[Home](https://stoneforgegroup.com/) · [Research](https://stoneforgegroup.com/research/) · Florida Owner-Led Firms

06 – Research

# Florida owner-led firms

Florida’s owner-age gaps against the national picture, in all five sectors checked, are too small to distinguish from sampling error. In construction, finance and insurance, real estate, professional services, and administrative and support, the share of owners aged 55 or over in Florida runs from 50.8% to 57.1% – and every one of those figures sits inside its own standard error of the matching national figure. That means no difference was detected, not that the two are shown to be the same.

## Executive summary

This study reads the Census Bureau’s 2022 Annual Business Survey for Florida specifically – Characteristics of Business Owners (abscbo) for owner age and tenure, and the core establishment survey (abscs) for firm counts, revenue and payroll – and asks a single question of each owner-age figure: is Florida’s gap against the national figure bigger than the sampling error on the comparison. In every sector checked, the answer is no. Florida professional-services owners aged 55 or over are 53.3% of the total, against 52.6% nationally; construction is 50.8% against 49.5%; finance and insurance 55.1% against 55.4%; real estate 57.1% against 58.8%; administrative and support 50.9% against 50.3%. Every one of those five gaps is smaller than the combined standard error of the two figures being compared. That is a failure to detect a difference at this sample size, not proof the two populations are the same – and Florida is part of the national figure, not an independent draw from it, so the two were never fully separate comparisons to begin with. Florida’s demographics lead many people to assume its business ownership skews older than the country’s; this data does not detect that difference, though it is not powered to rule out a small one either.

Florida does show one genuine, measurable difference at the sector-wide level: its professional-services firms generate less revenue per employee than the national figure, $214,895 against $234,354, about 8.3% lower – a gap large relative to this dataset’s reporting-level standard errors. That aggregate gap does not repeat band by band, though: of Florida’s seven employment-size bands, three run higher than the national figure and four run lower, from 11.4% higher in the 10-to-19 band to 25.1% lower in the 100-to-249 band – see the band-level table below. The aggregate 8.3% gap is real; it is not evidence of a uniform Florida pattern across firm sizes. Payroll per employee runs lower too, $78,080 against $90,560, about 13.8% lower, proportionally more than revenue falls – which is why Florida’s revenue-per-payroll-dollar ratio, $2.75, is actually higher than the national $2.59, even though the absolute revenue figure behind it is lower. Florida also shows the same leverage decline the national data does as firms move from the 5-to-9-employee band toward the 250-to-499-employee band – down about 22.5% in Florida, against about 20.9% nationally – though only that widest comparison survives Florida’s sampling error; the narrower comparison to the state’s own 50-to-99-employee band cannot be made with this data, and that limit is stated in full below.

Three things this page does not and cannot say, stated here so they are not read into any figure below: owner age is not firm size, revenue and payroll figures are not owner age, and none of this data measures what an owner intends to do next. See Three claims this page does not make, below.

- Florida, share of owners 55 or over

- National, share of owners 55 or over

Share of owners aged 55 or over, Florida against the national figure, five sectors. US Census Bureau, Annual Business Survey, Characteristics of Business Owners, 2022.

## Every sector's gap against the national figure sits inside its own standard error

What follows is a check that each Florida-to-national gap is no larger than the reported error on the two figures being compared – not a formal equivalence test with a predeclared margin, which this study does not run. Finding no difference large enough to distinguish, given the reported errors, is not the same as showing the two distributions are the same; it means this comparison cannot tell them apart at this sample size.

Two dependence problems sit underneath every number in this section, and neither is corrected for, only disclosed. First, Florida is one state's share of the same national sample the "national" figure is drawn from, not an independently collected estimate – the two sides of every comparison below share respondents, so treating them as independent overstates how much the gap between them tells you. Second, each "55 or over" figure sums two published age-band shares (55-to-64 and 65-or-over) that come from the same OWNPDEMP_PCT distribution and share its denominator – the count of owners in that sector and geography – so the two bands are not independent draws either; an owner base that under- or over-reports one band tends to do so consistently across the other, in the same direction on both sides of every comparison.

The Annual Business Survey publishes a standard error alongside every percentage it reports (OWNPDEMP_PCT_S), in percentage points – an absolute margin, not a relative one, and not the same field as OWNPDEMP_S, which reports a relative standard error on the underlying owner count. Combining two age bands into one "55 or over" figure means combining their two standard errors – done here in quadrature, the same conservative, independence-assuming method this site's leverage study applies to its own combined errors, which overstates the true combined error given the shared-denominator dependence just described. Comparing Florida's combined figure to the national one means combining the two again, the same way, on top of the Florida-is-part-of-the-national-sample dependence above.

