How to think about Consumer Price Inflation

By Vedran Vuk

Dear Reader,

The Bureau of Labor Statistics’ Consumer Price Index (CPI) is probably the target of more anger than any government dataset. And certainly, it’s not great at measuring inflation. But gauging inflation can be tricky; so today, I wanted to discuss some reasons for the anger over the CPI.

In my opinion, the central problem comes down to the weights for different categories. The BLS aggregates spending across the economy and creates weights for a "representative" consumer. Unfortunately, no one actually spends their money like this representative consumer. As a result, when a consumer looks at the CPI, more often than not the number doesn’t match their reality and anger inevitably ensues.

Let’s take a look at some distortions in the CPI methodology. I’ll use my own experience to make the point. In the past year, I have not purchased a television, yet the BLS counts the price of televisions with a weight in 1.816% in calculating the total CPI. However, one of my primary outlays was college tuition. Yet, according to the BLS weights, college tuition and fees are weighed at only 1.53%. For any college student or recent graduate, college tuition will represent a much higher portion of expenditure than 1.53%. As tuition rises drastically, my personal well-being becomes far worse. But the CPI doesn’t reflect this.

Suppose 20% of a student’s expenditures is tuition. Next year, the tuition makes a big 20% jump. As a result, the student’s overall expenses for the year increase by 4% – just from the tuition hike. On top of that, all other prices have slightly increased. Since last year, the student’s expenses may have risen as high as 6% to 9%. But, with the CPI weight of 1.53%, a 20% tuition hike registers only as a 0.3% increase in the overall consumer price index.

This is where the frustration arises. The student turns on the TV and hears about low inflation with a CPI of 3.2%. Naturally, he’s angry as his prices have increased 6 to 9%, and he feels economically much worse off.

The student isn’t the only frustrated individual. Check out some of the other low weights that could represent much higher spending percentages for certain individuals:

Prescription drugs: 1.253%

Medical care services: 4.994%

Health insurance: 0.461%

Child care and nursery school: 0.807%

Elementary and high school tuition fees: 0.422%

Tobacco and smoking products: 0.906%

Here’s the entire list. Once again, the problem is obvious: the disparity between the CPI and one’s personal spending may be very significant. And in some cases, these weights don’t reflect one’s spending at all. My personal weights for tobacco, high school tuition, daycare, and prescription drugs are zero.

This disparity is definitely problematic, but does it mean the CPI should be completely ignored? Yes and no. One probably shouldn’t use the CPI as a comparison for personal expenses. Here, the CPI will not yield much valuable information. But, it still may do a halfway decent job of capturing inflation. The dispersed nature of the CPI may do a better job of measuring a general rise in the prices than our individual expenditures. Inflation is a general rise in the price level – caused by printing too much money. Hence, my jump in tuition isn’t exactly a good measure of inflation. It may matter to me personally, but it is not significant when analyzing monetary policy. The same goes for health care and certain commodities. These prices may be rising for reasons other than the expansion of the money supply. For an inflation measurement, we want to track a general rise in all prices.

So, what’s a good way to gauge inflation?  One way is to examine prices of goods completely unrelated to a commodity spike or other similar fluctuations. For example, if the prices of shampoo, toothpaste, bottled water, and books rise next year (unrelated to transport and oil prices), then this would seem like a clear indication of rising inflation.

The CPI fails both at tracking inflation and representing the consumer; the index is in a limbo between the two. And this needs to be remembered when criticizing the measure. It’s not supposed to represent an individual’s personal spending perfectly, and it’s not a perfect gauge of inflation. Hence, the measure should be approached with extreme skepticism, but not total indifference.

The BLS could really improve its reputation easily. Here’s an idea: Create a section of the BLS website where visitors can enter their own weights for each category. That way, everyone can see how their own personal expenses compare against the BLS data. This wouldn’t be difficult at all; the entered weights would be simply multiplied by the respective price increases.

Though this wouldn’t improve the BLS’s measurement of inflation, it would be a good tool to illustrate the BLS system. If more people understood that the CPI doesn’t actually measure their own personal expenses, much of the anger would dissipate. Well actually, everyone would be still frustrated – but not for the wrong reasons.