When President Trump’s job approval numbers began their nose dive in earnest in mid-May, a friend asked me what I thought was going on. After all, at that point things seemed to be going well for the president. COVID-19 case counts were declining amid the country reopening, and polling suggested that the public believed the worst was behind us. We didn’t know it yet, but the economy was on track to reduce the unemployment rate. The George Floyd protests were still a few weeks away. Why were Trump’s job approval numbers in decline now, of all times?

The answer was straightforward. For years political science models have focused their attention on job approval and economic numbers from the second quarter (April through June of a year), because it takes a while for good news (or bad news) to sink it. Consider the 1992 presidential election. Unemployment peaked in June of that year, and then gradually began to trickle downward. October saw a stellar jobs report, with the number of unemployed declining by almost 400,000. But it was too late for the incumbent.

This isn’t to say that there is no room for events to affect things if they take place in September or October. It is only to say that for most people, it takes a while for news to sink in, so news in September or October matters less than in news from April or May. 

So the swoon in Trump’s job approval numbers probably reflects the news from two or three months ago, which was very bad indeed for the president. It also suggests, though, that there is some room for turning things around. Whether it’ll be enough for him to win is an open question, but the overall outcome likely comes down to two V’s.

The first V is the potential V-shaped recovery in the economy. The term itself, “V-shaped recovery,” is meant to describe charts about the economy during a recession. Some, like the recession of 1921, involve sharp dips with a quick snapback; you get a “V”-like shape in the graph of the economy.  Others, like the 2008-09 recession, produce something more like a distorted “U,” with the economy slowly grinding back into shape. You could theoretically get an “L.”  We don’t want an “L.”

On this front, the numbers have been good for Trump. Unemployment did not hit the 20% milestone people were fearing, and has dropped sharply over the past two months of the second quarter. Consumer spending is up, as are retail sales. The stock market has held steady, down from its peaks but up from its lows. For right now, it does look like Trump is getting his “V,” and is probably getting it early enough to make at least some difference.

At the same time, if you drop a cat from high enough it will bounce, as the saying goes. A lot of the jobs that were shed were temporary layoffs during the shutdown; as those businesses reopen they bring back workers en masse.  At a certain point, however, you begin to hit the workers whose companies collapsed during the recession. In other words, we’re picking off low-hanging fruit, but that is unlikely to last. Whether the recovery slows early because those workers represent just 5% of the economy or the slowdown keeps going because they are 10% likely makes a big difference in Trump’s reelection chances.

The second V is related to the first: whether there is a V-shaped recovery in COVID-19 cases.  If you’ve been paying attention to the news, many states are seeing an acceleration in cases, especially in the South and Southwest. The exact cause of this is unknown; it could be people retreating indoors to beat the heat, it could be a refusal to wear masks, but regardless, the acceleration is real.  If this continues through November -- if parts of the country begin to resemble New York last winter -- the president will likely feel the heat.

At the same time, some perspective is warranted.  This chart shows cumulative cases since mid-March in the United States (click to enlarge):

Clearly there is acceleration in some states, especially those in the South we mentioned, but we aren’t yet to New York levels.  This is true even on a per capita basis:

If we eliminate the Northeast, we can see the general spread, but again, at least for now, it is profound in only a few states:

It is important to bear in mind that, if anything, this understates the differences between states today and New York/New Jersey. We’re currently testing around 600,000 people per day, unlike the peak of the New York/New Jersey outbreaks, when testing was much more limited. We’re capturing more mild/asymptomatic cases now, which we would have missed in March.

Regardless, if the curve bends in these states soon, and then cases abate through the fall, the memory of these outbreaks will likely fade. However, if Arizona is the next New York, and things spread into Colorado and Nevada, the president will pay the price.

Right now, Trump is an underdog for reelection. It would be foolish to write him off, however.  At the same time, his fate is not entirely in his hands. Whether he still occupies 1600 Pennsylvania Ave. come Jan. 21 will depend on the trajectory of the economy and the epidemic.