A Union flag flies above an English St. George’s Cross flag near the Big Ben clock face and the Elizabeth Tower at Houses of Parliament in central London on June 29, 2016. (BEN STANSALL/AFP/Getty Images)

Two weeks ago today, Britain voted to leave the European Union. Why?

Many have said that immigration is the reason, including some here at the Monkey Cage. And in fact, that’s what many  Leave voters are telling pollsters; a large majority say they support restrictions to immigration from E.U. countries regardless of the consequences.

Others, including former prime minister Gordon Brown, say that the real reason is the displacement caused by globalization. Our evidence agrees.

Regardless of what voters or pundits might be saying, we find that Leave votes were systematically higher in regions more affected by the surge in Chinese imports over the last three decades. And we find no evidence that the presence or influx of immigrants correlates with a region’s support for Brexit.

Here’s how we calculated “import shock.”

Based on the official referendum results we computed the share of Leave votes in 39 administrative regions, which have on average around one million inhabitants each.

Then we constructed a region-specific indicator for how much the region’s employers were affected by Chinese imports, using a methodology proposed by economist David Autor and his coauthors that is now becoming standard in trade economics. The idea behind this indicator is straightforward. The Chinese shock has been stronger in those regions with a larger share of workers in industries competing more and more with imports from China, like textiles or electronic goods.

We measured industry specialization in 1989, before the emergence of China as a global manufacturing player. We then looked at import growth between 1990 and 2007, to avoid picking up the complicated ramifications of the 2008 global financial meltdown. This also reassures us that our analysis is picking up associations (or causal effects) that come from long-term shifts in the British and global economy, rather than simply from one particularly large shock.

We found a strong and statistically significant relationship between the strength of the import shock and the Leave share in the referendum.

How much of the actual difference in the Leave vote share between any two regions can be attributed to differences in intensity of the “import shock”? The effect is quite large.

Let’s look at Inner London, where 28 percent voted Leave, and Derbyshire/Nottinghamshire, where 56 percent did. That’s a difference of 28 percentage points in support for Brexit. According to the data and our statistical analysis, of these 28 points, at least 18 are attributable to the difference in the intensity of the “import shock” between the two regions.

Along the same lines, Merseyside, where 49 percent voted Leave, and Herefordshire-Worcestershire-Warwickshire, where 57 percent supported Leave, differ by eight percentage points. Of these, at least three points are explained by their difference in the intensity of the “import shock.”

Chinese import shock and support for the Leave option in the Brexit referendum, by region. Data: Eurostat Comext and Electoral Commission UK. Figure: Piero Stanig and Italo Colantone.

Significantly, we  obtained essentially the same results when we used a different method to calculate the “import shock.” In this case, as a measure of China’s manufacturing power in a given industry, we used exports from China to another advanced economy: the U.S.  By doing this we captured the effect of Chinese competition for British manufacturing not only within the U.K., but also on global markets.

We’ve done a more general analysis  for western Europe as well, calculating “import shock” for each region across 15 countries, and find that something similar to support for Brexit is happening. Wherever a region’s industries have been hit by increased Chinese competition, voters have drifted toward parties that appeal to national self-sufficiency, protection from international trade, and identity-based nationalism — like the National Front in France and the Northern League in Italy.

Here’s how we calculated “immigration shock”

Could immigration be another explanation? To check, we took data from the U.K.’s Office of National Statistics about both the number and the incoming flow of immigrants for each region in Great Britain. We computed the number of immigrants (defined as residents who were born outside of the U.K.) as a proportion of the total population, and the new arrivals in 2014, the most recent year available, based on registrations to National Insurance (for U.S. readers, that’s roughly equivalent to applications for a Social Security number). We then calculated the new arrivals as a share of the working-age population.

We found no evidence of correlation between support for Brexit and the proportion of either immigrants or new immigrants in a region. If anything, it seemed that areas with more arrivals were more likely to vote Remain, and areas with fewer arrivals supported Leave. For instance, in the region with the lowest rate of new arrivals — the Tees Valley and Durham – voters supported Brexit by a remarkable 60 percent, while London especially voted to remain in the E.U. and also contains the five regions with the most new immigrants.

But that’s probably not causal. New immigrants may settle in areas with more jobs, or areas that particularly welcome immigrants – and in either case, those areas might already be more likely to vote Remain.

In fact, when London (which contains five regions) is included in the analysis, we find a negative correlation between arrivals and support for Leave — meaning that on average, regions with more new arrivals have lower support for Brexit. But once we remove those five regions from the analysis, there is really no pattern left in the data. This means that the correlation is driven by London, which is historically more cosmopolitan and diverse. For this reason we do not want to make much of this negative correlation — but are, at a minimum, confident that there is really no detectable relationship between how many immigrants arrived in recent years and how much support the Leave option received in the referendum.

Immigrant arrivals and support for the Leave option in the Brexit referendum, by region.
Data: Office for National Statistics UK and Electoral Commission UK. Figure: Piero Stanig and Italo Colantone.

It’s globalization, not immigration

So what’s the most likely explanation for the Brexit vote? It’s economic globalization, not immigration. Displaced British manufacturing correlates far more strongly with a pro-Brexit vote than immigration does, no matter what voters say after the fact. They might not realize it, but voters may actually want to leave the World Trade Organization or the global economy more than they want to leave the E.U.

As Gordon Brown suggests, governments’ inability to compensate globalization’s losers seems to have pushed voters toward political entrepreneurs who cast themselves as isolationist and nationalist, and who blame immigrants for woes that probably have little, if anything, to do with immigration.

Italo Colantone is assistant professor of economics at Bocconi University. Piero Stanig is assistant professor of political science at Bocconi University.