What the record actually shows: crime reported to the state, how Croton compares to its neighbors, and how busy the police department is — from its own monthly reports. Every figure links to its source.
Index crimes — the seven serious offenses (murder, rape, robbery, aggravated assault, burglary, larceny, motor-vehicle theft) every police department reports to New York State. Croton's count has fallen from the high-100s in the 1990s to the 30s–60s today.
The orange line is arrests made, shown for the years the department has published them — from its annual reports (2019, 2021) and complete runs of its monthly reports (2022, 2024, 2025). The state publishes arrest counts only at county level, so there is no longer series. Arrests run two to nearly four times reported index crime because most are for offences the index excludes: driving offences, warrants, drugs, harassment.
Source: NYS Division of Criminal Justice Services, Index Crimes by Agency (exact agency Croton-On-Hudson Vg PD); arrests from Croton PD annual and monthly reports. Compare per 1,000 residents in the explorer →
Index crimes per 1,000 residents, 2024, against the other river villages. Two things are true at once here, and the second is the one usually left out. Croton's rate is well below Westchester County (11.2) and New York State (21.4). But among these seven comparable villages it is the second highest, behind only Dobbs Ferry — above Ossining, Tarrytown, Irvington, Hastings and Briarcliff. Croton is safe by the standards of the state; it is not the safest street on the river.
Village-to-village ranking is the least informative comparison available, because the whole cluster is quiet. The wider anchors matter more, and they separate the two halves of the total:
| Per 1,000 residents, 2024 | Croton | Westchester | New York State | United States | Croton as % of US |
|---|---|---|---|---|---|
| Violent offences | 0.12 | 1.88 | 4.05 | 3.64 | 3% |
| Property offences | 6.95 | 9.28 | 17.36 | 19.34 | 36% |
Croton's violent crime rate is about 3% of the national one. In counts rather than rates: one aggravated assault in 2024, no murders, no robberies. Property crime is the meaningful half at 36% of the national rate — 57 offences, of which 52 were larcenies, two burglaries and three stolen vehicles. Note also that New York State's violent rate (4.05) runs above the national one while its property rate (17.36) runs below it, so "below the state average" and "below the national average" are not the same claim.
Source: NYS DCJS (2024 offenses) ÷ U.S. Census ACS population (2019–23). Sleepy Hollow reported zero index crimes in 2024 and is omitted — that reads as non-reporting, not safety. County and state rates use 2024 population estimates. National rates: FBI Uniform Crime Reporting, 2023 — 363.8 violent and 1,934.1 property offences per 100,000. Note the cost tables further down use 2020 Census counts instead, as published in the state's own filings; for Croton the two differ by about 1.5% (8,205 against 8,327), which shifts its crime rate between 7.1 and 7.0 and changes no ranking on this page.
Some residents have asked whether the village massages its crime figures to look safer. It's a fair question — so here is the check. The short answer is no: the crime numbers aren't Croton's to edit, and every internal test points the same way.
Reported crime since 2000, each line set to 100 in its own first year. If Croton were quietly editing its numbers, its line would peel away from the region. Instead it falls in step with Westchester County and New York State — a gradual, shared decline, not a sudden unexplained drop.
Reported-crime series: NYS Division of Criminal Justice Services. Indexed to each jurisdiction's 2000 total = 100.
Almost all of it is property crime — larceny above all. Violent offenses (in orange) have stayed in the single digits to low teens for the whole period, as you'd expect in a village this size.
Source: NYS DCJS, violent vs. property index offenses, Croton-On-Hudson Vg PD.
Calls for service — every call the department is dispatched to (alarms, medical assists, traffic, disputes, welfare checks), from its own monthly reports. This is a measure of workload, not crime: most calls never become a reported offense. Calls ran about 1,500–1,700 for the first five months of each year from 2019 to 2023, then stepped up to roughly 2,000 in 2024–2026.
Source: Croton-on-Hudson Police Department monthly reports. 2020 and most of 2021 were not published (the COVID reporting gap). Open in the explorer →
To compare years fairly you have to hold the season constant — a summer month always runs hotter than a winter one. So this looks only at January–May, the same five months every year, using the department's own year-to-date figure. Recent years are shaded.
Source: Croton-on-Hudson Police Department monthly reports, January–May year-to-date "Total Calls." 2020–2021 omitted (not published).
