threshold of passing the vote is sufficiently high. Deviation by commitment, however, cannot be stopped. Since any joint deviation has to be mediated through communication, when the agents are less connected, they have less power in coordinating a joint deviation. I show that if the principal can limit the communication among the agents, then it is possible to deter joint deviation in both cases with less demanding conditions. The principal needs at least three agents to deter voting and six agents to deter commitment. The findings are useful in anti-corruption, institution design, and internal controls.
How Would European Firms Fund Themselves in the United States? (E6, F3)
AbstractI analyze the cross-sectional dimensions of firm debt choice (bank loans vs. bonds) in the Eurozone in comparison to the United States. For this, I compile a unique, extensive data set that covers around 60% of aggregate employment and revenues in the Eurozone. Based on the patterns observed in this data I show that firm size and collateral availability are significant predictors of a firm's debt choice. I introduce a theoretical model from the literature to estimate the extent to which the aggregate debt choice is driven by similar firms using different funding sources in the two areas, as opposed to being driven by fundamentally different firms operating in those regions. Based on these estimates, I present counterfactual scenarios for the aggregate funding choice of European firms if they were relocated to the United States. I find that the use of bond funding among them would remain significantly lower due to different firm fundamentals.
To my knowledge, this is the first paper to present counterfactual debt scenarios for the Euroarea. I thereby contribute to two strands of the literature: first, to the macro-financial literature on debt markets and, second, to the empirical literature on debt choices among heterogeneous firms. My results are also relevant for the policy discussion surrounding the European Union's capital markets union project. In related studies, the Eurozone is frequently compared to the United States as a more integrated market union with a more advanced financial market. I highlight in this context that the difference in aggregate funding choices between the United States and Europe and its negative implications for systemic risk are not exclusively a result of the different financial systems but are in part also caused by structurally different firm characteristics. My results thus provide an insight into the required nature of potential, targeted policies.
Hysteresis, the Big Push, and Technological Adoption (E3, O4)
AbstractI develop a supply-side theory of hysteresis—a phenomenon in which temporary shocks have long-run effects on output—based on coordination failures in the adoption and abandonment of a technology. Whenever the entry threshold exceeds the exit threshold for a sufficient mass of firms, hysteresis may appear in the aggregate. In sum, hysteresis is found to be non-linear on the size of the shock, and its emergence depends on the elasticity of substitution between input varieties, and barriers to adoption. This work bridges the gap between the macroeconomic business cycle persistence literature and the Big Push in development economics, by linking poverty traps to reverse hysteresis; i.e., a situation in which a positive shock might result in positive long-run effects. The model provides an alternative explanation for reverse hysteresis and the plucking model of business cycles that, unlike previous work, does not require downward nominal wage rigidities. Additionally, it defines a set of conditions for the possibility of reverse hysteresis, which could potentially inform how policy could counteract any long-term scarring in the productive potential of economies.
Identifying Multiple Shocks in Heteroskedastic Proxy Vector Autoregressions (C3, E3)
AbstractA test for time-varying impulse responses in structural vector autoregressions
is proposed when the shocks are identified by external proxy variables
which do not uniquely identify the shocks individually but need further restrictions
for full identification. The asymptotic analysis is supported by small sample
simulations which show good properties of the test. An investigation of the impact
of productivity shocks on a small macroeconomic model for the U.S. illustrates the
importance of the issue for empirical work.
Impact of the Opioid Crisis on the Elderly and Children (I1, I3)
AbstractThis paper examines the interdependence between individuals directly affected by the ongoing Opioid Crisis in the US and their parents and children. Social and financial interactions between individuals and their parents depend on both children’s and parent’s physical and financial status. A drug addiction can severely impair an individual’s financial and physical health. This in turn could increase dependence on their parents even in adult life. The age group most affected by the Opioid Crisis is ages 25-54. Consequently, their parents would belong to the age group between 55-75. This makes seniors a high-risk category for physical and financial stress due to the crisis.
Similarly, there is a significant impact of one or both parent’s drug addiction on their young children. It contributes to a child’s quality of care, possible removal from a birth parent and subsequent transfer into foster care. This makes young children a vulnerable group affected indirectly by the crisis.
I use restricted data from the Health and Retirement Study (HRS) to examine various channels such as the mental stress of having an addicted child, the physical strain of looking after grandkids for long hours, or the financial burden of having to support a child who lost employment due to addiction or having to pay for his/her rehabilitation and so on. I use publicly available data from AFCARS to study children’s foster care related outcomes with respect to opioid addiction.
The primary methodology is a difference-in-differences exploiting state level variation in drug overdose deaths as an indicator for the differential impact of the crisis.
The important policy implications of this study lie in identifying dependence that may be mentally or financially debilitating to the elderly or young children and seeking alternative ways to supplement and support the families of opioid addicts.
Impacts of Childhood Disability on Family: Labor, Marriage, Fertility, and Depression (J2, I0)
AbstractYoung children's long-term care (LTC) needs greatly impact their family members. Compared to the LTC needs of the elderly, children's LTC needs typically last longer, come as unexpected shocks, and arrive at a critical time in caregivers' life choices such as fertility and career development. Despite its significance, LTC literature focuses mostly on labor outcomes and the elderly primarily due to data limitations.
We fill this gap in the literature by studying one of the most prevalent (1 in 345 births) causes of children's LTC needs -- cerebral palsy (CP). We move beyond labor market outcomes and provide novel estimates on the causal effects of having a CP child on family members' marital status, fertility, health expenditure, depression diagnosis, and siblings' education outcomes. Moreover, we examine the role of social welfare policies and how they insure families against LTC risks.
Using the administrative record from the National Health Insurance Research Database from Taiwan, we examine more than 15,000 children diagnosed with CP and their families from 2000 to 2018. We exploit CP's unexpected nature and take an event study approach. We find that having a CP child with a disability status ends mothers' careers and reshapes the family structure. It decreases mothers' probability of work by 10 pp and income by 5%. It increases divorce by 2.5 pp and enrollment in low-income programs by 10 pp. Impacts on fathers' labor market outcomes are much smaller. We also find a precise null result in fertility outcomes. Current social welfare programs provide only minimal insurance to these families. Families with CP children enrolling in disability programs still suffer from significant impacts. Reforms to caregiver hiring, disability, and low-income programs further provide casual estimates of their values for CP families.
Impacts of Double-Fortified Salt on Anemia and Cognition: Four-Year Follow-up Evidence from a School-Based Nutrition Intervention in India (I1, O1)
AbstractLong-term follow-up of early childhood health interventions is important for human capital accumulation. We provide experimental evidence on child health and human capital outcomes from the longer-term follow-up of a school-based nutrition intervention in India. Using panel data, we examine the effectiveness of the use of iron and iodine fortified salt in school lunches to reduce anemia among school children. After four years of treatment, treated children, on average, have higher hemoglobin levels and a lower likelihood of anemia relative to the control group. Interestingly, the intervention did not have any impact on cognitive and educational outcomes.
In Utero Exposure to Radiation Fear and Birth Outcomes: Evidence from Fukushima Nuclear Power Plant Accident (I1, I3)
AbstractWe study the effects of exposure to radiation fear in utero on health at birth and five years later, using universal birth records linked to Censuses in Japan. We are the first to assess maternal stress due to a risk factor that is intangible and uncertain—radiation exposure. We leverage the Fukushima nuclear power plant accident in 2011 and create a quasi-experimental setting by focusing on children whose expected birthdates were within a 280-day window to the accident. We find radiation fear in utero causes a 30-gram decrease in birth weight, with a 19% and 38% increase in the risks of low birth weight and preterm delivery, respectively. The effects are more pronounced if expectant mothers are less educated or have toddlers, and if expectant fathers work in agriculture, indicating that the ability to collect information is important for stress alleviation and that food contamination is the main source of radiation fear. The findings have immediate implications for offspring health loss owing to maternal stress from intangible risk factors, such as infectious viruses.
Income Inequality and COVID-19 Mortality: the Early Evidence (J1, I1)
AbstractEmerging studies suggest that existing health disparities and variations in COVID-19 induced mortality around the world may be shaped by income inequality (Davies 2021, Wildman 2021, Sepulveda and Brooker 2021). Income inequality – health connection in the context of COVID-19 pandemic is still at a fairly early stage of scholarship, therefore, understanding of the complex nature and multivariate dynamics of the pandemic is very timely. Building on existing evidence of association between income inequality and COVID-19 mortality, this paper aims to examine the role of income inequality as a possible driver of health inequities during the pandemic and to identify the possible pathways that link income inequality with COVID-related mortality across nations. Using data from the World Mortality Dataset (Karlinsky and Kobak, 2021), the Standardized World Income Inequality Database (Solt, 2020) combined with several other sources for the period from March 2020 through January 2021, I empirically evaluate whether income inequality measured by Gini coefficient is associated with the number of excess-deaths as the most comprehensive measure of the human cost of pandemic. Preliminary results from multivariate regressions show positive and statistically significant associations between the country Gini index and logtransformed excess deaths across various model specifications. The results are robust to adjustment for the set of potential confounders including demographic, regional, economic, government and health controls. The results also suggest that income inequality might manifest itself through unequal access to public services including healthcare, thus supporting neo-materialist approach toward the mechanisms behind income inequality – health link.
Individual Experience and Expectations: Evidence from Housing Markets (D1, R2)
AbstractHeterogeneity in market experience plays a key role in explaining heterogeneity in expectations. We examine the role of idiosyncratic variation in experience. Combining household survey data, administrative data and housing transaction data in the Netherlands from 1996 to 2017, we find that households’ expectations about future price growth on an aggregate level may vary with their idiosyncratic, or individual, experiences. Our results are robust to alternative measures for individual experience, and to specifications that further reduces the issue of unobserved heterogeneity across households. When comparing the effects of experiences obtained from different sources and time horizon on expectations, we find that the effect of individual experience is greater than that of market experience, and the effect of recent individual experience is greater than that of distant individual experience. We further interpret these findings by referring to the role that information plays in expectation formation and potential underlying behavioral mechanisms.
Insurance Quality under Managed Competition: Evidence from Star Hospital Coverage in Medicare Advantage (I1, H4)
AbstractAccess to an academic medical center or “star" hospital is highly valued by nearby residents. However, in many insurance markets, star hospitals can be largely excluded from networks or covered only with high cost sharing. This paper analyzes why this is the case. I gather inclusion data for 71 star hospitals in Medicare Advantage. Star hospital coverage in Medicare Advantage is high --- 70% of plans cover their county's star hospital compared to 34% in ACA exchanges. I examine the degree to which Medicare Advantage's high coverage is the result of two unique institutional features. First, I study how inclusion is impacted by the presence a powerful public option, Traditional Medicare, in hospital price negotiations. Using the Massachusetts APCD, I show that Medicare Advantage plans pay rates close to Traditional Medicare rates, even at star hospitals, leading to much lower star hospital prices in Medicare Advantage. Using this model, I show that if insurers paid star hospitals relative prices similar to rates in commercial markets, coverage of star hospitals would drop by a third. Next, I examine the role of risk adjustment. Using descriptive exercises as well as estimates from my demand and cost models, I show that individuals with a preference for star hospital coverage tend to have higher expected un-adjusted healthcare expenditures. Using my structural model on network inclusion, I show that reductions in the fit of Medicare Advantage's risk adjustment system would dramatically reduce coverage of star hospitals.
Interest Rates, Competition, and Complexity: Demand and Supply of Retail Financial Products (G5, L1)
AbstractWe study the post-Great Recession market for retail investment products. With an experiment, we show that low interest rates drive investment demand but not product differentiation. Elicited margins go hand in hand with investors’ (under)estimation of complex risk exposures. We empirically document that (i) rising complexity follows market growth, (ii) issuer margins increase in complexity, and (iii) simpler products first-order dominate more-complex products. Furthermore, biased dependency perceptions predict margins in the cross-section. Banks employ strategic price complexity to mitigate competitive pressure, consistent with limited buy-side learning and growing sell-side competition. Our findings showcase how monetary policy can fuel excessive risk-taking.
Is the European Union More Unequal than the Habsburg Empire? Examining Regional Inequalities in Habsburg Regions from 1870 to 2018 (O4, N1)
AbstractThis study examines regional growth and convergence among Central and Eastern European (CEE) regions from the 19th century to today by using Geographic Information Systems (GIS) software to recreate historical regions in present-day projections and develop a new dataset of long-run growth in the CEE region. The findings suggests that regional disparities are markedly higher today than in the 19th century, despite rapid convergence in the past two decades. The study thus suggests that although regional EU policy has been successful over the past two decades, further policy measures are needed to make up lost ground. For the 1867-1913 time-frame, the study finds two regional convergence clubs. Over the entire 1870-2018 period under review, the study finds no evidence of convergence.
Is women’s competitiveness expressed vicariously through their husband’s income? (J0, D1)
AbstractPrior research on the contribution of competitiveness on the gender income gap has focused on the effect of individual competitiveness. We investigate the influence of heterosexual individuals’ own and cohabiting partner’s competitiveness on their own and partner’s income using a recently validated self-reported measure of competitiveness incorporated in 2017 into a Dutch representative survey. The present (2017) and future (after 2017) income levels of single and cohabiting men and single women are positively associated with their own competitiveness. Single men’s are not. Consistent with competitive women selecting high-income men, women’s competitiveness is positively associated with their male partner’s income, but the men’s competitiveness is not significantly associated with their female partner’s income. However, controlling for 2017 income as a proxy for unobserved factors across individuals and couples, only single men’s competitiveness increases their future income. Single women’s and cohabiting men’s and women’s do not. Remarkably, only men’s female partner’s competitiveness, not their own, causally increases the cohabiting men’s future income. Neither cohabiting women’s own nor their male partner’s competitiveness increases the women’s income. In short, our evidence suggests that whereas men’s competitiveness increases their income only as singles, women’s competitiveness never increases their income. However, they do match with higher potential income men as spouses and motivate these men to earn a higher income, increasing the income of their household.
It Is Not Only up to You! The Effect of Retirement on Healthcare Utilization-the Role of Physician Incentives (I1)
AbstractWe examine the effect of retirement on healthcare utilization using a unique administrative data set from a tertiary hospital in southeast China, with a particular focus on the role of physician incentives. We use a fuzzy regression discontinuity design that exploits the discontinuity in retirement rates at statutory retirement ages. We find that retirement has a significantly positive impact on outpatient care expenditures, and the effect is more pronounced when the physician has more incentives. We also examine the heterogeneous effects of other physicians’ characteristics. For example, young and male physicians are more likely to exaggerate the retirement effect on health care utilization.
Job Match Quality of College Graduates: The Effect of Major-Specific Market Size (J2)
AbstractResearch has shown that college graduates enjoy higher quality job matches
in larger cities in the sense that they are more likely to work in an occupation
that (1) requires a college degree ("degree match") and (2) is closely related to
their fields of specialization ("major match"). One possible explanation is that the
extent of the market -- the number of jobs in a city in occupations related to the
field of specialization -- is greater in larger cities. Using data from the American
Community Survey, I calculate the extent of the market for college graduates in the
city where they live, which varies both across cities and within cities, among
individuals with different fields of specialization. I estimate linear probability
models for match quality as a function of the extent of the market and city
Although the extent of the market is strongly positively correlated with city size,
the magnitude of this relationship varies considerably across college majors.
Elasticity ranges from a low of 0.5 for Environment and Natural Resources majors
to 1.27 for Law majors. As a result, there is ample independent variation between
the extent of the market and population to distinguish their effects empirically.
I find that the effects of market extent on the probability of a major match are
heterogeneous -- stronger for men than for women and strong for younger men and
women. This is consistent with the notion that career concerns are more powerful
for men than for women on average -- especially married women.
I also find age effects that the probability of a match is declining in age.
This is consistent with the notion that job promotion trajectories often take
workers away from their initial specialization.
Labor Exposure to Climate Change and Capital Deepening (M0, J0)
AbstractRising temperatures induced by climate change pose severe threats to the health of outdoor workers. The threats generate two types of climate risks that increase labor costs of firms relying on outdoor workers: 1) physical risk - lower labor productivity in high temperatures; and 2) regulatory risk - the possibility of introducing regulations to protect workers against heat-related illness. Building on each occupation’s exposure to weather, I show that firms adapt to the labor-channel exposure to climate change by adjusting their production functions towards higher capital-labor ratios - an increase in capital expenditure and a decrease in employment. The two risks also incentivize firms to increase innovation, especially innovations that facilitate automation and reduce labor costs. Further tests show that the results are stronger when a firm’s managers have strong beliefs in climate change or when a firm’s jobs can be easily automated, but weaker when a firm’s managers lean towards the Republican party or when employees are protected by the labor union.