Share of owners aged 55 or over, Florida against the national figure, 2022

Sector
Florida
Florida combined SE
National
National combined SE

Construction 50.8% 1.1 49.5% 0.4
Finance & insurance 55.1% 2.1 55.4% 0.9
Real estate 57.1% 1.5 58.8% 0.4
Professional services 53.3% 1.3 52.6% 0.6
Administrative & support 50.9% 1.4 50.3% 0.7

Take construction as the worked example, since its gap is the largest of the five in percentage-point terms. Florida's 55-to-64 band is 31.1% of owners (standard error 0.9 points) and its 65-or-over band is 19.7% (standard error 0.7 points); summed, that is the 50.8% in the table, with a combined standard error of about 1.1 points (sqrt(0.9² + 0.7²) = 1.14). The national figure, 49.5%, carries a combined standard error of about 0.4 points. Adding each figure's own combined error as a range around it: Florida's construction figure spans 49.7% to 51.9%, and the national figure spans 49.1% to 49.9%. The two ranges overlap, from 49.7% to 49.9% – narrow, but real. This is not a formal significance test, the same caveat the leverage study states for its own range checks; it is a check that the gap is no larger than the combined error, not a p-value.

Real estate is the next-closest call, and the same check applies: Florida's 57.1% carries a combined standard error of about 1.5 points, spanning 55.6% to 58.6%; the national 58.8% carries a combined standard error of about 0.4 points, spanning 58.4% to 59.2%. Those ranges overlap too, from 58.4% to 58.6%. Finance and insurance, professional services, and administrative and support all have more room still – their Florida-to-national gaps are smaller in percentage-point terms than construction's or real estate's, against combined errors of similar or larger size. Every one of the five sectors checked passes the same overlap test. This is not a formal equivalence test with a predeclared margin, and failing to find a difference is not the same as showing there is none – it means this data, at this sample size, cannot distinguish Florida's owner-age structure from the country's.

## Florida professional services, in numbers

Before turning to what Florida's owners look like, the underlying population: in 2022, Florida had 84,127 employer firms in professional, scientific and technical services (NAICS 54), employing 637,226 people and generating about $136.9B in receipts. This is the "all firms" row of the same Annual Business Survey establishment table (abscs) the leverage-gap study reads nationally – here pulled with for=state:12 in place of for=us:*, and with no employment-size filter, so it is every employer firm in the sector, not one band of them.

## Florida generates less revenue per employee than the national figure at the sector-wide level – a genuine difference, not sampling noise

Dividing that same all-firms row's receipts and payroll by its employment gives three sector-wide figures, computed the same way the leverage study computes them by band: revenue per employee, payroll per employee, and revenue per dollar of payroll. Florida's professional-services firms average $214,895 of revenue per employee, against $234,354 nationally – about 8.3% lower. Payroll per employee is $78,080 in Florida against $90,560 nationally – about 13.8% lower, a larger proportional gap than the revenue figure. Because payroll falls further than revenue does, Florida's revenue-per-payroll-dollar ratio comes out higher, not lower: $2.75 per dollar of payroll in Florida, against $2.59 nationally.

Unlike the owner-age comparison above, this is not a standard-error question – these are sector-wide averages across the entire firm population, not one age band's share of it, and the gap (8.3% on revenue per employee) is large relative to the reporting-level standard errors this dataset publishes on its aggregate rows. Florida firms in this sector generate less revenue per person than the national figure, genuinely, while paying those people proportionally even less – which is exactly why the ratio between the two runs the other way.

## The band-level pattern is not uniform, even though the aggregate gap is real

The 8.3% aggregate gap comes from dividing Florida's total receipts by its total employment, across every professional-services firm in the state at once. Broken out by the same seven employment-size bands the leverage-gap study uses, Florida runs higher than the national figure in three bands and lower in four:

Revenue per employee, Florida against the national figure, by employment-size band, NAICS 54, 2022

Band
Florida
National
Difference

1–4 $206,000 $202,000 2.1% higher
5–9 $182,000 $195,000 6.8% lower
10–19 $226,000 $203,000 11.4% higher
20–49 $221,000 $212,000 4.1% higher
50–99 $173,000 $227,000 23.7% lower
100–249 $183,000 $245,000 25.1% lower
250–499 $226,000 $247,000 8.8% lower

Three bands higher, four lower, and Florida's band-level relative standard errors on receipts run wide at this granularity – from 3.3% in the smallest band to 20.0% in the 100-to-249 band, confirmed live against the Census API on 7 August 2026, several times wider than the state's own all-firms figure of 3.3% used for the sector-wide comparison above. The aggregate 8.3% gap is not evidence of a uniform Florida effect by firm size; it is a fact about the sector-wide total, and the band-level detail does not repeat it.

