Price Theory and the Price Level

From Covid lockdowns to recent discussions of tariffs, there have been a lot of discussions about the price level and inflation. At times, these discussions focus on theoretical issues and different theories of price level determination. At other times, these discussions tend to focus on the measurement of the price level. Unfortunately, this commentary occasionally runs into a problem that I rarely see addressed. That problem is that theoretical arguments about price level determination necessarily assume a theoretically correct measurement of the price level. However, in reality, our measures of the price level are flawed.

This isn’t some conspiracy theory about the government purposely misleading the public or adjusting the measurement of some particular price index to get a desired result. Rather, my claim is rooted in price theory. Empirical analysis requires the use of a price index. A price index is an estimate or approximation of what economists mean by the term “price level.” The proper measure of the price level can be determined by price theory. However, a number of practical considerations make it difficult for our estimates of these price indices to perfectly replicate the theoretical measure. At best, we can get approximations of the theoretical ideal.

This means that in some sense, all of our price indices are flawed. Some of these flaws are more important than others. Nonetheless, all is not lost. Understanding the precise ways in which measurement might differ from theory can be useful in interpreting the data.

Measurement Requires Price Theory

Let’s start with this. People like to draw clear distinctions between theory and empirics. The current trend in the economics profession is not only to draw a clear distinction, but also to elevate empirical work to a higher status than theory. There is a “let the data speak” mentality prevalent in the profession. However, data do not speak. For data to be useful, the data must be used to test a hypothesis. In order to have a hypothesis to test, you need some theory that generates that hypothesis. Furthermore, you might want to be aware of alternative theories that might generate the same hypothesis, lest you might incorrectly conclude that your chosen hypothesis is the only explanation.

However, the point is deeper than that. If you are going to do good empirical work, you need to make sure that you are measuring things properly. To measure things properly, you also need ...