[Home](https://stoneforgegroup.com/) · [Research](https://stoneforgegroup.com/research/) · Who Owns These Firms

05 – Research

# Who owns these firms

Among professional, scientific and technical services firms, owners aged 55 or over run from 39.4% in computer systems design to 62.8% in accounting, tax preparation, bookkeeping and payroll services – a spread of more than 23 points inside the single NAICS 54 code that reports 52.6% overall.

## Executive summary

This study reads the Census Bureau's 2022 Annual Business Survey Characteristics of Business Owners (abscbo) – the ABS component that surveys the people who own respondent employer firms, not the firms themselves – for two questions: owner age (QDESC O09) and the year an owner acquired or established their ownership stake (QDESC O02). It measures owner age and tenure. It does not measure, and does not claim to measure, anything about what happens to a firm afterward.

The flat NAICS 54 figure – 52.6% of owners are 55 or over – is not this study's headline, because it hides a spread the underlying data resolves. Broken out by four-digit subsector, the same 55-or-over share runs 62.8% in accounting, tax preparation, bookkeeping and payroll services (NAICS 5412), 58.6% in architectural, engineering and related services (5413), 56.7% in legal services (5411), 50.1% in management, scientific and technical consulting services (5416), and 39.4% in computer systems design and related services (5415) – a spread of more than 23 points among subsectors that all sit inside one two-digit industry code. Across five broader sectors, the same 55-or-over share runs from 49.5% in construction to 58.8% in real estate, with professional services in between at 52.6%.

Two properties of this dataset bound every figure on this page. OWNPDEMP counts owners of respondent employer firms, not firms – a firm that reports several owners is counted once per owner, so a percentage of owners is not a percentage of firms. And abscbo carries no firm-size variable of any kind: nothing here can be joined to employee count, revenue, or any other measure of firm size. What this does not show, below, states the full limits of this measurement, including what the data cannot say about firms with more than one owner and about geography.

Share of owners aged 55 or over, by professional-services subsector. Full NAICS titles in the table below. US Census Bureau, Annual Business Survey, Characteristics of Business Owners, 2022.

## Owners aged 55 or over run from 39.4% of owners in computer systems design to 62.8% in accounting, tax preparation, bookkeeping and payroll services, both inside the single NAICS 54 code that reports 52.6% overall.

NAICS 54 – professional, scientific and technical services – is not one industry. It is a code that bundles legal services, accounting, architecture and engineering, computer systems design, management consulting, advertising and several smaller lines of work into a single two-digit number. The 2022 Annual Business Survey's Characteristics of Business Owners component publishes owner age at the four-digit level within it, which is finer detail than either of the two studies published alongside this one could get for firm size or firm age – both of those had to stop at the two-digit sector and disclose that they could not rule out subsector composition as the real driver of what they measured. This page can show the owner-age breakdown directly because this one dataset does not share that limitation – but owner-age composition is not firm-size or firm-age composition, and this page's subsector table says nothing about whether the leverage-gap or firm-size-and-age studies' own findings are driven by subsector mix. Their gap stays open; only the owner-age question this page asks is resolved at this level of detail.

Share of owners aged 55 or over, professional-services subsectors, 2022

NAICS
Subsector
55 to 64
65 or over
55 or over

5412 Accounting, tax preparation, bookkeeping & payroll services 31.9% 30.9% 62.8%
5413 Architectural, engineering & related services 32.4% 26.2% 58.6%
5411 Legal services 27.1% 29.6% 56.7%
54 **All of NAICS 54** 29.1% 23.5% 52.6%
5416 Management, scientific & technical consulting services 28.6% 21.5% 50.1%
5415 Computer systems design & related services 27.5% 11.9% 39.4%

The 55-to-64 band moves comparatively little across these five subsectors, from 27.1% to 32.4%. Nearly the entire spread sits in the 65-or-over band, which runs from 11.9% in computer systems design to 30.9% in accounting – close to three times as large a share of owners are past 65 in the older-skewing subsector as in the younger one. Whatever separates these five lines of work, it shows up mainly among the owners who have already passed 65, not the ones approaching it.