## The leverage comparison that still holds at the state level

The leverage-gap study's own headline, at the national level, is that professional services' revenue per dollar of payroll falls as firms get larger. Florida shows the same direction at the widest comparison this state-level data can support: from the 5-to-9-employee band to the 250-to-499-employee band, Florida's ratio falls from $3.01 to $2.33, about 22.5% – close to the national decline over the identical two bands, about 20.9% (from $3.01 to $2.38, (2.38-3.01)/3.01 = -20.93%, rounded). Florida and the national figure start at the identical rounded ratio, $3.01, in the 5-to-9 band; they diverge only slightly by the 250-to-499 band.

This comparison survives Florida's sampling error – see Give the RSEs, in What Florida's data cannot support, below, for the combined standard errors behind both ends of it. The narrower comparison the leverage-gap study actually leads with nationally, from the 5-to-9 band to the 50-to-99 band, cannot be made with Florida's data at all; that limit is stated in the same place.

## How long Florida's owners have held their firms

The same survey's tenure question (QDESC O02) asks the year an owner acquired or established their ownership stake. In Florida, professional-services owners who acquired their stake before 1980 are 1.1% of the total; 1980 to 1989 adds 4.0%; 1990 to 1999 adds 11.8% – summed, 16.9% of Florida's professional-services owners acquired their stake before 2000. Nationally, the identical three bands sum to 19.5% (1.9% + 5.4% + 12.2%, the same published figures the who-owns-these-firms study reports for the country as a whole). Florida's owners are, on this reading, very slightly newer to their ownership stakes than the national population – though this page does not run the standard-error check on that gap the way it does for owner age above, and does not claim the two are statistically distinguishable.

10.2% of Florida's professional-services owners answered "don't know" rather than naming a year or a band – a share that has to be disclosed alongside the 16.9% and 19.5% figures above, since it is not folded into either sum and is large enough to matter: if even a modest share of Florida's "don't know" owners in fact acquired their stake before 2000, the true pre-2000 share could sit meaningfully above 16.9%. Neither this study nor the underlying survey can resolve that.

## What Florida's data cannot support

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

### Give the RSEs: the in-market comparison cannot be made

The leverage-gap study's national headline compares the 5-to-9-employee band to the 50-to-99-employee band – the range that spans this market's own $1–20M revenue footprint. At the state level, that comparison cannot be made. Florida's 5-to-9 band carries relative standard errors of 10.8% on receipts and 6.6% on payroll, combined in quadrature to about 12.7% on the ratio; applied as a range around the $3.01 ratio, that spans $2.63 to $3.39. Florida's 50-to-99 band carries relative standard errors of 15.1% on receipts and 15.8% on payroll, combined to about 21.9%; applied to that band's own ratio, the range spans $1.84 to $2.87. Those two ranges overlap, from $2.63 to $2.87 – Florida's data cannot tell the 5-to-9 band's ratio apart from the 50-to-99 band's, so the comparison this market most wants – the range it actually operates in – is not one Florida's state-level sample size can support.

The 250-to-499-employee band is different: relative standard errors of 3.4% on receipts and 3.3% on payroll, combined to about 4.7%, applied to the $2.33 ratio give a range of $2.22 to $2.44. That range does not overlap the 5-to-9 band's $2.63-to-$3.39 range. Only this widest comparison – 5-to-9 employees against 250-to-499, a band this market rarely reaches – is decisive enough to publish from Florida's own data; it is the comparison reported in The leverage comparison that still holds at the state level, above. The narrower, more useful comparison is the one Florida's sample size cannot resolve.

### No firm size by firm age

The Business Dynamics Statistics time series, which the firm-size-and-age study reads nationally, publishes firm counts crossed by both firm age (FAGE) and employment size (EMPSZFI) – but only at the national level. Run the identical cross-tabulated query at the state level – FIRM,EMP,FAGE,EMPSZFI for Florida, NAICS 54, 2023 – and it returns 28 rows, and not one of them carries a genuine, non-aggregate code in both fields at once: every row is either a real age band against the aggregate "all sizes" code, or a real size band against the aggregate "all ages" code. There is no Florida version of the firm-size-and-age study, at any granularity this dataset offers – the crossed cell simply is not published below the national level.

