Former Treasury Secretary and longtime Democrat Larry Summers recently co-authored an important paper analyzing long-term inflation trends and statistics. The paper demonstrates that changes to the way the federal government measures inflation via the Consumer Price Index since the Carter era understate the current scope of the problem—and the challenge the Federal Reserve faces in getting inflation under control today.

Two changes to the inflation measure—one a one-time methodological change, and the other a long-running trend—explain much of the apparent difference in CPI rates between the late 1970s and today. The first comes from a 1983 move by the Bureau of Labor Statistics to remove homeownership costs from the CPI measurement and replace them with a metric called owners’ equivalent rent.

The new metric quantifies what homeowners would receive for their homes on the rental market. As one might expect, the metric closely tracks the rental market. (Rent is a separate component of the CPI.) Most importantly, shifting from homeownership costs to owners’ equivalent rent to calculate homeownership costs eliminated the direct effect of mortgage rates—and therefore interest rate policy—on calculating the rate of inflation.

Prior to the 1983 methodological change, the very direct link between interest rates and the homeownership component of the CPI magnified the effects of efforts to combat inflation. Consider the two possible scenarios:

1. The Federal Reserve raises interest rates to combat rising inflation. When interest rates go up, so do mortgage costs—and therefore the cost of homeownership, as reflected in the CPI. Paradoxically, a measure designed to mitigate the effects of rising prices actually increases them, at least with respect to one component of the CPI.
2. Conversely, when the Federal Reserve lowers interest rates, mortgage rates also fall in tandem, so the homeownership costs component of the CPI also declines. Rather than the vicious cycle of scenario one, this scenario would lead to a virtuous cycle, one in which declining inflation would allow the Fed to lower interest rates—which would lower CPI still further.

Summers and his co-authors argue that the “interest rate CPI ratchet” (my words, not theirs) of scenario one helped lead to the double-digit inflation rates of the late 1970s and early 1980s. While the 1983 change to the methodology means we will no longer see this “ratchet” in the monthly inflation statistics—which explains why inflation hasn’t risen above 10 percent—it also means we won’t benefit from the benefits of scenario two (i.e., a downward “ratchet”) once inflation starts to get under control.

Summers and company also note that, compared to past decades, a smaller portion of the Consumer Price Index consists of goods with volatile prices. This suggests that combating inflation will require a longer and more sustained effort.

For instance, in the early 1950s, food and clothing comprised roughly half of the total Consumer Price Index, as opposed to approximately 17 percent today. The shift means that more elements of the CPI come from “sticky” industries and sectors—ones less amenable to sudden price shifts.

While a grocery store or clothing retailer changes its prices quite often, for instance, manufacturers of computers or other durable goods alter their prices less frequently. The fact that the latter types of sectors dominate the CPI compared to prior decades suggests that wringing inflation out of the economy will not happen overnight, nor very easily.

Worse Dangers Ahead

Summers famously predicted last February that inflation would accelerate if Democrats rammed through their $1.9 trillion “stimulus” legislation. Sure enough, it did. His newest analysis therefore bears watching, as does one ominous conclusion: that bringing inflation down to the Fed’s desired 2 percent level “will…require nearly the same amount of disinflation as achieved under [Federal Reserve] Chairman [Paul] Volcker.”