Labor Market Institutions, Fiscal Multipliers, and Macroeconomic Volatility (E6, J3)
AbstractWe study empirically how various labor market institutions – (i) union density, (ii) unemployment
benefit remuneration, and (iii) employment protection – shape fiscal multipliers and macroeconomic
volatility. Our theoretical model highlights that more stringent labor market institutions attenuate both
fiscal spending multipliers and macroeconomic volatility. This is validated empirically by an interacted
panel vector autoregressive model estimated for 16 OECD countries. The strongest effects emanate
from employment protection, followed by union density. While some labor market institutions mitigate
the size or frequency of exogenous shocks, they, however, reinforce their propagation mechanism. The
main policy implication is that stringent labor market institutions render cyclical fiscal policies less
relevant for macroeconomic stabilization.
Labor Market Selection and the Dynamics of a Recovery (J6, E2)
AbstractI propose a modeling framework to resolve the puzzle of slow and near-linear recoveries. A key feature of this model is a novel two-sided ranked many-to-many matching mechanism in an otherwise standard framework. Early in the recovery, composition effects and separations depress job creation incentives and therefore job finding rates for all workers. However, under slack markets the drop in job finding rates becomes much stronger for workers in unemployment, as unemployed workers consistently get outranked by their employed peers when markets are slack. This effect leads to slow recoveries that mirror exactly the near-linear shape that is observed in the data. The model is able to match the paths of all transition rates among unemployment, non-participation and employment for the last 5 US recoveries closely, producing a realistic evolution of the unemployment rate and reproducing the shape of these recoveries. The model also provides a theoretical explanation for the observed higher success rate of employed job seekers relative to those in unemployment.
Labor Supply Response to the Elimination of the EITC for Undocumented Immigrants: Evidence from the 1996 Welfare Reform (J3, H2)
AbstractThe paper examines the effects of ending the Earned Income Tax Credit (EITC) for undocumented immigrants in the United States. The Welfare Reform Act of 1996 required that tax filers have a valid Social Security number to claim the EITC, thus disqualifying undocumented immigrants from receiving the EITC benefits. Using the SIPP data, I find that the policy led to a 23 percent decrease in the likelihood of Hispanic non-citizens receiving the EITC. Exploiting variation in the share of undocumented immigrants across counties reveals that counties with higher Hispanic non-citizens shares experienced a relatively large decrease in EITC per person after the year 1996 when SSN was made compulsory to claim the EITC. The difference-in-difference and a triple-difference that exploits variation across citizenship status and the number of children of single mothers show that ending the EITC sharply reduced the labor force participation of undocumented single mothers but did not affect undocumented immigrants in general. The results suggest that ending the EITC for undocumented immigrants makes the American children of single-mother undocumented immigrants doubly disadvantaged because they not only lack a tax credit but their mothers’ labor supply is also reduced as a result.
Leaning Against the Data: Policymaker Communications under State-Based Forward Guidance (E5, G1)
AbstractThe state-based forward guidance on the interest rate is often touted as allowing the private sector to adjust the expected policy stance automatically in response to economic news. However, implicit in this view is the assumption that the policymakers can commit to their promised policy actions tied with economic conditions. If policymakers change the way that they discuss
the economic outlook after introducing the state-based forward guidance and the market participants adjust their behavior following policymakers' communications on the economic outlook, the financial market will respond to economic news in a way that is not fully compatible with the original intention of the forward guidance. We use the forward guidance based on inflation and unemployment thresholds introduced in the December 2012 FOMC meeting as a testing
example by identifying the policymakers' speech tone about the outlook using text-analysis. Our results show that policymakers downplayed the significance of the fast decline in the unemployment rate when it approached the threshold earlier than expected. In addition, financial markets became more sensitive to the policy shock component of economic news rather than the growth shock component. Our findings highlight challenges in designing the effective state-
based forward guidance and suggest a caution in providing "too precise information" in forward guidance.
Lease-Adjusted Productivity Measurement (O4, E2)
AbstractWe document that leased capital accounts for more than 30% of the total productive physical assets used by US public firms. We develop an analytical framework to show how ignoring this important proportion generates an overestimation of productivity. Such overestimation can be decomposed into two distinct channels: one is from the mismeasured factor share, which we denote as the factor-share channel; and the other originates from the intensity of leased capital utilized by a firm, which we denote as the omitted-leased-capital channel. Empirically, we find that the overestimation in aggregate productivity is large, growing over time, and exhibits strong countercyclicality. In the cross-section, the overestimation presents asymmetric patterns for firms with different characteristics (i.e., leased capital ratio, size, financial-constraint level, and age). We also present strong empirical support for our decomposition of the overestimation. Our results emphasize the importance of explicitly adjusting for the “unmeasured” leased capital in studies related to firm productivity measurements, particularly those in the cross-section and in the time series.
Leverage and Rate of Return Heterogeneity among U.S. Households (D3, G5)
AbstractThis paper proposes measures of panel data for returns to U.S. households wealth and documents new facts on the heterogeneity in returns to wealth between households and across the wealth distribution. First, household leverage exhibits permanent heterogeneity and explains most of the permanent heterogeneity in the returns to wealth. As such, the permanent heterogeneity in returns to assets understates the permanent heterogeneity in returns to wealth, with standard deviations of 3.8 and 9.2 percentage points, respectively. Second, returns to wealth decline as households become wealthier and exhibit declining returns to scale and specialization. Third, household-specific returns to wealth and assets are correlated with the persistent component of labor earnings. Fourth, housing and the regressivity of after-tax mortgage rates are critical to explaining the permanent heterogeneity in returns. These findings inform the nature of scale and type dependence of returns heterogeneity for the study of portfolio allocation, wealth inequality, social mobility, and corresponding policies.
Liquidity Premium, Tariffs and Currency Internationalization (E6, O2)
AbstractThis paper develops a two-country, two-goods model to investigate the implications of international currency liquidity premium and trade tariffs on local currency internationalization. Results show that during the beginning of currency internationalization, the local currency should be overvalued against its balanced level but meanwhile, the misalignment rate does not exceed a certain level. After that, the local currency must be devalued in line with its increasing international liquidity. In particular, if factors are not perfectly mobile across the border, the final terms of trade are not only dependent on the long-run exchange rate levels but also affected by the exchange rate adjustment trajectories.
Long-Run and Heterogeneous Effects of Maternity Leave Expansions (J0)
AbstractUsing an administrative data set on pension claims, we analyze the impact of seven parental leave reforms in Germany on mothers’ earnings trajectories for up to 20 years after childbirth. The reforms are rolled out from the late 1970s to early 90s and each gradually expand the duration and financial support of the parental leave programs. The monthly data allow us to identify transparently causal effects of the reforms from the short to the long term using both, event-study and regression discontinuity methods.
We make two contributions to the existing literature. First, we take the long-run perspective, answering how family policies affect the long-term gender inequalities, decades after childbirth. Second, we explicitly assess the selection patterns that the reforms induce. Because with each additional reform, mothers with different childcare and labor market preferences leave the labor market longer, we can assess potential effects of a further extension of parental leave more credibly.
We find that each reform considerably affected the short-term employment rate of mothers. Moreover, we find that early reforms that extend maternity leave shortly after childbirth persistently reduce the earnings trajectory of mothers. Our results show that women who comply with the reforms have above average pre-birth earnings and experience. As these women may better compensate for lost labor market experience, this may explain our lacking long-term effects for those reforms.
Loss Aversion and Focal Point Bias: Empirical Evidence (D9, R3)
AbstractWhile focal point bias and loss aversion are associated with different families of behaviors, bounded rationality and prospect theory, recent research suggests a link between the two because both biases involve anchoring either to a previous price for loss aversion or to salient left digits or round numbers. First, we use data from an earlier consumption experiment that 1. measured loss aversion parameter by presenting the subjects with a series of lotteries, and 2. surveyed the subjects concerning how much they spent for lunch. We find that borrowers who reported integers amounts score 24% of a standard deviation higher on their loss aversion parameter estimate than those who do not. Second, we estimate traditional reduced form models of the effects of loss aversion on housing sales prices in repeat sales housing transactions, using a sample of buyers who did not appear credit constrained at purchase in the first transaction. We extend those models by comparing the effect of expected losses on sales price in the second sale for purchasers in the first sale who selected a round mortgage amount to those who selected a nearby continuous mortgage amount using a regression discontinuity design. We first find that having an initial mortgage that is a round number is persistent raising the likelihood that the next home purchase mortgage is round by 11 percentage points relative to a mean of 16%. We then show that sellers who initially selected a round number mortgage capture 12% more of the expected loss than sellers from the continuous mortgage amount subsample, relative to a base effect of 10% for the continuous amount subsample.
Low Interest Rates and the Distribution of Household Debt (E4, D1)
AbstractWe study how changes in interest rates affect the borrowing of households and the distribution of debt within the population. In a model of household borrowing with credit constraints and endogenous house prices, we show that less constrained households with more pre-existing housing wealth increase their borrowing most when interest rates fall. We then use unique loan level data on the universe of household credit in Belgium to document a shift in the distribution of debt over age, with older households borrowing more as interest rates fell in the last decade. First-time borrowers, who are more likely to be constrained, do not contribute to the rise in household debt. To identify the elasticity of household debt to the interest rate, we use regulatory data on foreign exposures of banks and on the location of bank branches. We find that a 1 percentage point fall in the interest rate is associated with a 15% growth in household debt.
Macroeconomic Effects of Government Spending Shocks: New Narrative Evidence from the United Kingdom (E6, H3)
AbstractThis paper examines the macroeconomic effects of government spending shocks in the
UK for the period of 1955 - 2003. We construct a novel measure of news about exogenous government spending changes identified through the narrative approach. We use government documents, mostly the budget speech, to identify the size, timing, and principal motivation for all planned major federal government spending changes. We study the motivation of each announced spending change to classify them as endogenous or exogenous. Endogenous spending changes are taken in response to current happenings of the economy whereas exogenous changes are uncorrelated to the current state of the economy. We construct a news variable about upcoming spending changes classified as exogenous. We find that the implied government spending multiplier for the UK using the news variable is between 1.46 and 1.93. We also find large and significant increase in consumption following a shock to the news variable.
Macroprudential Reciprocity in the Euro Area (E5, G2)
AbstractIn this paper we provide an analytical framework to study coordination and reciprocity on different macroprudential instruments in the Euro area. We consider both CCyB and borrower-based measures. We use a DSGE model with housing, and two types of agents; entrepreneurs and lenders. Entrepreneurs can borrow from domestic and foreign lenders and face collateral constraints when doing so. We incorporate a banking sector to account for capital requirements as a macroprudential tool. Within this setting, we compare the dynamics of the model for the cases of reciprocity and non-reciprocity in borrower-based measures (voluntary reciprocity). Then, for given reciprocity in the CCyB, we analyze the welfare gain of reciprocating the LTV. Finally, we compute the optimal parameters in both rules to assess to which extent reciprocating is welfare enhancing (both for the CCyB and the LTV). We find that, especially for borrower-based measures, voluntary reciprocity is welfare enhancing and it contributes to global financial stability without compromising macroeconomic stability.
Maintaining the Secondary Prevention against Breast Cancer for Women in Emilia-Romagna (IT) and External Social and Economic Shocks (2002-2016) (I1, K3)
AbstractI presented studies on the effects on secondary prevention against breast cancer for women in the city of Bologna after the 2010 changes in the offering of breast cancer early detection services within the National Health Service at the AEA Annual Meetings in the last years.
The results showed that starting from these changes a significant part of the women in this city, who previously received mammogram for the early detection of breast cancer, gradually moved away from mammogram (at least received within the National Health Service) and, in particular, this happened for women who previously accessed mammogram via spontaneous access with the prescription of the general practitioner (Gatti S., AEA Poster 2022).
One of the essential elements to achieve a high level of effectiveness of early detection of breast cancer lies in maintaining the recommended regular intervals for mammograms and other diagnostic investigations, and is therefore centered on the perseverance of women and their full participation in the interventions of early detection.
Now, in this paper I present on the basis of the same data provided to me (as a citizen in Generalized Civic Access) by the Regional Health Service (about 5,000,000 bilateral mammographic service records) and for the whole Emilia-Romagna region an analysis of the effects of social and economic shocks, external to the National or Regional Health Service or Local Health Authority, on the perseverance of women in Emilia-Romagna in maintaining the secondary prevention path against breast cancer. The time period of the study is that between 2002 and 2016.
Indicators of measurement of these effects on the secondary prevention against breast cancer for women linked to the most significant social and economic shocks in Italy and Emilia-Romagna in those years will be presented. The shocks that essentially undermine certainties and points of reference are considered.
Market Efficiency In A Zero-Price Environment: Renwable Energy and Electricity Markets (Q4, L1)
AbstractBeyond a certain quantity, renewable intermittent electricity generation creates a fundamental problem for every electricity market ever designed and currently in operation. All electricity markets in operation are based on the assumption that marginal costs increase at some point over the relevant range of production. This assumption is not true for renewable intermittent resources and this has significant effects. All electricity markets in the United States (and elsewhere) are based on the knowledge that electricity as a commodity cannot, nor should not, be separated from the path it flows from source to sink, i.e., electricity and transmission are entirely interdependent. Accordingly, what we call an "electricity market" is not, in fact a market for electricity at all. Rather every "electricity market" is actually a market for short-term transmission capacity. That is, the current electricity markets are not allocating electricity but rather (very) short-term transmission capacity. The allocation process for transmission capacity relies on, among other things, generator offers to find a least cost solution to the power balance equation. However, when all offer prices are zero, under the current market design, there is no market-based methodology for allocating short-term transmission capacity. In these situations the dispatcher must step in and provide an administered solution. We evaluate the efficiency of these solutions and provide recommendations for improvement.
Marriage Market Signaling and Women’s Occupation Choice (J1, J2)
AbstractDespite the general closure of gender disparities in the labor market over the past half century, occupational segregation has been stubbornly persistent. I develop a new model that explains these occupational outcomes through marriage market signaling. Vertically differentiated men have preference over women’s unobservable caregiving ability. Heterogenous women choose caregiving occupations to signal their ability to be caregivers. My model generates unique predictions on the influence of marriage market conditions on women’s occupational choices. I find empirical support for these predictions using longitudinal data on marriage rates, policy shocks to divorce laws, and shocks to the marriage market sex ratio driven by waves of immigration.
Measuring Dietary Diversity with High Frequency Mobile Phone Interviews in Ethiopia (I1, Q1)
AbstractWith a long reference period, researchers can measure diet diversity more accurately, especially for households that consume some items infrequently. Longer recall periods, however, are associated with a higher cognitive burden on respondents and higher recall errors. To address the fundamental trade-off between cognitive burden and recall period, we conducted an experiment to validate a novel method of constructing dietary diversity measures using high-frequency phone surveys in Ethiopia. Our novel survey method extends respondents’ reference period without exacerbating cognitive burden associated with a longer recall period. This frequent and bounded recall survey method collects diet data through phone calls twice a day over a 7-day window, with each call corresponding to a short, bounded recall period. We randomly assigned households to report their diets through either a traditional survey - single interview (SI) recalling the respondent woman’s 24-hour diet and the household’s 7-day diet - or through the frequent and bounded recall (FBR) method. We prepared list of 20 food groups based on which households report their diet. Our FBR method results in significantly different reporting of food item consumption over the 7-day reference period. For 13 of the 20 food groups, households were more likely to report them with the FBR method than the SI method, suggesting that households tend to forget reporting consumption of these food groups in the in-person surveys. For 7 foods groups, households were more likely to report consumption with the SI method compared to the FBR method, suggesting forward telescoping by which respondents include consumption episodes that occurred outside the 7-day reference period. When constructing a 7-day household dietary diversity score, the forgetting and forward telescoping errors offset each other, though the similarity masks important nutritional differences across methods in which foods are reported. Over a 24-hour reference period, our FBR method results in similar reporting of food item consumption compared to the SI method, suggesting that respondents do not have difficulty recalling consumption over a short period (i.e., “yesterday”) and respondents report consumption similarly in person and over the phone. We add new experimental evidence confirming the cognitive burden respondents face in reporting dietary intake data over a 7-day recall period, which is also highlighted by others. We shed light on the specific mechanisms (forgetting vs forward telescoping) that contribute to reporting differences between the FBR and SI methods. We offer a promising approach to extend respondents’ reference periods without exacerbating recall biases, which can help reduce within-person measurement errors of programmatic outcomes such as dietary diversity.
