I want to tell you the story of a president. This man came into office during a massive financial crisis, with the economy losing 800,000 jobs per month, and he turned things around. He created more than 5 million new jobs and saved the jobs of teachers who were at risk of being laid off. The manufacturing industry roared back, adding 460,000 workers. And after its worst decline ever, the housing market started to rebound, with home prices rising and the government helping millions of people avoid foreclosure.
Now I’d like to tell you the story of another president. This man also came into office during a massive crisis, but he hardly made things better. He created just 325,000 jobs over his term. More than 100,000 teachers were laid off, while the manufacturing industry lost 635,000 positions. Home prices slumped to a nine-year low, and 3 million homes went into foreclosure.
You’ve probably caught on to my gimmick: Both presidents are the same man, Barack Obama. And while those two descriptions of his economic record are based on facts, neither of them is really true. In politics and beyond, we often use numbers to validate our intuitions and to measure the vagaries of life, yet we know that these numbers are often imprecise. And perhaps more than any prior election, this year’s race has shown how easily numbers can be abused, exploited or misconstrued.
The campaigns of Obama and Mitt Romney have used numbers as weapons dressed as incontrovertible facts, often pulling them out of context and ignoring contrary evidence. But the misuse of numbers goes beyond political machinations. Numbers, particularly economic statistics, are often not as precise as they appear, especially when reported in real time. And yes, the news media (myself included) are guilty, too, with our relentless focus on any given month’s jobs number and what it means for the campaign. We’ll surely do that again when the nation’s third-quarter economic growth is reported on Friday. And again on Nov. 2, when we’ll get a look at how many jobs were created and what the unemployment rate was in the final month before the election.
It is not surprising that numbers are so central to this year’s political campaign, since the nation’s chief challenges are economic. Numbers offer candidates a way to provide evidence of their achievements or concrete plans for the future. They convey facts, not spin.
For voters, numbers may take on emotional significance — a couple of digits that can, simply and straightforwardly, quantify their feelings about the economy. If they’re down about their personal financial situation, they can ask: How was unemployment above 8 percent for so long under Obama? Or if they think they aren’t getting a fair shot, they can wonder: How can Romney promise a $5 trillion tax cut for the rich?
Even more than winning over new voters, though, numbers may energize supporters. “The numbers represent an opportunity to seize on facts and to prove to oneself you’re not just being biased, but the situation, as you perceive it, has been shown to be true by government statistics,” said Jeffrey Berry, a political scientist at Tufts University.
Several scholars have shown that voters tend to view the economy not through the reality of data but through the prism of partisanship. Let me use a few numbers to make my point. In February, a Washington Post-ABC News poll showed that 56 percent of Republicans saw the economy in poor shape, more than double the 25 percent of Democrats with that view. Larry Bartels, one of the country’s top scholars of the intersection of economics and politics, in March noted a similar mismatch.
A survey by the online polling organization YouGov asked whether the unemployment rate was higher now than when Obama became president. Sixty-two percent of Republicans said it had increased a lot, while 50 percent of Democrats said it had decreased a little. (The rate was 7.8 percent in January 2009 and 8.3 percent in February, when the survey was conducted.)
Properly used, numbers have enormous benefits, even when politically motivated. They attach something measurable to what a politician says. While it can be difficult for journalists to fact-check vague claims, it can be easier to evaluate how numbers do or don’t add up.
The most abused numbers of this election cycle — by politicians and journalists alike — are the jobs figures. Most people may not appreciate that even a healthy economy is a huge job-creating and job-destroying machine. A president’s role in creating jobs tends to be limited, but we need some way to judge his tenure.
Obama is fond of saying, as he did in this past week’s presidential debate, that the nation created 5.2 million jobs over the last 31 months during his term. This is a key argument in favor of his re-election, a description of what a recovery in the jobs market looks like. It is a big number, and it has problems.
The president is starting his count with the lowest low he can find. That is from February 2010, more than a year after he took office, and it includes only the private sector. In contrast, the public sector lost 537,000 jobs over the same period — a rich irony since the government has far more control over its own workforce than the private sector’s. And the number lacks context. Over 2 1/2 years, 5.2 million jobs isn’t that much — an average of 167,000 per month — given how low we’d fallen. During Obama’s first full year in office, the economy lost far more jobs each month, an average of 356,000, for a total of 4.3 million. And that was on top of the 3.6 million jobs lost in 2008.
So what’s the fair number to use? That’s a tough question, a reminder that objective-seeming numbers can be quite subjective. (Any assessment of jobs data in near-real-time, including mine in this essay, is somewhat arbitrary.) For Obama, it’s probably not fair to start counting in January 2009, when Obama took office, or even in the first few months of that year, as the economy was in free fall. It wasn’t possible to stem the decline so quickly.
