On the evening of September 1, 2011, I had the unpleasant task of delivering the confidential August jobs report to President Obama before it would be made public the next day. “I don’t like the look on your face,” I remember the president saying to me in the Oval Office.

“The job number is zero,” I said.

“Zero?” he asked. “What do you mean, ‘zero’?”

“The payroll jobs data is zero,” I replied.

Incredulous, the president asked, “You mean it is exactly zero? Like, if I took that paper out of your hand and looked at it, I would just see a zero?”

“Yes,” I replied with a wince.

“Has that ever happened before?” he asked.

“I don’t think so, sir,” I replied.

This was obviously not happy news. After he’d helped pass a payroll-tax cut and an extension of unemployment benefits near the end of 2010, the economy was projected to grow more than 3 percent in 2011. But after chaotic global events such as the Arab Spring and the Fukushima nuclear meltdown, as well as the self-inflicted wound of the debt-limit crisis, the economy had once again weakened. And politically, the president knew that “zero” would be easy partisan fodder. Sure enough, when the official numbers were released the next morning, it took the Republican National Committee only 94 minutes to coin and publicize the nickname “President Zero.” Commemorative buttons soon followed.

The irony? When the revisions came in two months later, it turned out the economy had actually seen the addition of 107,000 jobs in August 2011—not spectacular, but not the disaster it appeared to us that evening, and to everyone else the following morning. This was hardly an isolated occurrence. In 2010, the job numbers were revised up in 11 out of 12 months, with the more-impressive revisions never getting the attention received by the less-impressive initial announcements, which clearly hurt Obama politically. Yet, however negative or seemingly unfair this trend was, there was one thing neither the president nor anyone on his team would consider doing: attacking the credibility of the workers at the Bureau of Labor Statistics (BLS). The same went for those at other agencies, such as at the Bureau of Economic Analysis, that are responsible for compiling and announcing America’s major economic indicators.

For every administration in the past four decades, any public comment from the president casting doubt on the legitimacy of the government’s jobs, GDP, or health reports would have been unthinkable, for a few reasons. One, it would have been considered unfair to career civil servants to even suggest, with no evidence or proof, that they were tampering with numbers. Two, it would have hurt global confidence in the United States to even imply that anyone in the government was in any way manipulating the government’s economic reports—doing so would risk pushing the U.S.’s reputation toward that of China, whose published economic numbers aren’t widely trusted.

And additionally, every administration since 1974 had been cautious about never again allowing the type of ugly political interference that President Richard Nixon became famous for. Near the end of his first term, Nixon became convinced that the BLS’s assistant commissioner, Harold Goldstein, was not putting enough positive spin on the monthly jobs reports. Livid with Goldstein, Nixon ordered his top aides to “reorganize” BLS, demanding a count of the number of Democrats and Jews in the Bureau. He ended up demoting Goldstein, replacing him with a “politically sensitive, loyal Republican,” as well as appointing a loyalist as the new deputy commissioner overseeing all data analysis. A year later, the president and his top aides also forced the removal of the BLS’s commissioner, Geoffrey Moore, before the end of his term, which came after months of trying to re-assign him to another federal post.

Goldstein and the two more-junior career BLS employees demoted during Nixon’s crusade were Jewish, and the publication of Bob Woodward and Carl Bernstein’s 1976 book The Final Days, as well as the Nixon tapes, made clear that Nixon’s beliefs about and actions toward the BLS were motivated by paranoia and anti-semitism. “There’s a Jewish cabal, you know, running through this,” Nixon was recorded as saying, adding, “Now how the hell do you ever expect us to get anything from that [BLS] staff, the raw data, let alone what the poor guys have to say [inaudible] that isn’t gonna be loaded against us?”

