Skip to main content
The White House Takes Its Attacks On Jobs Data To A New (And Dangerous) Level

President Trump’s budget director on Sunday accused the Obama administration of “manipulating” economic data to make the unemployment rate look lower than it really was. That claim isn’t supported by evidence and will likely contribute to fears about the Trump administration’s commitment to honest reporting of government data.

“We’ve thought for a long time, I did, that the Obama administration was manipulating the numbers, in terms of the number of people in the workforce, to make the unemployment rate — that percentage rate — look smaller than it actually was,” Mick Mulvaney, a former Republican congressman from South Carolina, told CNN’s Jake Tapper during an interview Sunday. (Erica Groshen, the head of the Bureau of Labor Statistics in President Barack Obama’s second term, denied the charge to CNN.)

Mulvaney isn’t the first member of the Trump administration to question the validity of the unemployment rate. Trump himself called it “phony” and “one of the biggest hoaxes in American politics” during the campaign last year. Gary Cohn, the former Goldman Sachs president who is now one of Trump’s top economic advisers, once called it “very, very fictitious.” White House press secretary Sean Spicer, when asked on Friday about February’s strong jobs numbers, said Trump told him, “They may have been phony in the past, but it’s very real now.”

Those comments, however, generally left at least some ambiguity as to whether Trump was accusing Obama of fraud or simply arguing that the unemployment rate overstated the improvement in the job market. Mulvaney, by using the word “manipulating,” left no such room for interpretation. (When pressed by Tapper, Mulvaney acknowledged that he didn’t think the Bureau of Labor Statistics had changed the way it collected jobs data since Trump took office.)

What makes Mulvaney’s more explicit accusation so dangerous is that it could undermine the public’s (already shaky) confidence in government statistics. That’s worrisome because those statistics are our best way of knowing whether Trump is keeping his campaign promises, such as his pledges to bring back manufacturing jobs and boost economic growth. Even worse, by implying that economic data has already been tainted by politics, Trump and his advisers could make it easier to manipulate the data themselves. Already, there have been reports that the Trump administration is considering changing the way it calculates the trade deficit, and Spicer has preemptively attacked the nonpartisan Congressional Budget Office over its forthcoming analysis of the Republican health care plan. Those efforts appear to be part of a broader pattern by the Trump administration of undermining confidence in public institutions such as the media and the judicial system.

Because the stakes are so high, it is worth taking Mulvaney’s claim seriously — and taking the time to explain why it is so off-base.

First, it’s important to understand where the government’s data on jobs comes from. The best-known and most timely data comes from a pair of monthly surveys: one of employers (Current Employment Statistics, or the “establishment survey”) and one of households (the Current Population Survey). The establishment survey, which is larger, is the source for the monthly payroll figures — the ones being cited when news stories say that “the economy added 235,000 new jobs last month” — as well as information on earnings and hours worked. The household survey is the source for the unemployment rate and for information on the demographic breakdown of the workforce.

Those two surveys are only the tip of the iceberg when it comes to government jobs data. The monthly Job Openings and Labor Turnover Survey provides figures on how many people were hired, fired or quit their jobs. The Quarterly Census of Employment and Wages uses data from state unemployment agencies to get a more precise (but less timely) look at jobs and earnings. The Internal Revenue Service and Social Security Administration release their own so-called administrative data on employment and income. And then there are numerous private data sources that try to predict or supplement the official statistics: ADP, Gallup, the Institute for Supply Management and other groups all release their own measures of the job market. The various measures can diverge for a month or two, but over the longer term, they all tell more or less the same story of a steadily improving job market.

Manipulating the jobs figures, then, would mean not just messing with one number but rather interfering with an entire ecosystem of statistics, all of them closely watched by investors and economists. That would require a conspiracy theory of massive proportions, involving hundreds if not thousands of people. As it happens, in 2014 the Census Bureau’s Office of the Inspector General investigated an alleged case of political interference in the Current Population Survey and found no evidence to support the claim. (There are numerous protections in place to guard against such interference.)

What is true, however, is that the unemployment rate can be a misleading statistic. The government considers people “unemployed” only if they are actively looking for work. During the 2008-09 recession, the job market got so bad that many laid-off workers stopped looking for jobs, meaning that they no longer counted as unemployed. That meant the unemployment rate was lower than it otherwise would — or perhaps should — have been. (That problem is much smaller than it used to be — the improving economy is now drawing people back into the labor force. Still, the job market probably isn’t quite as healthy as the 4.7 percent unemployment rate would suggest.)

But there is no conspiracy here. Obama didn’t change the definition of unemployment, which has been essentially unchanged for decades. The Bureau of Labor Statistics releases numerous other measures — the labor force participation rate, the employment-population ratio and multiple alternative versions of the unemployment rate, among others — that do account for people who stop looking for work. Many of those statistics are cited by the same conspiracy theorists who accuse Obama of cooking the jobs numbers.

There is a long-running and entirely legitimate debate among economists about the best way to measure the health of the job market. Mulvaney, in his CNN interview Sunday, argued that the best approach is to ignore the unemployment rate entirely and focus on the number of jobs created each month. That’s an imperfect solution in the long term (we would expect job growth to slow as a larger share of available workers find jobs), but not bad advice on a month-to-month basis.

There is a big difference, however, between arguing that the unemployment rate is a flawed measure and accusing the prior administration of data manipulation. The latter is an extraordinary claim requiring extraordinary evidence — something no one in the Trump administration has yet provided.

Ben Casselman is a senior editor and the chief economics writer for FiveThirtyEight.