## The five-sector comparison

Widening the lens from professional services to five broad sectors – construction, finance and insurance, real estate, professional services, and administrative and support – shows a narrower range than the professional-services subsectors do on their own: 49.5% to 58.8%, against 39.4% to 62.8% within NAICS 54 alone. The subsector spread inside one industry code is wider than the spread across five different industries.

Share of owners aged 55 or over, five sectors, 2022

NAICS
Sector
55 or over

53 Real estate and rental and leasing 58.8%
52 Finance and insurance 55.4%
54 Professional, scientific and technical services 52.6%
56 Administrative and support services 50.3%
23 Construction 49.5%

Professional services sits close to the middle of this range, not at either end of it. The wider finding on this page is inside NAICS 54, not between it and its neighboring sectors.

## How long these owners have held their firms

The same survey asks a separate question – QDESC O02 – for the year an owner acquired or established their ownership stake, grouped into bands. In professional services, 11.8% of owners answered "don't know" rather than naming a year or a band; in real estate, 24.5% did, more than double the professional-services share and the highest of the five sectors this page covers.

Owners who answered "don't know" when asked the year they acquired their ownership stake, 2022

NAICS
Sector
Don't know

53 Real estate and rental and leasing 24.5%
23 Construction 15.2%
56 Administrative and support services 15.3%
52 Finance and insurance 13.3%
54 Professional, scientific and technical services 11.8%

The rest of professional services' owners – 88.2% of the 681,766 who reported an answer to this question at all – gave a year or a named band. Grouped into the bands the survey publishes, 42.4% acquired their stake before 2010 and 45.7% acquired it in 2010 or later, with the 2010s alone – 2010 through 2019 – the single largest band at 36.8%.

Professional services (NAICS 54): year ownership was acquired, 2022

Year band
Share of owners

Before 1980 1.9%
1980 to 1989 5.4%
1990 to 1999 12.2%
2000 to 2009 22.9%
2010 to 2019 36.8%
2020 4.1%
2021 3.6%
2022 1.2%
Don't know 11.8%

## What this does not show

Stated plainly, because a number without its limits is not evidence.

**This dataset has no firm-size variable of any kind.** abscbo does not carry employment size, revenue size, or any other measure of firm size – requesting EMPSZFI on this dataset returns HTTP 400, "unknown variable," not a partial or suppressed result. There is no query shape that joins owner age to firm size in this data, so no figure on this page can be read as "X% of firms with N employees have an owner 55 or over." The two studies published alongside this one measure firm size and firm age; this one measures owner age and tenure; nothing here connects the two.

**Owners are not firms.** OWNPDEMP counts owners of respondent employer firms, and a firm that reports more than one owner is counted once per owner it reports, not once per firm. Per the Annual Business Survey's Characteristics of Businesses component (abscb), about 67.3% of NAICS 54 firms report exactly one owner and 28.6% report two to four – so "52.6% of owners are 55 or over" is what this data supports, and "52.6% of firms" is not, because a two-to-four-owner firm can be counted several times in the owner-level figures above while it is one firm in any count of businesses.

**No federal dataset measures what an owner intends to do next.** Checked across abscbo, abscb and absmcb (Management and Business Characteristics) for 2022 and the years around it, none carries a variable asking whether an owner plans to sell, transfer, wind down, or continue running their firm. An ownership base concentrated at 55 or over does not, by itself, indicate a coming wave of business sales any more than it indicates acquisition, consolidation, or firms simply continuing under the same owner – all of those are equally unmeasured here, and this page asserts none of them.

**This is a snapshot, not owners followed over time.** Every figure on this page describes owners as the 2022 survey found them in a single year, not the same owners observed as they age. It cannot show an owner's age or tenure changing, only what owners of each kind looked like in 2022.

**A meaningful share of owners could not or did not answer the tenure question.** 11.8% of professional-services owners and up to 24.5% of real estate owners answered "don't know" when asked the year they acquired their ownership stake – see How long these owners have held their firms, above. Those owners are counted in the tenure figures as "don't know," not folded into any year band, and are excluded entirely from the age figures only if they also left O09 unanswered, a much smaller share (see Methodology and sources, below).

