Easy to lose, hard to restore: US data trust on the line

Reuters

a day ago

WASHINGTON/NEW YORK, Aug 5 (Reuters) - Donald Trump's move to fire the head of the U.S. Bureau of Labor Statistics has put trust in U.S. data reporting mechanisms on the line just as demand for reliable diagnoses of the health of the world's largest economy is bigger than ever.
Examples from elsewhere show credibility is easily lost and hard to restore. A first test will be the choice to replace Erika McEntarfer, accused without evidence by Trump of manipulating U.S. job numbers after weaker-than-expected growth and large downward revisions were reported last week.
"Imagine if one of your concerns is that there's a lackey in charge of the agency and the numbers are fake," said Michael Strain, director of economic policy studies at the conservative American Enterprise Institute, of an appointment Trump has said to expect within days.
"That's a whole other level of problems."
Policymakers, businesses and investors are scrambling to understand how Trump's attempt to up-end the global trade system will affect prices, employment and household wealth. Central banks, which once tried to guide market bets on rate moves months down the line, now say decisions are "data-dependent."
The rub is that data collection is proving to be harder. Debt-laden governments have, as McEntarfer experienced, cut resources in their data departments; phone surveys, the go-to method for much macro research, are struggling to produce adequate samples as many households do without fixed lines.
Trump's implicit accusation of partisanship by "this Biden Political Appointee" adds the troubling factor of a political dimension usually indicative of countries dogged by wider doubts over their democratic checks and balances.
The key lesson from past examples of loss of data confidence is that it can take years for trust to be restored.
When Argentina last year reported its first single-digit inflation in months, sceptics questioned the data and recalled the massive underreporting of inflation in the 2000s and 2010s for which it was censured by the International Monetary Fund.
"They manipulated the data for a long time," said Aldo Abram of libertarian think tank Liberty and Progress Foundation in Buenos Aires. "It's logical people remember this and continue having doubts."
Turkey has changed the head of its TUIK statistics institute four times since 2019, with opposition parties arguing the changes were political. Roger Marks, fixed income analyst at asset manager Ninety One, said the result for investors has been a "gradual erosion of our trust in the numbers."
For Greece, whose efforts during the 2000s to conceal mounting public deficits fed that decade's sovereign debt crisis, it has been an equally long haul back to credibility.
It required the overhaul of its ELSTAT statistics agency in 2016 and the creation of an international panel of experts to appoint its chief statistician - steps that have meant its hard-fought efforts to improve its budget are now unquestioned.
It also prompted European governments to grant the Eurostat statistics arm of the European Union powers to check suspect national statistics reported to it.
Longstanding doubts over the accuracy of Chinese statistics - with even former Premier Li Keqiang acknowledging in 2007 the country's output figures were man-made - have obscured genuine efforts to improve data quality, such as a new measure of youth unemployment excluding students that was released early in 2024.
"There were genuine methodological reasons for the change, but because of the history around Chinese data a lot of people, particularly foreign investors, just didn't really trust that," Julian Evans-Pritchard, an analyst with Capital Economics.
"That underscores to me that once you undermine confidence in the data, it is quite hard to restore that confidence."
Faced over the years with patchy official data, watchers of emerging economies have long sought to corroborate those numbers with other datasets. Capital Economics' China Activity Proxy is based on 18 indicators from freight traffic to electricity consumption.
Another metric is found in the data and sentiment surveys provided by independent researchers in all the big economies. But they can only sketch in one perspective on the picture.
"An awful lot of it is soft data: 'How do you feel? What do you think's going on?'" said Erik Weisman, chief economist and portfolio manager at MFS Investment Management in Boston.
"They're not asking for specifics. They're not asking, how many widgets did you produce? How many insurance policies did you produce? How many hours worked?," he said, adding that the concerns raised by the sacking of McEntarfer could nonetheless force analysts to turn increasingly to those other sources.
The most urgent question now is whether the breach in credibility which the Trump intervention has opened is now widened further or mended.
Enrico Giovannini, former chief statistician for the Paris-based Organisation for Economic Co-operation and Development (OECD), said there was more scope in the U.S. for political appointments of key statistics roles than in other advanced economies which tended to make long, fixed-term appointments.
"So the incoming government has to wait (to replace them), said Giovannini, who has also served in two Italian governments. "In the U.S., the spoils system works," he said of the practice of party supporters getting rewarded with government jobs.
The International Statistical Institute, a professional organization for data collectors, issued a statement late on Monday that said Trump's move violated U.N. principles aimed at protecting fact-based statistics and called on his government to take steps to restore public confidence in U.S. federal data.
William Wiatrowski, the BLS' deputy commissioner, will serve as acting commissioner until a successor to McEntarfer is named.
Beyond that choice, some fear that further dangers may emerge from a Trump executive order on federal hiring intended to reserve posts for candidates who can prove they are "dedicated to the furtherance of American ideals, values, and interests."
Aaron Sojourner, a senior researcher at the W.E. Upjohn Institute for Employment Research, said such a move would, if passed by Congress, apply to many jobs in federal economic statistical agencies.
"This proposal would convert many of those jobs into political jobs where people can be fired for any reason if they displease a political leader," Sojourner said.