PCE vs CPI: Which Inflation Measure Does the Fed Target?

The Federal Reserve officially targets personal consumption expenditures (PCE) inflation, not the Consumer Price Index (CPI), as its measure of price stability. Both track consumer-price changes, but PCE is broader, weights goods more flexibly, and includes services outside the scope of CPI. The Fed chose PCE because it better reflects how spending patterns shift when prices change, making it a more responsive indicator of true purchasing power than CPI’s fixed market basket.

Why the Fed switched from CPI

The Federal Reserve did not always target PCE. Historically, the Federal Reserve cited the Consumer Price Index, particularly the “headline” CPI that includes food and energy. But in 2000, the Fed formally adopted PCE as its reference measure.

The reason: PCE better captures how households actually behave. The CPI is built on a fixed market basket. If the basket includes 10 pounds of beef per year, and beef prices spike 50%, the CPI records a large increase even if households respond by buying 5 pounds of beef and 5 pounds of chicken instead. PCE adjusts the basket month-to-month, so if consumers substitute chicken for beef, the index reflects that shift.

This “substitution bias” in CPI means it overstates inflation in a world where prices change constantly. A family’s actual purchasing power may decline less than CPI suggests, because they can trade down or shift to alternatives. PCE’s chained weights minimize this problem.

The Fed also preferred PCE for breadth. CPI covers urban consumers’ out-of-pocket purchases. PCE includes all household consumption, including healthcare and housing services imputed from home values (not just rent). For a central bank concerned with the economy-wide price level, PCE’s broader base matters.

The two indexes diverge more than they converge

Over the long run, PCE inflation runs about 0.3–0.5 percentage points below CPI inflation. This reflects substitution bias: CPI overstates by roughly that amount over multi-year periods.

But in any given month or quarter, the divergence can swing wildly. If energy prices explode, headline CPI surges more than PCE (because CPI weights gas and crude oil more heavily in some methodologies). If healthcare costs jump, PCE may outpace CPI (because PCE’s health-services coverage is broader). During the pandemic, the divergences were pronounced, with PCE sometimes above and sometimes below CPI as different categories of demand and supply shocks shifted.

The Fed monitors both, along with trimmed-mean inflation (excluding volatile food and energy) and the core versions of each (headline excludes food and energy; core includes them). But the official 2% target is always stated in PCE terms. Market participants and policymakers must remember this: when the Fed says “inflation is at 2%,” it is speaking of PCE, not the CPI figure the public often hears.

The technical difference: chained vs. fixed weights

CPI uses a fixed basket updated once a year (or in some measures, every two years). The Bureau of Labor Statistics decides that the basket includes, say, 100 pounds of gasoline, 50 pounds of milk, and 10 doctor visits. Prices of those items are collected, and the index is recalculated.

PCE uses chained weights. Each month, the weights shift based on the previous month’s spending patterns, drawn from national accounts data. If households spent relatively more on services in month one, the weights on services rise in month two. This is more computationally complex and introduces a small lag (PCE data are released later than CPI), but it tracks actual behavior.

The trade-off is accuracy vs. timeliness. CPI is out faster and more transparent (the public can see the basket). PCE is slower but more adaptive.

How the indexes respond to shocks differently

During the 2022–2023 inflation surge, energy and shelter (housing) accounted for much of the jump in both CPI and PCE. But their relative weights differed. CPI’s energy weight spiked higher in the initial months because headline CPI is constructed to give energy a larger role. PCE’s energy weight was material but, thanks to monthly reweighting, smoothed out faster as consumers reduced driving and heating use.

Shelter—the largest component of both indexes—contributed more to PCE overall because PCE’s shelter measure is broader, including imputed rent on owner-occupied homes (valued at what the home would rent for). CPI’s housing measure relies more on actual rents paid by renters, which respond more slowly to price changes in the housing market.

By late 2023, core PCE (excluding food and energy) had declined below core CPI, suggesting that once energy volatility was removed, the substitution bias reasserted itself. PCE’s flexibility was allowing the index to show slightly slower underlying inflation than CPI’s fixed basket.

The implication for monetary policy

The Fed’s 2% PCE target is not arbitrary. The Federal Reserve believes that 2% PCE inflation is consistent with price stability and maximum employment. If PCE is running above 2%, the Fed tightens monetary policy—raising interest rates and tightening credit. If PCE is below 2%, the Fed may ease.

Because PCE typically runs below CPI by 0.3–0.5 points, a 2% PCE target is not the same as a 2% CPI target. If the Fed had chosen 2% CPI, it would effectively be targeting closer to 1.6–1.7% PCE over the long run. That would mean permanently lower inflation, lower nominal interest rates, and a different risk-return trade-off for savers and borrowers.

The choice to target PCE reflects the Fed’s desire to keep inflation measured by true purchasing power—how much more (or less) consumers can buy—not by a statistical artifact of fixed-basket methodology. It is one reason why Fed communications are sometimes confusing to the public: when headlines report CPI inflation at 3.5% while the Fed says “inflation is under control,” the Fed is often looking at PCE near 2% or core PCE falling.

Headline vs. core in both indexes

Both CPI and PCE come in headline and core flavors. Headline includes all goods and services, especially food and energy, which are volatile. Core excludes food and energy, showing the underlying trend.

The Fed watches core PCE closely because it is more stable and reflects monetary conditions rather than temporary supply shocks (like a bad harvest or an OPEC output cut). But the Fed also publishes projections and statement language around headline PCE, because headline inflation ultimately affects household welfare.

During volatile periods (like 2021–2023), the Fed emphasized that headline spikes were transient and core was the better guide. The public disagreed: their grocery bills were rising whether core inflation was “sticky” or not. This tension—between statistical measures and lived experience—is why both metrics matter.

Why some debate persists

Not all economists accept that PCE is uniformly superior to CPI. Some argue that CPI’s fixed-basket method, though biased upward, is more conservative and transparent. Others contend that CPI’s focus on out-of-pocket consumer spending, rather than imputed values like owner-occupied rent, is more intuitive.

Internationally, central banks use different inflation measures. The European Central Bank targets Harmonized Index of Consumer Prices (HICP). The Bank of England uses CPI. Each chose their measure for regional or methodological reasons.

The Fed’s choice of PCE is settled policy and unlikely to change. But it remains an important distinction: when comparing inflation across countries or across time before 2000, applying PCE definitions retroactively is necessary to get a fair comparison to Fed policy intent.

See also