Was it inflation that swung it for Trump?

Unlock the White House Watch newsletter for free

Your guide to what the 2024 US election means for Washington and the world

The ‘it’s inflation, stupid’ thesis is strong. As MainFT highlighted in a big piece days before the vote and again on election day. Inflation was clearly having a major effect on Kamala Harris’s chance of being elected.

But an analysis of geographically granular US inflation and the actual election results suggests that inflation didn’t have a straightforward effect on the swing to Trump.

If it did, places which suffered higher inflation would have been more likely to see increased Trump share of the vote over 2020. This is not what the data shows, however.

Cumulative CPI inflation since 2020 at the metropolitan area — the most granular measure of inflation released by the Bureau of Labor Statistics — is not associated with the 2024 Trump swing.

Instead, there is a slight negative relationship: cities which saw higher inflation experienced smaller swings to Trump.

The picture changes when you instead plot cumulative CPI against Trump’s vote share. Here, there’s a clear positive association — cities that saw higher inflation also saw a higher vote share for Trump in 2024.

But the same relationship holds for Trump’s 2020 vote share, when he lost to Joe Biden.

Granted, the sample size is small. We only have both CPI data and voting results for 21 metropolitan areas.

A larger sample comes from the Bureau of Economic Analysis, which releases implied inflation rates for over 300 metro areas in the US, albeit with a long time delay (the latest data is for 2022, when inflation peaked in the US).

Just as with CPI, cumulative PCE inflation by metro area shows no relationship with the Trump swing, and there’s also no relationship with Trump’s vote share in 2024.

This is puzzling. Exit polls and pre-election surveys show inflation was a major concern, and academic evidence is clear that people truly hate inflation and tend to blame the government for higher price levels no matter who or what is actually to blame.

So why didn’t areas that saw higher inflation swing more to Trump?

It’s not as if variation in inflation across areas is insignificant. Inflation differences are remarkably large — Tampa Bay saw close to 30 per cent cumulative CPI inflation since 2020, nearly double that of San Francisco, for example.

What then? The answer to this puzzle might be that objective, aggregate inflation — as measured by CPI or PCE at the city or country level — is a very different thing to both perceived inflation and experienced inflation.

First, on perceptions. Republican voters consumed a lot more negativity about the US economy over the last several years, and this showed up clearly in sentiment.

A survey by Pew ahead of the election found a 20 percentage point gap in concern around prices between people identifying as Republicans versus Democrats. And research shows that Republicans are much more likely to blame the government for inflation.

From this standpoint, the direction of causality runs in the reverse direction: support for Republicans shaped views on how bad inflation was and who was to blame.

But while perceptions might be 9/10th of reality, inflation was objectively high all over the US.

The other point is that aggregate measures of inflation obscure large differences in inflation experiences by households. We all have very different consumption baskets, and hence very different actual inflation rates.

One notable example are typical differences in inflation by income group. The poorest typically suffer the highest inflation pain, and the recent episode was no exception — the lowest income quintile saw two percentage points higher inflation on average since 2019.

Perhaps that helps explain why lower income voters were more likely to vote for Republicans over Democrats at this election, for the first time in decades.