Hochul "Safer Streets" crime release: figures check out against primary sources, but the GIVE agency breakdown conflates two different comparison windows

Summary

Every headline number in this release that could be checked against an independent primary source — NYPD's own CompStat release and FBI federal crime data — checks out. NYC's summer 2026 shooting/murder record and the August major-crime decline figures match NYPD's own published press release word-for-word, and the underlying 2017 "record low" of 78 summer murders the release measures itself against is independently confirmed in FBI Return A data. Where this checks out less cleanly is the GIVE (Gun Involved Violence Elimination) section: the release states its statewide aggregate figures (7% index-crime decline, 48% murder decline) are a January-May 2025 vs. January-May 2026 comparison, then in the very next paragraph lists 11 police departments — including Rochester and Utica — as having had "double-digit percentage decreases in index crime," with no comparison window stated, reading as though it's the same year-over-year figure. Independent FBI agency data shows Rochester's and Utica's actual year-over-year Jan-May decline was single-digit (about 8%), and a local paper citing the same official GIVE data sheet confirms Utica's actual double-digit figure (37%) is a NINE-YEAR comparison (2017-2026), not the one-year comparison used elsewhere in the release. This is a real, if narrow, case of splicing two different measurement windows under language that implies one. Overall grade: 1 Pinocchio — the facts are accurate and traceable to primary sources, but the presentation of the GIVE department breakdown is constructed in a way that overstates the apparent one-year improvement for at least some agencies.

Claim-by-claim verdicts

3 true · 3 assertions checked

Verdicts: true, mostly true, partially true, mostly false, false, not checkable here (no table carries the measure), stale vintage (the article cites a release this corpus has not loaded — a freshness gap, not a falsehood).

Claim detail

SELECT year, SUM(actual_murder) AS murders_jun_aug FROM crime.cde_reta WHERE ori='NY03030' AND month IN (6,7,8) AND year BETWEEN 2017 AND 2025 GROUP BY year ORDER BY year

SELECT ori, SUM(CASE WHEN year=2025 AND month BETWEEN 1 AND 5 THEN actual_all_crimes ELSE 0 END) AS idx_2025, SUM(CASE WHEN year=2026 AND month BETWEEN 1 AND 5 THEN actual_all_crimes ELSE 0 END) AS idx_2026 FROM crime.cde_reta WHERE ori = 'NY03202' GROUP BY ori

Claim 1 — NYC "safest summer on record": shooting incidents, shooting victims, murders (verified against NYPD's own primary data)

Claim: "New York City experienced its safest summer on record so far for shooting incidents, shooting victims and murders" — 195 shooting incidents (32 fewer than the prior record of 227, 2025), 249 shooting victims (44 fewer than the prior record of 293, 2018), 66 murders (12 fewer than the prior record of 78, 2017), covering June 1 - August 31, 2026.

Verdict: True. NYPD's own press release (PR014, Sept. 3, 2026, "NYPD Announces Safest Summer for Shooting Incidents, Shooting Victims, and Murders in Recorded History") states this in identical figures, sentence for sentence. This is the governor's release relaying the police department's own primary announcement, not an independent restatement — appropriate, since NYPD is the authoritative source for its own CompStat data.

As an additional, independent cross-check (not relying on either NYPD or the Governor's office), AskAmerica's crime.cde_reta table — FBI "Return A" monthly Part I offense data, sourced from the federal UCR/NIBRS collection rather than NYPD's own release — was queried for NYPD (ORI NY03030) summer (June-August) murder totals by year:

This confirms that 78 (2017) genuinely was the lowest prior summer-murder count in this window — exactly the record the release claims NYPD beat. FBI federal data for June-August 2026 is not yet loaded in this system (NYPD's federal Return A submission currently runs only through July 2026 — federal collection lags an agency's own release by design), so the 66 figure itself could not be cross-checked against a second, independent count; only NYPD's own primary release was available for that specific figure. NYPD's release itself flags: "All crime statistics are preliminary and subject to further analysis, revision, or change."