### No subsector detail

The who-owns-these-firms study reports professional services' owner-age spread down to the four-digit subsector nationally – 39.4% of owners 55 or over in computer systems design against 62.8% in accounting, tax preparation, bookkeeping and payroll services, a 23-point spread inside one two-digit code. That spread cannot be tested for Florida. Requesting the identical owner-age query with Florida's geography and a four-digit subsector code in place of NAICS 54 itself – for example NAICS 5411, legal services – returns HTTP 204, an empty response, the same way it does for every subsector code checked. State-level detail exists only at the two-digit sector; subsector detail exists only at the national level. Whatever composition of legal, accounting, engineering, computer-systems and consulting firms makes up Florida's professional-services owner base, this data cannot separate it out.

## Three claims this page does not make

**Owner counts are not firm counts.** Every age and tenure figure above counts owners, not firms – a firm reporting several owners is counted once per owner it reports. Nationally, about 67.3% of NAICS 54 firms report exactly one owner and 28.6% report two to four, so "53.3% of Florida's owners are 55 or over" cannot be reread as "53.3% of Florida's firms." No Florida-specific owner-count figure exists in this dataset; the 67.3%/28.6% split is the national figure, the only one published.

**Nothing here joins owner age to firm size.** The Characteristics of Business Owners survey (abscbo), which every age and tenure figure on this page comes from, carries no firm-size variable of any kind – requesting EMPSZFI on it returns HTTP 400, "unknown variable," confirmed live the same way the who-owns-these-firms study confirms it nationally. The revenue and payroll figures elsewhere on this page come from a different survey component (abscs) entirely, and nothing joins the two: this page cannot say whether Florida's older or younger owners run larger or smaller firms.

**No dataset here measures what an owner intends to do next.** Checked against the published variable lists for abscbo, abscb and absmcb for the 2022 ABS cycle, none carries a variable naming succession, sale, exit, transfer, or any owner's stated plan for their firm's future – the same finding the who-owns-these-firms study reports nationally, and it holds identically for Florida, since no state-level version of any of these three datasets carries the variable either. An ownership base with roughly half its owners 55 or over says nothing, on this data, about a coming wave of sales, of consolidation, or of firms simply continuing under the same owner. All three are equally unmeasured, and none is implied by anything on this page.

## Methodology and sources

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

### Data sources

US Census Bureau, Annual Business Survey (ABS), 2022 data year, two tables. Owner age (question O09) and tenure (question O02) come from abscbo (Characteristics of Business Owners), at https://api.census.gov/data/2022/abscbo – the same table the who-owns-these-firms study reads nationally, here queried with Florida's geography. Firm counts, employment, receipts and payroll come from abscs, at https://api.census.gov/data/2022/abscs – the same table the leverage-gap study reads nationally, here also queried with Florida's geography. Both are drawn from the same stratified sample of roughly 230,000 employer businesses nationally; Florida's share of that sample is smaller than the national total, which is exactly why several of the comparisons above carry wider standard errors at the state level than the identical comparison does nationally.

### 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 abscbo – it returns HTTP 204, an empty response with no error message, indistinguishable from "this geography has no data" to a script that does not check the status code. An earlier scoping pass against this exact dataset made that mistake and concluded Florida had no data at all. It does; the query was malformed. The working shape, confirmed live:

https://api.census.gov/data/2022/abscbo?get=QDESC,OWNCHAR,OWNCHAR_LABEL,OWNPDEMP,OWNPDEMP_PCT,OWNPDEMP_PCT_S&for=state:12&NAICS2022=54&QDESC=O09&key=REDACTED

Every query behind this page requests OWNCHAR alongside OWNCHAR_LABEL for exactly this reason.

### The exact queries

Owner age and tenure, run once per sector (five NAICS codes: 23, 52, 53, 54, 56) for O09, and once for NAICS 54 alone for O02, all with for=state:12 in place of the national for=us:*. Firm counts and dollar figures, one call against abscs with no employment-size filter for the all-firms totals, and one call per employment-size band (EMPSZFI codes 620, 642 and 652 for the 5-to-9, 50-to-99 and 250-to-499 bands) for the leverage comparison, again with for=state:12:

https://api.census.gov/data/2022/abscs?get=NAICS2022_LABEL,FIRMPDEMP,EMP,RCPPDEMP,PAYANN,EMPSZFI,RCPPDEMP_S,PAYANN_S,EMP_S&for=state:12&NAICS2022=54&key=REDACTED

All figures were pulled live against the Census API on 7 August 2026. The pull script is scripts/research_pull.py, function owner_rows(), extended for this study to run the O09 loop against Florida's geography for all five broad sectors rather than NAICS 54 alone, and to add the NAICS 54 O02 Florida pull. The state-level abscs base and leverage-band figures are not written to a CSV by that script – they are a one-off geography pull, documented here with their exact query and re-verifiable by anyone, the same convention this site's other studies use for figures outside their regular CSV output (see, for example, the leverage-gap study's NAICS 524 and 523 exclusion figures).

