Have we failed to correctly measure inflation?

I suspect that inflation has been undercounted for decades. That would help explain things like the falling birth rate, which for centuries, in all countries for which there is reliable data, has been tied to the male median wage — when men have more money, they have more children, and when they have less money, they have less children.

The most important insight in their work is that for decades the inflation measures skipped over the cases in which a tenant moves out. Think about that for a second: A landlord avoids raising the rent for a decade for a good tenant, then when the tenant moves, ups the rent to a market rate. Or, as is common in New York, rent regulations keep rent from rising until a tenant moves out. Those are precisely the units in which the rent is most likely to rise, and for some four decades they weren’t counted in inflation.

Crone, Nakamura and Voith estimate that this and other problems bring down the government’s measure of rent increases by about 1.4 percent a year for the whole period that runs from 1942 to 1985. Nakamura outlines these findings in a very readable paper published by the Philadelphia Fed. Over such a long period, 1.4 percent into a really big number. Add that in, and instead of falling 20 percent in real-dollar terms over six decades, rents rise 50 percent.

Think about what that means for the middle class. Since 1970, the average hourly earnings of American workers (excluding managers) have stayed almost exactly flat by the official inflation measure. If, however, the cost of housing has risen faster than those measures say, then that means that many folks are actually worse off than they were then. You can see how that matters, and not just to the people in the San Remo with Central Park views…..

With rent, however, the intuition that prices have risen faster than the inflation data appears very robust not only to laypeople, but to professional economists as well. Robert Gordon is an economist known for his pioneering work on how we’ve overstated inflation in durable goods, such as cars, which have improved over time. He suspects that housing may be different. As Gordon writes in an e-mail, ”the low prices my parents paid in the 1950s for a lot, a house, my Harvard education … seem unbelievably cheap by today’s standards.”

If rent has indeed risen more than the usual inflation measures indicate, it would help explain why many families have the sense that they are working ever harder to afford to live in what they think of as a nice neighborhood. More generally, understanding what’s happened to the price of shelter is key to the current debate on inequality and how we answer the question that gets asked (rightly) in every election year: “Are we better off?”

The other takeway from New York rents is that when you think about inflation, you may consider asking “inflation in what?” and “inflation for whom?” We are so used to talking about the headline inflation number that comes out each month that we tend to forget it includes changes in literally thousands of prices. The increase in rent at the San Remo is not the same as the increase at a new college grad’s shared apartment across the river in Brooklyn, or at a public housing complex in the Rockaways—or in a different city. Nor is it the same as the change in the price of steak or childrens’ clothes.

And because inflation isn’t the same for different items in the consumer basket, it’s not the same for different people, either. It may, for instance, be higher for the wealthy, who are competing for the same few ultraluxury properties. It may be lower for older people, who don’t worry about moving. And it may be higher for those who have to move often. Measuring inflation accurately is a hard problem, and over the years economists both inside and outside the government have made great efforts (and strides) to get it right. The evidence of New York is that we’re not there yet.