Medicaid Generosity and Food Hardship Among Children (I3, J1)
AbstractWe explore the role of the largest non-food support safety net program, Medicaid, on multiple measures of food hardship among households with children, including measures that capture hardship explicitly experienced by children. Using data from the 2001-2020 waves of the December Current Population Survey, we identify the effect of having a Medicaid eligible child on household food hardship by exploiting between-state, over-time, and between household income eligibility criteria. We find that having an eligible child reduces rates of household food insecurity and very low food security by 19% \& 24%, respectively. Among children themselves, eligibility reduces rates of food insecurity and very low food security by 20\% each. The effects are stronger for Black and Hispanic households as well as households that have children under 6 years old.
Medical Marijuana Legalization and Parenting Behaviors: An Analysis of the Time Use of Parents (I1, K3)
AbstractDoes access to medical marijuana improve parenting? Medical marijuana may improve the health of parents and allow them to spend more quality time with their children, but it can also have detrimental effects if parents abuse marijuana. We examine the consequences of medical marijuana legalization (MML) across US states on parenting behaviors as measured by time use in the American Time Use Survey. We find that access to medical marijuana increases the time parents spend in childcare, but with considerable heterogeneity across parental characteristics and childcare activities. We also find that MML reduces leisure time while increasing sleep time, consistent with the beneficial effects of medical marijuana. Our results suggest that MML may improve children's outcomes through increased parenting engagement.
Migration Barrier Relaxation and Entrepreneurship: Evidence from the Hukou Reform in China (J2, L2)
AbstractWe study the impacts of relaxing internal migration barriers on entrepreneurship by exploiting China’s 2014 nationwide hukou reforms and the administrative firm registry. Our difference-in-differences estimation finds that treated counties meet sizable increases in entrepreneurial activities relative to counties without reforms. The effect is most prominent in establishments of firms with a smaller scale and a lower likelihood of survival, indicating moderate expansions in labor demand. Migrant workers’ wages decline and entrepreneurial activities improve most in labor-intensive industries, implying that increased labor supplies serve as one underlying mechanism. Our findings highlight the important role that removing domestic labor market frictions plays in promoting entrepreneurship.
Migration: Impact on Taxes and Social Benefits (H2, H2)
AbstractWe provide novel and comprehensive evidence on the net fiscal contributions of natives and migrants to governmental budgets of EU countries. We account for income taxes and cash benefits, alongside with indirect taxes and in-kind benefits, that are often missing in standard data sets. We find that on average migrants are net contributors to public finance over the period of 2014-2018 in the EU-14, and moreover they contribute about 1.5 thousand euro per year per capita more than natives. We also show that this difference is partly due to selection on characteristics that make migrants net fiscal contributors, such as demographic factors and employment probability.
Miners' Reward Elasticity and Stability of Competing Proof-of-Work Cryptocurrencies (L1, G2)
AbstractProof-of-Work cryptocurrencies, such as Bitcoin and its forks, employ miners (freelance contributors) to sustain the system through algorithmic adjustments to the expected reward. The efficacy of the algorithm, the labor supply of miners, and the competition among cryptocurrencies are crucial factors in determining the stability of the system. In this study, we present a short-run supply-side model of the multicurrency mining market and identify the conditions under which cryptocurrencies can maintain a stable transaction speed. We estimate the elasticity of miners' labor supply using a discontinuity caused by an event known as ``halving,'' and conduct counterfactual simulations to analyze the factors that contribute to the stability of Bitcoin. Our findings suggest that the stability of Bitcoin is heavily dependent on low hash-supply elasticity and interactions with other cryptocurrencies, despite the fact that their market capitalization is substantially smaller than Bitcoin's. However, upgrading the algorithm can stabilize Bitcoin irrespective of these factors.
Minimum Wages, Health and Fertility (J0, I0)
AbstractMost minimum wage (MW) research has focused on high-income countries and its direct impacts on wages and employment. Relatively little is known about a) the effects of MWs in low- and middle-income countries in which most people affected by MWs live; b) the broader impacts of MWs, including on fertility, health, and households. This study contributes to filling these two gaps by studying the effects of minimum wages (MWs) on employment, fertility and adult and child health outcomes in Indonesia. Our analysis is based on the Indonesian Family Life Survey, a rich longitudinal dataset, merged with province-specific MWs, and uses individual and biological-siblings fixed effects and distributional effects based on lagged panel data. We find that MWs increase men’s earnings and labor supplies, with small and insignificant results generally for women. This is true especially for the lower parts of earnings distributions (where MWs are more likely binding). In our intergenerational analyses, the estimated effects of MWs in pregnancy on birth weights, standardized height scores and pregnancy complications suggests MWs significantly improve child health. Moreover, the effects are strongest for children whose fathers are in the lowest part of the earnings distribution. We also find evidence consistent with a Quantity-Quality fertility model with costly birth control: MW reduces fertility, increases contraceptive expenditure, and reduces unwanted births among parents who have achieved desired fertility. Our study highlights the nuanced but positive roles MWs could play in improving maternal and child health and in reducing fertility, despite not directly affecting women’s earnings and labor supply.
Misallocation under the Shadow of Death (L1, E2)
AbstractThis study focuses on a slow exit process, known as a shadow of death, as a new factor of inefficient resource allocation in the macroeconomy. First, we develop an endogenous growth model that incorporates firms’ R&D investment and distorted exit decisions. The model shows that the exit of firms in the market equilibrium is inefficiently slow from the social viewpoint even without distortions such as corporate subsidies, which further impede the exit process. Second, our empirical analysis using Japanese firm-level data confirms that exiting firms exhibit the shadow of death in a manner that is consistent with our model. Further, the degree of the shadow of death is related to our distortion measures such as corporate subsidies. Third, our simulation based on the calibrated model suggests that an increase in subsidies can help explain recent firm dynamics in Japan and worsen productivity
growth and welfare, although the quantitative impacts of subsidies are limited.
Monetary Policy and Inequality. The Case of Finland (E5, D3)
AbstractWe use Finnish household-level registry and survey data to study the effects of ECB’s monetary policy on the distribution of income and wealth. We find that monetary easing has a large positive effect on aggregate economic activity in Finland, but its overall net impact on income and wealth inequality is negligible. Monetary easing increases households’ gross income by reducing unemployment and leading to a general rise in wages, while at the same time it boosts asset prices. These different channels have counteracting effects on income and wealth inequality, as measured by the Gini coefficient and the ratios of income and wealth of the 90th percentile to the 50th percentile. The reduction in aggregate unemployment benefits especially households in lower income quintiles, where the initial rate of unemployment is high. Households in the upper income quintiles, where the rate of employment is higher, benefit relatively more from an increase in wages. An increase in house prices benefits all homeowners. In terms of net wealth, households with large mortgages, in the lower wealth quintiles, benefit the most from an increase in house prices due to a leverage effect. An increase in stock prices, in turn, benefits mainly households in the top wealth quintile.
Monetary Policy and Redistribution in Open Economies (F4, E3)
AbstractThis paper examines how monetary policy affects the asymmetric effects of globalization. We build a quantitative open-economy model in which households differ in their integration with international markets, as well as their income and wealth. We use the model to reassess the classic questions in international macroeconomics from a distributional perspective: What are the effects of monetary policy and external shocks in open economies? And how do alternative exchange-rate regimes compare? Our analysis yields two main takeaways. First, heterogeneity in households’ international integration plays
a central role in driving the unequal consumption responses to external shocks, more so than income and wealth. Second, households’ heterogeneity reveals the presence of a stabilization-inequality trade-off for the conduct of monetary policy in open economies, with fixed-exchange-rate regimes leading to amplified
but less unequal consumption responses to external shocks.
Monetary Policy with Non-Ricardian Households (E5, F4)
AbstractThis paper analyzes how the presence of non-Ricardian households can alter the dynamics in a New-Keynesian Dynamic Stochastic General Equilibrium (NK-DSGE) model. The model is calibrated using data for the US economy. We focus on two key areas. The first is the link between monetary policy and consumption inequality in the presence of non-Ricardian households. We find that a contractionary monetary policy shock increases consumption inequality. Part of this increase is due to a novel government transfer channel. This channel becomes significantly stronger when steady state debt is positive. We also find that because of this link, the presence of non-Ricardian households amplifies the impact of monetary policy on output and inflation. The second area relates to the choice of monetary policy rule. We compare four monetary policy rules: the Taylor rule, inflation targeting, nominal GDP targeting, and a new labor income targeting rule. We find that labor income targeting outperforms all other rules for most parameter values. Nominal GDP targeting is better than labor income targeting only when the share of non-Ricardian households is large or when the households exhibit low habit persistence.
Monetary Transmission under Heterogeneous Exchange Rate Exposure (F3, F4)
AbstractThis paper empirically studies the transmission of U.S. monetary policy to emerging market economies (EMEs) when EME firms are heterogeneous in exposure to exchange rate risk. I find that firm size plays a crucial role in international monetary transmission. In response to a U.S. monetary tightening, large firms in EMEs reduce the share of dollar debt, as well as investments in capital and financial assets. This pattern is salient in countries with relatively flexible exchange rate regimes. As a future research, I plan to study the contagion of U.S. monetary policy to local credit markets via bond and equity channels.
Mortality Beliefs and Saving Decisions: The Role of Personal Experiences (D1, G5)
AbstractThis paper is the first to non-experimentally establish a causal relationship between households’ mortality beliefs and subsequent saving and consumption decisions. Motivated by prior literature on the effect of personal experiences on individuals’ expectation formation, I exploit the death of a close friend as an exogenous shock to the salience of mortality of a household. Using data from a large household panel, I find that the death of a close friend induces a significant reduction in saving rate of 1.1 percentage points that grows to 1.7 percentage points over the following 6 years. I show that the incorporation of personal experiences in mortality beliefs can be explained by the canonical consumption life-cycle model augmented by the experiencebased learning model. The saving response to the shock strongly depends on households’ age, emotional involvement, risk aversion, and decays over time. Overall, this paper provides novel insights into whether and how mortality beliefs are incorporated into households’ financial planning.
Multi-tiered Interbank Market and Monetary Policy Implementations (E5, G2)
AbstractI offer a theoretical framework to compare the relative efficacy between different monetary policy instruments, featuring a multi-tiered interbank market and financially constrained heterogeneous firms. The open market operation (OMO) injects liquidity to small banks via the large bank through the interbank market, imposing a trade-off between the large bank's information advantage and monopoly power in credit allocation. Structural monetary policy instruments, like special lending facilities and reserve requirement ratios (RRR), bypass the multi-tiered interbank market, and distribute liquidity equally among small banks. I find a small deterioration in economic conditions can trigger the large bank's sudden dysfunction, leading to a sharp drop in the interbank market allocative efficiency and the aggregate output. Structural instruments are relatively more effective to OMO in implementing monetary policies in the dysfunction regime. In a dynamic general equilibrium setting, the large bank's sudden dysfunction can serve as a novel mechanism to generate endogenous boom-bust business cycles. I establish supporting evidence in the context of China.
Need for Speed: Broadband and Student Achievement (I2, O3)
AbstractThis paper studies the broad effects of the introduction of fiber broadband, through the lens of student achievement. I link granular data on new fiber construction and advertised download speeds with administrative test score data and local labor market data. Exploiting variation in the introduction of fiber at the census block group level, I implement a difference-in-differences design and find a modest effect on educational outcomes, roughly on par with lowering class sizes by one student. In addition, I show fiber increases local employment and search intensity for supplementary educational materials (e.g., Khan Academy). Last, I show that increased competition from fiber providers drives quality improvements in other available technology.
New Technology, New Hierarchy? Implications of Technology Change for the Division of Problem Solving (O3, J2)
AbstractWe measure how different technologies alter the structure of problem-solving and the division of labor across occupations, generating technological skill bias across occupations. We focus on automation versus consolidation of parts in the optoelectronic semiconductor industry as examples of innovations that change the inputs to production and the structure of production, respectively. We collect novel data from nine manufacturers in the optoelectronic semiconductor industry on skills, process structure, problem referrals to other occupations, and the distribution of production time per process step for lead operators, technicians, supervisors and engineers involved in more than 90 production steps, and engineers and managers involved in more than 100 process or product design activities. Firms divide problem solving across direct production workers, supervisors and managers and a body of staff (engineers, technicians) intervening “as-needed” with specific problem-solving expertise. Our early insights suggest that in cases of higher automation, skill heterogeneity increases for production supervision roles; in contrast, as designs become more consolidated (so that different strands of development must be more closely coordinated), designers and especially design managers must increase their breadth of skill.
Non-compete Agreements and Labor Reallocation Across Product Markets (J2, J4)
AbstractI analyze the effect of non-compete agreements (NCAs) on career trajectories of 600,000 inventors in the US. NCAs limit the choice set of inventors as they are less able to move to competitors. Inventors bypass their NCAs by moving to a new employer in a more distant product market. I identify causal effects using staggered changes in NCA enforceability across US states. Inventors are 25% more likely to move to another industry after higher NCA enforceability. Inventors move to new employers who are less likely to rely on NCAs and patent in unfamiliar technologies. There is a lower quality match between inventors and new employers. Consistent with this evidence, affected inventors are subsequently less productive. Regulation in the form of non-compete agreements put a constraint in the industry choice set of inventors, which leads to some detrimental reallocation of human capital in the economy.
Non-linear Dependence and Portfolio Decisions over the Life-Cycle (G5, D1)
AbstractUsing the Panel Study of Income Dynamics Survey, we reveal the non-linear dependence, between-squares correlation, between stock returns and earning risk exists. To understand how this non-linear dependence affects household life-cycle profile, we develop a life-cycle model that incorporates between-squares correlation and shows that this non-linear dependence can explain low participation rate and moderate risky asset shares. Empirical studies support the model’s predictions that households with higher between-squares correlations are less likely to participate in the stock market and lower their risky asset holdings conditional on participation.
Nudging or Nagging? Conflicting Effects of Behavioral Tools (I2, D9)
AbstractThe gap in reading skills between low-income children and their higher-income peers emerges very early in life. To help close this gap, we conducted an RCT with low-income parents of young children in Chicago, with the aim of increasing parental reading time through behavioral tools. Parents were given an electronic tablet with over 200 books, and this was used to track the total number of minutes parents read to their children using the tablet for 11 months. We also measured child literacy skills. Parents were randomized into 4 groups: 1) a control group, and groups that received 2) just the tablet, 3) tablet with reminder texts, and 4) tablet with goal-setting texts. Relative to the group that only received the tablet, we found that the goal-setting group read significantly more (0.5 SD) but had no significant difference in test scores. Unexpectedly, the reminders group scored significantly lower than the group that just received tablets, despite no significant difference in reading time. This demonstrates that nudging might have an unintended consequence of reducing the quality of task (reading) that people are nudged to do. Despite this, technology such as a tablet with books may help close the achievement gap in reading: the pooled average test score across all 3 treatment groups that received tablets was significantly higher (0.19 SD) than the control group that did not receive a tablet. All results are robust to controlling for baseline literacy scores, age, and school fixed effects.
Number of Creditors and the Real Effects of Credit Supply Disruption (E2, G2)
AbstractThis paper examines the effect of shocks to credit supply on real decisions in a firm by exploiting the exposure of firms to a bank involved in a massive fraudulent Ponzi scheme in 2011. Using a hand-collected dataset on bank-firm relationships for Iranian public companies it compares the employment outcome for the impacted firms (connected to the troubled bank) vs. non-impacted ones (connected to a non-troubled bank). Using a difference-in-difference approach, the paper finds that a sudden dry up in the credit supply is followed by up to 2.9% drop in annual growth rate of employment. The cross-sectional analysis shows that the effect is more pronounced in more financially constrained firms. The results highlight the importance of the credit supply channel and the indirect costs of financial scandals on the real side of the economy in emerging markets.
