Did the trade war flip congressional districts in 2018?


The pains of the US-China trade war are now showing up in prices and employment around the country. (A useful state-by-state tracking device is provided the US Chamber of Commerce on their "Wrong Approach" site.) 

But how much effect, if any, did the economic impact of the trade war have on the midterm elections? 

There have been varying views expressed. One analysis pointed out that 10 of America's top 19 soybean districts were held by Republicans. Others have noted that a Democratic House is likely to delay, or veto, USMCA.

To resolve this, we first have to get a sense of how relevant the trade war is for a given district. One way to measure relevance is to look at how many employees are implicated directly in the tariffs. A previous post here mapped affected employees by US county. We can convert those county lines into congressional districts. That allows us to see whether the districts that flipped from Republican to Democrat had more employees named in China's retaliation. The logic is that those voters, directly subject to trade retaliation, should be more likely to sanction the Republican Party at the voting booth.

The following table lists 30 districts that flipped from Republican to Democratic, as reported by Vox and Axios.


There is very little evidence that "flipped districts" were out of step with their states or with the country. At least not according to this one measure. 

Nationwide, the average district employment affected by China's tariffs was 1.25 percent (SD: 1.3). In flipped districts, the average is a slightly lower 1.13 percent (SD: 3.01).

The actual number of employees affected is also comparable across the two groups. The average district had 3,492 employees named. Flipped districts average 3,744. Not a huge difference. 

Can we conclude the trade war didn't matter in the midterms? There are at least four things to consider.

1. The lack of clear trend is partly due to the small numbers. Even using the best available data on industry of employment, we see that relatively few voters are directly affected by China's tariffs. The nationwide average is only 1.2 percent, or about 3,400 workers per district. That doesn't translate into many votes. 

2. At the same time, there are employees enjoying protection from US tariffs. Their preferences, presumably, are to support Trump and Republican Party.

3. Employment is only one measure. It fails to capture the full economic impact of the trade war on these districts. Number of employees does not necessarily capture the economic importance of a given industry. And it certainly does not capture the costs incurred by local consumers.


4. It's possible that the election results had nothing to do with the trade war at all.

Whatever the explanation, it will be interesting to see how trade policy moves forward. Trump's position has been more consistent with traditional Democrats on trade. Now, the House of Representatives will have to decide whether to stand up to Trump on trade.