Why Averaging BLS Earnings Medians Can Miss the Headline

The Bureau of Labor Statistics' 21 October earnings release reports three weekly medians for full-time wage and salary workers: $1,232 overall, $1,352 for men and $1,127 for women. A spreadsheet that averages the last two numbers, even with workforce weights, cannot reliably reconstruct the first. Group medians omit the earnings distribution needed to find the combined middle.

These are U.S.-dollar estimates for the third quarter of 2026, not seasonally adjusted. The dated BLS release covers July through September. Its publication on 21 October does not make the figures October earnings.

For someone using wage data to assess consumer conditions, the aggregation rule matters. A calculation can use the correct group labels and weights yet still produce the wrong statistic.

Keep the published medians separate

BLS Table 2 supplies the following selected records. Each concerns full-time wage and salary workers aged 16 and over, uses current U.S. dollars per week, and is a third-quarter 2026 average without seasonal adjustment.

Total, men and women combined: median usual weekly earnings of $1,232.

Men: median usual weekly earnings of $1,352.

Women: median usual weekly earnings of $1,127.

The technical note defines the median as the midpoint of an earnings distribution. BLS estimates it by grouping weekly earnings into $50-wide intervals centered on multiples of $50, then interpolating within the interval containing the middle. The published amount therefore need not equal any particular respondent's reported pay.

The total median comes from the combined distribution. Subgroup medians describe the middle within each subgroup. Worker counts tell a reader how large the groups are, but leave the earnings above and below those separate middles unknown.

Six fictional workers show the problem

Consider six fictional workers split into two groups of three. Each worker has equal weight. All amounts are nominal U.S. dollars of usual weekly earnings. These invented observations are not calibrated to BLS data and do not reproduce its survey weights or interpolation procedure.

Group A's complete earnings record is $600, $700 and $2,000 per week. Its middle observation, and therefore its median, is $700.

Group B's complete record is $800, $1,200 and $1,300 per week. Its median is $1,200.

Each group contains half the workers. Weighting their medians by those correct shares gives (0.5 × $700) + (0.5 × $1,200) = $950 per week.

Now put all six observations in order: $600, $700, $800, $1,200, $1,300 and $2,000. Using the ordinary sample-median convention for this fictional even-sized group, average the third and fourth observations: ($800 + $1,200) / 2 = $1,000 per week.

The weighted average of group medians is $50 below the combined median. Every observation and both group sizes were available; the failed calculation threw away information when it kept only the two medians. No seasonal adjustment, unequal group size or rounding is needed to produce the mismatch.

That example establishes a mathematical limitation. It supplies no explanation for the particular distance between the actual men's, women's and overall BLS medians. Reconstructing the published survey estimate would require the relevant weighted earnings distribution and BLS's estimation method.

Keep a weekly median out of a wage-bill calculation

Multiplying a median by a worker count does not generally recover total earnings. Workers on either side of the middle can have very different pay while leaving that middle unchanged. A median alone therefore cannot establish the aggregate dollars available for consumption, much less a particular company's customer revenue.

The BLS measure also has a defined population. Its technical note covers employees' usual earnings before taxes and deductions at their main job, including usual overtime, commissions or tips. Full-time means usually working at least 35 hours per week at the sole or principal job. Self-employed workers are excluded. The weekly unit is an earnings rate, not the amount paid across the entire quarter or a measure of household wealth.

These are survey estimates subject to sampling and nonsampling error. Keep the 21 October release vintage with the figures rather than silently replacing them when a rolling source page changes. The release separately prints a seasonally adjusted overall median of $1,239; combining that total with the unadjusted subgroup figures above would add an adjustment mismatch to the aggregation error.

Our earlier real-average-earnings article concerns payroll averages and changing job mix. The issue here arises when combining subgroup medians within the same quarterly release, before any comparison over time.

The release clock supplies no option valuation

BLS's October calendar and the dated release identify 10:00 a.m. Eastern Daylight Time on Wednesday, 21 October 2026, equivalent to midnight at the start of 22 October in Brisbane. This article was prepared at 07:06:59 Australia/Brisbane on 23 October.

The latest completed scheduled NYSE core equity session at preparation was Thursday, 22 October. Its scheduled 4:00 p.m. EDT close was 06:00 Brisbane on 23 October. That is a calendar boundary, not an observation of a closing price or a claim that every venue or extended session had closed. This article measures no stock-price, option-price or implied-volatility response to the earnings release.

An option still needs its own underlying, strike, expiration, multiplier, exercise style, settlement terms and timestamped bid, ask and available size. Our options-chain guide explains that contract record. A wage median cannot determine an individual option's price or profitability.

Options Matrix Pro publishes this article and offers commercial options-analysis software. This is general education, not personal financial, legal or tax advice or a recommendation to trade. Buyers can lose the entire premium; some short positions can lose much more, including unlimited loss on an uncovered call. American-style short options can be assigned before expiration. Exercise and settlement can create share-delivery or cash-funding obligations. Check the specific contract and the broker's approval requirements, deadlines and liquidation policies. Bid-ask spreads, limited liquidity, commissions and other charges can change an exit or its cost; concentration and taxes require separate consideration. Options are unsuitable for someone unable to tolerate the losses or fund the obligations. Read FINRA's options guidance and the current OCC disclosure document before trading.

Sources and method

Primary sources were accessed on 23 October 2026 Australia/Brisbane. The 21 October BLS archive, Table 2 and technical note supply the published figures, population and estimation method. The BLS calendar and NYSE hours and holidays establish the stated time boundaries. FINRA, the OIC's liquidity and account FAQ and OCC provide options-risk references. The six-worker example is fictional author arithmetic, with no market model or forecast.

When recording the third-quarter earnings headline, retain the published overall median. Use the subgroup medians for comparisons within their stated populations, and require a combined earnings distribution before attempting to calculate a new overall median.