About this page
This is the summary of an independent review of revision 1 of the report, reproduced unchanged. The full review package (findings and method, recalculation script, evidence downloads and its own checksum manifest) is in independent_review_2026-09-23/. Revision 2 of the report adopted its corrections; see the audit summary, Section 14.
Independent review of Harris County’s FY2027 budget investigation¶
Prepared September 23, 2026. FY2027 runs October 1, 2026–September 30, 2027. Dollar amounts below are millions unless stated otherwise.
My conclusion: the original investigation identifies a real operating-budget sustainability problem, but it needs material corrections before its detailed calculations and comparisons can be relied upon. The budget increase, reliance on one-time funding, employee-benefit pressure, and difficulty reconciling public disclosures are supported. Several cost-driver totals are overstated through double counting, and some reserve, tax, and compensation interpretations go beyond the evidence. The records reviewed do not establish concealed expenditure, misappropriation, or fraud.
I read the 29-page investigation, examined its supporting analysis and scripts, checked the supplied records, and made independent calculations from the county’s financial documents. All 984 entries in the original checksum manifest passed. Seven important PDFs—including both FY2027 budget volumes, the FY2025 audited financial report, the June 2026 financial report, and the county tax worksheet—matched newly downloaded official copies byte for byte. I rebuilt all 83 department change tables: each reconciles exactly, and the 81 county department changes tie to the General Fund increase. This is a documentary review, not an audit opinion based on transaction testing or external confirmations.
The budget is substantially higher, and one-time resources are financing part of it. Departmental General Fund appropriations rise from $2,769.425 million to $3,095.400 million: an increase of $325.975 million, or 11.77%. The larger resources-based budget, which includes working capital, rises 7.61% to $3,351.376 million. These measure different things and should remain separately labeled. County budget, Appendix A, printed pp. 103–105.
OMB projects $3,079.4 million of revenue, including $82.9 million it identifies as one-time revenue, and plans a further $16 million use of prior surplus. Together, these represent $98.9 million, or 3.20% of departmental appropriations, that must be replaced, offset by lower spending, or financed again if the same spending level continues. Property sales account for $41 million of the one-time revenue. This supports concern about sustainability; it is not a precise forecast of next year’s deficit. Some expenditures may also be temporary, and future recurring revenues may grow. The original “55% of the gap” is a ratio to the $181 million planning shortfall, not a fully verified reconciliation of how that shortfall was closed. County budget, printed pp. 27 and 40–44.
The most consequential error is in the original cost-driver parser. It counts department totals and the underlying program components together. Removing identical dollar rows cannot fix that overlap. My reconstruction uses only department summary change tables and retains their continuation pages.
| Item | Original investigation | Independently checked result |
|---|---|---|
| Named group-health increases across department summary tables | $156.5 | $87.536 |
| Named sustainable law-enforcement staffing increases | $44.7 | $30.770 |
| Named year-two law-enforcement parity increases, excluding HCTRA’s separately described contract costs | $44.6 | $44.578, confirmed |
| Pay-equity treatment | Adds or juxtaposes $55.6, $21.4 and $30 | Department allocations and central offsets must be netted; identified net incremental change is $30.000 |
The $87.536 million is the sum of specifically labeled rows in the 83 summary tables, not a replacement estimate of the entire health plan’s funding need. OMB’s separate $100 million employer-funding increase and approximately $89 million General Fund estimate require a scope reconciliation to those rows. The corrected tables are saved with this review. Department budget detail; especially PDF pp. 178, 855, 886–887 and 954–965.
A second parser defect omits three departments from the original FY2026 appendix. Its adjusted-budget total is understated by $22.880 million and its spending-plus-encumbrance total by $16.422 million. The corrected 81-department table reconciles to the Auditor’s totals.
Reserve use needs a different explanation. The original report infers an $89.1 million reserve drawdown from the decline in the working-capital budget line. That inference is unsupported. The Auditor certifies $387.046 million of beginning cash plus $2,964.330 million of revenue and transfers, totaling $3,351.376 million. Deducting departmental appropriations leaves the $255.976 million working-capital line. It is a budget calculation, not an actual cash-flow statement.