The other numbers the department reports each month, by full year. Arrests and fire/ambulance assists have risen; motor-vehicle accidents have drifted down. (Only complete years are plotted.)
Source: Croton-on-Hudson Police Department monthly & annual reports. Full-year figures use the department's annual report where published (2019, 2021), otherwise a complete 12-month sum.
Croton has 22 sworn officers today by the village's own roster and staffing page; it reported 21 to the state for 2024, before a lieutenant's post was filled that June. The chart below is the police department's own comparison of Westchester villages, often quoted to show Croton spends less per resident than its neighbours. Read it with the caution in the next section — the budgets in it are not measured the same way from village to village. Every per-officer figure elsewhere on this page uses the state's annual personnel return rather than this study's counts, for a reason given below: the study overstates headcount in nine of the sixteen villages it lists.
This chart is the 2024 study's numbers, and they are wrong. We reproduce it only because it is the chart the argument has been built on. Its budget figures are not measured the same way from village to village — some include employee benefits, most do not — and its officer counts run high in nine of the sixteen villages it lists. Both faults are documented below, with corrected tables.
The original is published by the Village: Comparison of Village Police Departments (PDF), linked from its Police Department reports page as "Comparison to Other Village Departments."
Source: Village of Croton-on-Hudson Police Department, Comparison of Village Police Departments (2024) — reproduced here as the subject of this analysis, not as evidence. The corrected figures are in "The comparison, rebuilt" and "What policing actually costs" below.
This is the part that trips up almost everyone who quotes those numbers, and it is worth spelling out.
New York does give every village a standardized chart of accounts. The State Comptroller assigns the same code to the same thing everywhere: A3120 is the police department, A9015 is police and fire pension, A9060 is health insurance. A dollar coded A9015 means the same thing in any village in the state.
What is not fixed is where a village puts employee benefits in its own budget document. The uniform system permits two presentations, and both are entirely proper:
We checked this against Briarcliff's actual adopted budget. Of its $5,477,182 police appropriation, $2,225,435 — 41% — is employee benefits sitting inside the police line. Strip those out and Briarcliff's police department costs $3,251,747.
So the peer study places a benefits-included number (Briarcliff) beside a benefits-excluded number (Croton) and treats them as equivalent. They are not. Here is the same pair of villages measured three ways:
Left: as the study compares them — Briarcliff looks far more expensive. Middle and right: both villages on the same basis. The gap closes, then reverses.
Croton figures: Adopted Budget FY2026-27. Briarcliff figures: Village of Briarcliff Manor Adopted Budget FY2022-23, account A3120 detail (printed total $5,477,182; the 2024 study lists $5,740,000). The two budgets are four years apart, so read this pair as illustrative — the properly normalized comparison for all villages is below.
Think of sixteen shops each reporting "what staff cost me." One includes payroll tax and health cover in the figure; another quotes wages only and keeps the rest in a separate ledger. Neither is lying. But you cannot rank them until you decide what the number includes and apply that decision to all sixteen. Normalizing is just that: forcing every village onto one definition before comparing.
This page normalizes twice, because there are two fair definitions:
1. Department operating cost — pay, equipment and supplies, with employee benefits excluded for everyone. That's the chart immediately below. It is the cleanest measure, because no estimating is involved, but it is not the real bill.
2. Full cost — the same, plus each village's employee benefits added back. That's the section after it, and it is the one that answers "what does policing actually cost here."
Croton moves between the two, which is not a contradiction — they measure different things. The study did neither: it printed whatever number each village happened to publish.
Every village in New York files an Annual Financial Report with the State Comptroller, on one prescribed structure: a "Police" expenditure figure, with employee benefits reported separately, the same way for everybody. Those filings are comparable by construction — exactly what the study's figures were not. Below, each village's police spending per resident: the study's number, and the normalized one.
A fair question: does the state figure match the village's budget? No, and it isn't meant to. Three things differ. A budget is what a village plans to spend; the state filing is what it did spend. They cover different years. And they cover different things: the state's police rollup for Croton also sweeps in the auxiliary unit, traffic control and $442,952 of police equipment bought through the capital fund, which no village's police budget line contains. For that reason the tables on this page are built from the general fund only, gathering every police-function account a village uses, and excluding capital purchases — which are lumpy, and which nine of the fifteen make and six do not.
