Economists’ inflation conundrum

Print More

For economists, inflation statistics are like a tape measure. Rising inflation tells us that demand for stuff is increasing faster than we can make it or that demand is constant but the supply is shrinking. 

We use inflation figures to measure changes in well being. Say your monthly rent was $1,000 in 2010 and $1,200 today. If general price inflation was more than 20 percent, your rent went down. Less than 20 percent? Your rent went up. 

Imagine a carpenter with a tape measure that gives erratic results. Measure at 9 a.m. and the board is 12 inches long. Following the “measure twice, cut once” maxim of all carpenters, it measures only 10 inches a few minutes later. Or the tape measure always reads 12 inches to you, but your buddy insists that it says 10 inches.

That’s economists’ problem with inflation. And like the carpenter with an unreliable tape measure, it makes it harder to do the job. 

Inflation is calculated by comparing the cost of what we buy in one year to the cost of the same “basket” of goods in another year. Sounds easy, right? Goldman Sachs economist Spencer Hill points to official statistics that say that the growth in consumption of personal electronics, communications and media during the 2010s was the slowest in five decades. I don’t believe that. Do you? That’s like measuring a piece of wood that looks to be a foot long—but the tape measure reads only 7 inches. An experienced carpenter will find another tape measure. The jumble of rapidly changing digital devices and services seems to confound our efforts to describe them or measure how much of them we “consume.”


New products and services have long been a problem for economists. I loved my 1964 Volkswagen. For fleeting instants, I consider finding an old one to fix up, only to remember its nearly-useless “heater” or the brakes that required monthly attention. We get a lot more car for our money today than we did in 1964. 

Photo credit: Cars Remember When Auto Sales

A paper published in 2018 reported that 44 percent of products sold online in 2015-16 were not available in 2014-15. Now they define “product” as a “product-merchant” pairing, so this may exaggerate a common language understanding of the problem. But still, that’s a lot! The paper concludes that this “new product” bias means that inflation statistics are 1.5 to 2 percent higher than they should be. And if measured inflation is only 1.5 to 2 percent, then prices may have remained unchanged or even gone down.

The cost of some goods and (mostly) services has actually been falling even without an adjustment for quality. A famous case is Verizon’s decision in 2017 to reinstate its unlimited data plan, effectively a significant cut in price. As Verizon has a large share of the market, the wireless services portion of the Consumer Price Index fell by nearly 9 percent from March to April in 2017. This is partly due to a real reduction in the cost of delivering wireless services and partly due to cutthroat competition among the big carriers. The wireless industry has a big “network” effect—once the infrastructure for wireless services has been built, the cost of servicing an additional subscriber (the “marginal cost” in econspeak) is well below the average cost per subscriber. The owner of the largest network has a big advantage.

Finally, what do we do with stuff that’s free? Take Google Maps. I carry a pretty good map of Rochester around in my head. I know that I can find my way from Irondequoit to Spencerport or Mendon without Google’s help. But I still ask. And if I’m in New York City or Washington, D.C., where traffic can be horrendous and there are a half dozen plausible ways to get anywhere, I’m careful to keep my phone charged! 

Without pointing at any family member in particular (!), some members of a younger generation seem unable to get to the closest Wegmans and back without help. How do we price something that’s “free” (in financial terms, at least)? Erik Brynjolfsson and colleagues at MIT found that subjects of a study were willing to pay $64 per month for mapping services and much more for e-mail and search engines. See the graphic below of their findings, via The Economist. 

In this post I’ve focused on economic changes that make our “tape measure” of price movements—traditional inflation statistics—unreliable. Moreover, we strongly suspect that the traditional statistics are biased upward. Correctly measured, inflation is lower than the statistics suggest.

Note: In this piece I’ve channeled an excellent article in The Economist. In my next inflation post I’ll discuss how persistently low inflation, however measured, poses a problem for policy makers. 

Kent Gardner is Rochester Beacon opinion editor.

6 thoughts on “Economists’ inflation conundrum

  1. A reasonably intelligent person, I am aware of my limitations in statistics and Economics. I think I am vaguely aware of what you are saying. While I enjoy your thoughts on the matter and respect your conclusions, I leave it to my very reliable financial advisor to keep me on the right track. He lives and breathes this “stuff”. I’m certainly glad someone like you and my advisor do. Thanks for your thoughts.

  2. Kent, Grapes currently cost $1.99/lb. In February probably $3.49/lb. Simple supply and demand. If I buy grapes today and keep them in my fridge until February — well, they’re worthless then, too old, too rotten. Unless I make wine from them, at which point they take on a new and differentiated value, and thus a new and different contribution to the economics of my life. Basing inflation commentary on “new product bias” (which has a short term life) and on-line choice experiments fails to account for all the market changes over time that serve to influence our assessment of inflation/deflation, which itself is a time related and, in many instances, retrospective measurement.

  3. Doctor, not economist here. Check the cost of the big ticket items like housing, cars, health insurance, etc. Also, please look outside the stagnant Rochester markets. Additionally, while a gallon of milk or gas is the same size, just about anything else comes in a smaller volume or weight while the price remains the same. My 40-lbs bag of dog food that costs around $50 eight years ago now weighs 28-lbs and costs the same. As well, my Hostess Twinkies are lighter and the good ingredients are missing. This seems to me to be hiding real inflation.

    • All good points! Another famous example is the “half gallon” of ice cream–which became 48 oz seemingly overnight! That said, the smart economists at the Bureau of Labor Statistics & Census have decades of experience adjusting prices for changes in packaging. Now the Twinkies example is different-Like my wife, you believe that Twinkies just aren’t the same as when we were kids! I wonder if our memories are reliable on this score. More generally, however, adjusting for changes in product quality (up or down) is devilishly difficult. I think that the evidence that inflation is very low is overwhelming. But whether you agree or not, changes in consumer markets has made inflation much more difficult to measure accurately.

  4. Agreed, Neil. You’ve addressed a number of challenges to the measurement of inflation. The point of my post is that major changes in consumer markets has multiplied these challenges significantly. Economists’ confidence in our ability to measure well being today v. well being in the past has fallen, which creates some significant policy challenges.

Leave a Reply

Your email address will not be published. Required fields are marked *