The Federal Reserve is preparing to declare “mission accomplished” on inflation, but the story is more complicated. As gauged by the core Personal Consumption Expenditure (PCE) Price Index—the Fed’s preferred measure—inflation has stabilized at an annualized rate near the Fed’s 2 percent target for the past six months. Other readings are higher, though. Core Consumer Price Index (CPI) inflation has been running above 3 percent, and services measures near 5 percent. If underlying inflation is closer to these other figures, doves at the Fed could be making a policy error that risks creating fresh upward pressure.
There are numerous drivers of the gap between the PCE and CPI numbers, but health-care inflation plays an outsize role since the measures’ accounting methods for health costs differ in scope and size. CPI seeks to measure costs paid by consumers; it captures direct expenditures from individuals to medical providers and insurers. By contrast, PCE’s wider scope tries to capture all expenditures for consumption purposes, even those made on behalf of consumers but not by consumers. Thus, PCE includes things like Medicare payments that go directly from the government to medical providers.
Using PCE for the Fed’s monetary-policy purposes involves several oddities. First, the government sets many of these medical expenditures by fiat. Monetary policy’s goal is to balance supply and demand in the economy to prevent over- or underheating, but prices that the government sets itself are not indicative of supply-demand balances and are therefore misleading. Since private insurers tend to follow Medicare on pricing schedules, changes to Medicare reimbursements drive price changes in the entire sector.
Second, these expenditures don’t correspond to consumers’ experience of reality, and thus don’t influence their inflation psychology. The primary objective in pursuing a price-stability mandate is to keep inflation expectations anchored—to avoid high inflation becoming embedded in consumer thinking. Responding to fluctuations in government-set health-care remuneration to medical providers isn’t conducive to this end, as most consumers don’t even know what reimbursement rates are.
According to the Bureau of Economic Analysis, such differences in health-care scope alone caused 0.5 percent of the 1.1 percent total annualized difference between CPI and PCE during the fourth quarter. Adding the 0.4 percent back in changes the story meaningfully.
Taking the Fed’s preferred PCE numbers at face value implies that lower PCE inflation from health-care costs should be offset by higher inflation in other goods and services, in order to compensate for declining growth in payments to medical providers from the government. Indeed, progressive economists advocate precisely that: one think tank recently proposed lowering Medicare reimbursements to reduce measured PCE inflation and “secure the soft landing”—as though a soft landing were a function of how we measure inflation rather than the price changes experienced by consumers. By this logic, Congress can achieve any inflation number it wants by appropriately setting Medicare rates. Such a conception of inflation is useless.
Moreover, not even CPI is doing a good job calculating the inflation in health-insurance costs that consumers are experiencing. The Bureau of Labor Statistics (BLS) estimates health-insurance costs based on the lagged profitability and earnings of insurers, looking at their corporate reporting rather than directly measuring insurance transactions with consumers. Nor does BLS examine the split in how much of those premia are paid by employers or households. Due to earlier decreases in insurer profit margins—which could owe to their labor costs as much as to prices paid by consumers—BLS calculated that health-insurance costs fell 27 percent in 2023, but most Americans were probably paying more for health insurance. The annual report of United Health Group, one of the nation’s largest insurers, shows that revenues per domestic-employer-sponsored and individual plan member (a proxy for insurance premia) grew by about 4.5 percent in 2023—a far cry from a 27 percent decline. Measurement error is severe, and the Fed should not be basing policy on it.
The other issue at play in the CPI–PCE gap is one of size. Various aspects of consumption get weighted differently in the two methods, partly because of the differences in scope of the two measures. For instance, medical care commands 8 percent of the CPI bundle but 19 percent of the PCE bundle. When speaking of core inflation, these numbers become, respectively, roughly 10 percent and 22 percent of the bundles. This is the logical outcome of including a large amount of Medicare spending in one bundle but not the other. Variations in the scope of what is measured become magnified by the weighting discrepancies.
Because methodological quirks are so subjective, measuring inflation precisely is impossible. While it looks like the Fed has hit its 2 percent target based on PCE inflation, other measures are running higher. If the Fed puts all its eggs in one basket and chooses incorrectly, it is likely to lead to renewed inflation pressures. It would therefore be much better for the Fed to treat its price-stability mandate as it treats its employment mandate: by taking a constellation of data, instead of focusing exclusively on PCE inflation. Doing so would tell us to be cautious about declaring victory over rising prices.