In April 2021, price inflation reached 4.2%, followed by May reporting a 5.0% increase. Both represent high points last seen in 2008. As the economy reopens, consumers released their pent-up demand after a year at home on an unprepared supply and production chain. As a result, the inflation reemergence has received a lot of attention from investors, consumers, and policymakers alike. Most observers wonder whether these price increases are transitory and moderating quickly. Or are they sticky, staying with us long enough to become uncomfortable?
Inflation is hard to calculate.
Lost in the transitory/sticky debate is whether we are looking at the correct data in the first place. Calculating price indices is hard work. Price inspectors must avoid sampling error, picking representative items. They must not assign product quality improvements where none exist, and they must account for cases where today’s six t-shirts cost what yesterday’s eight did.
We calculate inflation in many ways, but two methods are used and referenced more heavily than others. The Bureau of Economic Analysis calculates the personal consumption expenditures index (PCE), while the Bureau of Labor Statistics calculates the consumer price index (CPI). For reasons due to formula, scope, and weight, they often disagree.
|Formula||Considers substitution decisions||Uses an average price for representative items|
|Scope||Measures expenditures by all households||Measures expenditures by urban households|
|Weight||Derived from business surveys||Derived from household surveys|
Different formulas underpin the PCE and CPI indices. PCE uses a procedure that reflects consumer substitution decisions among detailed items as prices change. If a rise in beef prices led to consumers choosing chicken instead, the PCE formula wants to consider that decision. In contrast, CPI doesn’t account for those substitutions and relies on an average-price approach for the representative items.
PCE measures expenditures by all households, whereas CPI measures the expenditure change for all urban households. As a result of using a broader base, some expenses in the PCE index are outside the scope of the CPI. For example, medical expenses for PCE include services paid for on behalf of consumers by employers, employer health plans, and medical care paid for by the government. CPI only considers out-of-pocket costs paid by consumers.
The PCE and CPI indices are based on different data sources and apply different weights to the respective categories. PCE weights derive from business surveys conducted by the Census Bureau each month. In contrast, household surveys conducted by the Census Bureau determine CPI weights.
For example, shelter accounts for the most significant weight difference between the indices. The share of personal spending devoted to housing is smaller in the PCE, 16%, than in the CPI, 33%. The fact that shelter spending in the business surveys used for PCE is less than those from household surveys used for CPI explains the gap.
Until 2000, CPI was the prevalent inflation measure noted in contracts, invoked in wage agreements, applied to Social Security benefits, followed by casual observers, and used by the Federal Reserve to set monetary policy. After 2000, the Federal Reserve switched to the PCE, preferring the historical consistency so valuable for research. Unlike CPI, which is written into contracts and cannot be changed once published, PCE data allows revisions to reflect updated information and refined measurement techniques.
Beyond the specifics of inflation calculations, the important conclusion we wish to draw are inflationary trends, for which we can review more than just PCE and CPI. Both have versions that exclude food and energy costs on the theory that those two categories are more volatile than the others. Their removal allows for a more stable average that better reflects inflationary trends. Various other inflation iterations, both core and otherwise, strip out the most extreme monthly price changes.
One or two high inflation numbers don’t suggest a problem. But if the numbers above continue through 2021, that outlook would change. The upshot is that right now all these measures are heading upward.