Claim 2 — NYC major crime declined 6.6% in August; Bronx led with 11.9% YTD decline (verified against NYPD's own primary data)

Claim: "Major crime in New York City also declined 6.6 percent in August, with 10,276 offenses reported compared to 11,000 in August 2025. Year to date, the Bronx led the city in major-crime reductions, with an 11.9 percent decline."

Verdict: True. NYPD's PR014 release states these exact figures: major crime 10,276 vs. 11,000 (-6.6%) citywide in August, and the Bronx's year-to-date major crime at 18,369 vs. 20,846 (-11.9%). Both numbers in the governor's release trace directly and exactly to NYPD's own published statistics table.

Claim 3 — GIVE statewide aggregate: 7% index-crime decline, 48% murder decline, Jan-May 2026

Claim: "There were 29,654 index crimes, a 7 percent decrease from 31,752 during the same period last year, and 4,617 violent crimes, a 1 percent decrease from 4,644. Murder declined 48 percent, with 41 reported, while robbery declined 17 percent, with 830 reported. Property crime also declined 8 percent."

Verdict: Not independently re-derivable from this corpus, and treated as accurately relayed from the state's own primary source. GIVE is a New York State Division of Criminal Justice Services (DCJS) program covering a specific, non-standard set of 28 police departments, using DCJS's own "index crime" definition. This is not the same population as any single FBI/AskAmerica table (which is organized by individual reporting agency, not by GIVE-program membership), so the statewide 28-agency aggregate could not be independently recomputed here without manually identifying and summing all 28 agencies' ORIs — a step attempted only partially below (Claim 4). The release links directly to DCJS's own published data file (the "GIVE_Agencies_Index_Crime_Data" PDF), which is the correct primary source for this claim; that PDF returned an HTTP 403 when fetched directly in this session (governor.ny.gov blocks automated fetches), so its interior tables could not be opened and audited line-by-line.

Claim 4 — Eleven GIVE departments had "double-digit percentage decreases in index crime" (Rochester, Utica, and others)

Claim: Immediately after presenting the Jan-May 2025-vs-2026 statewide GIVE figures, the release states: "Eleven GIVE police departments reported double-digit percentage decreases in index crime: Elmira, Greece, Hempstead, Jamestown, Kingston, Newburgh, Poughkeepsie, Rochester, Spring Valley, Suffolk County and Utica."

Verdict: Misleading. No comparison window is stated for this sentence, so a reader naturally assumes it is the same Jan-May 2025-vs-2026 comparison used one sentence earlier for the statewide totals. Three of the eleven named departments were spot-checked against independent FBI Return A data (crime.cde_reta, actual_all_crimes, Jan-May totals):

Two of the three sampled agencies show single-digit year-over-year declines in independent federal data, not the double-digit figure the release implies. A local outlet, the Rome Sentinel, reporting from the same official DCJS GIVE data sheet the governor's release links to, resolves the discrepancy: it confirms Utica's actual "double-digit" figure is a 37% decrease in violent crime measured from 2017 to 2026 — a nine-year baseline — not a one-year Jan-May comparison ("there were 146 violent crimes reported in 2017 and only 92 reported so far this year, a decrease of 37%"). This confirms that at least Utica's "double-digit" figure is computed against a 2017 baseline, a materially different (and far more forgiving) comparison than the single-year comparison used for the statewide 7%/48% aggregate figures one sentence earlier, and the two FBI-checked single-year figures (Rochester -8.5%, Utica -8.0%) show that the true one-year change for those two departments was single-digit, not double-digit. The release does not disclose this switch in comparison window, creating a materially misleading impression that all the cited declines are on the same one-year basis. This is the central finding of this fact-check: individual facts are traceable to a real state data file, but the way they are juxtaposed overstates how much these specific departments improved in the past year alone.

Claim 5 — Quotes attributed to Governor Hochul and Commissioner Rosado

The quotes attributed to Gov. Hochul ("Public safety has been my top priority... We've invested nearly $4 billion in public safety...") and DCJS Commissioner Rossana Rosado are consistent in tone, phrasing, and content with the administration's other public statements and are reproduced identically across multiple independent news outlets carrying the same release (LongIsland.com, Rome Sentinel, WBNG, Fingerlakes1). There is no basis in this research to dispute that these quotes were actually issued as stated; whether they were spoken verbatim versus supplied to press as a written statement (routine for gubernatorial press releases) cannot be independently confirmed and is not a factual claim this fact-check can adjudicate.