Sorted by the normalized figure. Croton in dark red.
| Village | Study $/resident | Rank | Normalized $/resident | Rank | Move |
|---|---|---|---|---|---|
| Pelham Manor | $938 | 1 | $883 | 1 | = |
| Ardsley | $745 | 2 | $788 | 2 | = |
| Irvington | $627 | 7 | $730 | 3 | +4 |
| Larchmont | $673 | 5 | $699 | 4 | +1 |
| Bronxville | $669 | 6 | $656 | 5 | +1 |
| Tuckahoe | $699 | 4 | $632 | 6 | -2 |
| Tarrytown | $569 | 10 | $621 | 7 | +3 |
| Pleasantville | $604 | 8 | $563 | 8 | = |
| Sleepy Hollow | $498 | 11 | $548 | 9 | +2 |
| Pelham | $593 | 9 | $546 | 10 | -1 |
| Croton-on-Hudson | $441 | 12 | $531 | 11 | +1 |
| Dobbs Ferry | $379 | 14 | $512 | 12 | +2 |
| Rye Brook | $440 | 13 | $490 | 13 | = |
| Briarcliff Manor | $718 | 3 | $469 | 14 | -11 |
| Hastings-on-Hudson | $373 | 15 | $395 | 15 | = |
Everything above measures departments with employee benefits stripped out — consistent, but not the real bill. So we built the full one from the account-level filings, applying a single rule to all fifteen villages: use each village's own reported police benefits where it books them to the police account, and allocate its share of the central benefits pool where it doesn't. Only 1 of 15 villages (Briarcliff Manor) book them directly; the rest pool them. Each bar below is one village's total police cost per resident, split into pay-and-supplies and benefits.
These figures use the State Comptroller's own "Police" classification, which gathers up every police-function account a village uses, whichever number it files them under. That detail matters more than it sounds — see the section immediately below.
This deserves stating plainly, because it cuts both ways. The residents who argued that the two departments cost far more alike than the published table suggested were right — the filings say so. What they were wrong about was the mechanism: nothing was concealed and no money was moved. Briarcliff's published figure simply carries its employee benefits and Croton's does not, and that single accounting difference accounts for nearly the whole $2.08m.
Against the fifteen-village median of $917 per resident, Croton comes in 9% below, 11th of 15. So Croton does spend less per resident than the typical village here — but it is a mid-table position, not the standout economy the study implied, and it is arrived at for a different reason than the study gave.
Croton employs 22 sworn officers today, and its FY2026-27 budget funds 22 posts. Every per-officer and per-1,000 figure on this page nonetheless divides by 21 — the number Croton reported to the state for 2024, before a lieutenant’s post was filled that June.
That is deliberate. The spending being divided is fiscal 2025 spending, and 2024 is the only year for which every one of these villages has filed a personnel return. Pairing 2024 money with a 2026 headcount for Croton alone, while every peer stays on 2024, would manufacture exactly the mismatch this page criticises the 2024 study for.
For completeness, on today’s 22 officers Croton would be 2.64 per 1,000 (13th of 16 rather than 14th, level with Briarcliff) and $307,802 per officer (7th of 15 rather than 5th). Below the peer median either way; the conclusions do not turn on it. Where this page tests structure rather than spending — the patrol share of 17 of 22, and the payroll composition — it uses the current 22-post roster, because that is the only roster broken down by rank.
Splitting the per-officer cost answers a question the totals cannot. Croton's benefits per officer are $121,457 — 7th of 15, below the median of $120,947. Its officers are not unusually expensive to insure or retire. The difference is in pay: $196,367 per officer, 1rd highest against a median of $181,276. Excluding the civilian staff Croton's police budget also carries, it is $182,719 per sworn officer against a peer median of $171,614.
Same basis as the table above: state account-level filings, fiscal 2025, divided by each village's sworn strength from its state personnel return.