## Methodology and sources

This study is meant to survive an adversarial read. Every figure below is reproducible from the steps stated here, and every judgment call is disclosed rather than left for a reviewer to guess at.

### Data source

US Census Bureau, Annual Business Survey (ABS), 2022 data year. Table abscbo (Characteristics of Business Owners), at https://api.census.gov/data/2022/abscbo. Owner age is question O09; year ownership was acquired is question O02. The universe is owners of respondent employer firms – people, not businesses – drawn from the same stratified sample of roughly 230,000 employer firms the companion leverage study draws its own figures from, so every figure on this page carries sampling error for that reason before any judgment call documented below.

Owner-count figures (the "67.3% report one owner" comparison in What this does not show, above) come from a different ABS component, abscb (Characteristics of Businesses), question B01, which counts firms rather than owners.

### The query trap: a missing base variable returns 204, not an error

Requesting a _LABEL variable without its corresponding base variable does not fail loudly on this dataset – it returns HTTP 204, an empty response with no error message. Confirmed live against the Census API on 7 August 2026:

https://api.census.gov/data/2022/abscbo?get=OWNCHAR_LABEL&for=state:12&NAICS2022=54&QDESC=O09&key=REDACTED
→ HTTP 204, empty body

https://api.census.gov/data/2022/abscbo?get=OWNCHAR,OWNCHAR_LABEL&for=state:12&NAICS2022=54&QDESC=O09&key=REDACTED
→ HTTP 200, eight rows

The two calls differ only in whether OWNCHAR, the base variable, is requested alongside OWNCHAR_LABEL. Nothing in the 204 response distinguishes "this geography has no data" from "this query is malformed" – both look identical to a script that does not check the status code, and a scoping pass that made this mistake against a state-level query would conclude that state's data does not exist. It does; the query was wrong. Every query in this study's own pull requests OWNCHAR alongside OWNCHAR_LABEL for exactly this reason.

### The exact query

The literal call, with the key redacted:

https://api.census.gov/data/2022/abscbo?get=QDESC,OWNCHAR,OWNCHAR_LABEL,OWNPDEMP,OWNPDEMP_PCT,OWNPDEMP_PCT_S&for=us:*&NAICS2022=54&QDESC=O09&key=REDACTED

Run once per NAICS code for owner age (ten values: 23, 52, 53, 54, 56, 5411, 5412, 5413, 5415, 5416), and once per NAICS code for tenure with QDESC=O02 in place of O09 (five values: 23, 52, 53, 54, 56). Geography is for=us:* for every one of those calls; the Florida figure quoted in Geography, below, is the same query with for=state:12 in place of for=us:* and NAICS2022=54 only. The script issuing these calls is scripts/research_pull.py, function owner_rows().

### Variable definitions

- **QDESC** – Census name "Question description code." Which survey question a row answers: O09 for owner age, O02 for year ownership was acquired.

- **OWNCHAR** – Census name "Owner characteristic code." Which category within a question a row describes – an age band for O09, a year band for O02. Units: a categorical code, not a quantity.

- **OWNCHAR_LABEL** – The plain-text label for OWNCHAR, e.g. "55 to 64" or "Don't know." Requires OWNCHAR in the same call – see The query trap, above.

- **OWNPDEMP** – Census name "Number of owners of respondent employer firms." A count of owners, not firms – see What this does not show, above.

- **OWNPDEMP_PCT** – The share of OWNPDEMP's "Total reporting" row that a given OWNCHAR category represents, as published by Census, already rounded to one decimal place. Every percentage on this page is this published figure, not a value recomputed from raw counts.

- **OWNPDEMP_PCT_S** – Standard error of OWNPDEMP_PCT, in percentage points, as published by Census.

### How each figure is computed

Every 55-or-over percentage on this page is the sum of two published, already-rounded OWNCHAR categories – "55 to 64" and "65 or over" – added directly, not recomputed from the underlying owner counts. For NAICS 5412: 31.9% (55 to 64) + 30.9% (65 or over) = 62.8%. This matches the convention the leverage-gap study uses for its own band-to-band percentages: adding published rounded figures, not re-deriving a finer one that could disagree with them by a tenth of a point through independent rounding.

