“Expert failure” clearly has its moments. Pollsters, Wall Street analysts, technology futurists…all are being asked to consider that it’s wrong. However, economics seems to be receiving particular attention. Lately, this has metastasized into an Orwealian skepticism about government data itself. It’s another thing to claim that economists are misreading the numbers. It’s quite another to claim that the numbers themselves are lies.
Trust me, I understand reflexes. If it’s true that governments are failing at many things they try to do, why trust those statistics? But this cynicism makes the categorical error of conflating government’s inability to solve economic problems with its ability to solve technological problems. Understanding this difference explains why we can distrust economic planning efforts and at the same time trust employment statistics provided by, for example, the Bureau of Labor Statistics.
Simply put, economic problems involve mutually exclusive objectives and trade-offs. Should titanium be used to make railroad tracks or golf clubs? Should corn be turned into ethanol or used to feed cattle? The market resolves these through prices, profits, and losses. As FA Hayek demonstrated, governments are fundamentally incapable of assessing the trade-offs involved. In contrast, technical problems have a single goal in mind. Build railroad tracks, feed cows, and count the total number of jobs in the United States. There are no trade-offs involved in these problems, it’s just a matter of implementation.
It is clear that market participants can solve technical problems, but so can governments. For example, the Soviet Union defeated the United States in space, but was unable to stock grocery store shelves. This was no coincidence. Technical issues have clear and specific endpoints. Economic problems require the evaluation of infinite trade-offs that make market prices understandable.
Note that there is nothing here about the cost-effectiveness of the government’s solution, nor does it even suggest that it was worth solving the problem in the first place. In 1961, reaching space was an incredible feat. But an even greater feat is feeding people. As it turned out, the Soviet Union achieved the former, but not the latter. Result: collapse.
What does this have to do with government statistics? In short, everything. Data collection and analysis is a technical problem with a clear, single objective: accurate measurement. For example, there is no need to evaluate BLS, no resource allocation problems to solve, and no price signals.
Consider in particular the track record of BLS. To take another example, unlike the National Bureau of Statistics of China, which corresponds directly to the State Council and is more accurately described as a “propaganda department,” the BLS operates with statutory independence. The much-deprecated downward revision of total nonfarm payrolls by 911,000 people meant the U.S. now had more than 150 million nonfarm employees and the bureau’s accuracy was still well above 99%. The 2020 Census estimated that up to 782,000 people were missing. The U.S. population is over 330 million people, and the Census Bureau was accurate to within 0.25%.
Does this mean that the data collected completely corresponds to reality? Of course not. There are serious and legitimate debates about what should be counted in GDP, how CPI should be adjusted for aspects such as quality change, what the threshold should be for someone to be considered “unemployed,” and many other measures. All of these debates are about what to measure, not whether the measurements themselves are accurate or technically competent.
This distinction is important to classical liberals. We rightly don’t trust the government’s ability to pick winners, allocate resources, and plan the economy. But to dismiss government statistics as inaccurate statements is to confuse technical ability with economic planning. Could the private sector collect this same data in a more efficient way? Probably, but keep in mind that the Bloomberg terminal, which costs more than $24,000 per user per year, uses government data.
Should we trust the government to plan the economy? Absolutely not. But should we trust government statistics, at least US statistics? The evidence suggests we should. We should trust governments not because they are good, but because measuring and deciding what to do with that measurement are fundamentally different.
