Wage Measures and the Employment Cost Index

Average hourly earnings, a wage figure released monthly alongside the jobs report, has a quiet flaw: it can rise simply because the mix of who's working changed, not because anyone got a raise. If a wave of low-paid restaurant workers gets laid off while higher-paid office workers keep their jobs, average hourly earnings ticks up mechanically — the average shifted upward toward higher-paid survivors, even though not one person's actual pay changed. This is exactly the problem economists worried about during the 2020 pandemic layoffs, when average earnings spiked even as actual wage growth was murkier.

The Employment Cost Index (ECI), published quarterly by the Bureau of Labor Statistics, is built specifically to avoid that trap. It fixes the industry and occupation weights used to combine wage data — so it measures how pay changes for a constant basket of job types, holding the composition of the workforce fixed, the same way a fixed-basket price index avoids being distorted by consumers substituting toward cheaper goods.

The ECI also separately tracks wages and salaries versus total compensation (including benefits like health insurance), which matters because benefit costs can rise even when cash wages are flat — a distinction the Federal Reserve watches closely, since a wage-price spiral concern is specifically about the cash-wage component feeding back into prices.