The "before 2010" and "2010 or later" tenure figures in How long these owners have held their firms, above, are sums of the published year-band percentages the same way: 1.9% + 5.4% + 12.2% + 22.9% = 42.4% before 2010; 36.8% + 4.1% + 3.6% + 1.2% = 45.7% in 2010 or later. Neither sum includes the 11.8% who answered "don't know," so the three figures – 42.4%, 45.7% and 11.8% – add to 99.9%, not 100.0%, purely from rounding each published band to one decimal place before summing.

### Owner age published at the subsector level – unlike firm size or firm age

Confirmed live against the Census API on 7 August 2026, abscbo returns real OWNCHAR rows for owner age at the four-digit NAICS level: NAICS2022=5411, 5412, 5413, 5415 and 5416 each return the full eight-row age breakdown, not a single "all firms" aggregate. This is the opposite of what the companion leverage-gap and firm-size-and-age studies found in their own datasets – abscs and the BDS time series both publish size or age detail only at the two-digit sector level, which is why both of those studies had to disclose that they could not rule out subsector composition as the real driver of what they measured. This dataset does not carry that limitation for owner age specifically, which is why this page reports the subsector breakdown directly – but that closes the subsector question only for owner age. It says nothing about the size or age composition behind the leverage-gap and firm-size-and-age studies' own findings; their gap remains open.

### Geography: national and state work; metro, county and state-by-subsector do not

Confirmed live against the Census API on 7 August 2026. National totals (for=us:*) and state totals (for=state:, two-digit FIPS) both return real data for owner age at the two-digit NAICS level. Florida, NAICS 54, O09:

Owner age, Florida, professional services (NAICS 54), 2022

Age band Share Standard error

Under 25 0.3% 0.1
25 to 34 3.9% 0.4
35 to 44 16.0% 0.6
45 to 54 26.5% 0.8
55 to 64 30.2% 0.9
65 or over 23.1% 1.0

30.2% + 23.1% = 53.3% of Florida's professional-services owners are 55 or over, close to the 52.6% national figure. Each component standard error sits at or under one percentage point. Combined in quadrature, the same conservative method the leverage-gap study applies to its own combined errors – treating the two bands as independent, which overstates the true combined error since a firm that under- or over-reports tends to do so consistently – the standard error on the 53.3% figure itself is about 1.3 percentage points, not under one point on its own; the individual bands behind it are.

Two geography combinations that might seem like the natural next step return nothing at all. Metropolitan and micropolitan statistical areas return HTTP 204 for professional services even for the largest US metros – confirmed live against New York-Newark-Jersey City (CBSA 35620). Counties return HTTP 204 the same way – confirmed live against Los Angeles County. And crossing a state with a four-digit subsector, rather than the two-digit sector, also returns HTTP 204 – confirmed live against Florida, NAICS 5411 (legal services). State-level detail exists only at the two-digit sector; subsector detail exists only at the national level. The two cannot currently be combined.

### No dataset measures intent to exit

Checked against the published variable lists for abscbo, abscb and absmcb (Management and Business Characteristics) for the 2022 ABS cycle: none of the three carries a variable naming succession, sale, exit, transfer, or any owner's stated plan for their firm's future. Owner age and tenure are the only things this page measures.

### Owner count, by number of owners a firm reports

US Census Bureau, Annual Business Survey, table abscb (Characteristics of Businesses), question B01 ("Number of owners"), 2022. Confirmed live against the Census API on 7 August 2026, NAICS 54:

Number of owners per firm, professional services (NAICS 54), 2022

Owners reported Share of firms Standard error

1 person 67.3% 4.0
2 to 4 people 28.6% 0.9
5 to 10 people 1.4% 0.5
11 or more people 0.6% 0.6
Owned by a parent company, estate, trust or other entity 1.6% 1.1
Unknown number of owners 0.5% 0.3

This table is what supports "67.3% report exactly one owner" and corrects a looser reading: the second-largest category is firms reporting two to four owners, 28.6% – not firms reporting exactly two. No published Census category isolates exactly two owners on its own. Note also that this question's own standard error on the largest category, 4.0 percentage points on a 67.3% share, is the widest margin published on this page – wider than any error on the owner-age or tenure figures above.

