ITALY CRIME: CALCULATIONS AND METHOD
Evidence checked 11 October 2026.

Source: Istat public SDMX service, flow IT1,73_230,1.0; 2024 Italy.
Downloaded 11 October 2026 with Accept: text/csv.
Query: https://esploradati.istat.it/SDMXWS/rest/data/IT1,73_230,1.0/A.IT.OFFEND..9.TOTAL..?startPeriod=2024&endPeriod=2025
Query returned 2024 only, despite requesting through 2025.
Published-total flow update: 4 February 2026, 15:39:09.284 UTC.
Demographic subflow update: 19 December 2025, 09:06:34.862 UTC.
These are flow update timestamps, not exact extraction dates per observation.

UNIT AND FILTERS
OFFEND: presumed offenders reported or arrested by police. Attached metadata
says a person can be counted whenever reported in the year for each offence.
The data do not follow subsequent judicial proceedings. They do not count
unique people, convictions, all crime incidents or solved cases.

Filter the raw totals in this order:
FREQ=A; REF_AREA=IT; DATA_TYPE=OFFEND; TIME_PERIOD=2024; SEX=9;
AGE=TOTAL; COUNTRY_CITIZEN=WORLD; CITIZENSHIP=FRG or TOTAL.
Then choose TYPE_CRIME separately. Read counts from OBS_VALUE.
FRG means foreign citizenship. TOTAL means all citizenships in the same row.
Do not add individual countries to WORLD or add overlapping offence categories.
UNIT_MEAS and UNIT_MULT are blank in returned rows; counts are used as published.

FORMULA AND EXPECTED OUTPUT
Percentage = FRG / TOTAL * 100. Round only the final result.
Use a General/Number spreadsheet cell with *100, not Percentage formatting.
STREETROB: 6948 / 11571 * 100 = 60.046668%; one decimal 60.0%.
PICKTHEF: 5057 / 7257 * 100 = 69.684443%; one decimal 69.7%.
ROBBER: 12073 / 23058 * 100 = 52.359268%; one decimal 52.4%.
RAPE: 2783 / 6360 * 100 = 43.757862%; one decimal 43.8%.
THEFT: 48258 / 101023 * 100 = 47.769320%; one decimal 47.8%.
BAGTHEF: 1327 / 2177 * 100 = 60.955443%; one decimal 61.0%.
The selected rows retain ITL too, for the reconciliation check.

MINISTRY REPORT, DISTINCT SOURCE AND VINTAGE
https://www.camera.it/temiap/2025/11/18/OCD177-8589.pdf
2024 annual report, Doc. XXXVIII n. 3. Printed page 8, PDF page 9.
Foreign reports: 230812 without arrest + 56892 arrest reports = 287704.
All reports: 672208 without arrest + 156506 arrest reports = 828714.
287704 / 828714 * 100 = 34.716923%; one decimal 34.7%.
Printed page 5, PDF page 6, reports the rounded share.
This is NOT a corrected share of crimes committed, or an arrests-only share.
Do not combine its counts with a different Istat extraction.

FINAL CENSUS RESIDENT COUNTS, NOT A CRIME-RATE DENOMINATOR
https://www.istat.it/wp-content/uploads/2025/12/Censimento-e-dinamica-della-popolazione-Anno-2024.pdf
Final release 18 December 2025. Page 2 Prospetto 1 and page 6 Prospetto 4.
31 December 2023: 5253658 / 58971230 * 100 = 8.908849%.
31 December 2024: 5371251 / 58943464 * 100 = 9.112547%.
Both round to 9%. End-of-2023 is the opening population stock for 2024.
The police citizenship table is not restricted to registered residents.
No matched per-resident rate or demographic adjustment was calculated.

AUTOMATED RECONCILIATION
Extract the complete ZIP, then run: python verify_istat.py
No extra packages or network calls are needed. It recreates the selected CSV
and calculation JSON from raw files. Every selected citizenship total matches
the 16 observed detail cells: sex 1 and 2; ages Y_UN13, Y14-17, Y18-24,
Y25-34, Y35-44, Y45-54, Y55-64 and Y_GE65. Direct published totals govern
the calculation. Missing unknown cells are not assumed zero. FRG + ITL = TOTAL
does not establish how unknown citizenship is treated in the source system.

ACCESS AND INTERPRETATION LIMITS
The original caption's source and year and full speech remain unverified.
The precise older press extract is unrecovered. The 43% all-crime assertion
is unsupported by matching recovered evidence, not disproved for every possible
year or measure. No all-offences total was manufactured from categories.
The substantial recorded disparity remains real. No evidence here establishes
that statistical adjustment removes it, or that nationality causes it.

PROVENANCE
Raw CSV, XML and JSON files are preserved statistical service responses from
Istat. source-manifest.json lists official queries, retrieval date and hashes.
Selected rows and percentages are TALLYCAT01 derivations, not another official
release. Research and production assisted by AI.