Veterans have lower unemployment overall, but that reverses once you compare like-for-like ages

Summary

It depends entirely on whether you match for age. Raw and unadjusted, veterans have a LOWER unemployment rate than nonveterans — 3.5% vs 4.2% in 2025 (a statistically significant 0.7-point gap, Welch t-test p<0.0001 on 2022-2026 monthly data) — which is the opposite of the premise in the question. But that raw comparison is almost entirely a composition artifact: veterans are a much older population (labor force participation 51.9% vs 64.1% for nonveterans, ages 18+), and unemployment falls with age for everyone. Once you compare veterans and nonveterans of the SAME age, the gap flips: in June 2026, veterans ran higher than nonveterans in every working-age band except 45-54 — most starkly among 18-24 year-olds (17.4% vs 9.1%) and 25-34 year-olds (5.7% vs 4.5%). Post-9/11 (Gulf War-era II) veterans as a whole — the cohort closest in age to the general working-age population — ran higher than all nonveterans in June 2026 (4.8% vs 4.1%). So: no, it is not harder for the average, all-ages veteran to find work; but yes, for working-age and especially younger/post-9/11 veterans, it measurably is.

Why the raw comparison is misleading

The recipe risk here is well documented: comparing two population groups' aggregate rates without checking whether they differ in age composition can manufacture (or hide) a gap that has nothing to do with the thing you're actually asking about. Veterans 18+ numbered 16.6 million against 243.5 million nonveterans in 2023 (Census ACS), and only 51.9% of veterans 18+ are in the labor force at all, versus 64.1% of nonveterans — because a large share of the veteran population is well past typical working age and has retired out of the labor force entirely. Unemployment rate is calculated only among people IN the labor force, so it doesn't directly measure this retirement effect — but it is still an average across an age distribution that skews far older for veterans, and older workers of any background have structurally lower unemployment. Averaging that in makes the 'veteran' aggregate look better purely because it contains more people in their 50s and 60s.

The age-matched picture

BLS's own June 2026 release (Employment Situation of Veterans, published via CPS and summarized by Syracuse University's IVMF) breaks out unemployment by age band for veterans and nonveterans side by side:

Every working-age band except 45-54 shows veterans doing WORSE, not better, than nonveterans the same age. The youngest veterans — who are almost entirely post-9/11 — face by far the largest gap. This is consistent across the post-9/11 subgroup specifically: post-9/11 veterans as a whole ran 4.8% unemployed in June 2026 versus 4.1% for all nonveterans, even though post-9/11 veterans are themselves younger on average than the total veteran population.

Caveat: these are CPS unpublished, not-seasonally-adjusted cells, and single-month age x veteran-status breakdowns carry real sampling noise (BLS itself flags this for subgroup cells). The direction — veterans doing worse within age bands, especially younger ones — is corroborated across multiple 2026 monthly releases (March, May, June all show the pattern), not a one-month fluke, but exact point values will move month to month.

Other confounds worth naming

Sex is a second real confound layered on top of age, and it does not move in one consistent direction: in some 2026 months female veterans ran far higher than nonveteran women (March 2026: 7.1% vs 4.0%), in other months close to even or slightly lower (June 2026: 3.6% vs 4.1%) — volatile because the female-veteran sample is small. Race/ethnicity cells are similarly noisy month to month (e.g. Asian veteran unemployment jumped from 3.1% to 12.1% between May and June 2026 in the same release, a sample-size artifact BLS explicitly warns about) and are not a reliable basis for a headline claim in either direction.

Every query behind this report

9 warehouse calls ran in this session, in order. Each is reproducible against the same snapshot.

SELECT series, "period", "year", value FROM econ.employment_statistics WHERE series IN ('LNU04049526','LNU04049601','LNU04066408') ORDER BY "year" DESC, "period" DESC LIMIT 30

SELECT series, "year", avg(value) AS avg_rate, count(*) AS n
FROM econ.employment_statistics
WHERE series IN ('LNU04049526','LNU04049601','LNU04066408')
GROUP BY series, "year"
ORDER BY "year", series

SELECT series, "year", "period", value
FROM econ.employment_statistics
WHERE series IN ('LNU01049526','LNU01049601','LNU02049526','LNU02049601')
ORDER BY "year" DESC, "period" DESC
FETCH FIRST 20 ROWS ONLY

SELECT "year", sum(veterans) AS total_veterans, sum(nonveterans) AS total_nonveterans
FROM census.acs_veterans
GROUP BY "year"
ORDER BY "year" DESC
FETCH FIRST 5 ROWS ONLY

SELECT value, series FROM econ.employment_statistics WHERE series IN ('LNU04049526','LNU04049601') AND "year" >= 2022

SELECT series, avg(value) AS avg_level_thousands
FROM econ.employment_statistics
WHERE series IN ('LNU01049526','LNU01049601','LNU02049526','LNU02049601')
AND "year" = 2023
GROUP BY series

SELECT DISTINCT geography FROM census.acs_veterans

SELECT "year", sum(veterans) AS total_veterans, sum(nonveterans) AS total_nonveterans, sum(civilian_population_18_plus) AS total_pop
FROM census.acs_veterans
WHERE geography = 'state'
GROUP BY "year"
ORDER BY "year" DESC
FETCH FIRST 3 ROWS ONLY

SELECT series, "year", avg(value) as avg_level
FROM econ.employment_statistics
WHERE series IN ('LNU01066408','LNU02066408')
GROUP BY series, "year"
ORDER BY "year" DESC, series
FETCH FIRST 6 ROWS ONLY