If Americans vote with their pocketbook, a presidential election pitting Hillary Clinton versus Donald Trump will be a squeaker, some economists think.
To figure out why, it's necessary to zero in on three key benchmarks of how the U.S. economy is faring: growth in real disposable income, unemployment and inflation, said Gregory Daco, head of U.S. macroeconomics at Oxford Economics.
In other words, as James Carville famously put it during Bill Clinton's 1992 presidential run, when it comes to winning elections, it's the economy, stupid.
Most Americans' views on the economy boil down to how fast their pay is rising, their job security and the cost of living. That gives economists like Daco, who spend most of their time thinking about things like GDP and productivity, a useful framework for predicting elections.
Of course, people choose to support a given candidate for all sorts of reasons, not just economic ones. That could include a nominee's position on hot-button issues, such as trade or immigration, or less tangible qualities such as the person's general likability. Dyed-in-the-wool Democrats or Republicans might blanch at the idea of voting for the other side. Or gender could come into play.
But plenty of research, notably the work of Yale University economist Ray Fair, suggests that voters pick a presidential candidate largely based on whether they think their economic prospects are likely to improve with that person ensconced in the Oval Office. Because the future is forever hard to predict, meanwhile, people tend to focus instead on the recent past, evaluating the economy based on how it has performed over the last year.
Another important factor is which political party is in office. A strong economy tends to benefit incumbents, while a recession or other economic calamity can lead to a changing of the guard (as presidents ranging from Herbert Hoover and Jimmy Carter to George H.W. Bush and George W. Bush can attest).
Oxford, which has correctly called 13 of the last 17 national elections in the U.S., finds that Clinton currently holds what the research firm calls a "razor-thin" edge over Trump (Oxford whiffed in 1968, 1976, 2000 and 2012. In 2000, it correctly predicted Al Gore's winning the popular vote, but didn't foresee Bush ultimately becoming president after the Florida recount and intervention by the Supreme Court.)
Nothing affects people's pocketbooks quite so directly as their pay, and on that front Clinton is benefiting from a pickup in wage growth this year, Daco said. "Although wage growth is muted, hovering around 2.5 percent, it could be just enough to tip the balance in favor of the incumbents."
The presumptive Democratic nominee also could get a boost from the job market. Research shows that when unemployment is below 8 percent, voters tend to reward the incumbent party. The jobless rate is now 4.7 percent, the lowest since November of 2007.
But any lift for Clinton would be minor because unemployment is already low and unlikely to tick down much further. And the recent slowdown in job growth, if it persists, could dent Clinton's chances.
"A worsening employment picture would weigh negatively on the incumbents' likelihood to be reelected," Daco said. "If it leads to an uptick in the unemployment rate, it could weigh on Clinton's odds of winning the election."
Not surprising, then, that Trump recently described the government's latest labor report as a "terrible" and a "bombshell."
To be sure, other economists do not foresee a close election. Mark Zandi, chief economist with Moody's Analytics, said the firm's election model forecasts a "sizable win" for Clinton. But he noted that prediction depends on a rebound in the job market, gas prices remaining low and house prices continuing to rise.
As for Oxford, it predicts Clinton would eke out a victory at the polls with 50.05 percent of the vote, compared with 49.95 percent for Trump.
Daco emphasizes that this tenuous margin could easily change as the economic winds blow between now and November. It's also worth noting that Oxford's election model only forecasts the outcome of the popular vote, not Electoral College results.
Indeed, if the current campaign proves anything, it's that political predictions have a way of making even the savviest prognosticators look stupid.