MODERN PRESIDENTIAL races are remarkably steady affairs. Once one candidate establishes a lead, the other finds it hard to reduce it by much, especially once autumn starts. This year, extraordinary events have left their mark on the contest: the spread of the coronavirus pandemic in the spring and protests for racial justice in the summer. The main parties’ poll numbers also shifted a little after their online conventions. But since Labor Day support for neither President Donald Trump nor Joe Biden has budged by more than a percentage point from the average: 54% of the two-party vote for the Democratic challenger and 46% for the Republican incumbent, according to The Economist’s number-crunching.
The televised debates between the two candidates, the first of which takes place in Cleveland on September 29th, offer Mr Trump a chance to make up the ground. But just two days before the face-off, the New York Times published the findings of a long investigation into the president’s tax returns. Among other things, the Times reports that he paid only $750 in federal income taxes in 2017, the year he moved into the White House, and in several years paid nothing at all. The Times also says Mr Trump or his various holding companies owe hundreds of millions of dollars in debts for various properties, including a $100m mortgage on the commercial space in Trump Tower in New York.
The revelations, which the president characteristically dismisses as “fake news”, are sure to give Mr Biden a handy line of attack in the first debate. But the really bad news for Mr Trump may be that, even without the disclosures, he had little chance of using the debates to turn his campaign around. History suggests that they seldom shift voters’ intentions.
According to The Economist’s analysis of polling data compiled by Robert Erikson and Christopher Wlezien, political scientists at Columbia University and the University of Texas at Austin, past presidential debates have mostly had only small, temporary effects on the polls. Since 1960, when the first nationally televised debates between presidential candidates were held, the incumbent party’s share of the vote in opinion polls has, on average, not changed at all between two weeks before the first debate and two weeks after the last one. Only rarely have the polls moved rapidly during the debate season. The sharpest change came in 1976 when Jimmy Carter, the Democratic challenger, slipped from a robust 58% of the vote for the major parties to a much more precarious 51% (though he still beat the incumbent, Gerald Ford).
The lead has changed hands after the debates only once in the 12 campaigns since 1960 that have featured televised matches. That was in 2000, when George W. Bush briefly took the lead from Al Gore—only to lose it swiftly, as well as the popular vote in November. (That said, Mr Bush still took the White House because he won a majority in the electoral college.) Even in 2016—when there were far more undecided and third-party voters than there are today—the debates did not “matter”. Much of this stasis is attributable to the fact that the candidates are well known by the time debates roll around, and that political polarisation has forced most partisans to choose their sides. We estimate, from past contests, that the chances of an eight-point swing to the incumbent after the debates have been roughly one in 20. In other words, there is only a 5% chance that the debates could erase Mr Biden’s lead.
For all that, presidential debates do not occur in a vacuum. On the day before the first debate in 1976, Playboy magazine published excerpts of an interview with Mr Carter. That called into question his reputation as a pious southern Baptist. It helped Gerald Ford eat into his lead; it may have dogged him for the rest of the campaign. This year, Mr Trump is facing a similar problem, in the shape of the Times’s revelations.
It is hard to forecast how they will shape the race. On the one hand, most voters have already taken Mr Trump’s financial dealings into account. The president’s supporters long ago accepted his flaws, whereas his critics have always regarded him as unfit for office. On the other hand, a report that the president owes hundreds of millions he may not be able to repay could hurt his standing with the 6% of likely voters who are still undecided (and who usually pay little attention to the regular news cycle, but might take note of a torrent of big stories about the president’s taxes).
The main consequence of the Times’s bombshell may be to force the president to defend himself against attacks from Mr Biden and others, taking up time that he does not have. With early and postal voting under way and only five weeks left before election day proper, every day that the president’s polling numbers merely hold steady—or, worse, fall—is a day that decreases his chances of retaining the White House.
To be successful on Tuesday night, Mr Trump must not only see off Mr Biden’s criticisms of his handling of the pandemic and America’s racial strife, as well as of his taxpaying record, but also start to convince more voters that he deserves to keep his job. The statistical evidence from past debates suggests that is unlikely. And a performance that only reinforces impressions that voters have already formed will do nothing to dent Mr Biden’s lead.