What Economic Forecasts of Trump’s Re-Election Are Missing
(Bloomberg Opinion) -- Election forecasts that use economic conditions rather than polls of voters suggest that President Donald Trump’s re-election chances are much higher than his abysmal approval rating suggests. Take them with a grain of salt. The U.S. economy is steady overall, but conditions in several battleground states continue to deteriorate — and so will Trump’s chances if those trends persist.
The underlying premise of these models is that the economy matters a lot more, and candidates a lot less, than most people think. That’s probably correct. Trump was unpopular with much of the electorate in 2016, and virtually all conventional polls predicted Hillary Clinton would win. By contrast, economy-centric models predicted a Republican victory based on a mini-recession in 2015.
The intuition is straightforward: Most voters are not up for grabs and simply support their party’s nominee. The few voters who are genuinely undecided are either unaware of or immune to political rhetoric, and cast their ballot based on what they see in their own lives. That view is dependent on the state of the economy.
The flaw in this set of models is that they fail to account for regional variations in the economy. They also tend to focus on a narrow snapshot of conditions or ignore the importance of the Electoral College.
So what are the models saying now about the 2020 election? Two look only at the popular vote share between the two major parties. One, by the economist Ray Fair, plugs in the latest economic data to give a sort of horse-race analysis. That shows Trump winning by four percentage points. The other, by Oxford Economics, projects what economic conditions will be in fall 2020 and projects a five-point win for the president.
A third model, from Moody’s Analytics, gives a more detailed analysis of the Electoral College vote but relies heavily on such measures as gas prices and the stock market, which can change dramatically. This model currently predicts that Trump could win as many as 351 electoral votes.
Again, these models are a valuable addition to pure political polling. That said, a look at trends in crucial battleground states shows the outlook for the president is far murkier. If he is going to retain the White House in 2020, Trump will have to hold on to at least three of six battleground states that he won in 2016: Michigan, Pennsylvania, Wisconsin, Florida, Arizona and North Carolina.
In Arizona and Florida, the economy is doing well. Both states have seen job growth over the past year that was significantly higher than the national average. In Florida in particular, job growth has been trending upward.
Things are a bit less bright in North Carolina. The latest reading on job growth was strong, but it appears to be an outlier. Overall, North Carolina’s job growth is tracking the national average — declining but strong enough to support the case for re-election. Crucially, while North Carolina’s economy has been traditionally dependent on manufacturing, it has made the transition to finance and technology, areas that are less affected by global trade.
It’s in Michigan, Pennsylvania and Wisconsin that real trouble awaits the president. Job growth is not only slower there than the national average, but it is also declining steadily. Those economies are still heavily exposed to trade, in terms of both sales and investment. Just as important, Trump’s razor-thin margins in those states means that he has little room to spare.
These economy-centric models are a useful corrective to the view that Trump’s poor polling and the results of the 2018 midterms mean that a Democrat will win the White House in 2020. It would be a mistake, however, to underestimate the president’s weakness on the economy in several crucial battleground states.
To contact the editor responsible for this story: Michael Newman at mnewman43@bloomberg.net
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Karl W. Smith is a former assistant professor of economics at the University of North Carolina's school of government and founder of the blog Modeled Behavior.
©2019 Bloomberg L.P.