### Inclusion and exclusion rules

- **"Total reporting" is the denominator for every percentage.** OWNPDEMP_PCT is Census's own published share of each category against that question's "Total reporting" row (OWNCHAR code EF for O09, CO for O02), which already excludes owners who left the question blank ("Item not reported"). This study does not recompute that denominator.

- **Ten NAICS codes for owner age, five for tenure.** Owner age (O09) is pulled for all ten codes this page reports: the five NAICS 54 subsectors, NAICS 54 itself, and the four other broad sectors in the five-sector comparison. Tenure (O02) is pulled only for the five broad sectors – the subsector tenure breakdown is not published on this page, though the same query pattern would retrieve it.

- **Florida is the only state-level pull.** This study runs one state-level query, Florida NAICS 54 O09, to establish what geography this dataset does and does not support (see Geography, above). No other state is pulled.

### Reproducing this study

The pull script is scripts/research_pull.py, functions owner_rows() and owner_count_rows(). Run from the repo root, with the Census API key passed as an environment variable, never hard-coded – the script aborts immediately if CENSUS_API_KEY is unset:

CENSUS_API_KEY=your_key_here python3 scripts/research_pull.py

It writes [assets/data/owner-age-tenure-2022.csv](https://stoneforgegroup.com/assets/data/owner-age-tenure-2022.csv) (owner age and tenure, every NAICS code and both geographies this study pulls) and [assets/data/owner-count-2022.csv](https://stoneforgegroup.com/assets/data/owner-count-2022.csv) (the number-of-owners breakdown, above), alongside the companion studies' own CSV output in the same run.

scripts/research-provenance.py runs on every deploy and checks every figure published on this page against a manifest of declared sources, offline – it never calls Census, so a deploy never depends on the API being up. A separate --refresh mode, run manually and not part of the deploy, re-pulls from Census live and reports any drift between the declared figures and what Census now returns.

### Known limitations

The limitations that apply to every study on this site – this is a snapshot rather than the same owners followed over time, and the underlying survey is a sample, not a census – are stated in full on [the research hub](https://stoneforgegroup.com/research/) and in What this does not show, above, and are not repeated here.

Specific to this methodology:

- **No firm-size join is possible.** abscbo carries no firm-size variable; requesting EMPSZFI on it returns HTTP 400. See What this does not show, above.

- **Owners are not firms.** A percentage of owners is not a percentage of firms wherever a firm can report more than one owner, which 32.7% of NAICS 54 firms do. See Owner count, above, and What this does not show.

- **No exit-intent variable exists in any dataset checked.** See No dataset measures intent to exit, above.

- **Margin of error stated for the headline comparisons only.** This study pulled OWNPDEMP_PCT_S for every row it publishes and reports the individual component errors directly in the tables above; the one combined (quadrature) calculation on this page is the Florida 55-or-over figure, in Geography. A reviewer who wants a combined error on any other pair of bands can compute it the same way from the published component errors.

- **Subsector detail exists for owner age, not for tenure.** This study did not pull O02 (tenure) at the four-digit level; only the five broad sectors are published for tenure. Nothing here indicates whether the tenure spread would resemble the age spread inside NAICS 54.

- **The CSV stores figures as published, not further rounded.** owner-age-tenure-2022.csv and owner-count-2022.csv store OWNPDEMP_PCT and FIRMPDEMP_PCT exactly as Census publishes them – one decimal place – the same figures shown on this page, not a finer or coarser rounding.

## Citing this work

Stoneforge Consulting Group, *Who Owns These Firms*, 2026. Published 7 August 2026. Analysis of US Census Bureau Annual Business Survey, Characteristics of Business Owners, 2022. https://stoneforgegroup.com/research/who-owns-these-firms/

## More research

This study is part of Stoneforge's ongoing analysis of federal business data. See [all published studies, sources and limitations](https://stoneforgegroup.com/research/).