Taking government out of GDP makes no sense

It's a trivial idea from people who don't understand economic measurement. But attention must be paid as there's something very worrisome potentially going on here.

There are at least two worrisome aspects of the idea from Trump economic officials to redefine Gross Domestic Product, the dollar value of the economy, by subtracting out the government sector (Commerce Secretary Howard Lutnick and Musk have suggested this change). First, it makes no sense and is inconsistent with the concept. The government sector is part of the definition of GDP just like “stove” is part of the definition of “things in my kitchen.” The idea is trivial and silly, it doesn’t change anything re how we understand growth, and not worth spending much time on.

Except for worry #2, which is that I cannot dismiss the nagging fear that this could be a prelude to an administrative attempt to mess with the underlying data produced by government statistical agencies, and that would be an unmitigated disaster. I don’t mean that nerds like me won’t have the numbers we love to crunch. I mean that investors, businesses, firms, central banks, markets—globally—won’t have the input they need to do their jobs. Real human suffering would ensue, especially if we don’t have the data we need to determine that the economy has significantly slowed.

The reason why this shift is trivially wrong is that it intentionally leaves out goods and (mostly) services that are obviously part of GDP. Under the proposed change, a public-school teacher’s output doesn’t count; a private-school teacher’s does. Same with a VA doctor, military cop, or state trooper (if the latter two were private security guards, their output would count).

Axios got this right:

When the government buys a fighter jet, or builds a road, or educates a child, it reflects the production of goods and services. So if you exclude government spending from GDP, you aren't getting a full picture of U.S. output.

I guess the motivation is DOGE-like antipathy toward the public sector, and I also strongly suspect they think gov’t’s share of the economy has grown gangbusters over time and must be stopped. Nope. The figure below shows GDP without gov’t (private GDP) as a share of total GDP. It fell a lot in the 1950s (real defense spending almost tripled in the early ‘50s) but has been secularly climbing since and is about back where it started. The cyclicality here is a reminder of the important role gov’t spending plays in downturns. Since Keynes, it has been recognized that when the private sector contracts, the government sector needs to step in.

The two series—real overall and private GDP—don’t reveal anything shocking; neither do their relative growth rates. As you’d expect, private GDP grows a bit faster over expansions and, because of gov’t’s countercyclical function, overall GDP does better (less bad) in downturns.

There’s just nothing to see here folks.

However, if this proposal is designed to soften the ground for a broader takeover of government statistical agencies, there will quickly be a whole lot to see. I emphatically do not sit around thinking about new things to worry about. And I’m confident that the economists I know in the administration would stand firmly against any musking around with the numbers.

But blocking releases of unfavorable data is an old ploy by authoritarian governments; see China for contemporary examples. And given rising concerns about a forthcoming Trump Slump, the administration’s rhetoric about doubling down on harmful tariff policies, tanking consumer/biz confidence, and recent market volatility, we’d be foolish not to consider this possibility.

The take-gov’t-out-of-GDP is just a silly idea from people who don’t think about economic measurement. Manipulating the economic numbers is a whole other ballpark. I don’t expect them to go there, but I’ll be watching.