July 2, 2024

The political consequences of NAFTA

Jiwon Choi and Gavin Wright discuss the economic and political fallout of the North American Free Trade Agreement.

President Bill Clinton participates in a NAFTA products event.

Source: National Archives and Records Administration

In 1993, the North American Free Trade Agreement (NAFTA) was passed with bipartisan support and near-universal endorsement by economists. In hindsight, the economic costs and political consequences were far greater than many contemporary observers would have imagined.

In a paper in the American Economic Review, authors Jiwon Choi, Ilyana Kuziemko, Ebonya Washington, and Gavin Wright found that US counties most exposed to NAFTA and Mexican import competition saw their total employment drop by roughly 6 percent compared to those with little exposure to the trade deal. However, workers in these communities didn’t respond by moving away to find better opportunities, and many, feeling betrayed by the Democratic party, embraced the Republican party instead. 

Choi and Wright recently spoke with Tyler Smith about the economic and political history of NAFTA and what economists have learned since its passage.

The edited highlights of that conversation are below, and the full interview can be heard using the podcast player.



Tyler Smith: Can you explain what NAFTA was and what its backers hoped to achieve at the time?

Jiwon Choi: The North American Free Trade Agreement or NAFTA was a trade agreement between the US, Canada, and Mexico that was passed in 1993 but enacted in January 1994 under the Clinton administration. I want to also highlight that because the US–Canadian trade was already mostly tariff free with a preexisting free trade agreement, the biggest impact NAFTA had was the change in trade relations with Mexico. The agreement included a lot of elimination of tariffs on Mexico's exports to the United States—more than half of the export items. NAFTA was a pretty salient political topic during the 1992 presidential campaign.

Gavin Wright: NAFTA was a big deal. It was very controversial. It was at the heart of Ross Perot's third-party candidacy in 1992, and he got 20 percent of the vote, which is an extraordinary development in American history. I also remember the famous debate between then Vice President Al Gore and Ross Perot. I think all observers agreed that Al Gore, the pro-NAFTA candidate, won handily. And almost all professional economists were solidly in favor of NAFTA, including the idea that it was not just good for the nation as a whole, but that it was going to be good for virtually everybody, allowing a little time for adjustment.

Smith: One issue here is that employment can change for many reasons. The economy is always going through business cycles and technology is always changing. How do you isolate the impact of a trade deal?

Choi: It is definitely a challenging task, but we try to put our best foot forward. Basically we use an event-study approach where we compare US counties with more and less exposure to NAFTA. What this means is that we first create a county-level measure of how much each US county is exposed to Mexican import competition. And then, once we have that measure, we are able to do the comparison. We create this measure using two aspects of counties and industries. One is the exposure to import competition, incorporating how each industry experienced a decline in tariffs against Mexican imports with NAFTA. And the second is using information about how each county relied on different industries. 

Smith: Sticking with the economic side, when you applied this methodology, what was the impact of NAFTA on these counties that you compared?

Choi: Focusing on the employment results, we compared counties in the top quartile of trade vulnerability to counties in the lowest quartile of vulnerability. And we found that counties that were more exposed to Mexican import competition experienced a significant decline in employment, roughly 5 to 7 log points by year 2000 compared to the least exposed counties. That was our main employment result.

Smith: What kind of counties were impacted the most? In particular, what kind of industries did they have?

Wright: I think the big industries are textiles and apparel. It’s a little hard to describe, but there may be four or five states in which that was the largest employing industry. I think it had been threatened by foreign competition for decades, but it had been protected against that foreign competition. It was the kind of thing that economists would say is not sustainable. The economy was moving more toward new industries. It was understood that there were going to be job losses. But what we bring out in the paper as well is that these areas not only suffered, they continued to suffer. We do not observe significant outmigration. We've long thought that America has a high-mobility population. But it does seem that at least workers in these particular old-line industries were not able to move very easily. That's part of the reason why I think the political implications were so strong. These people really didn't have alternatives out there.

Smith: How does this economic shock fit into US political history? How are people in these communities responding to NAFTA and this loss of employment?

Choi: In terms of the political effect, we could also measure the impacts using a similar approach. We could look at these counties and see how they changed their support for Republicans and Democrats or third-party candidates before and after NAFTA. And we find that counties that were more affected or more exposed to NAFTA were likely to increase their Republican support around the time of NAFTA, so from 1994 on.

We found that counties that were more exposed to Mexican import competition experienced a significant decline in employment, roughly 5 to 7 log points by year 2000 compared to the least exposed counties.

Jiwon Choi

Smith: Why do the Democrats get blamed for NAFTA?

Wright: You really have to embed the NAFTA episode into a larger narrative in which the Republican Party was adapting its rhetoric in the South to better match the historical prejudices of White Southerners. Black Southerners did not become Republican as a result of this. Also, the people associated with organized labor felt betrayed. Labor was opposed to NAFTA. And they felt they finally got their guy, Clinton, into the White House, and they felt that he let them down. They were angry. So we distinguish between two kinds of effects. There was an immediate effect, which we understand as people who were watching the politics, who were very angry at the NAFTA vote. They were going to either just not vote for Democrats, stay home, or maybe even stick it to them by voting for Republicans. And then there was a more gradual effect as the economic impacts worked their way out over time. In political science, the simplest kind of voting model is one that says if voters are in hard times, they blame the people who are in power and vote for whatever the alternative is. I'm sure real life is much more complex than that simple model, but it captures quite a bit of people's response. As I see it, probably the idea to actually identify as a Republican was a more gradual process. But NAFTA was an immediate shock.

Smith: You mentioned at the beginning of this conversation how overwhelmingly pro-trade economists were during the time of NAFTA. Do you think there are any lessons for economists from this research?

Choi: Our paper does not answer whether NAFTA was good for the US economy as a whole. Actually, there are other papers that already documented that NAFTA may have had an overall, but small, net positive impact on the US economy. We're really measuring the relative growth and changes across counties. What we're documenting is a potential increase in spatial inequality in the US that connects to these political responses. I think our work supports this consistent pattern that when there is a local economic downturn, we see that people do not adjust as much as economists traditionally think.

Wright: I would say probably the single most important component of our answer is we are specifically not trying to draw direct policy conclusions about trade policy from this one episode. In that respect, I think we are following in the footsteps of David Autor, David Dorn, and Gordon H. Hanson, who wrote about the China shock. We're saying that they’re right about the pattern of that shock, but it started earlier. It really started with NAFTA. And NAFTA was far larger in terms of the loss of manufacturing jobs than economists predicted at the time and in terms of the length of the effect. But if there were any policy conclusion, it would be that if you're going to do a policy of this kind, you should be much more sensitive to the downside to those who are on the losing end. Society as a whole should be more aware of these costs and their significance, and more prepared to do something about it.

Local Economic and Political Effects of Trade Deals: Evidence from NAFTA” appears in the June 2024 issue of the American Economic Review. Music in the audio is by Podington Bear.