Croton is the only village whose post-by-post detail we were able to find, so its composition can be read directly. Others may publish the same somewhere we did not look. Of about $3.87 million budgeted for police personnel in FY2026-27:
| Component | Amount | Share |
|---|---|---|
| Base salaries, all 22 posts | $3,071,453 | 79% |
| Overtime — patrol $350,000, investigations $65,000, training $50,000, community $40,000 | $505,000 | 13% |
| Vacation, longevity, holiday and sick incentives | $284,618 | 7% |
| Night differential | $9,814 | 0.3% |
Night work is not the driver, though it is the intuitive answer: the differential is a quarter of one per cent of payroll. What the round-the-clock commitment drives is structure and overtime. Every shift needs a sergeant, and covering all of them with rotation and leave takes five squads — five patrol sergeants supervising twelve officers, one supervisor per 2.4, where a larger agency might run one per six or eight. A sergeant is budgeted at $157,523 against $136,977 for a first-grade officer, and the five patrol sergeants alone take 36% of patrol pay — 51% once the lieutenant and detective sergeant are added. Holding a minimum shift strength from twelve officers then forces the overtime: at $505,000 it is more than twice the $233,000 by which Croton's payroll exceeds the median once both are put on a per-officer basis ($196,367 against $181,276, across 21 officers).
The limit on that conclusion: Croton is the only village whose post-by-post detail we could locate, and the state filings do not separate base pay from overtime. So we can show what Croton's payroll is made of and that it sits above the median — but not that its neighbours spend less on overtime or supervise more thinly, because we could not find enough published by them to check.
Village of Croton-on-Hudson adopted budget FY2026-27, account A3120 personnel detail.
The comparison that circulated locally was per officer: Briarcliff at $287,000 against Croton's $174,286, about 65% more, offered as proof that Croton must be hiding costs. Fully loaded, that gap closes almost entirely — and tips the other way.
Croton spends $332,088 per sworn officer, 5th of 15; Briarcliff $312,014. Croton runs a comparatively small force — 2.52 officers per 1,000 residents — whose officers cost somewhat more each than most. That combination is the honest version of what people have been arguing about.
These use the state's headcount, not the study's — deliberately. Every village files an annual personnel return to the Division of Criminal Justice Services giving sworn and civilian staff on one statewide definition. Set against it, the study's officer counts are too high in nine of the sixteen villages it lists, by one or two apiece: Hastings 21 against 19, Pelham 24 against 22, Tuckahoe 23 against 21, and so on. It agrees in six and runs low in one. Dividing by too many officers makes a department look cheaper per head than it is, so we do not use those counts anywhere in our own arithmetic. The state return also independently corroborates Croton's civilian staffing, reporting 2 full-time and 6 part-time civilians against a police account that pays for a dispatcher, a parking-enforcement officer, crossing guards and seasonal rangers.
| Village | Population | Officers (2024) | Police payroll | Benefits | Total cost | Per resident | Per officer | Index crime |
|---|---|---|---|---|---|---|---|---|
| Pelham Manor | 5,752 | 27 | $4,780,382 | $2,995,031 | $8,071,398 | $1,403 | $298,941 | 214 |
| Ardsley | 5,079 | 20 | $3,711,004 | $2,542,231 | $6,543,276 | $1,288 | $327,164 | 38 |
| Irvington | 6,652 | 23 | $4,506,491 | $3,083,493 | $7,940,857 | $1,194 | $345,255 | 26 |
| Larchmont | 6,630 | 24 | $4,406,144 | $2,723,915 | $7,360,561 | $1,110 | $306,690 | 67 |
| Bronxville | 6,656 | 22 | $3,762,537 | $2,553,899 | $6,921,517 | $1,040 | $314,614 | 32 |
| Tarrytown | 11,860 | 34 | $6,626,252 | $4,719,438 | $12,087,736 | $1,019 | $355,522 | 55 |
| Tuckahoe | 7,084 | 21 | $3,894,719 | $2,603,197 | $7,081,317 | $1,000 | $337,206 | 9 |
| Pleasantville | 7,513 | 22 | $3,993,961 | $2,660,840 | $6,887,948 | $917 | $313,089 | 50 |
| Pelham | 7,326 | 22 | $3,728,302 | $2,637,041 | $6,640,402 | $906 | $301,836 | 26 |
| Sleepy Hollow | 9,986 | 30 | $5,119,168 | $3,345,712 | $8,814,212 | $883 | $293,807 | 0 |
| Croton-on-Hudson | 8,327 | 21 | $4,123,712 | $2,550,588 | $6,973,847 | $837 | $332,088 | 58 |
| Briarcliff Manor | 7,569 | 20 | $3,347,591 | $2,686,710* | $6,240,278 | $824 | $312,014 | 15 |
| Rye Brook | 10,047 | 26 | $4,664,954 | $3,102,455 | $8,020,991 | $798 | $308,500 | 66 |
| Dobbs Ferry | 11,541 | 26 | $4,713,187 | $3,133,176 | $9,039,743 | $783 | $347,682 | 191 |
| Hastings-on-Hudson | 8,590 | 19 | $3,151,664 | $2,414,963 | $5,806,438 | $676 | $305,602 | 32 |
* benefits reported directly to the police account by that village; all others are its share of the central pool, allocated on police payroll. Source: NYS Comptroller, account-level annual financial report data, villages, fiscal 2025 (account A3120 and the A90xx benefit accounts); population 2020 Census; index crime from NYS DCJS, 2024; officer counts from the state's Law Enforcement Personnel return, 2024 — not from the 2024 study, whose counts run high (see below). Limits: these are actual expenditures, not budgets; allocating pooled benefits on payroll share under-credits villages whose officers sit in the costlier police pension tier; Sleepy Hollow last filed a personnel return for 2022, so its per-officer figure uses that year; Elmsford has filed no financial report with the state since at least 2021 and cannot appear.
