Taking the Price Temperature?
The development and use of the price index is analogous to the development and use of the thermometer. Thermometers do not measure temperature perfectly; to some extent pressure changes and other factors bias the temperature reading. Also, thermometers only measure temperature "objectively"; they do not show how an individual reacts subjectively to whatever the temperature is. So are we to throw away all our thermometers because they are technically imperfect and non-subjective, that is, "non-Austrian"?
Timberlake, in his defense of price indices against Austrian criticisms, from which I quote above, voices a common anti-Austrian complaint. Austrians disdain econometric measures because they are inaccurate, but, of course, all measurements, even in physics, are inaccurate. Therefore, Austrians are being "unscientific."
It is important to address this issue, because, on the surface, Timberlake's analogy is appealing. Are Austrians anti-measurement sentiments due, perhaps, to some inappropriate feeling of reverence for their founding principles, or, more sinisterly, because they are afraid that measurement would disprove their theories?
To answer this question, let us see whether the consumer price index (CPI), and other measures like it, are really analogous to a thermometer. A thermometer is a measuring device that we can place in an environment where an independent phenomenon (in this case, molecular motion) is taking place. The measuring device responds to that phenomenon in a predictable manner, most notably, by some sort of (usually) visible display that can be interpreted to correspond to the quantity we want to measure. For example, the level of mercury in the thermometer corresponds to the temperature in the area where it is placed, and the position of the needle on a voltmeter corresponds to the voltage applied to the terminals of the device.
When it comes to price levels, what is the phenomenon in question, and what is the measuring device? Our "readings" are all in terms of prices, in other words, the exchange ratio of money to other goods. For the thermometer analogy to hold we must take money itself as the measuring device, where it is, perhaps, measuring the value of goods. (Austrians have refuted the theory of prices as a measure of value, but we'll ignore this objection and see if we are yet able to construct a coherent description of what a price index is.)
A price index is not measuring the "monetary temperature" of a good, or indeed, all goods, at a point in time, but is instead an attempt to track this temperature over time. However, a price index is not concerned with, for instance, whether the total "real" value of all goods has increased. Instead, it tries to track changes in the money price of the "same" basket of goods over time. Therefore, since money is our "thermometer," a price index is an attempt to gauge the stability of our measuring device.
Now we can construct a more accurate thermometer analogy than Timberlake. We have a device, money, which we take to measure the value of goods. This device gives us various readings -- gold traded for $275 today, bread is $1.19 a loaf, and so on. What we are interested in is finding out whether, over time, this thermometer is drifting -- is it giving us generally higher or lower readings for all goods on average, and not just for some particular good?
If we had some measure of value besides money prices, this venture might make more sense. However, we don't. So those advocating price indices are recommending that we check for drift in our thermometer by wandering from place to place, checking the temperature from time to time -- with the very thermometer whose accuracy we are testing!
We can see that we have hit upon an insurmountable obstacle. There is simply no way to determine which changes in measured temperature are real, and which due to drift in our thermometer. We'll look at one real world example of these difficulties, which should suffice.
Let us say that we are trying to determine whether the cost of computer programming has risen in the last thirty years. Now, the services of today's programmers, armed with significant advances in software engineering and capitalized with better tools, are simply not the same good as those of the programmers of thirty years ago. (This is no comment on the people themselves -- any particular person programming today should be vastly more productive than he was thirty years ago.) From this we can see that an hour of programming today should be more valuable than an hour of programming was thirty years ago. In order to determine if the cost of programming has risen, we will have to establish some ratio -- for example, one hour of 1970 programming is worth fifteen minutes of 2000 programming -- and compare the cost after applying this ratio.
However, it is hard to see how we can measure this change in value other than by comparing what employers are willing to pay for programmers now with what they paid in 1970. This presents advocates of price indices with a rather embarrassing problem. The only reasonable thermometer for measuring this change in value -- money -- is precisely the device whose accuracy we wish to check.
We must grant that the attempt to measure "price levels" is not completely mad, as long as it is taken as an extremely rough approximation of monetarily induced changes. To return to the thermometer analogy, if, in our peripatetic attempts to check its accuracy, we read 80 degrees, walked five feet, and then read 40 degrees, we would be justified in suspecting that something is up with the thermometer. Similarly, Austrians do not dispute that when the CPI shows 20% inflation, prices are probably undergoing a general rise. What is absurd is to publish CPI figures that show that inflation has "ticked up" from 2.4% to 2.6%. As Mises said:
The pretentious solemnity which statisticians and statistical bureaus display in computing indexes of purchasing power and cost of living is out of place. These index numbers are at best rather crude and inaccurate illustrations of changes which have occurred. In periods of slow alterations in the relation between the supply of and the demand for money they do not convey any information at all. In periods of inflation and consequently of sharp price changes they provide a rough image of events which every individual experiences in his daily life. A judicious housewife knows much more about price changes as far as they affect her own household than the statistical averages can tell. She has little use for computations disregarding changes both in quality and in the amount of goods which she is able or permitted to buy at the prices entering into the computation. If she "measures" the changes for her personal appreciation by taking the prices of only two or three commodities as a yardstick, she is no less "scientific" and no more arbitrary than the sophisticated mathematicians in choosing their methods for the manipulation of the data of the market. (Human Action, XII.4)
As an Austrian, I have no objection to measuring anything that can meaningfully measured. However, the problem with price indices is not that they are inaccurate, but that the idea of a price index can't even be defined coherently. Moreover, such indices are used to justify any amount of economic tinkering on the part of the government, whereby it displays its ability to "control" numbers that are its own arbitrary creation. The most beneficial change that the government could initiate in this field would be to fire all its econometricians, lowering the "price level" of our taxes.
Note: The views expressed on Mises.org are not necessarily those of the Mises Institute.