I’d suggest June 2009, when a national panel of economists declared that the recession had ended. It’s also when Obama’s main policy response — the $831 billion stimulus — started to have a real effect. From that date, just over 3 million total jobs, and 4 million private-sector jobs, have been created during his term.
What can we conclude? During Obama’s tenure, the job market has climbed about halfway out of the economic abyss — perhaps a little more, perhaps a little less. But “my policies helped reverse half of the pain of the recession” is not a very effective debate line.
Romney plays a different game with jobs figures. He picks a really big number about the future of a really big economy and makes it sound, well, humongous. In Tuesday’s debate, Romney said his economic plan would generate 12 million jobs in four years. That sounds dramatic, especially in contrast to the lackluster job growth of the previous years.
But several top forecasters expect the economy to create that many jobs without any changes in policy — a good, but not spectacular, showing. If the economy continued to grow at its current, snail-like pace, it would probably add about 8 million jobs over the next four years. However, as the scars of the recession fade, the economy is likely to pick up speed. The Congressional Budget Office estimates 9.6 million jobs will be created; Moody’s Analytics suggests 12 million.
The centerpiece of Romney’s campaign is not just what he would do for the economy, of course, but his critique of Obama’s performance. And that has taken the form of a list of economic statistics. It doesn’t matter if we remember the particulars of these numbers — the impression counts.
At Tuesday night’s debate, Romney bemoaned the fact that unemployment was “so high,” above 8 percent, for so long. He was capitalizing on our sense that Obama should have done better. But logic and history suggest that, whatever our intuitions, it could have been a lot worse. This past week, Kenneth Rogoff and Carmen Reinhart, who are among the top authorities on financial crises, wrote that “in the five years since the onset of the financial crisis the United States has performed better in terms of . . . unemployment” than in previous crashes.
Obama and Romney have a great abettor in the misuse of economic statistics: the media. We need numbers just as much as they do. But this year, we have often read too much into the latest economic statistics and their implications for the election.
One of the great media memes of the 2012 campaign has been that since 1940, when Franklin D. Roosevelt won a third term, no president has won re-election with unemployment above 7.2 percent. (Ronald Reagan won in 1984 with unemployment that high.) On the surface, this seems deeply worrisome for Obama. But there is less here than meets the eye. Although no president has won, only three — Gerald Ford in 1976, Jimmy Carter in 1980 and George H.W. Bush in 1992 — have lost with unemployment above that threshold. So there simply aren’t enough examples to make much of that conclusion.
This year, another piece of conventional wisdom is that if the economy adds an average of 150,000 jobs each month, Obama is likely to win re-election. Much less would favor Romney. Nate Silver, the writer of the FiveThirtyEight blog for the New York Times, ploughed through the historical evidence and conducted a statistical analysis suggesting that 150,000 is in the ballpark of what Obama needs to win re-election. It makes intuitive sense, too: If that many people get jobs each month, the unemployment rate will come down at a steady, though not rapid, pace. This is, however, far from a science.
The media’s incessant focus on each month’s jobs report tends to overstate what we know at any given moment. On the first Friday of each month, when the government announces the prior month’s jobs numbers, we’re getting just an initial snapshot based on about 70 percent of a sample of businesses and government agencies responding to a survey. The Labor Department continues to collect and then refines the data for more than a year.
The first snapshot — the one that gets all the attention — is almost always off by tens of thousands of jobs and sometimes much more. As a result, you could come up with a different assessment of Obama’s odds based on the revisions. This year, the initial reports have averaged about 136,000 jobs created, below the 150,000 threshold. But when you consider the revised data — something we usually ignore — it looks like an average of roughly 157,000 jobs have been created each month, a more positive sign for the the president.
There may come a time when numbers fade from the forefront of public debate. If Obama wins, we may say Americans bought the notion that, even if the economy isn’t in stellar shape, the president has pushed it in the right direction. But if Romney wins, we may say it was because the economy wasn’t improving quickly enough.
But for now, the candidates are likely to focus on one number above all others: 270. That’s the unfungible number of electoral college votes needed to win. Later, Obama and Romney will probably view the election outcome in binary terms — I won, he lost.
On the surface, we want our world divided into simple outcomes, winners and losers. But the candidates would be well advised to look at another number: the popular vote, which is likely to be extremely close. Winning slightly more than the other guy does not equal a mandate. If in 2013 politicians show a little more modesty about what numbers really mean, they may see a particularly important number improving: their approval rating.
Zachary A. Goldfarb is an economics reporter for The Washington Post.