The only positive legacy of this sad history is that it drove home for future administrations the importance of avoiding even the perception of political interference in the collection and production of U.S. economic numbers. In fact, following the Nixon scandal, the Office of Management and Budget issued a number of directives more strictly mandating the separation between releases of technical economic news and statements about politics and policy. Put simply, it would have been considered unthinkable in any of the post-Nixon administrations for anyone in the White House to seek to impede or publicly question the production of economic numbers.

On the eve of the release of the first jobs report during Trump’s presidency, this leads to a pressing question: Will he revive the Nixonian tradition of interfering with federal economic agencies if they announce numbers that do not please him?

Perhaps President Trump, who inherits a strong and stable economy, will be greeted by only positive economic news going forward, and will have routine cause to boast about government economic releases. And perhaps, even if economic numbers come in that do not bolster his message, he will handle it maturely, without berating government officials or trying to subvert the collection and posting of economic data.

However, with regard to the latter possibility, recent events are not encouraging. First, Trump has shown an almost compulsive unwillingness to accept any number that does not align with a combination of his self-image and gut instinct. His insistence that the crowd at his inauguration (which was estimated to number 160,000 to 250,000) was larger than that at Barack Obama’s (1.8 million)—or even the 450,000 to 500,000 who showed up for the Women’s March the next day—is the latest and most disturbing example of this behavior. It is not hard to imagine Trump refusing to accept a weak manufacturing jobs report because it contradicts his gut view that his bully-pulpit efforts to encourage domestic car production are working. Similarly, if GDP doesn’t reach his promised 4 percent growth per year, will he simply dismiss the validity of such numbers?

Second, President Trump seems not to care much for the norms dictating that presidents shouldn’t berate everyday citizens or castigate career government officials. Recent events—including the president’s personal attack on the local union head for the air-conditioning company Carrier and his phone call to the National Park Service’s acting director pressuring him to find proof that the crowd at the inauguration was larger than their photos implied—suggest a president who is not beyond intimidating anyone who does not provide him with what he wants. Similarly, in 2013, he tweeted, with no proof, that “census workers cooked the job numbers for Obama right before the election.” It is not encouraging that Trump has already imposed a ban that prevents career government officials and scientists at the Environmental Protection Agency and the Department of Agriculture from speaking to reporters or posting on social media about their work.

Third, Trump has already made a habit of trying to delegitimize the Bureau of Labor Statistics. He has made absurd economic statements, as when he argued that the “real” unemployment rate is 42 percent (at the time, it was actually 5.1 percent). More generally, he has called the BLS’s employment numbers “total fiction,” “one of the biggest hoaxes in modern American politics,” and “phony.” During a recent media briefing, his press secretary refused to answer when he was asked to simply state the unemployment rate.

While it's very unlikely that any president could succeed in getting BLS or the Census Bureau or the Centers for Medicare and Medicaid Services to actually juice up their numbers, there is certainly reason to worry that Trump, like Nixon, might bring about the demotion of officials whom he feels are not on his team and replace them with political loyalists, which itself would cast doubt on the objectivity of published statistics. Even more likely is the possibility that he would undermine the credibility of these agencies by publicly attacking their officials or telling the public to ignore their statistical findings.

Another troubling outcome would be if he sought to defund economic agencies already low on resources, or called for ending surveys in areas he didn’t think would play well to his agenda. For example, if Trump comes to fear that his eventual replacement for Obamacare will not meet his commitment to lower costs, lower deductibles, and lower premiums without hurting coverage, will he seek to simply terminate any health-care surveys that would reveal its shortcomings?

As the first few jobs reports come out, Trump will likely have no qualms about blaming Obama should they turn out to be poor, just as he would probably not hesitate to take credit for any positive numbers coming out early on in his tenure. But as later jobs reports (and any other future government findings that could displease him) are published, it will fall to Trump’s advisors and to Republican Congressional economic leaders to warn him not to politicize official economic data. If they can’t—or worse, won’t even try to—convince him, his attacks will damage global confidence in the American economy and public confidence in the rule of law.