Croton's budget carries a second police account alongside the main one, headed "Auxiliary Police Services." The name is misleading. It holds no volunteer officers. Opened up, it is the department's civilian staff — the police dispatcher, five school crossing guards, the parking-enforcement officers, the seasonal park rangers — about $234,413 a year in wages. Whether or not you would call that "police," the village does: they work for the department, and the state files the account under Police.
Now the awkward part. Croton is the only one of these fifteen villages that uses that account — and one of just 8 in the entire state, out of 335 villages that report any police spending at all. Had we built the comparison around the obvious account, A3120, we would have deleted a quarter of a million dollars from Croton's column while leaving every neighbour's equivalent staff — filed under whatever number that village prefers — sitting quietly in the totals. Croton would have looked cheaper than it is, for no reason but bookkeeping.
And the trap runs the other way too. Checking the reverse — whether any neighbour keeps police costs somewhere our definition would miss — turned up Pleasantville, the one village of the fifteen that files a separate "Public Safety Administration" account, worth $626,631. Left out, Pleasantville looked like one of the cheapest departments in the county. Added back — and it plainly belongs, because doing so reproduces Pleasantville's own published police budget almost exactly, which the police account alone does not — the village moves up the table by several places, from near the bottom to ninth. Four villages also run small lockups booked under "Correctional Services"; those are counted too, though at $122 to $2,518 they change nothing. Building and code inspection, which most villages file under public safety, is excluded throughout: it is not police work.
This is not a quirk of one village. Across New York, police spending is spread over four different accounts, and 110 of 335 villages — 33% — use more than one:
| Account | What villages put in it | Villages using it |
|---|---|---|
| A3120 Police | the main department account | 283 (84%) |
| A3310 Traffic control | signals, signage, school crossings | 113 (34%) |
| A3320 On-street parking | meters and parking enforcement | 64 (19%) |
| A3189 Other police services | Croton files its civilian staff here | 8 (2%) |
Set the fifteen side by side and no two keep their books quite alike. This is every village's police filing for 2024 — which accounts it uses, where its employee benefits sit, and whether it buys police equipment out of the operating budget or through the capital fund:
| Village | Police accounts used | Benefits | Equipment via operating | Equipment via capital |
|---|---|---|---|---|
| Ardsley | 3120, 3310, 3320 | central | $74,790 | $54,927 |
| Briarcliff Manor | 3120 | in police account | $70,545 | $34,348 |
| Bronxville | 3120, 3310, 3320 | central | — | $183,320 |
| Croton-on-Hudson | 3120, 3189, 3310 | central | $133,869 | $442,952 |
| Dobbs Ferry | 3120, 3150, 3310, 3320 | central | $649,393 | — |
| Hastings-on-Hudson | 3120, 3150, 3320 | central | $70,808 | — |
| Irvington | 3120, 3310, 3320 | central | $63,322 | $128,298 |
| Larchmont | 3120, 3310 | central | $34,551 | — |
| Pelham | 3120, 3320 | central | $32,657 | $206,468 |
| Pelham Manor | 3120, 3150 | central | $88,989 | — |
| Pleasantville | 3010, 3120, 3310, 3320 | central | $16,876 | — |
| Rye Brook | 3120 | central | — | $549,766 |
| Sleepy Hollow | 3120 | central | $24,190 | $268,025 |
| Tarrytown | 3120, 3150, 3310 | central | $59,895 | — |
| Tuckahoe | 3120, 3320 | central | $90,899 | $214,201 |
Three villages run everything through a single account; Dobbs Ferry and Pleasantville use four. Fourteen pool their benefits centrally, Briarcliff books them to the police account. And equipment splits both ways: Dobbs Ferry spent $649,393 on police equipment out of its operating budget while buying none through capital; Rye Brook did the reverse, $549,766 through capital and nothing through operating. A comparison that reads one account, in one fund, catches some of these villages and misses others — which is why the fifteen figures above are built from all of it.