Organisational Leadership: How Much Does It Matter? (I3, J5)
AbstractWe study the influence of organisational leadership on worker wellbeing and organisational performance using data from the 2004 and 2011 British Workplace Employment Relations Survey (WERS) and alternative empirical approaches including the method of IV. Virtuous leadership is significantly and robustly associated with: (i) job satisfaction positively, (ii) job anxiety negatively, and (iii) an upbeat assessment of organisational performance. The results found lend strong support for the hypothesis that good leadership is vital for worker wellbeing and the success of business, which organisational policy makers ought to heed. There is a dearth of evidence on organisational leadership as an institution and its influence on organisational outcomes, which this paper aims to contribute to. Economists had shunned research on the role of organisational leadership until Hermalin (1998; 2012) developed an economic theory of leadership, where he showed how leaders may get followers and highlighted about the important role leadership plays in shaping the fate of organisations. More recently, Martinez et al. (2015) stressed on the important theoretical and empirical work economists can do on organisational leadership and its link with performance, while Gibbons and Roberts (2015) identified the nature and role of leadership in organisations as one of the key outstanding issues for organisational economics to address. This paper addresses the gap in the literature these influential studies have underscored.
Pandemic Effects on the Pricing Function for Residential Buyers (R3, C5)
AbstractThis study looks at the effect of the onset of the COVID-19 pandemic on house pricing and probability of sale using local Multiple Listing Service (MLS) data. The pandemic not only directly impacted average prices and probability of sale, it also caused the coefficients of some features of the property and sale to change while others were stable to the exogenous shock of the pandemic. The hedonic pricing and probability of sale coefficients that did change did not do so instantaneously, however; the impact evolved over several months at the beginning of the pandemic. The results should be of interest to buyers and sellers of residential properties, agents specializing in residential properties, and researchers looking to better capture the impact of exogenous events on housing prices.
Parental Health, Children’s Education and Unintended Consequences of State Support: Quasi-experimental evidence from KwaZulu-Natal, South Africa (I1, I2)
AbstractPrevious evidence showed that parental health is correlated with children’s education, but causal evidence on this relationship remains scarce. This study investigates whether parental eligibility for HIV treatment (ART) improved children’s educational attainment in KwaZulu-Natal, South Africa using a quasi-experimental design.
We combine clinical and household panel data between 2000 and 2017. During our study period, ART eligibility was based on a biomarker in the blood, allowing us to employ a regression discontinuity design. We further analyze how different types of state support might mediate this relationship.
We find that eligible parents just below the threshold initiate ART much faster and are less likely to report a clinic visit in the past 12 months than their ineligible counterparts just above the threshold. However, the transmission into gains in children’s education is mediated by the type of state support that parents received: Previous recipients of health-contingent state support can lose the state support after initiation of ART, as their health improves after ART is initiated. Indeed, we see a negative impact of ART eligibility on the household asset index and on children’s education for these parents. In contrast, there is a positive impact of ART eligibility on children’s education for fathers who received non-health-contingent state support.
Our findings underline the role of the household’s economic situation for the transmission of health improvements into children’s education, and show that the design of state support can mediate this relationship.
Party Lines or Voter Preferences? Explaining Political Realignment (D7, H1)
AbstractWhat has triggered recent changes in the structure of political cleavages? Is it that political parties have shifted ideologically or is it voter preferences that have changed? This paper approaches this question by quantifying the extent to which shifts in political cleavages can be attributed to the evolution of party platforms, voter demographics, and citizens' preferences, respectively. Using websites of candidates for the United States House of Representatives and applying an unsupervised probabilistic topic model, I estimate each candidate’s position on a series of key topics for congressional elections between 2000 and 2016. I find that the distance between the average position of the two main US parties has considerably widened during the twenty-first century. I show that candidates strongly accommodate their positions to the demographic composition of their congressional districts but do so differentially per topic. Expanding upon the results from Gethin et al. (2021), which reveal the emergence of the so-called Brahmin Left, I use election results at the precinct level starting in 2000 to show that the relationship between the precinct share of college graduates and Democratic vote exhibits a U-shape but has slid down over time. Leveraging exogenous variation in candidate policy propositions, I estimate a heterogeneous demand function for policy positions. I show that college graduates favor liberal positions on cultural questions such as reproductive rights or military expenditures. This demand has not changed since the early 2000s but do not have any different preferences for economic policies than non-college graduates. Specific demand for environmental policies from college graduates emerged only recently and did not exist in the early 2000s. Ultimately, these results indicate that most of the change in political cleavages that has occurred in the twenty-first century can be attributed to oscillations in party positions and not to variation in voters' preferences.
Political Power-Sharing, Firm Entry, and Economic Growth: Evidence from Multiple Elected Representatives (D7, O1)
AbstractDoes the extent of power-sharing among political representatives shape local economic activity? Using a novel dataset on the universe of private firms in India, we find that increasing the number of politicians governing an area increases firm entry by 3%. This is associated with higher employment and improvement in real economic activity measured through satellite nightlight intensity. The identification strategy exploits uneven overlap of electoral and administrative boundaries, leading to a quasi-random variation in the number of politicians governing adjacent administrative units. This setup allows us to implement a geographic regression discontinuity design across boundaries separating a split (multiple politicians) unit from an unsplit (single politician) unit. Greater state efficiency, lower regulatory costs, and reduced cronyism due to increased checks and balances among multiple non-aligned politicians is the primary driver of higher firm entry.
Presence of Food Retailers and Nutrition Security in the United States (D1, Q1)
AbstractMany studies find evidence of nutritional inequality in the United States—that is the association between low access to healthful food and low income or non-white population. However, the findings are not causal: we do not know whether the residents of the low-income non-white neighborhood are choosing unhealthful diet because of the lack of access, or large food retailers with a variety of healthful options choose not to locate in the neighborhood due to lack of demand.
The study finds the impact of the presence of various food retailers (e.g., large and small groceries, fast food and full-service restaurants) and their composition in the neighborhood on household nutrition. We utilize COVID-19 lockdown and store closures for identification. Household food purchase information is obtained from scanner data, and food retailers' presence and operation were tracked using cellphone mobility data.
First, we use machine learning to derive the conditional probabilities of a food retailer type locating in a neighborhood, given the demographic and regional features. Second, we calculate the rate of substitution between two types of food retailers to understand the dietary preferences of residents. If one type of food retailers is present or absent, we assess how it affects residents’ nutritional choices, such as fresh vegetable consumption.
Unlike previous studies that show nutrition choices are mainly driven by demand, we find evidence of supply playing a critical role. The intrinsic food choices change due to the neighborhood food environment, e.g., black households buy fewer added-sugar items compared to white in general, but they buy more when a small grocery is in the neighborhood. Hispanic consumers buy fewer chips but buy more when a convenience store is nearby. Thus, the access to and share of healthful retailers matter, and it has a heterogeneous impact across races.
Price Duration Using Online Data: The Case of Mexico (D4, E3)
AbstractIn this paper, we use a novel microdata set of web scraped prices gathered on a daily basis to investigate big retailers' price-setting decisions before and during the Covid-19 pandemic in Mexico. To that end, we first provide a statistical description of price spells, highlighting (i) the different types of price spells (such as complete, left-censored, right-censored and interval-censored) and (ii) the hazard rate of price changes. Second, using semi-parametric duration models with time-varying regressors, we analyze the duration of price spells and its determinants. Preliminary results suggest strong presence of intra-week and intra-month seasonal patterns, as well as a non-negligible role of macroeconomic conditions, like inflation and the exchange rate, affecting the probability of price changes. As expected, since the start of the Covid-19 pandemic, it is more likely to observe a price adjustment. Our results are important as they highlight that purely time- or state-dependent pricing models might neglect relevant price-setting features.
Procurement Institutions and Essential Drug Supply in Low and Middle-Income Countries (I1, O1)
AbstractInternational procurement institutions have played an important role in drug supply. This paper studies price, delivery, and procurement lead time of essential drugs supplied in 106 developing countries from 2007-2017 across four procurement institution types. We focus on four major therapeutic areas for infectious diseases that disproportionately affect people living in developing countries: antiretrovirals, antimalarials, tuberculosis drugs, and antibiotics. We find that pooled procurement institutions lower prices: pooling internationally is most effective for small buyers and more concentrated markets, and pooling within-country is most effective for large buyers and less concentrated markets. Pooling can reduce delays, but at the cost of longer anticipated procurement lead times. Finally, pooled procurement is more effective for older generation drugs, compared to intellectual property licensing institutions that focus on newer, patented drugs. We corroborate the findings using multiple identification strategies, including an instrumental variable strategy, the Altonji-Elder-Taber-Oster method for selection on unobservables, and reduced-form demand estimation. Our results suggest that the optimal mixture of procurement institutions depends on the trade-off between costs and urgency of need, with pooled international procurement institutions particularly valuable when countries can plan well ahead of time.
Product Variety and Retailer Profitability in Two-Sided Market (L1, L2)
AbstractThe paper will examine the effect of coffee pod variety and ownership of coffee pods by coffee machine makers on market outcomes, firm profits, and consumer surplus. In a complementary good market, most firms are following razor-razorblade model where they gain the most profit from the high markup consumable goods. Product variety is important for consumers’ platform adoption decisions. Platform would like to form partnerships with national or local brands to provide product variety. With more brands joining the network through licensing, consumers are more willing to adopt the initial durable platform due to brands royalty and more selections available. This demand-increasing effect boosts the profit for its own consumables if the firm is selling both durable platforms and its own consumable goods. But introducing variety may also cannibalize sales of its own consumable goods. Nevertheless, the larger network size coming from the varieties results in larger market presences, which also gives the company a better bargaining position during the bargaining with various brands and retailers. And this bargaining position depends on the demand increasing effect and cannibalization effect together. This may increase distortion due to double marginalization in the vertical supply chain if the manufacturer has a higher bargaining position. Therefore, the total effect of variety on consumers is not clear. I build a three-party bilateral bargaining model to study a unique setting in the single-serving coffee industry, where Keurig, coffee brands and retailers interact with each other. And this paper answers the following questions: Is the cannibalization effect stronger than the demand-increasing effect when the platform company adds more licensees? How does it vary among different licensees? Will a better bargaining position caused by a larger network size lead to higher double marginalization distortion? What’s the effect of vertical integration in a two-sided market on consumer welfare?
Progressive Pension and Optimal Tax Progressivity (E6, H3)
AbstractWe examine the implications of progressive pensions for designing an optimal progressive income tax code. In a simple analytical model, we first show that optimal tax progressivity is negatively linked with pension progressivity. This relationship is then analysed further using a stochastic dynamic general equilibrium overlapping generations model calibrated for Australia, where pension payments are universally means-tested to target low income retirees. Importantly, these payments are financed directly by progressive income taxes. We find that flattening the income tax code and tightening means-testing rules for pension payments improve overall welfare. The optimal design consists of a flat tax rate and a strict means-tested pension scheme. Hence, reforms that shift the social insurance/redistribution role embedded in the tax system to the means-tested pension system can yield better welfare outcomes. Our results also highlight that more generally, a redistributive tax and transfer system can be improved by addressing redistribution concerns directly through more progressive transfers while improving efficiency by reducing tax progressivity.
Prospering through Prospera: CCT Impacts on Educational Attainment and Achievement in Mexico (I2)
AbstractThis paper develops and estimates a dynamic model of student enrollment, school choice, academic achievement and grade progression to evaluate the impacts of Mexico’s conditional cash transfer program Prospera on educational outcomes over grades 4-9. Academic achievement is measured by nationwide standardized test scores in mathematics and Spanish. Enrollment decisions are the outcomes of sequential decisions at each age from individuals’ feasible choice sets, determined by the types of schools locally available and local-labor-market opportunities. The achievement production function has a value-added structure. Model parameters are estimated by maximum likelihood using nationwide administrative test-score data (the ENCEL data) combined with survey data from students and parents, census labor-market data, and geo-coded school-location data. The estimation approach controls for selective school enrollment in different types of schools, grade retention and unobserved heterogeneity. The results show that the Prospera program increases school enrollment and academic achievement for program beneficiaries in lower secondary school grades (grades 7-9). The average test-score impacts are 0.09-0.13 standard deviations in mathematics and 0.03-0.05 standard deviations in Spanish. Students from the most disadvantaged backgrounds experience the largest impacts. The availability of telesecondary distance-learning schools is shown to be an important determinant of the Prospera program’s impacts on educational outcomes.
Racial Discrimination in Entrepreneurship: Decomposing Demand and Supply Constraints (E2, J7)
AbstractUsing micro-data on startups, we find that Black-owned (relative to White- owned) startups operate with lower capital intensity and have lower average returns to capital. We formulate a theoretical framework showing that these findings are consistent with the hypothesis that Black entrepreneurs face both consumer (demand) and financial (credit supply) discrimination. We further show that the differences in capital returns are persistent over time, whereas the differences in capital intensity disappear after four years. This indicates that any negative effects of financial discrimination are transitory, but demand-side discrimination can permanently reduce the profitability of Black-owned firms. Our results provide a complementary explanation for the persistent racial wealth gap despite efforts to reduce discrimination in credit markets.
Randomization of Class Schedules and Education Outcomes at a Large Public University (I2, J0)
AbstractIn this paper, I intend to examine how early morning classes affect students’ education outcomes. More specifically, I investigate the impact of early morning classes on students' propensity to take STEM classes and students' choice of a major with a unique course registration process at a large public university in Indiana, USA. Implementing instrumental variable approach, I find that assignments to early morning classes increase non-STEM participation by roughly 57%, and students are less likely to choose a major directly corresponding to the assigned early morning classes by about 52%. To better understand the mechanisms of the findings, I explore two potential mechanisms: (1) academic performance and (2) non-academic reasons. I find that non-academic reasons are the major driving force to the results in this paper. One of the potential non-academic reasons is attribution bias. The contribution to the existing research is twofold: first, the student data provides a more suitable comparison with other university students in the United States; and second, the institutional structure and student body allow me to minimize the peer effect from the treatment effect.
Recipes for Female Success: Becoming a CEO (J1, J2)
AbstractIn this paper, we focus on the determinants of female leadership in the corporate setting. We
employ machine learning algorithms to identify the most important predictors of first-time female
chief executive officers (CEOs). With a comprehensive dataset of over one hundred characteristics
of thousands of executives, we find that female executives who are well-connected in social
networks and have diverse professional experiences, including board exposure, are more likely to
become CEOs. Interestingly, advanced education and international experience may have the
opposite effect. Our study provides a novel approach to analyzing career determinants of women
in corporate America, contributing to the ongoing discussion on leadership and gender.
Reconciling the Effects of Government Spending: The Role of Information Frictions (E6, H3)
AbstractCan the effects of fiscal policy be state-dependent due to changes in information frictions? Yes. I empirically document a novel result that reconciles the Keynesian and neoclassical predictions of fiscal policy: by emphasising the importance of information frictions. I use a non-linear local projections framework with forecasters’ disagreement as a measure of information frictions. The key finding is that during periods of high information frictions, households act less Ricardian (as they are less forward-looking), and thus government spending increases consumption. In contrast, during low information frictions, households act sufficiently Ricardian such that consumption falls in response to government spending rise. I provide a theoretical quantitative framework to the empirical findings on how information frictions could affect the consumption response to a government spending shock. I build on a general equilibrium model with sticky information, and add households with limited asset market participation. When information frictions are not severe, many households are able to identify a government spending shock and thus, their Ricardian effects dominate the rule-of-thumb households leading to a fall in aggregate consumption. In contrast, when information frictions impede their ability to identify the shock, only few households save in advance of higher future taxes and therefore, aggregate consumption rises.
Regret Searching for Jobs (J3, D9)
AbstractI introduce time-varying search costs to classical Sargent and Ljungqvist (1998). In this model, the cost of search consists of two parts, a time-invariant part that is linear in search intensity as in classical models, and a time-varying part that depends on the economic conditions. The time-varying disutility is motivated by regret theory: agents want to reduce regret by timing the search. This market timing in search intensity has two key implications: agents will search less (1) when they expect less likely to get a job; (2) when they expect a better job in the future.
Responsiveness of Consumers’ Medium-Term Inflation Expectations: Evidence from a New Euro Area Survey (E3, D8)
AbstractMedium-term inflation expectations of the public are a key variable of interest to any modern central bank responsible for price stability. Using the new ECB Consumer Expectations Survey, this paper investigates revisions of medium-term inflation expectations. We provide robust evidence that consumers adjust medium-term inflation views in response to changes in short-term inflation expectations and, to a lesser degree, to changes in inflation perceptions. We find that the strong adverse Covid-19 shock contributed to an increase in consumer inflation expectations. We show that both higher financial literacy and higher trust in the central bank reduce responsiveness of medium-term inflation expectations, which supports their stability. Our results increase understanding of expectations formation, which is essential for medium-term oriented monetary policy.
