Missing BLS Wage Data for Your Trade? OEWS Suppression Explained

The Friday you open the BLS page and the cell is blank

You have a commercial HVAC service tech candidate who wants an answer by Monday. You know the offer has to sit somewhere defensible, so on Friday afternoon you pull up the BLS wage page for your metro, find the occupation, and scan across to the percentile columns. The median is there. The 10th percentile is there. The 90th percentile is an asterisk, and the footnote gives you no reason you can use. You check the next metro over and the whole row is dashes.

The reflex is to assume you pulled the wrong file, or that BLS made a mistake, or that nobody in your area does this job. None of those is usually true. What you are looking at is an estimate that BLS chose not to publish, and the rules behind that choice are knowable. Once you know them, a blank cell stops being a dead end and becomes a signal about how thin the local data is and where to look instead.

This article explains how BLS OEWS suppressed wage data works in the Occupational Employment and Wage Statistics (OEWS) program, why small trade occupations in small areas get hit hardest, and a practical order of operations for pricing a trade offer when your metro's numbers are missing.

What BLS OEWS suppressed wage data actually means

OEWS is a federal survey of employer payroll records. The program covers over 800 occupations at the national, state, and metropolitan and nonmetropolitan area levels, and the wage files are public domain (BLS OEWS). The most recent reference period is May 2025, released May 15, 2026, after a delay caused by the October 1 to November 12, 2025 federal shutdown. BLS reported that the response rate stayed within its normal range and that no methodology changes were needed (BLS release archive).

The key point is that OEWS is a sample, not a census. According to the May 2025 technical note, the survey draws roughly 1.1 million establishments across six semiannual panels over three years, and it represents about 55% of national employment (BLS OEWS Technical Note). That is a large survey. But when you slice it by occupation and by metro, the number of responding employers behind any single cell can get very small.

"Suppressed" is the word for an estimate BLS holds back because publishing it would not meet the program's standards. The two common reasons are that the estimate is not reliable enough, or that publishing it could allow a reader to identify a specific employer's data. The exact thresholds are set out in the BLS technical notes, and they can change between releases, so we do not quote a number here. Read the current technical note on bls.gov for the criteria that apply to the release you are using.

Suppression is different from a true zero. A blank cell does not tell you the occupation does not exist in that metro. It tells you BLS could not responsibly publish a figure.

How to read the asterisks, dashes, and footnotes

BLS tables use symbols in place of numbers, and the legend sits in the footnotes of each table or file. A symbol can stand for an estimate that was withheld, or for a wage that is capped rather than missing. Those are different situations, and treating them the same leads to bad offers.

A capped value at the top of the range is not missing data. It tells you that your top percentile is high enough to be capped, which is useful if you are hiring a senior or supervisory role. A symbol standing in for a withheld estimate is the case this article is about.

Always check the legend for the specific file you downloaded. Symbol conventions can differ between the downloadable files, the data maps, and the occupation profile pages, and a quick read of the footnote saves you from misinterpreting a cell.

If you are still getting comfortable with the columns themselves, start with how to read BLS OEWS data for trades. It walks through the layout of the tables before you get to the gaps.

Why trade occupations in small metros lose data first

Suppression is not random. It concentrates where the sample is thinnest, and trade work hits that pattern in a few ways.

Narrow occupation codes. OEWS publishes by Standard Occupational Classification (SOC) code. A specialty like structural iron and steel workers (SOC 47-2221) or sheet metal workers (SOC 47-2211) employs far fewer people in a typical metro than a broad category does. Fewer workers means fewer sampled employers, and fewer sampled employers means a higher chance a percentile fails the publication test.

Small and nonmetropolitan areas. The program publishes roughly 530 metropolitan and nonmetropolitan areas (BLS OEWS News Release, May 2025). A rural nonmetropolitan area is a large geography with a small number of employers in any one trade, so the same problem shows up there.