Counts from the NYS Comptroller's account-level village filings for fiscal 2025: villages reporting any general-fund expenditure the state classifies under the Police function, tallied by which account codes they use. Croton's civilian-staff detail from the Village's own adopted budget line items.
Assembling the comparison surfaced findings that have nothing to do with Croton. They are worth recording because they affect how any of these numbers should be read — and in two cases because somebody ought to be asking about them.
Sources: NYS Comptroller account-level village filings, fiscal 2025 (financial reporting and account usage); NYS DCJS Law Enforcement Personnel by Agency (staffing); NYS DCJS index crime by agency (Sleepy Hollow series).
Every village’s filings run back decades, so the comparison need not be a snapshot. Here is police cost per resident for all 15 villages from 2018 to 2025, on the single rule used throughout — and it shows that rising police costs are not a Croton phenomenon. Every one of the fifteen spends more per resident than it did in 2018.
Croton went from $660 per resident in 2018 to $837 in 2025 — up 27%. It began the period in the lower half and ends it there: the line moves up with the pack rather than away from it. The steepest rises belong to Ardsley (+30%) and Tarrytown (+43%). The dashed final point is Croton’s budgeted 2026-27 figure, $845 — a plan, not an outturn, and Croton’s only, since no peer has filed or published anything for that year.
Source: NYS Comptroller account-level village filings, fiscal years ending 31 May. Population held constant at the 2020 Census for every year, so the series measures spending rather than blending it with population change. The 2026-27 point is computed from the Village of Croton-on-Hudson’s adopted budget by the same rule; note that Croton budgeted $4,391,168 of police operating for 2026-27 against $4,423,259 actually spent in 2025, so its adopted line has lately run below outturn — treat the projection as a floor rather than a forecast.
This is the question the cost argument is really a proxy for, and it has a professional answer that almost nobody quotes locally. Three approaches are set out below: comparison with peers, comparison with national norms, and the workload method the profession itself recommends.
Croton fields 21 sworn officers, 2.52 per 1,000 residents — 14th of 16, against a peer median of 2.98. Only two villages here police with fewer officers per resident. This chart covers sixteen villages, one more than the cost tables: Elmsford files its personnel return every year, so it belongs in a staffing comparison even though its missing financial filings keep it out of the cost ones. Including it is the harder test — at 4.01 officers per 1,000 Elmsford raises the median Croton is judged against. The dashed lines are the FBI's national averages: 2.8 per 1,000 for cities in the Northeast, and 4.2 for cities under 10,000 people — the most heavily policed size class in the country, and Croton's own. Croton is below both. That second benchmark should be read with care: 15 of these 16 villages are also below 4.2 — only Pelham Manor exceeds it — so "below the national average for towns this size" describes almost every village on the river, not something distinctive about Croton. The peer median is the more informative comparison, and Croton sits below that too.
The Village's staffing page sets out the model: round-the-clock cover, minimum two patrol officers and a sergeant on every eight-hour shift. That is a fixed obligation, and it can be costed:
| Step | Figure |
|---|---|
| Shifts to cover a year (3 × 365) | 1,095 |
| Personnel required per shift (2 officers + 1 sergeant) | 3 |
| Officer-shifts to fill a year | 3,285 |
| Shifts one officer can work, after leave, sickness, training and court | 210–250 |
| Patrol officers needed for coverage alone | 13–16 |
| Posts Croton fills outside patrol (chief, lieutenant, det. sergeant, 2 detectives) | 5 |
| Floor for running this model at all | 18–21 |
| Sworn officers employed (2024 return / today) | 21 / 22 |
Croton runs at or barely above the arithmetic floor of its own coverage promise. That promise — not crime, not call volume — is what sets the size of this department.
