Philip Cross: Trump’s conspiracy theory about ‘rigged’ U.S. jobs data doesn’t add up

U.S. President Donald Trump on the roof of the West Wing of the White House in Washington, D.C., on Aug. 5. Photo by Win McNamee/Getty Images

Article content

After last week’s jobs numbers from the U.S. Bureau of Labor Statistics showed a marked slowdown in job growth, Donald Trump fired its head, Erika McEntarfer, claiming “the numbers were phoney” and the process “rigged” to make his economic record look bad.

Article content

Superficially, it may seem Trump has a case. Although July job growth of 70,000 was close to expectations, unusually large revisions essentially wiped out most gains in the previous two months. Trump’s paranoia about the Washington bureaucracy no doubt reflects lingering memories of how in his first term some government officials openly created “The Resistance” to frustrate his actions — not to mention how 95 per cent of D.C. voters opposed him in 2024.

Article content

Article content

Article content

But it’s simply not believable that the jobs data are being manipulated to undermine Trump’s agenda. To start with, 0.4 per cent employment growth in his first seven months is not significantly different from the 0.6 per cent gain in the past seven months of the Biden/Harris administration — which Trump claims was inflated. The slight slowdown under Trump is understandable given: the massive uncertainty surrounding his erratic plans for tariffs; the DOGE job cuts, which doubtless had something to do with a decline in government jobs that was the chief source of weakness over the past three months; and the (until July) uncertain fate of his fiscal stimulus package in Congress. In quarterly earnings reports, many firms cite uncertainty as a factor in their decision to curtail hiring plans.

Article content

Article content

More broadly, however, Trump is articulating a widely held belief on the right that data are manipulated for political purposes. Jack Welch, General Electric’s former CEO, wrote in the Wall Street Journal about his suspicions regarding data showing lower unemployment during the 2012 election campaign. Shortly after leaving Statcan in 2011, I was asked on TVOntario — hardly a right-wing hub — whether data could be trusted.

Article content

Article content

But conspiracy theories about data don’t hold up to close examination. Biased analysis is one thing, and not hard to find. But statistical agencies systematically manipulating data would require many people to keep a secret that would be immensely valuable both politically and financially. The track record of bureaucrats guarding confidences is abysmal: as 1964 Republican presidential candidate Barry Goldwater once said about Washington, “There are more leaks here than in the men’s room at Anheuser-Busch.”

Article content

Article content

Even if a cabal wanted to manipulate data, how would it? The July jobs report that so upset Trump was published Aug. 1, an incredibly tight schedule considering the volume of data being released. To think statisticians could collect the raw data and then process, aggregate and seasonally adjust the estimates only to decide they disliked the result and were going to re-do the whole process with re-jigged data that maintained internal coherence across dozens of variables credits them with omnipotence no one possesses.

Article content

The reality is that statistical agencies are staffed by perfectly capable people — if not the cream of the civil service, whose most ambitious high-flyers aim for the centres of monetary and fiscal power. The litany of Statcan errors testifies to eyeball-rolling incompetence on the part of some employees, including: using the wrong spreadsheet for provincial GDP; overlooking thousands of import customs documents; entering the same price data for years on end; and deliberately releasing data early.  Most famously, Statcan confessed to making a 41,800-person mistake in job growth for July 2014 because staff failed to run a computer program when updating the labour force survey.  Bungling that bad suggests the elaborate planning, co-ordination and secrecy needed to cook the books are beyond the agency’s ability.