What the “Trump tax” cost Republicans

On Wednesday, Ezra Klein argued that the 2018 elections showed that Republicans are paying the “Trump tax,” suffering at the polls because they are attached to a president who is surprisingly unpopular given national political and economic conditions. I think this is right, but I wanted to expand a bit on what it means and try to put a “price” on that tax.

The Trump tax is conceptually the difference between where President Trump’s approval ratings are and where a more typical Republican president’s would be given national conditions. A Vox analysis in 2016 suggested that Trump was running several points behind a generic Republican presidential nominee in that election.

As John Sides, Michael Tesler, and Lynn Vavreck demonstrate in their new book Identity Crisis, this tax has persisted during Trump’s presidency. As of the beginning of 2018, they write, “His ratings were mired around 40 percent, whereas a president presiding over similarly favorable consumer sentiment would typically be polling near 60 percent.”

What would that have meant in terms of the 2018 midterm elections? My forecast model uses presidential approval, economic growth, and current party seat shares to predict how the president’s party will fare in House elections. It’s a noisy model (although I got pretty close this year), but it suggests that each additional point of presidential approval (as measured by the Gallup approval rating at Labor Day) translates to about 1.1 saved House seats for the president’s party.

This year, Trump’s Labor Day approval rating was 41 percent. What if, as the Sides et al. estimate suggested, it was at 60? If we take this estimate and the regression coefficients literally, that would translate to 21 saved House seats. In other words, instead of losing an estimated 37 House seats, Republicans would have lost just 16 and kept control of the House. Indeed, if Trump had just been 14 points more popular than he is now, that would have been enough for Republicans to hold the House.

Now, I can’t say with certainty that this is what would have happened if Republicans had nominated, say, Jeb Bush or Marco Rubio or someone else in 2016. The relationships between economic growth and presidential approval, and presidential approval and midterm seat losses, are likely quite real, but measuring them precisely is difficult, and a different president might have underperformed, or overperformed, historical norms.

Nonetheless, these estimates are pretty reasonable and backed by historical evidence, and they suggest that the tax that Republicans paid cost them dearly. The bulk of their legislative agenda has been put on hold, and Democrats now have the ability to subpoena Trump administration staffers and family members.

A different Republican presidential nominee in 2016 likely would have won (by a more comfortable margin) and likely would have delivered at least as much for Republicans as Trump has in terms of policy goals. And such a president likely would have cost Republicans a lot less.