Outer percentiles. OEWS reports five wage percentiles, the 10th, 25th, 50th, 75th, and 90th (BLS). Extreme percentiles can be missing even when the median is published, so you may see a published median with a missing 10th or 90th percentile for the same occupation and area. If you need a refresher on what each percentile tells you, see wage percentiles explained for trades.

Geography definitions. The metropolitan area boundaries matter too, because a boundary decides which employers are pooled together. If you are not sure what counts as part of your metro, what an MSA is and why it matters for wage data covers it.

None of this means the work is unpaid or unmeasured in your area. It means the federal sample cannot always say so at the local level.

What to do when BLS OEWS suppressed wage data hits your metro

Work down this list in order. Each step widens the lens a little, and you stop as soon as you have enough to anchor an offer.

1. Use whatever percentiles did publish. A missing 90th does not erase a published median and 25th. If the middle of the distribution is there, you can still anchor on it. Be honest with yourself that the top of your range is a judgment call.

2. Check the median versus the mean. Sometimes a mean is published when a percentile is not. They measure different things and skew differently in trades, so read mean vs median wage for trades before you substitute one for the other.

3. Move up to the state level. State estimates for the same occupation draw on far more responding employers than any one metro does, so they are often more likely to be published. Treat the state figure as a starting point and adjust for how your metro's cost of labor compares, using what you know about local competitors.

4. Try a neighboring metro with a similar labor market. If a comparable area has a published figure, it can bracket your own. Note the difference in size and industry mix before you lean on it.

5. Fall back to the national figure. The national estimate almost always publishes. For a sense of scale, BLS reported a $62,350 median for electricians (SOC 47-2111) and $62,970 for plumbers, pipefitters, and steamfitters (SOC 47-2152) in May 2024 (BLS Occupational Outlook Handbook). Those are national medians from the earlier release, shown here to illustrate the type of anchor available, not as a substitute for the current May 2025 files.

6. Step to a related SOC code. If your exact occupation is suppressed, a closely related code with a published estimate can serve as a cross-check. Document why you chose it.

7. Use your own payroll and the market. Your current crew's pay and the offers your candidates are actually fielding are real data. Federal figures are a floor for the conversation, not a replacement for it.

Whichever step you land on, write down the anchor, the release it came from, and the adjustment you made. A band you can explain is easier to defend than a number that appeared from nowhere.

Turning an imperfect anchor into an offer band

A worked example shows the method. Suppose the state median for your trade is $60,000 and your metro's median is suppressed. You decide the state median is your anchor, and you set a spread buffer of 10% either side because you know your market runs a little hotter than the state as a whole. The band is a minimum of $54,000, a midpoint of $60,000, and a maximum of $66,000.

Those inputs are round numbers chosen to demonstrate the arithmetic, not a real wage. The point is the structure: pick one percentile as the anchor, state the geography it came from, choose a spread, and let the band follow. SkilledMarkets joins BLS OEWS wage percentiles with O*NET occupational profiles by SOC code and generates min, midpoint, and max bands from a percentile anchor and spread buffer in this same way, but the method works in a spreadsheet too.

The same data is free. Free BLS data versus a wage tool compares doing it by hand with using software, and bls.gov remains the authoritative source either way.

Suppression means the estimate was withheld, not that the wage does not exist. Move to a wider geography, document the choice, and keep the release date next to every number.

Keep a record of the gaps

Because OEWS is released once a year, suppressed wage data can change between releases: a cell that is blank this year may publish next year, and one that published may disappear. Keep a short log of which trades and metros you checked, which cells were suppressed, and which fallback you used. When the next release arrives you can re-check only the gaps instead of starting over.

If you want the percentile, mean-versus-median, and suppression rules in one reference, the OEWS Percentile & Suppression Reader Guide explains how to read any BLS OEWS trade wage page for owner-operators. And if you price one trade across states, the State-by-State Trade Wage Atlas renders a single trade's cited OEWS wage distribution across all 50 states and DC, with suppression shown honestly.

If you would like these explainers as they publish, subscribe to the newsletter. Wage data is public, and the work is in reading it correctly.