Refresher: CPI and the PCE Deflator

On Thursday, the US Department of Commerce will publish the May estimate of personal consumption expenditure deflator. Last week, investors learned that the consumer price index rose at its fastest rate in two years.

Since extraordinary monetary policy measures were taken, some observers have warned of inflation risks lurking around every corner. With CPI rising above 2%, could they finally be right?

Not so fast. The Federal Reserve's 2% inflation target does not apply to the consumer price index. Since mid-2012, the Fed has identified the core PCE deflator as its preferred inflation measure. There is a significant difference. The PCE measure of inflation tends to be lower than the CPI measure.

In part, the two measure different things. The CPI is calculated by the Bureau of Labor Statistics of the Department of Labor. The PCE is calculated by the Bureau of Economic Analysis at the Commerce Department. An important methodological issue is the weights for the different components. The CPI uses the same weights for several years, while the PCE deflator uses current and preceding expenditures to calculate the weights. One implication of this is that the PCE measure also allows for substitution of goods with rising prices for similar goods with stable or falling prices.

When it comes to health care expenditures, which played a large role in the sharp downward revision to Q1 GDP, CPI and PCE measure somewhat different things. The CPI calculus includes only the expenses employees pay, while the PCE deflator also includes money that employers spend for their employee health care.

The CPI is calculated based on surveys of more than 14,000 families and 23,000 businesses they patronize. All told some 80,000 consumer items are included. Sales taxes are included.

The PCE deflator calculation is completely different, though both the PPI and CPI measures are used. Essentially, and at the risk of over-simplifying, the PCE deflator is derived from production, which is at the producer price level. The PCE methodology converts the producer priced goods/services into consumer prices, adding profit margins, taxes and transportation costs.

In addition, other data is incorporated. For example, it incorporates the monthly retail sales report, and the price of food grown and eaten on the farm comes from the US Department of Agriculture. The dealers' margin on used vehicles is pulled from the National Auto Dealers Association.

The core measure strips out food and energy, mostly. Recent definitional changes, however, means that the core PCE measure includes restaurant meals. These have been redefined to be food services. Pet food is also included, seemingly as part of the pet expenditures rather than as part of the food budget.

Over the past 10 and 20 years, headline CPI has averaged 2.4%. The PCE deflator has averaged 2.1% and 1.9% for the 10 and 20 year periods respectively, or 0.3 percentage points and 0.5 percentage points respectively.

It is widely documented that for at least the past half century, headline inflation converges to core inflation (and not the other way around). This may help explain why the Federal Reserve prefers core measures.

The chart at the top shows the core CPI (white) and core PCE deflator (yellow) over the past 20 years. Over the past five years, the core PCE has undershot core CPI by 0.2% on average. Over the past ten and 20 years, the undershoot has been 0.2% and 0.4% respectively. The persistent lower estimates of the PCE deflator means that the Federal Reserve is likely to tolerate what may appear to be an overshoot of CPI.

There appears to be two elements that produce a more subdued rise in the PCE measure. One is the substitution effect that is not incorporated into the CPI measure. The other is the way that hospital expenses and airfares are calculated.

The consensus calls for the May core PCE deflator to rise to 1.6% in May from 1.4% in April. If accurate, it will keep this measure 0.4 percentage points lower the core CPI. It is consistent with the Fed's assessment that the economy is evolving toward the Fed's mandates. Some observers will make hay of the substantial contraction in Q1 GDP, but this too should not alter the outlook for Fed policy. The Fed's economic mandates, as we have discussed, are on employment and prices and both did indeed move in the (Fed's) desired direction in Q1 and appear to have done so further in Q2.