Returns to Schooling: Credit Hours, Major, and Occupation Choice after a Kansas College (I2, J3)
AbstractThis paper uses an innovative new instrumental variables technique to estimate the returns to different
types of degrees and credits depending on the type of occupation chosen after college. The study
first examines whether there is a difference between credit hours from a community college compared
to a baccalaureate institution. Using a method proposed in a recent working paper by Weiss et al
(2021), we use transcript data to analyze job skills and how they provide a return in certain occupation
types after graduation. Briefly, we show that the wage gap for students who began at a community
college and completed at a baccalaureate institution is negligible compared to their peers who began
at a baccalaureate institution and completed their degree. We also show that the returns to math skills
are increasing and that remedial math classes are important to eliminate a wage penalty for those with
fewer quantitative skills that would otherwise exist in the absence of remedial math.
Risk-Based Regulations in Credit Markets: A Heterogeneous Financial Accelerator (E5, G2)
AbstractCredit markets in the U.S. are dominated by institutional investors, who are subject to various regulations limiting their risk capacity. I study the macroeconomic implications of such risk-based regulations in a general-equilibrium model featuring heterogeneous firms and a bond investor subject to risk-based constraints. During economic downturns, these risk-based constraints become a heterogeneous financial accelerator: It increases the debt financing cost for risky firms, amplifying their default risk, while generating convenience yields for the safest firms. In aggregate, these constraints significantly amplify the drop in investment and output. I also study the effects of credit market intervention policies using this framework. I find that credit policies successfully mitigate the initial drop and speed up the follow-up recovery.
Search Capital and Unemployment Duration (J6, E2)
AbstractIn the last recession, the increase in long-term unemployment has been higher for
younger workers than for older age groups. I propose a novel mechanism, search capital,
to explain long-term unemployment patterns across different ages along the business
cycle: ceteris paribus workers who have been successful in finding jobs in the recent
past become more efficient at finding jobs in the present. Search ability increases with
successful search experience and depreciates with tenure if workers do not search often
enough. In labour markets where short-term jobs are a significant share of employment,
this mechanism can explain cyclical bursts of long-term unemployment. Using Spanish
administrative data, I provide empirical evidence that search capital, as proxied by the
number of temporary jobs a worker has had, is negatively correlated with unemployment
duration. The addition of search capital to a standard search model manages to replicate
these empirical findings while also generating increases of long-term unemployment by
age and along the business cycle that are consistent with the data. Although workers
with stable jobs have higher welfare than workers with many employment spells when
the economy is booming, they suffer higher losses during recessions because of their lower
Seasoned Equity Offerings, Limited Investor Attention, and Post-Announcement Drift: Theory and Evidence (G3, G4)
AbstractWe develop a model of seasoned equity offerings (SEOs) under limited investor attention. While existing models of equity issues assume that all investors pay immediate attention to SEO announcements, we assume that only a fraction of investors pay immediate attention, with the remaining fraction paying delayed attention. We develop three testable predictions from our theoretical model not generated by existing equity issue models. First, in addition to an announcement effect, there will a post-announcement stock return drift following SEOs. Second, the announcement effect of an SEO will be increasing and the post-announcement drift will be decreasing in the fraction of equity market investors paying immediate attention to the SEO announcement. Third, both the announcement effect and the post-SEO drift will have predictive power for the post-SEO operating performance of firms. We test the above three predictions of our theoretical model using the media coverage of firms prior to SEO announcements as our proxy for investor attention and find consistent evidence. Our baseline empirical results are robust to making use of abnormal investor attention (instead of the actual investor attention) received by firms, allowing us to rule out the possibility that our results are driven by the characteristics of certain firms that receive greater investor attention compared to others. We also use an instrumental variable analysis to show that the above empirical relationships are causal. Lastly, we demonstrate the robustness of our results using SEC EDGAR filing searches by investors as an alternative proxy for investor attention.
Skill Hybridization and Technological Advancements (J2, J0)
AbstractI document that the U.S. economy has seen a substantial increase in the mixing of skill requirements, or “skill hybridization,” from 2000-2020. American workers increasingly work in occupations that demand mixtures of analytical, computer, and interpersonal skills rather than specializing in one of them. This change occurred primarily within the medium- to high-wage occupations, and the return to switching to occupations or college majors with more hybrid skills has also increased. To understand the sources of these shifts, I present an occupation formation model in which workers supply bundled skills and firms optimally choose occupations’ skill intensities, delivering endogenous specialization in skill demand. Counterfactual analysis shows that skill efficiency shifts account for the majority of hybridization changes.
Slow Violence of Waste: Evidence from Chinese Environmental Policy in Waste Trade (Q5, F1)
AbstractSince the 1990s, China has been the largest importer of waste in the world. However, China changed its environmental policy in 2018 not to import plastic, paper, and textile. This 2018 Chinese policy left the questions regarding where the waste is exported to and which groups of countries are more affected by importing more waste. This paper seeks to answer two research questions. First, how does the Chinese waste import ban affect the amounts of exports and re-exports of trash among high-, middle-, and low-income countries? Second, does the Chinese waste import ban increase the waste exports to countries with relatively weaker environmental regulations than their trading partners?
Based on the gravity model, I employ the Poisson Pseudo-Maximum-Likelihood (PPML) estimation methodology and use bilateral waste trade data obtained from UN Comtrade and an index of environmental stringency for 90 countries. The evidence is found that the Chinese waste import ban decreased waste exports to high-income countries by 11% but increased waste exports to middle-income countries by 71% and low-income countries by 44%. It is unexpected that middle-income countries import more waste than low-income countries, and this result suggests that middle-income countries need more recycled materials as an input to other production. For every 1% that a low-income country’s environmental regulations are relaxed relative to their trading partners, low-income countries experienced a 73% increase in waste imports after the Chinese waste import ban. Similarly, re-exported waste, which is low-quality and more contaminated, arrived more in low-income countries with having loose environmental regulations than in middle-income countries, implying the low-income countries are more likely to be a target of a destination of less preferred waste.
Starting Off on the Right Foot - Integrative Schooling and the Educational Success of Immigrant Children (I2, J1)
AbstractThis study is the first empirical analysis to identify the causal effect of an educational integration model which focuses on language acquisition for newly immigrated primary school-aged children on their academic success. Employing unique administrative panel data from German city states between 2013 and 2019, we use the random allocation of refugee children to schools to study the effect of attending a separate "preparatory class" for language learning on standardized test scores and the probability of attending an academic track in secondary school.
Our results show that primary school-aged refugees who visit a preparatory class do significantly worse in standardized test scores in fifth grade. The negative effect is particularly strong for Math, while we see no significant differences in English test scores and the probability of attending the academic track. Overall, our results indicate that separate classes for newly immigrated children focusing on language learning do not foster their academic achievement.
STEM Employment Resiliency During Recessions: Evidence from the COVID 19 Pandemic (J2, I1)
AbstractSTEM occupational employment suffered smaller peak-to-trough percentage declines than non-STEM employment during both the Great Recession and COVID-19 recession, suggesting a relative resiliency of STEM employment during recessions in the digital age. We exploit the sudden peak-to-trough declines in STEM and non-STEM employment during the COVID-19 recession to measure STEM recession-resiliency during the pandemic, decomposing our difference-in-differences estimate into parts explained by various sources including differences in demographics, educational attainment, job tasks, remote work capability, industry, and STEM knowledge importance on the job. We find that STEM knowledge importance on the job explains the greatest share of STEM employment resiliency, and that workers in non-STEM occupations who nonetheless use STEM knowledge experienced higher employment rates during the pandemic. We show that R&D expenditures and employment also remained resilient, suggesting only a mild effect of the COVID-19 pandemic on innovative activity. Altogether, our findings suggest that increasing opportunities for STEM training---including outside the college-track---may help improve the employment resiliency of workers during future recessions.
Sticky Leverage and Debt Overhang: Evidence from Foreign-Denominated Debt in Latin America (E4, G3)
AbstractThe sticky leverage theory proposes that when debt payments are fixed in nominal value, a monetary expansion reduces the real burden of existing debt boosts firm performance. This paper investigates this debt financing channel of monetary policy transmission through a novel setting: the impact of U.S. monetary policy shock on Latin American companies with different dollar denominated debt positions. Such a setting helps us distinguish the standalone debt financing channel from the stick price channel of monetary policy transmission. We first present a dynamic model of corporate capital structure with foreign debt to predict the response of firms. And consistent with our theoretical prediction, using the data of Latin American publicly-listed companies, we find that a more dollar-indebted firm experiences higher increases in equity value, capital expenditure, and sales after an unexpected U.S. monetary expansion, especially when the maturity of such foreign-denominated debt positions is long. Nonetheless, such effects are not found in the firms’ locally denominated debt positions. Finally, we document that a larger net export position amplifies the responses of dollar-indebted firms to the U.S. monetary shocks. Thus, the debt friction and the price friction reinforce each other through which monetary policy generates real effects. Finally, dissecting monetary shocks, we further notice that firms with foreign-denominated debt are mostly responsive to unexpected changes in the current target for the federal funds rate and the policy surprise factors of monetary policy surprises.
Subsidy-Driven Firm Growth: Does Loan History Matter? Evidence from a European Union Subsidy Program (G3, H3)
AbstractWe test whether firm-level subsidies work through reducing financing costs or easing credit constraints. Using credit registry data from Hungary, we classify those firms as credit constrained which applied for but did not receive a loan, and those firms as unconstrained which received a loan. We combine difference-in-differences methods with matching, and to further remove unobserved heterogeneity caused by selection into subsidy application, we use rejected subsidy applicants as the control group. We find that subsidised firms increase their assets faster regardless of whether they are credit constrained or not -- consistent with reduced financing costs. Moreover, credit constrained firms increase assets even faster than unconstrained firms -- consistent with easing credit constraints. However, this latter effect is transitory and fades away after about three years. Relatedly, asset growth does not translate into higher firm efficiency or more new loans. Credit constraints, therefore, may not be the primary hurdle for these firms, and likely reflect other shortcomings such as the lack of good management or viable projects.
Take it to the Bank! Local Discourse and Deposits (G5, G2)
AbstractUsing a hand-collected data set on almost one million local television news stories in the U.S., this paper shows that depositors respond to changes in the intensity of reporting about the COVID-19 pandemic by holding more demand deposits. Counties, where pandemic news stories are 10 percentage points more frequent relative to all news stories hold 1.3% more demand deposits after the onset of the pandemic. This effect holds when controlling for the intensity of the pandemic and several other alternative explanations. Further evidence shows that local news reflects the intensity of local discourse, which in turn causes a spike in deposits, especially in counties with a weaker social structure. The results suggest that the intensity of societal discourse around an event can have significant implications for banks and the real economy.
Technological Substitution of Jobs and the Transferability of Human Capital (J2, O3)
AbstractThere is abundant literature on the decline of routine occupations due to substitution by automation capital. Surprisingly, however, evidence on how routine workers fare after (technology-induced) displacement is lacking. We empirically investigate the labor market transitions of routine vs. non-routine workers after job displacement. More generally, our work addresses how technological change affects the labor market prospects for different types of workers and generates differential reallocation costs. We purport that specific human capital in combination with changing technologies is a major reason for routine workers’ particular difficulties in adjusting to labor market shocks.
We use German administrative data to identify job displacements due to plant closures as exogenous job separations. We extensively match on pre-displacement characteristics and outcomes to account for worker sorting into plants. Our dynamic difference-in-differences estimates show that routine workers face substantially larger displacement costs than non-routine workers. These stem from both the extensive and intensive margin: relative to their non-displaced statistical twins, displaced routine workers suffer from higher unemployment rates and incur larger wage losses than non-routine displaced workers. In line with the secular decline of routine occupations, we also find that routine workers switch occupations more often after displacement.
Consistent with decreasing employment opportunities for routine workers in their own occupation, implying larger losses in occupation-specific human capital over time, the relative wage losses of routine workers displaced in later periods are larger. Over time, displaced routine workers also tend to switch to occupations requiring less of their main tasks. This suggests that the loss of task-specific human capital is one major reason for the large displacement costs of routine workers.
Our results are robust to propensity weighting and entropy balancing between routine and non-routine workers to render these two worker types observationally equivalent. Moreover, our results are robust to recently developed robust difference-in-differences estimators.
Technology Undermining Education: The Case of Chegg (A2, I2)
AbstractChegg is an education technology company with a market capitalization greater than traditional textbook publishers. Its main product, Chegg Study, allows students to obtain homework solutions from "expert” contractors. I track the effects of Chegg over a five-year period for a large finance course of upperclassmen. I find that 25% of students—including 15% of high-scoring students—use Chegg blindly, copying obviously wrong answers. To study the causal effect of Chegg, I use variation in takedown requests for copyright violations as a quasi-experiment. These takedown requests remove answers from the website, making Chegg less tempting. I find that high-scoring students do particularly well on subsequent exams after the temptation to shirk on homework is removed.
Television and the Labor Market: Evidence from Natural Experiments in West and East Germany (J2, D1)
AbstractTelevision is a major spare-time activity with the potential to lower economic activity but also to manipulate behavior by changing preferences and social norms through role models. To provide a comprehensive investigation into the impacts of television on the labor market, we study two natural experiments providing variation in access to television broadcasts. First, we leverage a setting in West Germany, where individuals in some regions could watch commercial television at no charge via terrestrial frequencies, while others could only watch public TV. By analyzing rich panel data, we provide evidence that is inconsistent with the notion of negative impacts on labor market outcomes through, for example, changes in leisure preferences. Instead, we find positive effects on the labor supply among females. This employment premium has limited monetary consequences for females, which could be due to gender-specific occupational choices. To better understand the mechanisms and to examine long-run effects of television, we exploit a setting in East Germany, where for decades, citizens in most areas had access to Western public TV with its emphasis on conservative and family-oriented values, except for those areas where only state-run socialist television with a focus on full employment was available. By analyzing several datasets, we provide evidence on the beliefs of East Germans about the role of women in the labor market, which could explain the persistent effects of TV on the female labor supply and gender equality in labor market participation. It appears that Eastern socialist TV shares a surprising similarity with today's dominant form of Western free-market TV: they both encourage women to work.
Testing Rockets and Feathers Pattern in the Soybean Complex: A Multivariate Quantile Approach (C3, Q0)
AbstractA well-known empirical finding of many commodity markets is that output prices do not react symmetrically to changes in input prices, with output prices rising faster than they fall (Peltzman 2000). This asymmetric price response is termed first by Bacon (1991) as “rockets and feathers.” A long-standing debate, with mixed evidence, exists on the rockets-and-feathers hypothesis in output price responses to input price changes, especially for agricultural commodities. In this paper, we use a multivariate quantile framework to capture price asymmetries in output responses based on conditional quantiles, which provides a complete picture of price responses at different locations of conditional distributions. More specifically, we apply this method to investigate the price responses of two end products, soybean meal and oil, which are jointly produced when crushing soybeans. We use monthly cash prices of the soybean complex and develop the vector autoregressive quantile model with quantile cointegrating relations among three commodities. Our findings show mixed evidence of asymmetry in cumulative price responses of end products. Moreover, the rockets-and-feathers responses of any of the end products usually occur at the higher quantiles of its own price distribution when the quantile of the other product is low. Since the locality of the quantile approach is advantageous to cluster prices that reflect similar market conditions, our study will contribute to the literature by providing empirical evidence that the occurrence of rockets-and-feathers price responses is related to a specific market condition.
