Trump and Chetty

There’s been a lot of talk about whether Trump’s success in the primaries reflected genuine economic distress on the part of voters or mere racism and xenophobia.

It’s a complicated question, and my own tendency is to see Trump as the result of general cultural disruption and change—Coming Apart/End of the Bourgeois Era stuff, which is connected to economic changes within and between families as well as cultural changes around race but not exactly the same as economic distress per se. A Gallup study of Trump voters found that:

According to this new analysis, those who view Trump favorably have not been disproportionately affected by foreign trade or immigration, compared with people with unfavorable views of the Republican presidential nominee. The results suggest that his supporters, on average, do not have lower incomes than other Americans, nor are they more likely to be unemployed.

But I did remember this map of upward economic mobility among whites from Raj Chetty’s Equality of Opportunity project and how much it looked, with some exceptions, like a map of Trump’s vote share of the GOP primary.


So I downloaded some of the descriptive statistics Chetty got from IRS data and while there’s a lot of state-to-state variation, if you just look at county-level (within state) variation, there’s a pretty consistent story you can tell with just three economic variables, that explains most of the variation (see fixed effects regression below.)

  • In a county where kids who grew up in the 80s and early 90s at the 25th percentile of the national income distribution were doing better at age 26, Trump did worse. (This is more-or-less the variable that the map above shows.)
    • On average, for each $2000 more kids like this are making at age 26, Trump got about 1% less of the vote.
  • In a county where kids who grew up in the 80s and early 90s at the 75th percentile of the national income distribution were doing better at age 26, Trump did better.
    • On average, for each $6000 more kids like this are making at age 26, Trump got about 1% more of the vote.
  • In a county where parents were making more money in the 80s and 90s (the average household income of the parents of the cohort Chetty followed was higher), Trump did worse.
    • One average, for each $5500 more (in inflation-adjusted dollars) parents like this were making in the 80s and early 90s, Trump got about 1% less of the vote.

So you can tell a story based on this. Places where working class kids are doing pretty well are places Trump did comparatively badly. Places where middle and upper-middle class kids are doing well are places where Trump did well. Places that are and have been well-off for decades are places that Trump did badly.

Is Trump’s rise the result, in part, of working class parents who are not themselves badly off seeing their kids fall behind the kids of more well off parents? It doesn’t seem that far-fetched.

Note: You can reproduce the regression using Chetty’s data here (I rescaled the variables to be a bit more intuitive, but you just use kid_fam_inc_p25, kid_fam_inc_p75, and par_fam_inc_mean). I took the county-level results from the 2016 primaries from Kaggle (you have to make a Kaggle account to download.)


areg trumpperc  Kidfrom25thPercentile Kidfrom75thPercentile ParentIncome, absorb(state_abb)


Linear regression, absorbing indicators         Number of obs     =      2,389

                                                F(   3,   2342)   =     298.93

                                                Prob > F          =     0.0000

                                                R-squared         =     0.8700

                                                Adj R-squared     =     0.8674

                                                Root MSE          =     5.6040



            trumpperc |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]


Kidfrom25thPercentile |  -.5429087   .0552855    -9.82   0.000    -.6513223    -.434495

Kidfrom75thPercentile |   .1623044   .0262649     6.18   0.000     .1107995    .2138094

         ParentIncome |  -.1790755   .0076783   -23.32   0.000    -.1941325   -.1640184

                _cons |   56.72411   .9796469    57.90   0.000     54.80305    58.64518


   state_abbreviation |       F(43, 2342) =    346.872   0.000          (44 categories)

3 thoughts on “Trump and Chetty

  1. The exceptions are pretty important: Pennsylvania. New Jersey. Massachusetts. Connecticut. Rhode Island. Upstate New York. West Virginia+Eastern Kentucky.

    To understand Trump’s success, one must first understand White ethnic voting (especially in the Northeast), as well as this:

    Trump did well because he was viewed as best on the economy, immigration, the budget deficit, and doing what he sets out to do (winning).

    To Northeastern non-Hispanic Catholic Republicans, Trump is probably more popular than the Pope.


    1. Yeah, I wish I had better data available on the age distribution of Trump voters; from what I can see they are probably right around the median income for non-Hispanic whites of their age group and below the median if you control for marital status. I’ve heard arguments back and forth about Catholic and Protestant support. Overall, Trump clearly did extremely well in the Northeast. (One of the variables from Chetty’s data that has a lot of negative predictive power for Trump support if you look across states but not within states is % of girls who get pregnant before age 20; Trump does better in states without teen pregnancy, but within states it’s not predictive.)

      Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s