Should Uncle Sam Be Involved in Data Collection?

On the demands of statesmanship related to information

The Trump administration’s firing of the head of the Bureau of Labor Statistics raises an interesting question:

Should the federal government be our go-to-source for important data?

Washington as a Reliable Source

On one level it’s worrying that the president would dismiss the head of a data-reporting agency because the president didn’t like the data that had been reported. That comes across as petty. But it also implies that data that should be cold and hard—apolitical—is now subject to political vetting. Is it about to become the new normal for federal statistics to be massaged to fit the preferences of the current occupant of the White House?

It is true that some federal reports over the years have been political. But in my experience those are more exceptions than the rule. In general, we have been able to take most federal data to the bank. For example, for generations analysts depended on the Statistical Abstract of the United States. It was an annual collection by the Census Bureau. It included a staggering amount of information; hundreds of pages of tables and charts and descriptions. It had just about everything. And you could rely on it.

In the education world, we have used reports from the National Center for Education Statistics. Arguably the most valuable, consistent, and comparable student-achievement data of the last couple generations came from the “Nation’s Report Card,” the National Assessment of Educational Progress (NAEP). It has been produced by the federal government. Years ago, when I thought a particular philanthropic award was wrongheaded, I used a specific NAEP data set to make my point. Some folks didn’t like my conclusions, but people trusted the numbers.

Similarly, people have generally trusted federal data related to inflation, unemployment, labor-force participation, and more. Those producing and publishing these numbers were not seen as political actors. In fact, those numbers often frustrated the day’s administration—e.g., when inflation or unemployment was higher than the then-president wanted.

I have two personal experiences with this phenomenon. When I worked for the George W. Bush education department, an independent federal research entity published unflattering data on the DC voucher program—a program we in the administration really liked. But we knew those producing the findings were fair. We knew we had to accept the findings.

During the Obama administration, the education department desperately wanted good results from its super-expensive and high-profile School Improvement Grant program. But federal data kept showing and showing and showing that the program was a failure, which was obviously embarrassing for the administration. The final federal report showed that the program had spent billions but failed to produce positive academic results. Defenders of the program would’ve loved for this all to disappear, but the reports came out.

For the most part, then, I (and others) have believed that the production of reliable, non-politicized data can and should be a responsibility of the federal executive branch.

But is that asking too much of the federal executive branch?

The Political Unitary Executive?

I’ve argued (along with others) that our system of government requires the president to have control over the entire executive branch. That includes the ability to hire and fire senior executive-branch officials. If there are high-ranking officials in the executive branch independent of the president, the president’s authority is compromised, and we don’t really have three distinct branches.

Since the president (with the VP) is the lone elected official in the federal executive branch, shouldn’t we assume that s/he will always have political aims? So if s/he is in total control of the executive branch, won’t all executive-branch functions—including data production and dissemination—be at least somewhat political in nature?

And since we need non-politicized data, shouldn’t we remove all data production and dissemination from the federal executive branch? Maybe it should live outside of the government?

I think the answer to that is, “No.”

I think you can believe in the unitary executive AND believe the president will always be politically minded AND believe federal executive-branch data collection can and should be non-politicized.

But how???

The Head of State

As we all learned in high-school civics, the president wears two hats: head of government and head of state. In other nations, those jobs are typically separated. But here the president is the elected leader of the executive branch and clearly a political actor. But the president is also a non-partisan figure; the official who acts on behalf of the entire nation. That second, head-of-state role includes receiving ambassadors, leading national ceremonies, and serving as the symbol of the country as a whole. In other words, there are times when we recognize the non-political duties of our chief executive.

That can include the collection and distribution of trustworthy information. For instance, when the president named the Warren Commission, the Rogers Commission, and the 9-11 Commission, Americans understood their head-of-state president was acting. The president was gathering and sharing the most reliable information available on subjects important to the nation. The president’s nonpartisan behavior in such instances doesn’t preclude his/her political head-of-government behavior in other areas. It simply reflects the office’s limited but essential head-of-state responsibilities.

Weaponizing Weaponization

One final thought.

As a general rule, I like moving things from government control to non-government control. So why shouldn’t we simply hand all data-collection and -distribution tasks to outside researchers? Just get all of this out of Uncle Sam’s hands.

A part of me thinks that makes sense. I think most researchers take data collection very seriously. I think most researchers take fair, rigorous data analysis very seriously. And I think most researchers believe that straightforward data and fair, rigorous analysis should be widely disseminated.

But over the past decade, we’ve seen our nation grow polarized. We’ve seen many institutions shed their defined missions and become more politicized. I’ve grown concerned that some researchers and academics have stronger political commitments than scholarly commitments. If a huge data set or a major analysis supported heterodox political or policy conclusions, are we confident that a strong-willed researcher—were s/he in charge of that particular national data operation—would disseminate that information? Ask yourself, “Today, would I completely trust the data issued by a team of academics whose politics are diametrically opposed to mine?”

Of particular concern to me is the ascendance of the term “weaponize” in this context. In recent years, too many journalists, researchers, academics, commentators, and advocates have taken to saying things like, “They are weaponizing that data” or “He is weaponizing that report’s finding.” Often, this means that the person isn’t disagreeing with the data or the finding; instead, that person really doesn’t like how that data or finding is being used. I always find myself wondering if those charging others with weaponizing information believe that some data and analysis are so susceptible to weaponization that they would feel entitled to bury that data and analysis.

In short, we need those engaged in public life to recognize their obligation to produce and share reliable information even if they find that information unhelpful to their cause. That’s the duty of many of our institutions and those who work inside them.

Good reporters and editors understand that good journalism requires the sharing of information even if the reporter/editor dislikes it.

Good scholars understand that a strong academy requires the sharing of information even if the scholar dislikes it.

Good public servants understand that statesmanship requires the sharing of information even if the public servant dislikes it.