### Variable definitions

- **OWNPDEMP_PCT** – the share of owners in a category, as published by Census, already rounded to one decimal place. Every owner-age and tenure percentage on this page is this published figure.

- **OWNPDEMP_PCT_S** – standard error of OWNPDEMP_PCT, in percentage points – an absolute margin. This is a different field from OWNPDEMP_S, which is a relative standard error on the raw owner count and is not used on this page.

- **RCPPDEMP / PAYANN** – receipts and annual payroll of respondent employer firms, both in thousands of dollars. RCPPDEMP_S and PAYANN_S are relative standard errors, as percentages, on those two fields.

- **EMPSZFI** – employment-size band, on abscs: 612 is 1 to 4 employees, 620 is 5 to 9, 642 is 50 to 99, 652 is 250 to 499 – the same code space the leverage-gap study documents in full.

### How each figure is computed

Florida's base figures: FIRMPDEMP 84,127; EMP 637,226; RCPPDEMP 136,936,910 (thousands of dollars, so $136,936,910,000 – rounded to $136.9B in prose); PAYANN 49,754,905 (thousands of dollars). Revenue per employee is RCPPDEMP × 1,000 / EMP: 136,936,910,000 / 637,226 = $214,895, rounded. Payroll per employee is PAYANN × 1,000 / EMP: 49,754,905,000 / 637,226 = $78,080, rounded. Revenue per dollar of payroll is RCPPDEMP / PAYANN: 136,936,910 / 49,754,905 = 2.75, rounded to two decimals. The national equivalents use the identical formula against the national all-firms row.

Combined standard errors on the leverage bands are computed the same way the leverage-gap study computes its own: sqrt(RCPPDEMP_S² + PAYANN_S²), treating receipts and payroll as independent sources of error – a conservative assumption, since the two tend to move together in practice. Combined standard errors on owner-age figures use sqrt(ED_SE² + EE_SE²), the same quadrature method applied to the two age-band standard errors rather than to two relative standard errors.

### Known limitations

The limitations that apply to every study on this site – these are samples, not a census of every firm, and a snapshot in one year rather than the same owners or firms followed over time – are stated in full on [the research hub](https://stoneforgegroup.com/research/) and are not repeated here.

Specific to this study:

- **Florida's sample is smaller than the national one.** Every standard error on this page is wider than the identical comparison's national standard error, because Florida is one state's share of the same national sample rather than a separately drawn one. This is why the owner-age gaps in Every sector's gap, above, require the overlap check at all, and why the in-market leverage comparison in What Florida's data cannot support cannot be made.

- **No difference detected is not the same as no difference.** The owner-age overlap check in Every sector's gap, above, is not a formal equivalence test with a predeclared margin. It shows this comparison cannot distinguish Florida's owner-age structure from the national one at this sample size; it does not show the two are the same. Florida being part of the national sample, and each "55 or over" figure summing two age bands that share a denominator, are both disclosed there – neither is independent, and the combined standard errors are wider because of it.

- **The revenue-per-employee gap is a sector-wide average, not a uniform pattern by firm size.** Florida's 8.3% aggregate gap is real, but the band-level table in The band-level pattern is not uniform, above, shows three of seven bands running higher than the national figure and four running lower. The aggregate does not establish a Florida effect that holds across firm sizes.

- **No firm-size join, no subsector detail, no size-by-age crosstab.** All three limits are stated in full in What Florida's data cannot support and Three claims this page does not make, above, and are not repeated here.

- **The tenure "don't know" share is not folded into either sum.** 10.2% of Florida's professional-services owners did not name a year or a band; see How long Florida's owners have held their firms, above.

- **Margin of error stated for the comparisons this page makes, not for every possible one.** This study pulled standard errors for the owner-age comparisons and the three leverage bands it reports. A reviewer wanting a margin of error on a different cut of this data – a different band, a different sector pairing – would need to pull the matching standard-error fields themselves, the same reproducible way this study did.

## Citing this work

Stoneforge Consulting Group, *Florida Owner-Led Firms*, 2026. Published 7 August 2026. Analysis of US Census Bureau Annual Business Survey, Characteristics of Business Owners and core establishment tables, 2022. https://stoneforgegroup.com/research/florida-owner-led-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/).