More significantly, OMB’s revenue forecast exceeds the Auditor’s initial certification by $115.070 million. The book explains that OMB anticipates receipts the Auditor may certify later. This does not establish an unlawful or unfunded appropriation: beginning cash is included in certified resources. It does identify a collection and certification risk that deserves an itemized reconciliation. Do not add this $115.1 million to the $98.9 million above as if they were independent exposures; their underlying sources can overlap. The report’s claim that the FY2027 certification was unavailable is incorrect: it is included in the supplied book. Auditor’s September 17 letter, PDF pp. 148–150; OMB explanation, printed p. 44.
The tax increase is confirmed, but the tax narrative needs correction. The county’s total rate increases 9.59%, from $0.38096 to $0.41750 per $100. It comprises M&O of $0.36396 and debt service of $0.05354; the full rate should not be called the “General Fund rate.” The combined four-entity rate rises 7.58%. The statutory 3.5% framework concerns the voter-approval calculation for operations, with new property and other adjustments; it is not a blanket ceiling on total revenue or individual bills. County tax information, Texas Comptroller explanation.
There is also a $35.756 million unresolved difference in the county’s own disclosures: the statutory M&O and debt revenue tables imply a $246.455 million increase, while the budget’s front notice states $210.699 million. The investigation should disclose both and seek a reconciliation. Likewise, its $128 million comparison belongs to OMB’s May revenue scenarios; it cannot be reproduced using the final worksheet inputs as the original audit notes claim. Neither discrepancy alone establishes an improper tax calculation.
Healthcare is a genuine risk; some related allegations are too broad. The employer contribution per budgeted employee rises from $20,228 to $25,342, or 25.28%. The June financial report shows the Health Insurance Management fund with negative net position of $31.342 million and $48.970 million of incurred-but-unreported claims. The proposed $6 million transfer from Workers’ Compensation merits particular scrutiny because that donor fund also shows negative net position, $0.610 million, although both deficits are described as timing differences. Cash availability and reserve adequacy need separate examination. June report, printed pp. 30–31.
The budget describes the $100 million contribution increase in response to the health-fund funding gap. The original report should not present the $92 million gap as an established additional deficit remaining after that increase. Premium-free employee-only coverage on some plans also does not mean all employee healthcare is free.
The compensation comparison overstates its precision. Houston’s official civilian schedule, effective July 4, 2026, has higher ranges than the classification PDF used in the original comparison. Applying that schedule to the original 34 pairings reduces the median county minimum advantage from roughly 31% to 26.3%; the median maximum advantage falls to 1.4%. For the 69 exact-title pairs, the minimum advantage becomes 22.6%, with the county higher in 61 pairs, rather than 64. These remain selected salary-range comparisons—not measurements of actual pay for equivalent work. Job duties, seniority, effective dates, and total benefits remain unmatched. Houston civilian pay structure.
The $786,188 annual base-pay increase for the 12 elected officials is supported, but represents only about 0.24% of the $326 million departmental increase. Its governance significance should be evaluated separately from its limited explanatory power for the overall budget increase.
Contract patrol and audited control weaknesses deserve more precise attention. Of the $90.8 million county-funded portion of contract patrol, $51.0 million relates to contracts purchasing less than 100% of an officer’s service and $39.8 million is the difference between contract rates and officer costs. The latter is the clearer pricing shortfall. Describing all $90.8 million as a private-customer subsidy ignores the county’s retained service share; deployment records are needed to test whether that share is delivered. County budget, printed pp. 20–21.
The FY2025 audit identifies stronger evidence of reporting-control problems than the budget’s presentation alone: a corrected $10.1 billion overstatement on both sides of fiduciary activity, a corrected $341 million OPEB-liability overstatement, capital-asset classification problems, and deficient review of indigent-defense reporting. These are material control findings, not evidence that those amounts were stolen or are additional FY2027 spending. The financial statements received an unmodified opinion. FY2025 ACFR, printed pp. 296–301.
My recommended priorities are to correct the investigation’s calculations, publish reconciled revenue and reserve schedules, and test execution of the adopted plan. The county should identify the collection dates and certification status of anticipated supplemental revenue, the recurring funding needed after FY2027, health-fund reserve requirements, contract-patrol cost recovery and service delivery, and remediation of the external auditor’s findings. The evidence supports concern about sustainability and financial controls. It does not support treating every reserve, transfer, or public disclosure problem as concealed spending.
The accompanying Findings and method provides report-page corrections, an additive budget-growth reconciliation, source details, and the remaining records needed. Recalculated results and the reproducible calculation script document the arithmetic.