A potentially more appropriate way to ascertain the true degree of inflation was suggested by Richard Ruggles of Yale University. I learned about this test in graduate school many years ago, and Ruggles may have made his suggestion many years before that.

What inflation rates should really measure is the decline in the utility value of a nation's currency. In principle the right test is to offer a sample of the population a constant amount of currency, say $1,000, along with the opportunity to spend it on a menu of all goods and services available this year or a menu of all goods and services available a while back, say five years ago. If this sample of the population votes in equal numbers for this year's menu and for that available five years ago, one can conclude that prices have been stable over the five-year period. If the majority vote is for the earlier menu of goods and services, one can conclude that prices have risen, or that the utility value of the $1,000 has decreased. If the majority vote is for the recent menu, one can conclude the reverse--that true prices have actually declined.

I have seen no rigorous polling evidence on this question. But for years in teaching college macroeconomics courses, and recently at the Fed, I have conducted such a poll among my audiences. All audiences have reported that their understanding of what inflation is all about was much improved by this thought experiment. Generally, college students have voted for the current menu even in times when the aggregate rate of price increase averaged 3 percent or more, implying that they felt that true prices had actually declined.

College students may be unusually influenced by fads that do not truly improve goods (narrow or wide ties, etc.), and the implicit bias in measured price indexes may well be overstated by collegiate polls. Since coming to the Fed, I have had the opportunity to talk to and poll many banker groups about inflation, and as one would expect, they are generally more inclined to vote for the earlier menu of goods than were my college students, at any given rate of inflation. But these days, when measured rates of inflation are running at 1.5 percent to 2 percent, even bankers consistently vote for the current menu of goods by fairly wide margins. If even bankers feel that the implicit measurement bias in price indexes exceeds 2 percent, that may be a phenomenon worth noting. The upshot of this highly anecdotal test is that I have long suspected that true price stability might really be achieved in the vicinity of measured inflation rates of 2 percent or even more.