The Agglomeration of Urban Amenities: Evidence from Milan Restaurants (R3)
AbstractWe test for the presence of agglomeration externalities in Milan's restaurant sector
and estimate their effects using the abolition of a unique regulation that restricted
where restaurants could locate.Before 2005, Milan mandated a minimum distance be-
tween restaurants which kept the number of establishments articially constant across
neighborhoods.The regulation was abolished in 2005.Using administrative data, we nd
that after 2005 the geographical concentration of restaurants increased sharply and at
an accelerating rate.Consistent with the existence of strong and self-sustaining agglom-
eration externalities, the city's restaurants agglomerated in some neighborhoods and
deserted others, leading to an accelerating divergence in local amenities across neigh-
The Big Short (Interest): Closing the Loopholes in the Dividend-Withholding Tax (H2, G3)
AbstractWe study the effect of reforms that close loopholes in the enforcement of the dividendwithholding
tax (DWT). We focus on a Danish reform enacted in 2016, and compare
Denmark to its Nordic neighbors. Our main outcome of interest is the quantity of stocks
on loan. Before the reform all Nordic countries have a strong spike in stocks on loan
centered around the ex-dividend day. The magnitude is large: on average excess stocks on
loan peak at around 4 percent of the public float. The spike in lending is consistent with the
most popular DWT arbitrage schemes. After the reform the spikes in Denmark disappear,
but they continue in the other Nordics. We interpret this as evidence that the reform was
successful at eliminating DWT arbitrage. We consider the welfare effects of the reform.
Using synthetic difference-in-difference we find that stricter DWT enforcement resulted
in a 130 percent (approx. 1.3 bln USD annually) increase in DWT revenue in Denmark.
We detect no changes in foreign portfolio investment or dividend policy. We also consider
DWT arbitrage among 15 European countries between 2010-2019. We find evidence of
DWT arbitrage in all countries that levy DWT, though there is strong heterogeneity across
countries. Importantly, similar to Denmark, Germany’s 2016 reform has eliminated the
spikes in lending completely. We validate our identification strategy by showing that we
find no evidence of DWT arbitrage in the UK, which does not levy a DWT.
The Complementarity of Field Observational and RCT Studies: Evidence from BT Eggplant in Bangladesh (Q1, C1)
AbstractGenetically engineered (GE) crops have seldom been adopted for human consumption. In 2013, Bangladesh introduced Bt eggplant, the first biotech vegetable crop, which was developed by South Asian scientists. Ahmed et al., 2020 conducted a randomized control trial (RCT) to assess the impact of Bt eggplant on traditional eggplant. In this paper, we present the econometric results of a field study using survey data of 481 farmers in Bangladesh. We find that Bt eggplant raises yield by 19% compared to non-Bt varieties, lowers total cost, and pesticide costs per hectare are 9 and 59 percent, respectively. These results are consistent with estimates of Ahmed et al. We also find that Bt farmers generated a 27% price premium compared to non-Bt eggplant, because of fewer blemishes, and, thus Bt eggplant adoption increased profits by 23 percent. In addition, we were able to assess factors contributing to the adoption of Bt eggplant that include better access to credit, larger families, farm sizes, and profitability. Our results show that field studies add information and can validate RCT findings, and the two methods are complementary.
The Effect of Authorized SNAP Store Entry and Exit on Program Participation (I3)
AbstractSupplemental Nutrition Assistance Program (SNAP) is the largest federal nutrition assistance program, but it does not reach all households who need it. In this paper, we use detailed panel data on entries and exits of authorized SNAP stores from 1990 to 2020 to estimate the effects on program participation. Our data allow us to identify the store type and location, and the precise authorization and end date. We apply a panel event-study design and exploit the timing of authorized store entries and exits while allowing for the possibility of confounding trends. We find that authorized SNAP stores have a robust positive effect on SNAP participation, with one additional store increasing SNAP participation rates for all eligible people by 0.4 percentage points. The average effect on program participation diminishes after the late 2000s, while the estimated effect remains similar in all years for rural areas.
The Effect of Cumulative Job Mobility on Early-Career Wage Development: Does Job Mobility Actually Pay? (J3, J6)
AbstractUsing a diverse sample of men and women drawn from the National Longitudinal Survey of Youth 1997, I analyze the early job mobility and consequent wage evolution of workers from the millennial generation. Millennial workers are found to hold an average of 5.23 total jobs and 3.74 full-time jobs over the first 10 years of the working lifecycle. This marks a notable decline from what has been reported by researchers working with data on earlier cohorts of workers. A series of panel wage models are then estimated which account for both the timing and frequency of job changes over the first decade of the working career as well as for complex interactions between job mobility, actual work experience, and job tenure. The wage estimates indicate that workers who demonstrate moderate job changing in the first two years after labor market entry but then taper their mobility thereafter actually raise their log-wage path above that of either immobile workers or persistent job changers. This finding is significant because previous studies have often identified a negative relationship between cumulative job mobility and wages with immobile workers typically earning the highest wages. The results from this study show that a poor job match at the start of the career need not permanently lower a worker’s wage profile but can be more than made up for through strategic early-career job mobility.
The Effect of Peers' Genetic Predisposition to Depression on Own Mental Health (I1, I0)
AbstractThe goal of this research is to examine whether peers’ genetic predisposition for depression affects one’s own mental health in the short- and long-run by considering same-gender grademates as peers. There has been growing concern about adolescent depression. Adolescents with depression are about three times more likely to be depressed in adulthood compared to non-depressed adolescents (Johnson et al., 2018). Therefore, understanding the factors that affect adolescent and adult depression is key for preventing and treating depression. Since peers have particularly strong impacts on each other during adolescence (Brown and Larson, 2009), peers’ mental health may be an important determinant of own mental health. In this study, I use data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), which follows a nationally representative sample of US adolescents starting in the 1994-95 school year. The genetic data in Add Health includes the polygenic risk score for major depressive disorder (hereafter, MDD score), a composite measure of genetic markers that are correlated with MDD. A higher MDD score means a higher genetic risk for depression. The individual MDD score is used to construct same-gender grademates’ average MDD score, which is the main explanatory variable of interest. I find that an increase in same-gender grademates’ average MDD score is significantly associated with an increase in the probability of being depressed in adulthood for both males and females, with weaker effects on adolescent own depression. I explore the mechanisms underlying the relationship, including substance use, educational performance and attainment, and labor market choices.
The Effect of Remediation on Two-Year College Students (I2, J0)
AbstractThree quarter of first year community college students take at least one developmental course and yet their effectiveness remain ambiguous. Developmental education increases years spent in the college and it also better prepares students for the college-level courses. This paper assesses these two opposing impacts by using state developmental education policies as an instrumental variable for developmental course taking. Results using the NLSY97 transcript data indicate that taking developmental courses significantly increases one’s probability of obtaining an associate’s degree while increasing the total years spent in the college. A sequential model of education is presented to understand more about the complementarities between development and college-level courses.
The Effect of Short-Term Rentals on Local Consumption Amenities: Evidence from Madrid (Z3, R1)
AbstractThis paper investigates the impact of the arrival of Airbnb on local consumption
amenities in Madrid. We exploit the exogenous variation created by the timing
and the unequal distribution of Airbnb listings across the urban geography to
identify its effects on food and beverage establishments. Using an instrumental
variable strategy, we find positive local effects on both the number of restaurants
and their employees: an increase in ten Airbnb rooms in a given census tract
translates into almost one more restaurant, and the same increase in a given
neighborhood generates eight new tourist-related employees. The results are
robust to specification and sample composition. This paper contributes to the
literature on the economic impacts of the platform economy on urban areas by
providing evidence of market expansion externalities from short-term rentals.
The Effect of SMS Reminders on Health Screening Uptake: A Randomized Experiment in Indonesia (I1, O1)
AbstractThe burden of non-communicable diseases is rising in low- and middle-income countries, but the uptake of free screening for these diseases remains low. This relates to the literature on underinvestment in preventive health - screening is a special case of preventive health behavior though as it is not the aim to avoid an illness altogether but to detect a prevalent condition early enough to avoid or postpone complications. In order to address information and behavioral barriers, we conducted a randomized controlled trial in Indonesia to assess whether personalized and targeted text messages can increase the demand for existing public screening services for diabetes and hypertension in the general at-risk population. Our intervention increased screening uptake by approximately 6.6 percentage points compared to the pure control group. Among those, who received and read the messages, the effect size is 17 percentage points. The intervention appears to work through a reminder rather than a knowledge effect. We conclude that text messages can be a cheap and easily scalable tool to reduce testing gaps in a middle-income country setting.
The Effect of the SSI Student Earned Income Exclusion on Education and Labor Supply (I3, I0)
AbstractYouth with disabilities face financial constraints to attaining post-secondary education and encounter strong labor market disincentives when considering employment opportunities. Encouraging human capital development through employment and education could help young Supplemental Security Income (SSI) recipients transition off SSI reliance and improve their long-run economic self-sufficiency. I study the effect of the Student Earned Income Exclusion (SEIE), an education- and work-incentive for youth with disabilities receiving SSI benefits. The SEIE enables SSI recipients under age 22 to exempt $1,930 of their monthly earnings from the SSI benefits determination if they are enrolled in school. Using the Survey of Income and Program Participation (SIPP), I compare changes in SSI recipients’ education and labor decisions in the months surrounding the strict age-22 SEIE eligibility cutoff. I find the SEIE causes SSI recipients to increase school enrollment by 8.6 percentage points and increase employment by 8.4 percentage points. The findings suggest that the SEIE helps relax binding financial constraints for SSI recipients to attend college while revealing a substantial preference for employment among these recipients.
The Effectiveness of Employment-Based Tax Credits Under Labor Market Frictions: Evidence from the Child Tax Credit (H2, J1)
AbstractOne of the central goals of employment-based tax credits is to incentivize work. This paper considers the effectiveness of these tax credits when individuals face labor market frictions. We analyze the labor market response of unauthorized single mothers to the 2008-09 Child Tax Credit expansion in an event study framework. Due to their lack of legal status, unauthorized single mothers face severe labor market frictions. Using data from the American Community Survey, we find large increases in the labor force participation and employment rate of single unauthorized mothers after the expansion relative to unauthorized childless women. The initial increase in employment is partially driven by employment into the informal labor market (proxied by self-employment). The increase in labor market participation coincides with a sharp decrease in school attendance among unauthorized single mothers relative to the control group.
The Effects of Post-Release Supervision on Crime and Recidivism (K4)
AbstractPrior to 2011, inmates who committed lower-level offenses in North Carolina were not subject to post-release supervision. The North Carolina Justice Reinvestment Act changed policy to require nine months of post-release supervision. Leveraging a discrete policy effective date in a regression discontinuity in time model and using administrative data from the North Carolina Department of Public Safety, I explore the effects of this legislative change on criminal outcomes. Evidence indicates that post-release supervision decreases property and violent crimes, but these changes do not persist beyond the supervision period. Results suggest that supervision leads to more individuals returning to prison at a faster rate due to technical, not criminal, violations; however, requiring lower-level offenders to undergo post-release supervision is a cost-effective program.
The Effects of Reducing Parent-Child Information Asymmetries on Students' Academic Performance: Evidence from a Field Experiment (I2)
AbstractInformation asymmetries between parents and children and lack of communication between schools and parents often create obstacles to parental involvement in children's education. This paper addresses two specific questions: (1) What is the effect of reducing information asymmetries on students' performance? (2) What is the mechanism that drives parents' and students' behavioral change? A school in Shenzhen, China, provided relevant data. Parents of a sample of 250 students were selected to receive information (both good news and bad news) about students' performance at school. Teachers of the treatment group sent a message about students' behavior at school to their parents every two weeks. The experiment continued for two semesters since the spring of 2019, and stopped in the spring of 2020 due to COVID-19. The Difference in Differences estimation method is used to evaluate the effects of the treatment. The results, however, are contrary to expectations and the literature: providing more feedback to parents is associated with lower test scores. Upward "belief bias" (parents having unrealistically favorable beliefs about their children's behavior at school) appears to explain this counterintuitive result. This result also adds to the literature of "good news-bad news" asymmetry with new supporting evidence from the perspective of economics of education.
The Global Transmission of Real Economic Uncertainty (F4, E3)
AbstractUsing a sample of 30 countries representing about 65% of global GDP, we find that real economic uncertainty (REU) has negative long-lasting domestic economic effects and transmits across countries. The international spillover effects of REU are both statistically significant and economically meaningful. Trade ties play a key role in explaining the spillover effects. Innovations to the foreign component of global REU can contribute up to 28% of the future variation in domestic industrial production, with the effect being disproportionately larger on its manufacturing component -- contributing the most to the tradable goods sector -- than on its retail sales component.
The Impact of Internal Oversight on Arrest and Use of Force (K4, Z0)
AbstractWhile there is much desire for holding police accountable for misconduct, there is little evidence on whether the most common accountability system, internal affairs, impacts police behavior in intended or unintended ways. Using data from a large city where there is conditionally random assignment of officers to 911 calls, I employ regression discontinuity and difference-in-differences methods to distinguish the impact of investigations from confounding factors. Results indicate that increased oversight from internal investigations does not change an officer's likelihood of making an arrest or using force. This is true across different types of allegations, including those that are sustained. Surprisingly, even imposing sanctions after a sustained allegation does not change police behavior, irrespective of the severity of the sanction. This has important policy implications, as it suggests that the current system of internal oversight has no impact on police behavior.
The Impact of Married Women's Property Laws on Female Labor Outcomes and Marital Status (J0, N0)
AbstractIn the 19th century, husband and wife were considered one person under the United States laws, which translated to the legal nonexistence of married women. Once women were married, they lost the right to own and control properties on their own and to earn wages. In contrast, single women had the same property rights as men. The Married Women's Property Acts are laws enacted by the individual states of the U.S. during the second half of the nineteenth century that granted economic rights to married women. While much of the research has focused on the effects of these laws on economic growth, fertility rates, and bargaining power, less work has been conducted to understand the impact on labor outcomes and marital status. I use spatial variation to identify the effects of this legislation on female labor participation and marital status. I also analyze heterogeneous effects across married and unmarried women. Preliminary results show that better economic rights increased the work incentives of all women. However, female labor participation marginally increased for married women. In contrast, the effect is more substantial on unmarried women. Additionally, marriage rates decreased, suggesting that women could have delayed their marriages to work more.
The Impact of Medicaid Expansion on Coverage among Those Lacking Housing Basics, 2010-19 (I1)
AbstractHousing quality plays a crucial role in that it can be harmful or beneficial
to our health in significant ways. Individuals with housing issues are typically identified as
having more healthcare needs such as infectious diseases. The purpose of this study is to
investigate disparities in health insurance coverage among individuals with various levels of
housing quality, excluding those who are experiencing homelessness. We use a difference-in-differences (DID) and DDD approach to estimate pre-post ACA insurance coverage between Medicaid expansion and non-expansion states, and estimate if such discrepancies vary across the following, according to if individuals live in homes that contain all basic necessities and those lacking at least one. For individuals with incomes below 138% of FPL, adjusted uninsured rates decreased by 6.68 percentage points more in Medicaid expansion states than in non-expansion states among those lacking at least one basic necessity. In addition, adjusted Medicaid coverage increased by 12.52
percentage points more in expansion states than in non-expansion states among those lacking at least one basic necessity. By comparison, there is a 2.69 percentage point greater adjusted increase in employer-sponsored coverage in non-expansion states than in expansion states. We also find that lacking at least one basic necessity individuals experienced a 1.43 percentage point larger decline in the probability of Medicaid, compared to those with complete necessities. It does not result in differential improvement for individuals lacking at least one basic necessity versus population with complete necessities for uninsured coverage, employer-sponsored and directly purchased private insurance for individuals with income up to 138% of FPL. Those living in homes lacking at least one basic necessity displayed significant improvements in insurance coverage.
The Impact of Sick Leave Mandates on Workers' Compensation Claims (I1, J3)
AbstractThe COVID-19 pandemic highlights a perennial policy debate over paid sick leave in the United States. This paper aims to provide new evidence on the spillover impact of sick leave mandates on workers’ compensation programs. Using administrative workers’ compensation claims data from California, we exploit county-level variation in the implementation of sick leave mandates. Specifically, San Francisco became the first city to implement a sick leave mandate in 2007, making 90 percent of employees with paid sick leave. We use a synthetic control model and a difference-in-differences model to estimate the causal effects. We find that access to paid sick leave reduces the number of workers’ compensation claims by about 6 percent, with significant decreases in the first two years of the mandate. Our results suggest that investments in workers’ health through paid sick leave programs could lead to positive spillover effects on another insurance program for employees.
The Impact of Social Security Wealth on the Distribution of Wealth in the European Union (D3, H5)
AbstractWe use a novel data source – the Eurosystem Household Finance and Consumption Survey – to compare the impact of the public pension system on wealth inequality in 19 European countries. In all investigated countries, social security wealth reduces wealth inequality, The augmented wealth inequality is ca. 30% lower than the private wealth inequality. Marginal effects of social security wealth identified through inequality decomposition are negative and statistically significantly different from zero.
