Lower-end wages rose faster, but the gain after inflation was small

On August 6, Treasury Secretary Scott Bessent rejected the claim that the US economy was improving only for high earners while lower-income households fell behind. That pattern is called a K-shaped economy. One arm of the K points up for people doing better, while the other points down for people losing ground. Bessent said some media outlets had chosen an anti-Trump conclusion first and then selected statistics to fit it. His evidence was the Bureau of Labor Statistics' report on weekly earnings for the second quarter of 2026. BLS is the federal agency that produces official US jobs, inflation, and wage data.

How to read the chart

The chart ranks workers by weekly pay and compares three positions. These positions are called percentiles. The 25th percentile is the cutoff below which 25% of workers fall. The median is the worker exactly in the middle. The 75th percentile is the cutoff below which 75% fall.

Bessent used the gap to argue that lower-paid workers were leading the recovery. He also said President Trump was reversing the loss of purchasing power under the Biden administration. The claim needs three separate tests. First, is the wage-growth arithmetic correct? Second, did workers gain purchasing power after inflation? Third, can the improvement be credited to Trump administration policy?

US weekly earnings, second quarter of 2026 (Bureau of Labor Statistics, 2026-07-21)

출처: bls.gov

The arithmetic is correct

The 25th-percentile weekly wage rose from $806 in Q2 2025 to $850 in Q2 2026. That was a $44, or 5.5%, increase. The median rose from $1,196 to $1,251. That was a $55, or 4.6%, increase. The 75th percentile rose from $1,887 to $1,915. That was a $28, or 1.5%, increase. Bessent was therefore right that the lower-pay cutoff rose more than three times as fast as the upper-pay cutoff. The 2025 figures come from the agency's report a year earlier. Both the 2025 and 2026 releases remain listed in the source section below.

The 25th percentile is not the average pay of the bottom quarter

The $850 figure is not what the lowest-paid 25% earned on average. It is one cutoff in the ranked wage distribution. It also does not mean that the same group of lower-paid workers received an average 5.5% raise.

This is not a same-person wage tracker

The BLS table takes one picture of the workforce in 2025 and another in 2026. The person at the 25th percentile need not be the same in both pictures. For example, if fewer young and relatively low-paid workers are in the sample, the cutoff can rise even when individual pay changes little. This is called a composition change. The Federal Reserve Bank of Atlanta's Wage Growth Tracker uses a different method. It matches the same person's hourly wage 12 months apart. In June 2026, the middle growth rate was 3.6% overall, 3.4% for people who stayed in the same job, and 4.1% for people who switched jobs. The two measures do not conflict. BLS tracks movement in a wage cutoff; the Atlanta Fed tracks pay changes for matched people.

Same-person wage changes (Federal Reserve Bank of Atlanta, 2026-07-09)

출처: atlantafed.org

Purchasing power rose only about 0.5% after inflation

Nominal pay is the dollar amount on a paycheck. The median nominal weekly wage rose $55, or 4.6%. Real pay removes inflation to show how much more a worker can actually buy. The Consumer Price Index rose 3.9% over the same period. Pay did beat inflation, but only by 0.7 percentage point. BLS also converts pay into 1982-84 dollars so the two years can be compared at the same prices. That measure rose from $372 to $374, or about 0.5%.

The table does not cover every worker

It covers 120.945 million wage and salary workers who usually work at least 35 hours a week. It excludes the self-employed and part-time workers. In the same release, median weekly pay for part-time workers fell from $396 to $391. The result should therefore not be applied to every US worker.

Another quarter is needed before calling it a policy effect

The table shows what changed. It does not show why. It cannot separate the effects of tax policy, hiring demand, industry mix, job switching, and labor supply. Housing costs, wealth, debt, and people without jobs are also outside the table. One wage report is too narrow to declare the K-shaped economy over.

Three questions matter next time. Does the 25th percentile again rise faster than the 75th? Does the inflation-adjusted median move above $374? Does part-time pay rebound? If all three improve, the policy-success case gets stronger. If only one improves, this quarter is more likely a temporary change than a durable trend.