Business Buyer School

Business Buyer School Research · Methods · Version 2026.1

Methodology: The Owner Age Study

Why the American Community Survey is used, how 265 industry groups were screened and scored, and what the study cannot show.

Business Buyer School calculation. Scores, ranks and rates are calculated by Business Buyer School from official Census Bureau data. These are shares of working people, not counts of businesses, and they include owners with no employees. Age does not show that an owner intends to sell: an owner aged 65 or older might sell, hand the business on or close it.

Version 2026.1 published
October 6, 2026
Analysis prepared
October 6, 2026
Data window
American Community Survey 2019 to 2023 5 year PUMS (self employed in own incorporated business) and Census SUSB 2022 firm sizes
Unit
People, not firms: ACS industry groups, with standard errors from the survey replicate weights

Short version · Full study · Methodology · Data · How to cite

Version 2026.1. Run date 2026-10-06. Calculation version succession-score-v1. Every number in the study is reproduced from the public Census files by the study's scripts.

1. Question

How old are the people who run incorporated small businesses in acquisition-relevant U.S. industries, and where are older owners combined with many small firms? This is a measure of age structure, not of owners' plans, listings, or transitions.

2. Material change from the brief, and why

The brief assumed owner-age shares from the Census Annual Business Survey. The ABS public API and tables publish owner sex, ethnicity, race and veteran status for employer firms, but not age (verified against the variable lists for the 2017 to 2023 owner-characteristics datasets). No current public ABS table gives owner age by industry for employer firms; older Survey of Business Owners releases are not used here. Chosen alternative: ACS 2019 to 2023 5-year PUMS, class of worker 7 (self-employed in own incorporated business), the standard public proxy. Consequences, all disclosed: (1) units are people, not firms or owners of employer firms; (2) solo incorporated owners are included; (3) industry is self-reported by the worker; (4) the finest industry detail is the ACS group (265 groups; construction is a single group); (5) no firm-count is multiplied by an age share anywhere in this study, and no "number of businesses with owners over X" is estimated. Counts of workers in the data file are ACS tabulations of people.

3. Sources

ACS 2019 to 2023 5-year PUMS (persons; 15,912,393 records read); ACS 2023 PUMS industry code list (composition of each industry group in NAICS 2022); Census SUSB 2022 (firms by enterprise employment size). SUSB 2022 is published on NAICS 2017 codes at the detail used here, so firm measures are used only for the 204 groups whose NAICS codes are identical in both vintages; for the 27 groups that fail the match (23 retail and 4 information groups) owner age is reported but no index is computed. Nothing is allocated or crosswalked.

4. Universe

Start: 265 ACS industry groups. Rules (structural, outcome blind): U1 agriculture, mining, utilities, postal, holding companies, regulated finance (22 groups); U2 nonprofit, government, religious, military, institutional-care (23); U3 infrastructure, telecom, broadcasting, common carriers (8); U4 mean employees per firm above 75 in SUSB 2022 (35); U5 owner-practitioner (1); U6 "not specified" or partial groups (18); U7 fewer than 100 ACS sample records of incorporated self-employed workers (18). Result: 125 excluded; 76 ranked; 37 unranked low confidence; 27 age only. Every row and reason is in the CSV. Note a difference from Report 1: the ACS group for securities and investment funds (which includes investment advice) is excluded here as a whole.

5. Confidence tiers

High: 1,000 or more ACS sample records of incorporated self-employed workers; Medium: 400 to 999; Low: 100 to 399 (unranked). A group is moved down one tier if the standard error of its 55-or-older share exceeds 4 percentage points. Ranked: High 47, Medium 29.

6. Formulas

Shares are weighted (PWGTP) shares of workers in each age band within group. Standard errors: ACS successive-difference replication with 80 replicate weights, SE = sqrt(4/80 times the sum of squared differences between each replicate estimate and the full-sample estimate). Reported intervals are 1.96 standard errors (a 95% interval). Context: all-employed-worker share aged 65 or older in the same group; gap = owner share minus all-worker share. SUSB: small-firm share = firms under 20 employees divided by all firms; firm counts add code totals within a group and subtract excluded codes (a firm in several codes of one group is counted once per code).

7. Index and normalization

Components are mid-rank percentiles (0 to 100) within the 76 ranked groups: owners 65 or older, owners 55 to 64, share of firms under 20 employees, count of firms under 20 employees. The brief's weights 35/20/20/15/10 included a mature-firm share (15). That component cannot be built below sector level, so Model A drops it and renormalizes: 35/20/20/10 divided by 85. Higher means older owners and more small firms. This is an index of age structure and firm population, not of expected sales.

8. Sensitivity (weights tested, not assumed)

Models: A balanced; B age-heavy 50/30/10/10; C supply-heavy 25/15/30/30; D equal; E age only 60/40; F age-gap-adjusted (owner share minus all-worker share in place of owner share). Plus 3,000 Dirichlet weight draws and 300 sampling draws that perturb the age shares by their standard errors. Minimum pairwise Spearman 0.45 (age-only against supply-heavy). Stability: 8 stable highest, 7 stable lowest, 38 method-sensitive, 23 moderate. 4 top-quartile groups fall out of the top quartile in more than one in five sampling draws. Because the index is not stable to weights, the report leads with the owner-age shares and treats the index as secondary.

9. Red-team review

  1. Is incorporated self-employed a fair proxy for owners? It captures owner-operators of incorporated businesses and also solo corporations. Unincorporated (sole proprietor) shares correlate 0.80 with the incorporated shares, and an unincorporated rescore of the index agrees at Spearman 0.82 (n=76).
  2. Is owner age just workforce age? Partly: correlation 0.61. The age-gap-adjusted model F keeps 16 of the top 20.
  3. Do small samples create false rankings? Intervals are shown; low-confidence groups are unranked; 4 top-quartile groups are sampling-unstable.
  4. Does age mean intent to sell? No. The report says so repeatedly and makes no estimate of listings or sales.
  5. Do old owners signal opportunity? Possibly supply of transitions; possibly closures. Unknown from this data.
  6. Could a single wave-year cohort effect distort results? The 2019 to 2023 pool spans COVID; no trend is claimed.
  7. Does pooling five years age the file? Yes; ages are self-reported at survey date across five years.
  8. Are buyer conclusions stronger than the evidence? Buyer section limited to what the data show.

10. Limitations

Workers not firms; solo owners included; no firm age; broad groups (construction is one); national only; ACS sampling error; no trend over time; SUSB firm counts overlap across codes within a group; 27 mostly retail groups have age data but no index; owner age is a structural fact of an industry as much as a sign of turnover.

11. Versioning

YYYY.N. New ACS or SUSB releases trigger a re-run and a version increment; any change in weights, thresholds or universe rules needs a calculation_version increment.

How to cite

Cite as: Business Buyer School, America's Small Business Succession Opportunity: Owner Age by Industry (methodology), https://businessbuyerschool.com/research/business-owner-age/methodology/. Version 2026.1. Published October 6, 2026.

A data license has not been published; short quotations with attribution and a link are welcome.

Take this into your own search

Industry research is context, not a verdict on any one business. These free parts of the school show how to use it.