The equalizing impact of social security wealth on wealth inequality is strongest in Austria, Estonia, Germany, Netherlands, and Cyprus. In Slovenia, Greece, and Croatia it is rather weak. The correlation between the position of the household in the distribution of private wealth and its position of the distribution of social security wealth is strongest in Latvia, France, and Italy, while it is weak in Lithuania, Hungary, and the Netherlands.
We investigate the determinants of equalizing power of public pension systems. The relation between the value of social security wealth and the value of private wealth, the correlation between both distributions, the share of single households, and homeownership taken together explain up to 60% of the variation in the equalizing power. A higher value of social security wealth relative to the value of private wealth increases equalizing power of the public pension system, while higher homeownership weakens the impact.
The Impact of Sodomy Law Repeals on Crime (J1, K0)
AbstractWe exploit variations in the timing of decriminalization of same-sex sexual intercourse across US states to estimate the impact of these law changes on crime through difference-in-differences and event study models. We provide the first evidence that sodomy law repeals led to a decline in the number of arrests for disorderly conduct, prostitution, and other sex offenses. Moreover, in line with the hypothesis that sodomy law repeals enhanced mental health and lessened minority stress, we show that these repeals led to a reduction in arrests for drug and alcohol consumption. Further analysis also suggests that our findings are aligned with a drop in the number of suicides among men following such repeals.
The Impact of Subsidies on Deductible Choice in Health Insurance (I1, H0)
AbstractThis paper analyzes how subsidies affect health insurance deductible choices among low-income adults in Switzerland. I use the presence in the kinked schedules in the relationship between previous earnings and subsidy level to identify the effect of subsidy in the regression kink design. Empirically, using the rich administrative data on health insurance choice, health care utilization, and socio-demographic factors, linked with a representative household survey data, I document varying demand for deductible coverages under financial constraints and subsidies. First, the result suggests that subsidies increase the demand for low deductible insurance contracts. I find that 40 percent of the subsidy recipients select the lowest deductible plan, compared to 30 percent in the high-income group. Second, exploring variation in the subsidy amount, driven by the policy changes across cantons, I use a differences-in-difference design to show that the effect of subsidies on deductible choice decreases dramatically as the subsidy amount decreases. And the impact becomes negligible when the subsidy is less than 40 percent of the premium. The results also show that the response to subsidies is not conditional on income level. In addition, I present evidence that in the absence of subsidy, when high-income adults face income shocks, they exhibit risk selection behavior. Furthermore, receiving subsidies does not affect the number of times of doctor visits.
The Increasing Penalty to Occupation-Education Mismatch (J3, I2)
AbstractCollege-educated workers whose job is unrelated to their field of degree typically suffer a wage penalty (Robst 2007). Using U.S. data from the National Survey of College Graduates, we document that, though the fraction of workers that work in a job unrelated to their field of degree has remained extremely stable over time (around 20%), there was a 57% increase in the wage mismatch penalty between 1993 and 2010, which narrowed to a 34% increase between 1993 and 2019. This increase in the mismatch penalty corresponds to an overall increase in the college wage premium during this time, consistent with mismatched workers having, in effect, a lower level of education. Results from occupational congruence confirm that mismatched workers are, occupationally, more similar to less-educated workers than to workers at their own education level but who are well-matched in their job.
The Lavish, the Wealthy, and the Healthy - Effect of Housing Wealth on Health Outcomes and Behaviors (I1)
AbstractHow does housing wealth affect people’s health outcomes and health behaviors? We study such effect on Chinese elderlies exploiting a discontinuity in housing wealth generated by two housing policies. These policies gave tax and down-payment breaks to owners of houses 90. Using objective biomarker information, we find that increased housing wealth has a deteriorating effect on lung functionality. Subjective health indicators also point to exacerbated cases of self-reported lung, stomach, and dyslipidemia diseases. One explanation for such discrepancy is that wealth not only has a direct impact on health, but also increases the possibility of having a health condition diagnosed. These results also echo our findings that increased housing wealth is accompanied by more frequent healthcare use and worsened smoking habits.
The Leverage-Liquidity Tradeoff of Mortgage Regulation (G5, E5)
AbstractWe evaluate the impact of mortgage regulation on household balance sheets, highlighting important trade-offs in terms of financial vulnerability. Using Norwegian tax data, we show that loan-to-value caps reduce house purchase probabilities, debt and interest expenses -- thereby improving household solvency. We show that the documented reduction in leverage is robust to also accounting for parental debt uptake, suggesting that concerns about regulatory arbitrage are unwarranted. However, the higher downpayment requirement leads to a persistent deterioration of household liquidity. Back of the envelope calculations suggest that the positive leverage effect is dominated by the negative liquidity effect, making household consumption less stable in response to shocks. We confirm that households with reform-induced reductions in liquidity have larger consumption falls upon unemployment. Our results indicate that while LTV-caps are effective in reducing household debt and aggregate credit growth, they are not effective in increasing household resilience.
The Liquidity Trap and a Price Dispersion Puzzle (E3, J3)
AbstractIt is well-known that price dispersion leads to inefficiency in the presence of sticky prices. In this paper, we report that in a textbook nonlinear general equilibrium model, the level of price dispersion rises by 100 times when the economy is in a liquidity trap, i.e., when the nominal interest rate reaches the so-called “zero lower bound.” This is sharply inconsistent with the data, which shows that price dispersion did not differ significantly before and after the Great Recession. We call this the “price dispersion puzzle.”
The price puzzle arises for the following reason. A large adverse shock puts the economy into a liquidity trap in which both prices and nominal wages fall sharply and, because the fall in nominal wages is even larger, real wages fall. A fall in real wages lowers firms' marginal costs, prompting them to further lower their prices, which results in huge price dispersion. This mechanism drives a large wedge between the transitional dynamics of GDP and labor. Specifically, the GDP decline is more severe, but the unemployment is less severe when compared with a recession without a liquidity trap.
We propose a simple approach to resolve or alleviate the puzzle: introduce downward wage rigidity. Empirical findings suggest that when the economy falls into a recession, nominal wages are extremely sluggish downward. When this feature is introduced into the model, it leads to a rise in real wages when prices fall, which raises firms' marginal costs, thereby easing deflation and resulting in lower price dispersion. We find that downward wage rigidity makes the size of price dispersion in normal times and the liquidity trap largely consistent, reconciling the model’s prediction with the relevant empirical literature. The reduction in price dispersion also removes the abnormal wedge between the dynamics of GDP and unemployment.
The Local Impacts of Chinese Infrastructure Investments in Africa (O1, F5)
AbstractChina's grand plan of infrastructure investment in Africa has been overshadowed by reports of civil protests and violent attacks recently. We present new evidence on the local economic and social impacts of Chinese infrastructure aid from 2000 to 2020. Exploiting panel data at the level of the 0.1° grid cell (around 11km×11km) over the entire African continent, our empirical strategy combines matching with an event study approach to trace changes in nightlight, protests and violent conflicts outcomes in gridcells with Chinese-financed infrastructure projects relative to suitable control ones. We find that the construction of these projects leads to an 11% increase in nightlight intensity, 0.6% decrease in violent conflicts incidence but a 5% increase in the incidence of protests. Further, we show that the protest effects are limited to regions occupied by ethnic groups less represented or excluded by the local governments, despite larger local economic benefits in these regions. The protest effects are also smaller for projects that hire more local workers, proxied by project type and local malaria exposure.
The Local Supply Channel of QE: Evidence from the Bank of England’s Gilt Purchases (E5, G1)
AbstractOne way QE purchases of government bonds by central banks may affect the yield curve is by creating scarcity in the purchased securities, leading to an increase in their prices or equivalently a reduction in their yields. We analyse and compare the importance of this so-called “local supply” (or scarcity) channel across all of the Bank of England’s QE government bond purchase programmes during 2009 to 2021. We find strong evidence overall for the role of the local supply channel in explaining gilt yield reactions both to QE announcements (“ex ante”), as well as after purchases have begun (“ex post”). The largest impact on the yield curve through local supply seems to have been in response to the initial QE1 announcements in 2009, both in terms of total impact (the impact of the announced programme), marginal impact (the impact of a given amount of purchases) and relative impact (the proportion of the total change in yields explained). Our findings also imply there may have been an increase in the relative importance of other channels and/or policies over time.
The Long Goodbye: the Economic Effects of a Parent’s Death (J1, J2)
AbstractEarly life events have the potential to divert individual life trajectories with irreversible consequences on child development. The death of a parent early in life is one of the most disruptive event for a child with at least two major consequences, such as prolonged grief and income losses. The aim of this paper is to analyze the effects of parental loss up to age 28 on longer-term outcomes such as education, mental health, and labor market across the different ages of the offspring. For doing this, we use the universe of the Danish population from 1980 to 2013, and focus on a cohort composed by those individuals born between 1980 and 1983 for whom we can identify parents in the data. This information is coupled with detailed information on the causes of deaths. We can probe the robustness of our results using accidental deaths.
We find that the death of a parent (either mother or father) has a substantial negative effect on income of the offspring at the age of 30-33 (typically considered a good proxy for permanent income) and we find that such effects are larger if parental death happened in the very early stages of life, suggesting a lower income of about 7,200 USD (13.4% of the mean income). The effects are then decreasing for later deaths.
We point out that such negative effects arise, and can on average be almost fully account for, by the loss of schooling. Furthermore, a deterioration in mental health seems to be extremely relevant for those who lost a parent early in life, with an increase in the probability of being admitted to hospital for mental health treatment being double that of the general population.
The Macroeconomic Impact of Euro Area Labor Market Reforms: Evidence from a Narrative Panel VAR (E6, C1)
AbstractLabor market reforms have been a central part of economic policies in the euro area during episodes of stagnation in individual economies. Important examples are the reforms to employment protection legislation after the 2008 Financial Crisis in countries facing fiscal stress and the German Hartz reforms a decade earlier.
The paper uses a narrative panel VAR to estimate the macro-economic effects of reforms in the euro area in between 1998 Q1 and 2018 Q4. It employs a new database that provides quarterly information on the legal adoption and implementation dates of 35 reforms events. Reforms are related to unemployment benefit schemes, employment protection legislation (EPL) of regular labor contracts, and legislation targeting temporary contracts.
The identification scheme is close to the proxy VAR approach used, for instance, by Mertens and Montiel-Olea (2018) for studying the effects of fiscal policies. However, given that labor market reforms are difficult to quantify, identification is solely based on the timing of reform events and neither requires the scale of events nor the precise timing of the resulting shocks to be known. Instead, the model provides estimates of the latter.
I find that unemployment benefit reforms lead to a quick increase in employment and a moderate decline in the real wage. In the medium term, the effect on employment remains persistent, while the real wage reverts back to baseline. The effects of regular contract EPL reforms are similar, but the response of employment builds up gradually and reaches its full scale only after about six years. The short-term effects of EPL reforms depend on the state of the business cycle: in states of low growth the response of real activity and employment is more delayed. Some reforms had sizeable medium-term effects, such as the German Hartz reforms and reforms in Portugal after 2010.
The Political Economy of Trade Deliberalization: A Social Identity Analysis of the US-China Trade War (F5, F6)
AbstractThis paper theoretically and empirically studies the political economy consequences of de-globalization in the context of the US-China trade war. I present a model of trade and social identity, which shows how Chinese citizens in high-trade regions may shift their identities from US-friendly to nationalistic in response to the trade war: the trade war reduced group economic interests and enlarge the perceived distance from being US-friendly. Exploiting variation in US-specific trade penetration across Chinese regions and the timing of the trade war, I find that the trade war had a larger negative impact on trust in Americans for Chinese citizens living in regions with a higher level of ex ante US trade exposure, and a larger positive impact on nationalistic sentiment, consistent with the theory. I also provide supporting evidence on the impacts of the trade war on the economic status of social groups (i.e., negative labor market shokcs), and the information search behavior of differentially affected groups (i.e., being exposed to more government propaganda or information not favoring the US).
The Shapley Value and the Nucleolus of a Two-Sided Platform Game (C7, D4)
AbstractThis paper introduces a new coalitional game with transferable utility — called a two-sided platform game. The participation of the platform user on one side benefits the other side, and the platform can be established if and only if there is more than one entrepreneur. Well-known point solutions and set solutions are investigated. It turns out that the kernel and the nucleolus coincide, and both the Shapley value and the nucleolus have simple expressions. The paper sheds light on platforms and antitrust issues. When there is more than one platform entrepreneur, the utility share of each entrepreneur is relatively low in both point solutions.
The Signaling Effects of Sovereign Borrowing (F3, H6)
AbstractWe provide empirical evidence for the signaling effects of sovereign borrowing on a country’s default risk. Using the S&P sovereign rating as a proxy for default risk, we find significant state-contingent effects of sovereign debt growth on the country’s rating, with the state being the country’s recent fiscal performance measured by its government operating balance. Conditional on a good fiscal state, higher sovereign debt growth significantly improves the sovereign rating, indicating a positive signaling effect of sovereign borrowing that more than compensates for its direct effect of increasing a country’s debt burden. Conditional on a poor fiscal state, higher debt growth significantly reduces the sovereign rating, even after the lagged rating, current government operating balance, sovereign bond yield, and other common determinants of sovereign rating are controlled for, which suggests a negative signaling effect of sovereign borrowing. We also provide a two-period model to rationalize these findings.
The Size and Census Coverage of the U.S. Homeless Population (J0, R0)
AbstractDespite widespread concern about homelessness, fundamental questions about the size and characteristics of this hard to study population remain unresolved, in large part because it is unclear whether existing data are sufficiently complete and reliable. We examine these questions as well as the coverage of new data sources that will allow ground-breaking new analyses of homelessness. We compare three largely unused, restricted use data sources to the less detailed public use data. In doing this triangulation of sources, we examine the completeness and accuracy of available data and improve our understanding of the size of the homeless population and its inclusion in widely used household surveys. Specifically, we compare restricted data from the 2010 Census, American Community Survey (ACS), and Homeless Management Information System (HMIS) to HUD's public-use point-in-time (PIT) estimates and the Housing Inventory Count (HIC) at the national, city and county, and person level. We explore the extent to which definitional differences, weighting methodology, frame completeness, and seasonality explain discrepancies between sources. We also link HMIS shelter use data to the Census to evaluate the usefulness of these microdata to study the homeless population. Our analyses suggest that on any given night there are 500,000-600,000 people experiencing homelessness in the U.S., about one-third of whom are sleeping on the streets and two-thirds in homeless shelters. Despite employing substantially different methods, the Census, ACS, and PIT arrive at similar estimates after accounting for definitional differences, ambiguity in the classification of certain facilities, and differences arising from the timeframe of Census response. The coverage of these sources is surprisingly good given the difficulties of surveying this population. This paper sheds new light on the usefulness of available data to learn about this extremely deprived group.
The Unintended Consequences of #MeToo: Evidence from Research Collaborations (I2, J7)
AbstractI study the impact of the #MeToo movement on research collaborations in Economics departments of U.S. universities. The intent of the #MeToo movement (2017) was to decrease tolerance for sexual harassment in the workplace. At the same time, a social movement like #MeToo may have unintended spillover effects that induce reluctance for women and men to interact with each other in the workplace. I hypothesize that the #MeToo movement will have a negative impact on collaborations between male and junior female faculty members. Research papers are outputs produced by a collaborative effort of researchers. Collaborations are formed on a voluntary basis, are of high intensity, and usually persist over long periods of time. As a discipline, Economics is still male-dominated. Professional customs have evolved that consider work-related interactions outside of the workplace and standard working hours as commonplace. This creates settings that are particularly high-risk for ambiguous situations to arise that may be viewed as sexual harassment. My results show that after the #MeToo, early-career junior female academics start 0.7 fewer projects per year than before. About 60% of this decline can be explained by a decline in collaborations with male co-authors. The most important driver is the decline in collaborations with new male co-authors in the same university who account for a total of 28% of the decline in new project initiations after #MeToo. This is the first study to provide evidence of unintended consequences of the #MeToo movement on the careers of women in a context where professional relationships are voluntary.