Police staffing has a recognised method: the Rule of 60, set out by Jim McCabe for the International City/County Management Association and used in staffing studies across the country. Its opening premise is that officers-per-resident ratios should not be used to set staffing — the very comparison this local argument has run on, and the one in section 1 above, which is why that section is the weakest evidence on this page rather than the strongest. The rule has two tests.
Test one: about 60% of sworn officers should be on patrol. Croton's roster puts five patrol sergeants and twelve officers there — 17 of 22, or 77%. Croton is patrol-heavy, carrying few specialists, which is the profile of a stretched small department rather than a padded one.
Test two: calls should consume no more than 60% of patrol time — the "saturation index." Croton's minimum coverage model provides about 26,280 officer-hours of patrol a year. Against 4,410 calls:
| If a call consumes… | Officer-hours a year | Saturation index | Verdict |
|---|---|---|---|
| 1.0 officer-hours | 4,410 | 17% | well under |
| 1.5 | 6,615 | 25% | under |
| 2.0 | 8,820 | 34% | under |
| 2.5 | 11,025 | 42% | under |
| 3.0 | 13,230 | 50% | under |
| 3.6 — the breach point | 15,876 | 60% | threshold |
Comparable studies generally put a call between one and three officer-hours including backup and paperwork. Croton would need calls averaging more than three and a half hours each to reach the threshold. Its patrol is not saturated on any plausible assumption — there is uncommitted time in the schedule, which is what proactive policing is meant to use.
Our arithmetic applying a published method to published figures — not a professional staffing study. Croton's shift minimum is as stated on the Village's staffing page; call-handling time is shown as a range because the department's average is not public. A consultant with dispatch records could compute the saturation index exactly. Method: McCabe, An Analysis of Police Department Staffing: How Many Officers Do You Really Need?, ICMA. Officer counts: NYS DCJS Law Enforcement Personnel by Agency, 2024. National averages: FBI Police Employee Data.
Having pulled the original document, we checked every column in it, not just the budgets. The arithmetic is sound — the rank breakdowns add up to the stated officer totals in fifteen of sixteen rows. The problems are with definitions and sources.
None of this suggests bad faith. It is an in-house table, assembled from whatever each village published, and it reads like one. The failure is that it has since been quoted — by the village's defenders and its critics alike — as though it were a normalized dataset. It was never built to carry that weight.
Study audited: Comparison of Village Police Departments (Croton-on-Hudson PD, 2024), all sixteen rows and every column. Checked against the 2020 Census, the Comptroller's village filings for 2022–2024, and Briarcliff Manor's own adopted budget.
The document itself is a single page. Its embedded file properties say it was created on 12 March 2024 with Acrobat's Word add-in, from a document titled "Mercy College," author Richard Corvinus — who is, per his own professional listing, Croton police. In other words the table circulating as the village's police-cost comparison is one page lifted out of a larger document that the village does not publish. We looked: every file linked from the Police Department and Police Advisory Committee sections is a bicycle-safety flyer, the 2019 and 2021 annual reports, the PAC survey decks, and this. Searches of the village site for "comparison of police departments," "police staffing study" and the author's name return no document, and it appears in no Board minutes we hold.
Its purpose is not mysterious, though. The village's own Staffing page makes the same argument in prose, under the heading "Why so many officers?" — that Croton's headcount "is in keeping with neighboring departments with a similar number of residents." The table is a staffing justification. Comparing headcount is what it was for; the budget column is incidental to that job, which is part of why it cannot bear the cost rankings later built on it.
There is also a predecessor. An earlier version, Comparisons 2-8-18, was published under the previous chief and survives in the Internet Archive. It is instructive in two ways. It was in one respect better: it distinguished authorized from actual staffing, printing Croton's as "21 (18)" — a distinction the 2024 version drops, so we cannot tell which its numbers are. And the Pelham land-area error is new: the 2018 sheet listed 1.0 square miles, about right for the village; the 2024 revision changed it to 2.17, the figure for the town — in the same pass that corrected Sleepy Hollow and Tuckahoe. The table has been getting less careful, not more.
