I had cataract surgery last week and it brought to mind a knotty problem the Bureau of Labor Statistics faces in tabulating the Consumer Price Index: How do we account for getting “more bang for the buck” when we measure inflation?
People my age might remember how difficult surgical removal of a cataract was only a few decades ago. The surgery was complicated, recovery took a long time and patients usually were condemned to wearing Coke-bottle glasses the rest of their lives.
All that changed with new technology. Now it is an outpatient procedure under local anesthesia involving 20 minutes or less of surgery. The following morning at my ophthalmologist’s office, I could read the 20/20 line on the eye chart without glasses for the first time in 50 years. I had no pain and no limitation other than having to put drops in my eyes four times a day for some weeks.
So in price-change calculations, how does one account for such dramatic changes over time in the quality of a good or service? If the price of something goes up, but the product also becomes better, what is the true increase in “price?” Does the price reflect the new value of what we’re paying for? And at what point is one really dealing with an entirely different product?
I don’t know what my surgery cost because we are fortunate to have one of those “Cadillac” health plans through my wife’s work. I am sure that the procedure cost quite a bit more than it would have in 1974 or 1964, especially in “nominal terms,” not adjusted for inflation. But my outcome was much much better than that of someone who had the nominally cheaper procedure back then.
That is true for a lot of things. In 1965, I remember the owner of a local filling station telling his coffee buddies the amazing news that “the Lotterman kids got 21,000 miles on the original tires of that Volkswagen of theirs.” At that time, 16,000 to 18,000 was probably the norm for a U.S.-made tire on a sedan. And if a car battery went three years before needing replacement, one felt lucky. Now tires can go four times as far, and I didn’t replace the factory battery in my pick-up until it was 8 years old.
Televisions don’t need a tube replaced every year or so, have much better picture quality and seldom go on the fritz. Modern paint sticks to siding longer. And how does your laptop compare to the Apple II you bought 35 years ago — probably for a lot more money?
So why does this matter to the Bureau of Labor Statistics?
When government statisticians try to measure changes in the general price level, they track the prices of a representative set of goods and services. They assign a numerical “weight” to each based on its relative importance — toothpick prices “weigh” less than gasoline. They then check the price of each item in this “market basket” at regular intervals. The percentage change in the weighted cost of all the items is an estimate of the change in the general price level.
It is vital that they compare the same items. They cannot use the price of tuna in spring water one month and in oil the next. Nor can they take the price of tuna at a discount supermarket one month and at a corner store in a high-income neighborhood the next.
They cannot check cotton underwear at Target one time and silk at Saks another, nor compare the price of gasoline at a station with two competitors on the same neighborhood intersection with that at a monopolistic highway “plaza” on some closed-access toll road.
Nor can they compare my cataract operation with that of the poor guy in my home town when I was young, who wore such thick glasses that people made fun of him.
How do they avoid this?
It is simple when technology clearly changes. Distributor breaker points are now out and crankshaft position sensors are in. Betamaxes are out and DVD players in — but even their days are limited. They have to take obsolete items out of the “market basket” over time and phase new items in. This often involves changing their weights over time.
Vinyl LPs still amounted for something before 1985 and compact discs for little. But over the next decade, the relative importance shifted sharply. This is not much different than accounting for changes in “tastes and preferences.” We eat 10 times as much pasta per person now as when I was a kid, but lamb consumption has fallen by more than 90 percent.
However, it isn’t so simple to adjust for modern-day paint that covers the garage very well after 20 years versus some that peeled badly at 10, or the battery that started my pickup well seven years on versus one that conked out after three.
Twenty years ago, it was increasingly clear to economists that failure to correctly adjust for quality changes in like products was causing the as-then-tabulated CPI to overestimate general price changes.
A commission, led by Michael Boskin, who had headed George H.W. Bush’s Council of Economic Advisors, was appointed to look at all sources of bias, both up and down, in the CPI. It found that the overestimate from unaccounted-for quality improvements was 0.6 percent per year.
One solution is “hedonic regression,” a sophisticated statistical technique too complex to explain here. (While Wikipedia should be used cautiously, its explanation of “hedonic regression” does a pretty good job if you remember your college math.)
Changes introduced in response to the commission’s findings mean that the CPI and other price indexes are a more accurate estimator of price level changes than they were 20 years ago. There are other factors that make price indexes underestimate true price level changes, but these must wait for another column.
Understand also that quality changes also bias Gross Domestic Product data. Adjusted for inflation, every $10 worth of paint provides more service for a homeowner than it did in 1960, and every $1,000 spent on cataract surgery produces a much greater improvement in daily life than it did only a few decades ago. But these go into GDP at their nominal prices.