Tilting the Wrong Firms? How Inflated ESG Ratings Negate Socially Responsible Investing Under Information Asymmetries (G3, Q5)
AbstractSocially responsible investing (SRI) reduces aggregate sustainable performance when investors use Environmental, Social, and Governance (ESG) ratings. Due to information asymmetries, socially responsible investors shift their portfolios towards firms with high ESG ratings rather than those firms with exemplar sustainable performance. We show that this provides incentives for firms to reduce their cost of capital by inflating their ESG ratings. Particularly, Refinitiv, MSCI IVA, and FTSE ESG ratings are inversely related to sustainable performance, in part because promises of sustainable performance improvements do not materialize up to 10 years in the future. Subsequently, we show that firms reduce their cost of capital by inflating their ESG ratings due to the presence of ESG-rating-based SRI. This inefficient capital allocation increases the threshold for new sustainable initiatives.
Time to Baccalaureate Degree in the Labor Market: Evidence from a Field Experiment (I2, J0)
AbstractAbout 42 percent of bachelor's degree graduates take longer than four years to complete their degree. In this paper, I study whether the amount of time students take to complete their bachelor's degree affects labor market outcomes after graduation using a resume-based field experiment. I randomly assign a time to degree of either four or six years, as well as the selectivity of the public colleges where the degrees were received, to fictitious resumes of recent graduates where all other resume attributes are equivalent on average. I send over 7,000 resumes to real job vacancy postings for entry-level business jobs on a large online job board and track employer response rates. In the full sample of jobs, resumes listing bachelor's degree completion in six years received about 3 percent fewer employer responses than resumes indicating graduation in four years, but this difference is not statistically significant. However, for jobs with relatively large applicant pools, resumes listing six years to degree receive 17 percent fewer responses. Meanwhile, I estimate that listing a relatively more selective college increases response rates by about 13 percent, and by about 33 percent among higher paying jobs.
Towns and Rural Land Concentration in India (O4, R1)
AbstractUsing the universe of land records from a large state in India, we document three empirical facts on rural land holding concentration at the village-level: 1) rural land holding concentration is higher close to urban areas and decreases with distance, 2) the increase in land concentration near urban areas is due to fewer medium sized farms (i.e. more small and large farms near urban areas), and 3) the distance to urban area-land holding concentration relationship depends positively on the size of the urban area. A simple model where individual farmers face financial frictions, a U-shaped production function in land size and farm productivity, and significant opportunity cost of farm production, explains these patterns. In this set up, farmers choose between continuing or exiting farm production in each period. Owners of medium sized farms near urban areas find it profitable to exit production on such farms. Financial and land market frictions are key factors behind small farms remaining small.
Trade Union Membership and Life Satisfaction in Germany (J5, I3)
AbstractIf individuals join a trade union, their utility should be expected to increase. Accordingly, union members should also exhibit higher life satisfaction than comparable non-members. However, empirical evidence provides mixed evidence on the correlation between life satisfaction and the trade union membership status. Our paper aims to further develop the understanding of the relationship between life satisfaction and trade union membership in Germany.
We use the rich German Socio-Economic Panel data, which gives information on the union membership status in eleven survey years from 1985 to 2019. The findings suggest a negative relationship between life satisfaction and union membership in OLS and FE models.
We attempt to find channels through which this counter-intuitive relationship is being built.
Trends in Earning Volatility for U.S. Men: 1979-2017 (J3, D3)
AbstractThis article answers the question of how U.S. male income evolves over the course of time, ranging from 1979 to 2017. The research aims to decompose the income volatility into the permanent component − the long-term average − and transitory component − the period-specific deviation from the average − since the two have different implications in practice. After constructing a pseudo panel using the Current Population Survey, I estimate the structure of income volatility using an extended semiparametric model proposed by Moffitt & Zhang (2018). On average, the transitory variance explains about 70% of the total variance. The transitory variance fluctuated through the mid-1990s and declined until 2002. Since then, the transitory variance increased through 2013 and almost recovered to the level in the mid-1990s. The different trends of the transitory variance between the mid-1980s and the late-1990s across studies are not significant since it merely depends on samples and volatility measures. Furthermore, the increase in the gross volatility around the Great Recession is primarily caused by the increase in the transitory variance, which is consistent with previous studies even though the exact year of the turning point differs. The result indicates that income mobility and uncertainty on income have risen for men in recent years. Estimating the female income volatility and comparing it with that of the male are left for future research.
Two-Sided Discrimination in an Entrepreneurial Financing Setting: Experimental and Theoretical Evidence (J7, G2)
AbstractWomen’s participation rate in the high-growth entrepreneurship has been consistently lower than in other high- skilled jobs for previous decades. This paper aims to explain this unique persistent gender gap in both the US venture capital industry and the US entrepreneurial activities from the point of view of gender discrimination on both sides of a two-sided matching market (i.e., the US entrepreneurial financing market). We invite real US startup founders to evaluate multiple randomly generated venture capitalists’ and angel investors’ profiles. Despite knowing all investor profiles are hypothetical, startup founders are willing to provide truthful evaluations to receive a data-driven investor recommendation list. This experiment provides the following findings. (i) Male entrepreneurs have implicit gender discrimination against female investors, who are perceived as lower quality and with fewer investment intentions. However, gender discrimination does not exist among female entrepreneurs. (ii) Implicit gender discrimination mainly exists among attractive investors and senior investors, suggesting the existence of a glass ceiling for women in the financial industry. (iii) The magnitude of implicit gender discrimination is stronger when startup founders’ internal thresholds are higher, and investors need to compete for great deals. Together with Zhang (2020), this paper completes an experimental system that identifies gender discrimination on both the investor side and startup side in the US entrepreneurial finance system. Built on this experimental evidences, we provide a theoretical framework to explain several novel findings in recent experiments and how this two-sided gender discrimination can lead to a long-lasting low participation rate for women in the US entrepreneurial community.
Uncertainty, Citizenship & Migrant Saving Choices (D1, J1)
AbstractIn most Western countries, migrants hold significantly less wealth than natives. Migrants also face significantly more uncertainty about their future. This paper examines the central role of uncertainty over citizenship prospects and future location in explaining their saving choices. Exploiting quasi-experimental variation and panel data from Germany, I show that migrants with a right to citizenship save as much as comparable natives, while migrants without this right save 30% less. This unexplained gap is closed completely when migrants in the latter group gain access to citizenship. The effect is not driven by changes in resources, but rather willingness to save. While standard theory predicts that saving increases in uncertainty, I show that the effect can reverse if utility is state-dependent, malleable, or resources are not equally accessible across states. I build a life-cycle saving model with uncertain retirement location and heterogeneous country preferences. The model shows that agents can have a “preparatory saving motive” that decreases in uncertainty. I confirm the importance of this novel motive empirically, showing that migrants become significantly more likely to invest in illiquid assets if they gain certainty about their right to stay.
Under Control? Price Ceiling, Waiting, and Misallocation: Evidence from the Housing Market in Shanghai (R3, L5)
AbstractThis paper develops a model to study the general equilibrium effect of price control policies in the housing market in Shanghai. The government has imposed a price ceiling on new houses in Shanghai since 2017 to control the housing price growth. The proposed framework extends the existing literature by allowing the consumers to be forward-looking. Consumers can choose to wait, pay the entry/waiting cost, and re-enter the market if the houses are not allocated to them currently due to the excess demand. I assemble a new dataset that contains information about the sales and characteristics of the new and existing houses and the households' new house lottery participation records. The structural estimation results suggest that the welfare loss associated with the price ceiling is around 10.4 billion US dollars from 2018 to 2020. Entry/waiting cost accounts for 37% of the total welfare loss. Richer households face a smaller waiting cost, which impairs the distributional benefits of the price ceiling policy. Comparing these estimates with those from the model with myopic consumers highlights several important biases that arise when the forward-looking consumer consideration is ignored. I also use the model to study the welfare implications of alternative policies, including increasing the new housing supply and imposing an additional price ceiling on the existing houses.
Undergraduate Gender Diversity and Direction of Scientific Research (O3, J1)
AbstractCan diversity lead to greater research focus on populations underrepresented in science? Diverse researchers can bring new questions and perspectives, but exposure to diversity may also inspire scientists, regardless of demographic identity, to pursue new topics. This paper studies a potential determinant of research ideas: the diversity of the academic environment. Between 1960 and 1990, 76 all-male US universities, including many elite and prominent research institutions, transitioned to coeducation. Using a generalized difference-in-differences design, we document a 42% increase in the number of gender-related research publications authored by scholars at newly coed universities. This increase is explained by a combination of a more diverse researcher pool in terms of gender and prior research interests, as well as a shift in the research focus of individual scientists towards more gender-related topics. A bounding exercise suggests that the direct effects of the policy on scientists' research focus can account for more than half of these gains. These findings suggest that a diverse academic environment can influence the direction of scientific research.
Unemployment and Health Heterogeneity: The Role of Personal Finances (I0, J6)
AbstractRecent studies using large panel data have reached a consensus that unemployment reduces mental health and health satisfaction (Bartelink et al. 2020; Cygan-Rehm, Kuehnle, and Oberfichtner 2017). Various explanations have been proposed: unemployment not only reduces income but also displaces other noneconomic benefits of employment (Marcus 2013).
However, heterogeneity has been less explored. Two questions remain unanswered. First, is unemployment generally bad for health? Luo (2020) found that life satisfaction for half of the unemployed did not decline or even rise. Similar findings can be expected, as life satisfaction is highly correlated with health satisfaction and mental health. Second, which subgroup did not experience a decline in mental health or health satisfaction? In the literature, all subgroups were found to experience a decline in health (e.g., Salm 2009).
This paper aims to fill this research gap. Based on German Socioeconomic Group (SOEP) data and using mental health and health satisfaction as health indicators, this paper finds that the health of half of the unemployed did not decline or even rose. The paper also explores various financial metrics and finds that unemployed individuals in better financial circumstances may not experience declines in health.
In terms of identification, this paper adopts different strategies to address endogeneity. Individual fixed effects are designed to account for the bias caused by the selection of time-invariant unobservables, matching strategies are used to account for the bias caused by the selection of observables, and unemployment due to plant closures (arguably exogenous) is explored to address possible reverse causality. Furthermore, in the matching process, machine learning algorithm LASSO is used for variable selection.
This paper has strong policy implications. The large health heterogeneity in financial status suggests the importance of unemployment benefits. Increasing unemployment benefits can increase social welfare (Schmieder et al., 2012).
US Monetary Policy and Global Banks' Optimal Leverage (E5, G2)
AbstractI propose a model of global banks’ exposure to unexpected US monetary policy decisions, and the resulting accumulated burden in their balance sheet that leads to suboptimal capital structure in the long run. Banks are actively adjusting their capital structure in response to US FOMC announcements given the level of interest rate and exchange rate exposure in their balance sheet. An unexpected decision by the FOMC leads to a surprise jump in banks’ leverage ratio with permanent and transitory effect depending on their beliefs of future policy action. The model proposes a monetary policy channel that works through the impact of accumulated burden in banks’ balance sheet on their leverage ratio target, providing key insights on the level of optimal leverage for banks. The model speaks to the empirical observations on the leverage ratio of banks in 72 countries including the US.
Using Vouchers to Incentivize Public Transport. Mental Account Bias among Public Transport Users in a Developing Context (R4, H2)
AbstractWe implemented an RCT intervention in Bogotá, Colombia, to provide evidence of the effects of subsidies on public transport. Differently from current subsidies in the system that lower the fare per travel, we provided a monthly top-up of the personalized cards of public transport users. This corresponds to 25% of the average user's monthly expenses in transportation. This transfer led to an increase in the number of travels made in the system (+8%) and a net reduction in their monthly travel expenses (-17%). The reaction of the users to the transfer differs in systematic ways from the action of rational economic agents. Specifically, participants behave as having a Mental Account Bias. This is evident in trip increases only occurring when nominally coming from the transfer, in the higher-than-expected response to the transfer, and in the reaction to a chain store voucher at the end of the period. The presence of a Mental Account Bias can be leveraged to design policies that increase the use of public transport.
Variation in Monopsonistic Power in the Labor Market for High-Skilled Workers of Different Immigration Statuses (J6, J4)
AbstractA sizeable body of evidence has accumulated supporting the case of monopsony in the labor market. However, there has been few attempts to apply the monopsony models to the native-immigrant labor market and the monopsony literature has tended to focus on low-skilled labor. In this paper, I investigate the variation in monopsonistic power in the high-skilled labor market by measuring the differentials in labor market outcome and behavior based on citizenship and immigration statuses. The estimates from fixed effects model and match event study design show large differences in mobility, wage premium from sector separation, and the value of naturalization in the event of sector separation. I further provide evidence that exogenous institutions may increase monopsonistic power in the labor market for high skilled immigrants. However, I do find some evidence that assimilation is likely the biggest factor based on the comparison of foreign-born and naturalized high-skilled workers.
Wages and Prices in the Euro Area: Exploring the Nexus (E3, C3)
AbstractWe investigate the link between wages dynamics and core inflation in the euro area using Structural Bayesian VAR models. We find that the pass-through from wages to consumer prices net of food and energy is less than unity and somewhat similar across several shocks driving the economy, but of opposite sign for the case of demand and supply shocks. The estimated impulse responses suggest that firms set pro-cyclical mark-ups in response to both aggregate demand and monetary policy shocks, inconsistently with the mechanism underlying the standard New Keynesian model. Financial shocks, captured by credit spreads and indicators of systemic stress, are instead associated with negative co-movement between wages and prices and generate firms’ countercyclical mark- ups, consistently with theoretical models featuring financial frictions and nominal rigidities. These findings (i) highlight the importance of jointly modeling financial and labour markets and (ii) provide a novel explanation for the lack of sustained path of core inflation in the missing inflation period, in spite of the robust pick-up in wages growth. Our findings inform policymakers on the post-pandemic high inflation associated to strong recovery of labour markets: wage-price spirals could be only activated by aggregate demand or wage mark-up shocks, in contrast to the current prevailing adverse supply shocks.
Waiting or Acting: The Effects of Climate Policy Uncertainty (G3, Q5)
AbstractThis paper studies the effect of climate policy uncertainty on firm environmental performance. Firms reduce toxic emissions and close polluting facilities when the climate policy uncertainty is high. Contrary to the prediction of the real options theory that uncertainty delays investment, I find that firms respond to climate policy uncertainty by adopting abatement technology to reduce pollution. Further analyses suggest that climate policy uncertainty increases the likelihood of institutional holdings declines and regulation penalties in polluting firms. Exploiting the number of Congressional voting on the climate change topics to instrument the climate policy uncertainty index, I argue that the effects of climate policy uncertainty on the reduction of toxic emissions and the closures of toxic facilities are likely causal.
Welfare Programs, Eligibility Tightening, and Financial Distress: Evidence from Automating the Indiana Welfare System (H4, I3)
AbstractThis paper investigates the effect of losing welfare benefits on local household financial distress and crime. We estimate this effect using a quasi-experiment in which the state of Indiana outsourced and automated the processing of TANF, food stamps, and Medicaid applications. The welfare automation policy was implemented in three waves for different counties prior to its cancellation in 2009 and before it reached all counties. Exploiting the staggered rollout of policy and consumer credit panel data, we find that the Indiana welfare automation program, which reduced enrollment in SNAP and TANF enrollment, significantly increased the number of accounts in collections, collections balances, bankruptcy filings, and decreased credit scores. Using data from the Uniform Crime Reporting series, we find that welfare automation policy has also led to increases in crime, primarily property crimes and crimes motivated by financial gain.
What Determines the Duration of Biologics License Application Approvals (I1, O3)
AbstractThe Biologics License Application (BLA) approvals differ significantly from New Drug Application (NDA) by nature, and progressed in a surprising speed. To better understand this quickly emerging field of chemical compounds, the effect of firm scale advantage and lobbying behaviors on the approval speed merits investigation. Overall, I look into the role that different factors play in affecting the approval speed for BLAs, and find that (1) priority review clock and orphan drug designations help reduce the waiting period; (2) no evidence of firm scale or rent-seeking effects; (3)past experience with FDA in NDA approvals is negatively associated with the approval duration when firms apply for BLAs.
What Do Bond Investors Learn from Macroeconomic News? (E4, G1)
AbstractMacroeconomic data releases drive US bond yields primarily through the term
premium instead of the expectation channel. The evidence exploits a monthly specifi-
cation for yields embedding the impacts of news identified from high-frequency data.
To match the facts, we develop and calibrate a no-arbitrage model where investors
use data releases with imperfect information to learn about future monetary policy. If
macro news carry perfect information, the model predicts that the bonds’ Sharpe ratio
decreases and the term premium declines by half for every maturity, suggesting that
central bank’s communication can lower the term premium and financing costs across