Trump’s Abuse of Government Data

The President seems to declare employment numbers “very real” or fake news based solely on how they reflect on him.
Illustration by Tim Lahan

On January 22, 1930, not quite three months after the stock-market crash and the ensuing economic collapse, the Times, in a front-page article, quoted President Herbert Hoover saying that “the tide of employment has changed in the right direction.” His Secretary of Labor, James J. Davis, citing reports on America’s industries, pronounced the country “well on the way to complete recovery.”

Reading the article, Frances Perkins, the industrial commissioner of New York State, was horrified. She knew that the President was relying on data from the U.S. Employment Service, a notoriously inaccurate source, rather than from the Bureau of Labor Statistics, which showed that huge numbers of people were being laid off. She imagined the psychological harm that Hoover’s words would inflict on the unemployed, writing in her memoir that she feared the jobless “would feel that there was something wrong with them personally. A great despair would enter their hearts.” Young people would “read the story and say, ‘Why doesn’t Papa work?’ ”

Perkins called a press conference, prompting a national scandal about statistical methodology. Within a month, the Times was condemning Hoover’s misuse of data and praising Perkins (who went on to become Secretary of Labor under Franklin Roosevelt). The economist overseeing the unemployment numbers quit when he learned that the Administration also was trying to alter Bureau of Labor Statistics data to make them seem sunnier. Congress commissioned the development of objective employment data, and, since the nineteen-forties, two government data sets have been essential measures of the nation’s economic health.

Those statistics have often played a role in politics. In good times—under Dwight Eisenhower, Ronald Reagan, and Bill Clinton—the data have helped an incumbent win reëlection. In bad times—Jimmy Carter, George H. W. Bush—the data have contributed to defeat. Still, until Donald Trump, no major candidate or President had publicly challenged the validity of even the grimmest numbers. Throughout the campaign, Trump openly mocked employment data released by the Bureau of Labor Statistics and the Bureau of Economic Analysis as “phony” and “totally fiction.” His tone changed when he wanted to take credit for a good jobs report. In March, Sean Spicer said at a press briefing that the President wanted to make clear that the unemployment rate “may have been phony in the past, but it’s very real now.” The White House press corps laughed, but, to government statisticians, the words sounded less like a joke and more like a threat. “When I saw that, I said, ‘Wow. You said the numbers were wrong,’ ” Kathleen Utgoff, who ran the B.L.S. under President George W. Bush, told me. “ ‘Now you are politicizing them.’ ” She said that she is “terrified” by the President’s willingness to declare the government data “very real” or fake news based solely on how they reflect on him.

Utgoff isn’t afraid that the Trump Administration will meddle with statistical methodologies or the numbers they produce; that would require co-opting thousands of career government surveyors, statisticians, and economists in an effort to alter data collected from hundreds of thousands of businesses and citizens. (Even Richard Nixon, who, in 1971, hatched a plan to rid the B.L.S. of what he thought was a “Jewish cabal” out to destroy him, was unable to undercut the bureau’s independence.) Nor does anyone object to the reasonable arguments about which unemployment rate (there are six of them) best reflects the true state of the economy. The danger is that a President who disparages the data might convince his followers that bad economic news is political propaganda, and offer numbers that have no statistical rigor behind them.

Good economic statistics benefit the left and the right, government and business. Without reliable data, businesses can’t take risks on investments. Boeing, for example, decides how many 787 Dreamliners to build and therefore how many people to employ based on its Current Market Outlook forecast, which is rooted in B.L.S. data and projects aircraft demand for the next twenty years.

On a visit to a Boeing plant in South Carolina, in February, President Trump made his first major speech about employment. “We’re going to fight for every last American job,” he told the crowd, which had gathered in a hangar, in front of a new Dreamliner. Rhetoric and anecdote are Trump’s preferred jobs data. Last month, he claimed credit for a new Intel chip-manufacturing plant in Arizona “that will result in at least ten thousand American jobs.” In December, he celebrated a deal to keep around eight hundred jobs at a Carrier plant in Indiana. Trump didn’t point out other, less convenient facts—that Boeing was laying off thousands of American workers; that Intel’s plant had been announced in 2011, and would employ three thousand people, not ten thousand; and that Carrier was eliminating twenty-one hundred jobs in Indiana. Were Frances Perkins still alive, she would surely be thinking about those laid-off workers, who, hearing how Trump had saved their jobs, might see the failure in themselves and not in the Administration’s false promises. ♦