A story in today’s New York Times discusses one of the ugly truths of economics: Sometimes people make mistakes.
The article, “On Economy, Raw Data Gets a Grain of Salt,” focused on the enormous faith that policymakers, economists and investors place in financial data, such as the rate of economic growth, despite the fact that such data is often based on crude estimates and gets revised after-the-fact to account for errors, faulty assumptions or new information.
As one expert put it: “People want the best information that we have right now. But people need to understand that the best information that we have right now isn’t necessarily very informative. It’s just the best information that we have.”
In fact, nearly all national economic data is subject to revision after it is released. The reason: the data is often based on estimates, surveys, and other imperfect measurements. For instance, the federal and state unemployment rates seem to offer a measure of precision. But in reality, we don’t know exactly how many people are in the labor force, either working or unemployed, at any one time,. The best we can do is estimate those numbers based on other measurements, such as unemployment insurance claims.
When it comes to economics, we have very few real counts of things. The U.S. Census, taken every ten years, is supposed to count every resident of the country, but even that is subject to error. One recent notable oversight: census takers mistakenly counted 600 people supposedly living on boats in Hampton, when they were actually in Portsmouth.
If actual counts are subject to such errors, imagine the possibilities when the data is based on estimates. Or just basic math or an analyst transposes two numbers, and someone catches the mistake after the data is released.
The only set of economic data that is not subject to regular revision is the Consumer Price Index. This is the most commonly used measure of inflation, and has multiple uses: to index Social Security benefits, to calculate cost-of-living adjustments in public employee contracts, and to adjust property tax caps – even in calculating rates for boat docks.
The CPI is not perfect. It is no more perfect than any other data series. So why doesn’t the CPI get revised to reflect more accurate data?
Because it is the basis for so many other calculations. Imagine, for instance the lawsuits if a union found out that its members were denied a wage increase because of a counting error. The ramifications of revisions to the CPI could ripple out in countless directions.