The final jobs report before the presidential election is here. In September, 661,000 more people were employed than in August, and the unemployment rate fell to 7.9 percent. That’s a slight — but not huge — improvement over last month, when the unemployment rate was 8.4 percent. (Compare this report’s 0.5-percentage-point decrease to the 1.8-point drop between July and August.)
So heading into the fall, employment is in a better position than it was in the spring and summer, but the recovery is also plainly slowing down. With less than five weeks to go before the election, how much will that affect President Trump or former Vice President Joe Biden’s chances?
Even in a normal year, our election analysis would be looking at the jobs report as just one economic indicator among many. But this year is even weirder than usual because there’s kind of a Rorschach test element to the report — people of different political leanings can see what they want to see.
Historically, jobs reports released shortly before the election have been reasonable bellwethers for how the incumbent party will do, although the last jobs report of the cycle isn’t necessarily the most reliable. In the 18 presidential elections since 1948,<a class=”espn-footnote-link” data-footnote-id=”1″ href=”https://fivethirtyeight.com/features/septembers-jobs-report-is-a-political-rorschach-test/#fn-1″ data-footnote-content=”
The earliest year for which we have seasonally adjusted data.
“>1 the unemployment rate in the September before an election has had little correlation with the incumbent party’s eventual popular vote margin. But when you frame that September number relative to earlier months — particularly the previous six to 12 months — it has traditionally been a much stronger predictor of electoral success for the incumbent party. (Though if you look at unemployment too far outside that window, the correlation dissipates.)
This suggests that voters care most about how unemployment is trending in the six months to a year leading up to the election and will give the incumbent party credit — or blame — for that trend.
This is all well and good under normal circumstances. But 2020, of course, is about as far from normal as it gets. And what the current economy means for Trump is very unclear. The unemployment situation, in an absolute sense, looks bad for an incumbent; a 7.9 percent September unemployment rate is the highest heading into any election since at least 1948. (It was 7.8 percent in September 2012.) But it could be much worse — things have been steadily improving over the past few months, with September’s unemployment rate coming in 6.8 percentage points lower than in April. So even though the numbers might be awful if you look at them in a vacuum, the fact that unemployment is trending in the right direction could be read as good news for Trump.
These varying interpretations are perhaps why Americans of different political stripes have wildly different views of the economy. According to polling by Civiqs, for a fleeting moment early in the pandemic, there wasn’t such a big gap between Republicans and Democrats in terms of how they saw the economy. But the gulf has widened considerably since then. A poll conducted at the end of September found that 80 percent of Republicans thought the economy was in at least fairly good shape, compared to only 8 percent of Democrats.
That dissonance might be related to those two distinct ways of looking at the employment situation right now: Do you focus on the absolute level of unemployment? Or do you hold tight to the trend of the past few months? Depending on which lens you choose, you’re looking at a completely different economy. If you look only at the past three or four jobs reports, the country appears to be making steady progress toward economic recovery. If you look at the change over the past year, unemployment still appears catastrophically high.
The trouble is that neither of these statements is wrong, exactly — it is true that workers are coming back on the job much more quickly than economists initially predicted, and it’s true that tens of millions of Americans are still out of work. But together, it doesn’t make for a very clean political message — which is why Republicans and Democrats can each look at this jobs report and see good news for their candidate and bad news for their opponent.
“Republicans are looking at the changes month-to-month and saying, “We’re creating millions of jobs every month; we’re getting the economy back on track,’” said Andrew Chamberlain, chief economist at the job search site Glassdoor. “And you have people on the left saying: ‘There are still millions of Americans out of work and we had Depression-level changes in employment in April, May and June.’” Partisanship being a helluva drug, there are ample opportunities for both sides to spin this most recent jobs report in whatever direction benefits them most, potentially rendering any historical comparisons irrelevant.
Making things even more difficult to interpret, unemployment is just one of many factors that would be considered when assessing the country’s economic situation heading into the election.
In our election forecast model, payroll jobs are included as part of the economic picture that contributes to the overall prediction. But we also incorporate a number of other economic indicators, and right now, some of those indicators are moving in opposite directions.
That’s because this is a really weird recession: Millions of people are out of work, but savings and income remained elevated for months, at least in part because of the federal supplement to people’s unemployment checks. Now that’s gone, but spending is still up a lot even though income has fallen a bit. And don’t even talk to us about the stock market, which rallied this summer despite the other economic turmoil.
So from the perspective of the election model, the economic situation has improved since June and we’re basically at an average economy — even though jobs are still quite bad in a historical sense.
The last wrinkle is that the picture could still change between now and Election Day, according to Nick Bunker, the director of economic research for North America at the Indeed Hiring Lab, a research institute connected to the job-search site Indeed. This year, the October jobs report will be released three days after the election. (To be clear: the Bureau of Labor Statistics, the federal agency that produces the jobs report, did not do this on purpose.) “In a normal year, you wouldn’t expect the economy to change all that dramatically over the span of a month and a half,” Bunker said. But over the next five weeks, anything is possible. And given that some voters are already casting ballots, any changes in October would mean that Americans who vote early could be making their choice for president in a fairly different economic reality than those who vote on Nov. 3.
All of this means that it’s not clear if the jobs report will help — or hurt — either candidate. In the end, it might just be a wash. After all, very few voters are undecided
. And this report might, more than anything else, just be a reminder of why it’s possible for Republicans and Democrats to see the economy in totally different ways.