Virginia Postrel’s Forbes column from the September 4 issue uses Adam Smith to great effect in analyzing how people complain that there’s economic stagnation, even though the evidence of growth and prosperity is obvious. We see the same phenomenon today.
Nowadays, candid and intelligent people–not to mention partisans–tell us that the average American’s standard of living has barely budged in decades. Supposedly only the rich are living better, while everyone else stagnates or falls behind.
She then asks the logical question: who’s buying all the stuff, then? The flat-screen tvs, the ipods, etc. Her comments highlight the research of my colleague Robert Gordon, who pays careful attention to productivity growth and income growth.
Gordon is the author of a much-cited study showing that from 1966 to 2001 real income kept up with productivity gains for only the top 10% of earners. What the pessimists who tout his study don’t say is that, while Gordon does find that inequality is increasing, he’s convinced that the picture of middle-class stagnation is false.
“The median person has had steadily improving standards of living,” he says. But real incomes have been understated. The problem lies in how the U.S. Bureau of Labor Statistics calculates the cost of living.
Do we want to know how much money it would take the typical American to buy today what the typical American bought 20 years ago? If so, what about all those things that didn’t exist back then–not just iPods and mobile phones but everyday items like wrinkle-free pants, effective sunscreens, prewashed salads-in-a-bag or comfy hotel beds?
In other words, the doom-and-gloom crowd looks at only one side of the equation, the income side. But innovation and technological change have meant two different effects on the cost side. First, innovation has reduced the cost of consuming some standard goods that are captured in government measures of standard consumption bundles. Second, what about all of those new products that have made such a difference in our lives that aren’t reflected in those “standard” consumption bundles?
The Swiffer is my favorite example of this phenomenon. The way we clean our houses, and the associated amount of time we have to spend doing it, has changed greatly in the past 20 years. Pre-moisturized towels for all kinds of uses, wet-jet brooms, and so on. But, as Virginia describes in her column, such changes aren’t reflected in the standard measures.
Is there a way to improve government productivity statistics? Part of the problem is backward compatibility; keeping the consumption bundle the same allows for longitudinal comparisions across time. But the tradeoff for that is that you lose information about changes in the composition of the bundle. Sounds to me like you need two statistics: a basic consumption bundle that does not change over time, and a consumption bundle that represents the actual weighted-average shares of what we actually consume.