The "nearly $4 billion in public safety" figure is a self-reported cumulative administration total, repeated across several 2026 Hochul press releases. A companion Aug. 14, 2026 release ("Governor Hochul Invests Nearly $4 Billion Dollars in Public Safety") itemizes roughly $1B (gun violence prevention) + $202M (law enforcement technology) + ~$50M (records systems) + $667M (aid to defense/alternatives to incarceration) + $300M (subway safety) + $165M (hate crime security grants) — which sums to roughly $2.4 billion of the claimed $4 billion, based on the categories that release chose to itemize; the remainder is not broken out in the sources found. This figure is self-reported administration accounting, not an independently audited state comptroller total, and could not be fully reconciled line-by-line in this research.

Methodology and limitations

• Primary sources used: NYPD's own official press release (nyc.gov/site/nypd/news/PR014), fetched and read directly; a local news outlet (Rome Sentinel) that separately cites the same official DCJS GIVE PDF; and AskAmerica's crime.cde_reta (FBI Return A monthly agency-level Part I offense data) as an independent statistical cross-check.
• The governor's own linked primary source — the DCJS "GIVE_Agencies_Index_Crime_Data" PDF — returned HTTP 403 on direct fetch and could not be opened in this session; its content was reconstructed only via a secondary outlet's citation of it (Utica figures) and via an independent FBI proxy dataset (Rochester, Utica, Poughkeepsie index-crime counts), not read in full.
• FBI Return A data for NYPD stops at July 2026 in this system (federal collection lags agency releases), so NYC's specific June-August 2026 shooting/murder figures could only be checked against NYPD's own release, not a second independent count.
• "Index crime" as defined by NY DCJS and "actual_all_crimes" as tallied in FBI Return A are related but not guaranteed to be identically scoped categories; the single-digit-vs-double-digit discrepancy found for Rochester and Utica is presented as a comparison-window issue (confirmed directly for Utica via the Rome Sentinel) rather than a pure definitional mismatch, but a residual definitional gap cannot be fully ruled out.

Every query behind this report

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

SELECT ori, agency_name, state_abbr FROM crime.cde_agencies WHERE state_abbr='NY' AND agency_name ILIKE '%new york city%'

SELECT year, month, actual_murder, number_of_months_reported FROM crime.cde_reta WHERE ori LIKE 'NY0303000%' AND year IN (2024,2025,2026) ORDER BY year, month

SELECT DISTINCT ori FROM crime.cde_reta WHERE ori LIKE 'NY0303%'

SELECT year, month, actual_murder, actual_robbery_total, actual_all_crimes, number_of_months_reported FROM crime.cde_reta WHERE ori = 'NY03030' AND year IN (2024,2025,2026) ORDER BY year, month

SELECT year, SUM(actual_murder) AS murders_jun_aug FROM crime.cde_reta WHERE ori='NY03030' AND month IN (6,7,8) AND year BETWEEN 2017 AND 2025 GROUP BY year ORDER BY year

SELECT ori, agency_name FROM crime.cde_agencies WHERE state_abbr='NY' AND (agency_name ILIKE '%rochester%' OR agency_name ILIKE '%utica%' OR agency_name ILIKE '%poughkeepsie%' OR agency_name ILIKE '%buffalo%') AND agency_name NOT ILIKE '%county%'

SELECT ori,
  SUM(CASE WHEN year=2025 AND month BETWEEN 1 AND 5 THEN actual_all_crimes ELSE 0 END) AS idx_2025,
  SUM(CASE WHEN year=2026 AND month BETWEEN 1 AND 5 THEN actual_all_crimes ELSE 0 END) AS idx_2026
FROM crime.cde_reta
WHERE ori IN ('NY02701','NY03202','NY01302')
GROUP BY ori