Finally, the village currently publishes two peer comparisons that disagree with each other. Against the study's figures, the Staffing page gives Briarcliff 21 officers and about 7,500 residents (the study: 20 and 8,000), Pleasantville 20 officers (the study: 23), and Bronxville 21 (the study: 23). Both documents are live on the same website.
Provenance from the PDF's embedded document properties; the 2018 predecessor via the Internet Archive; staffing figures from the village's own Staffing page, read July 2026.
Normalized figures: NYS Office of the State Comptroller, Financial Data for Local Governments — Annual Financial Report, Villages, fiscal year 2024, "Police" expenditure ÷ 2020 Census population as published in the same file. Study figures: Croton-on-Hudson PD, Comparison of Village Police Departments (2024). Read with these limits: the Comptroller's figures are actual expenditures while the study used budgets; the normalized "Police" line excludes employee benefits for every village alike, so it measures departments on a common basis rather than their all-in cost (adding benefits back raises every village, Croton included — see the section above); and Elmsford filed no expenditure detail for 2024, so it drops out of the fifteen shown.
The figure usually quoted as "the police budget" — the $4.1 million police-department line — is only part of what the village spends on policing. Pension, employee benefits and the auxiliary unit are budgeted in other sections of the same adopted budget (standard New York municipal accounting, not concealment). Add them back and the fully-loaded operating cost is closer to $7.4 million.
Center figure is the fully-loaded operating total. Debt-funded police capital (vehicles, radios, equipment) adds a further $1,050,937 this year and is not included above.
Why this differs from the $6.97m figure further down. This donut is the Village's adopted budget for 2026-27 and counts the auxiliary police unit and a payroll-share estimate of pooled benefits. The peer table below uses actual spending in fiscal 2025 as filed with the State Comptroller, and counts every police-function account a village uses, so that every village is measured identically. Two different questions — what Croton plans to spend on policing broadly, and what Croton's police department actually spent on a basis comparable to its neighbours.
Source: Village of Croton-on-Hudson Adopted Budget FY2026-27 (police department A3120, auxiliary A3189, police & fire retirement A9015, capital schedule). The health-insurance / payroll-tax / workers'-comp slice is budgeted as one pooled village-wide sum; the share shown is allocated by policing's 41% of village payroll and is an estimate, not a line item. See the whole village budget →
Reported crime is a third of its 1990 level and well below county, state and national rates — though it has roughly doubled since its 2015 low, and among the river villages Croton's is the second highest, not the lowest. Those figures belong to the state, not the village. The department is modestly busier: matched January-to-May, calls averaged about 1,600 across 2018–2023 and about 1,900 across 2024–2026, a rise of roughly 18% — a number the department publishes itself, every month. Policing costs the village roughly $7.4 million a year all in, about $3.2 million of it outside the line most people call "the police budget" — standard accounting, published in the same document, easy to miss.
Measured against its neighbours on a common basis, Croton spends $837 per resident on policing, 11th of 15 and below the median — a somewhat small force whose officers cost somewhat more than most. And Croton and Briarcliff Manor, the two villages this argument has centred on, turn out to spend within about five per cent of each other — a difference of $41 a resident, not $2 million.
Which leaves the thing that actually failed here, and it is neither of the accusations. No money was hidden — every dollar sits in the accounts the state prescribes, in documents anyone can download. And Croton is not the bargain the quoted table suggested. What was wrong was the comparison itself: a one-page in-house sheet, built for a different purpose, from sixteen documents that were never written to be compared, which both sides of a long local argument have been citing as though it settled something.
We came to this expecting to find the opposite of what we found, twice, and both are worth saying plainly.
Statewide count from the NYS Comptroller's account-level village filings for fiscal 2025 (villages reporting any expenditure on account A3120, and of those, any carrying an employee-benefits object on that same account). Reporting comparison from the peer villages' own police pages, checked July 2026 — twelve attempted, seven reached.
Reported crime comes from the NYS Division of Criminal Justice Services; workload figures from the Croton-on-Hudson Police Department's own monthly and annual reports; village-by-village cost from account-level annual financial reports filed with the NYS Comptroller; the Village's own budget documents for the Croton breakdown; population from the U.S. Census. Each panel names its source, and the underlying figures are all public — every one can be checked against the documents linked here.
We have corrected two of our own conclusions while building this page, and said so where they appear. If you find another error, tell us and we will fix it and note the change.