You can’t run a country, or any organization, without reliable data, and firing the head of a statistics agency because you don’t like the numbers it produces starts a path the United States does not want to go down. At the other end lies a ruined economy and a damaged democracy.
Is there room for improvement in how America produces economic and other statistics? Undeniably yes, in ways that would help bolster public trust in the data and also better protect statisticians from political interference.
Safeguarding the independence and objectivity of official statisticians is akin to a universal ethics rule, and it’s commonly understood that they can’t be fired in ways that would compromise that ideal. In the United States, it is even codified. “Federal statistical agencies must be independent from political and other undue external influence in developing, producing and disseminating statistics,” reads the ethics code written for the country’s statisticians. The principle is intended primarily for those outside the agency who might exert political influence. Statisticians themselves also have to resist such efforts.
This month, President Trump fired the Bureau of Labor Statistics commissioner, Erika McEntarfer, hours after the August jobs report showed weaker than expected new jobs figures and lower numbers than originally reported by the bureau for previous months. Mr. Trump publicly accused the bureau of producing inaccurate, politically biased data. “The Economy is BOOMING under ‘TRUMP,’” he wrote on one of his social media accounts.
It is clear that the president was unhappy that the data were not consistent with his narrative of the economy booming. But attacking the B.L.S. and firing its head undermines the professional independence of Dr. McEntarfer and the rest of the staff, which will now be more likely to censor its statistical releases. It also alters the incentive structures of future employees and leaders. In the language of the profession, people who make statistical decisions will be more likely to take into account nonstatistical considerations.
Unless the next head of the bureau is someone of high professional reputation and impeccable nonpartisan credentials, as well as capable of adhering to the statistical principles unswervingly in difficult circumstances, the problem will only get bigger.
I know the damage caused by attacks on statistics from experience. I ran Greece’s national statistical office, the Hellenic Statistical Authority, from 2010 to 2015 during the country’s crippling debt crisis. In that role, I supervised the revision of the government’s chronically misreported deficit and debt figures for the years from 2006 to 2009, the years leading up to the crisis, which started in early 2010. The revised, correct numbers challenged the narrative propagated by most political parties that Greece’s fiscal situation had been under control and thus did not cause its hardship.
As a result of telling the truth, I was subjected to seven criminal and civil legal cases, one of which remains open. While I managed to stay in my position for the duration of my term, these prosecutions and the attacks on the statistical agency prolonged the job of restoring trust in Greek data and its economy and undoubtedly damaged the incentive structures of statisticians for years to come.
Trust in official statistics is essential for many aspects of economic and social life, including the functioning of democracy itself. Official statistics inform decisions on whether and where to invest, whether to buy a home and where to seek employment. They inform the government’s decisions on whether to adjust spending and taxes, whether to undertake social and environmental policies and also the central bank’s call on whether to adjust interest rates. Importantly, statistics enable democratic accountability.
It is difficult to measure the negative impact of meddling with the Bureau of Labor Statistics, but it is likely to be significant if accompanied by other events that also infringe on the principles of producing official statistics. Unfortunately, they are already in motion: The Trump administration has suppressed the national greenhouse gas emissions inventory report produced annually by the Environmental Protection Agency.
The path the United States is testing is not an attractive one and can put it in bad company. In 2019, the president of Brazil at the time, Jair Bolsonaro, accused the National Institute of Space Research of falsifying data that showed that deforestation had accelerated in the year he took office. He promptly fired its head. In 2022, President Recep Tayyip Erdogan of Turkey fired the head of the Turkish Statistical Institute after it released data for 2021 showing that the inflation rate hit a 19-year high.
The B.L.S. firing has increased risks for official statistics production everywhere in the world. Countries that have a track record of attacking statistical independence will find it easier to legitimize their actions, because America does it. And countries that usually respect these standards will no doubt see some policymakers emulate Mr. Trump.
So what can be done? One core problem is how government statistics are produced in the United States, where the system is decentralized and statistical agencies are part of policymaking bodies. Such a system brings statisticians directly under the orders of political supervisors, giving them little recourse to protect their independence. This fundamental conflict of interest should be self-evident: The official statisticians produce information by which their political bosses are judged — by other branches of government as well as voters.
This is vulnerable to abuse. The system is not prepared to deal with non-benign conditions. Getting this right is the responsibility of the executive and legislative branches of government, and not of the statisticians. There has been, unsurprisingly, resistance over the years to making reforms. That should change.
Protecting statisticians’ professional independence does not render them above criticism. Critiques can lead to greater accountability and also improvement in the production of official statistics, if they are done in a thoughtful manner and based on scientific analysis.
One area for improvement is in how agencies sometimes communicate results, especially when they make revisions. Voters must be explicitly reminded that revisions are standard and be given a heads-up if they are likely to be significant. When large revisions are made, they should be highlighted in the opening paragraph of news releases and then explained with as much clarity as possible. Even statistical agencies in the United States can work on these issues.
Statistical agencies around the world need additional resources to improve their methods, adopt new technologies and explore new data sources. That’s even more pressing today, given the declining response rates to the statistical surveys that these agencies rely on for their data collection. Statistical agencies should take actions necessary to enhance the quality of their statistics. However, policymakers should first provide agencies with the necessary resources to do so and definitely not take resources away.
Official statistics are a tool for societies to know themselves, like a mirror. Those in power have always wanted to control the public’s understanding of reality, and official statistics give them a tool to do that. For America to keep its democracy operating smoothly, it has to ensure the mirror is reflecting reality.
Andreas V. Georgiou ran the national statistical office of Greece from August 2010 to August 2015 and now teaches statistical ethics at Amherst College.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.
The post I’ve Seen What Happens When a Country Messes With Data. America, Beware. appeared first on New York Times.
