Archived: DPWH NEP → HGAB pre-OO2 comparison

Potential repeated line items: FY2027 NEP → FY2026 GAA, then FY2025 GAA

Multi-purpose buildings, roads, bridges, and flood-control projects. The name screen found 283 FY2027 NEP lines with a 90% or higher title match in the FY2026 enacted GAA, and 454 with a 90% or higher title match in the FY2025 enacted GAA. Bridge candidates require overlapping plus-form chainage. Road matches are excluded when both stated ranges are disjoint; missing chainage is flagged for review. Different package, phase, stage, lot or segment labels require scope review. They may identify legitimate separate work or changes in how the same work is described; overlapping chainage still signals shared extent. Explicit opposing directions are excluded. For all categories, verify location, scope, phase, and completion status before treating a row as the same project. Explicit opposing directions are excluded as location conflicts. Amounts are in pesos. Title similarity is a screening lead, not proof that two appropriations cover the same scope.

PH location DB supplies current 20th Congress names through qualified geographic context and the strict local-map screening rules. Uncertain locations stay blank. Names describe geographic coverage, not sponsorship.

Where are the potential insertions and removals?

Our corrected comparison identifies 5,549 domestic insertion candidates and 1,002 domestic removal candidates. Another 4 unpaired FAP records are reviewed separately, giving 5,553 HGAB-only review records in these breakdowns. 0 FAP removal candidates remain. These are provisional line-item matches, not confirmed legislative changes.

The externally reported 5,837 insertions / 2,139 removals belong to a different comparison and are not the totals generated by this pipeline. Our removal shortlist excludes 252 unpaired NEP records with plausible HGAB counterparts. Retained projects with allocation increases or decreases belong to allocation adjustments below.

Implementing offices and programs come from each source schedule, not inferred project geography. Missing offices remain unspecified, including the 4 unpaired FAP records. Insertion amounts are HGAB allocations; removal amounts are original NEP allocations, not certified savings.

Potential insertions · domestic HGAB-only line items

The domestic 3rd-reading title screen leaves 5,549 HGAB lines without a plausible NEP counterpart, carrying ₱158.36B. Foreign assisted projects are excluded and reviewed separately below. Unmatched titles are review leads, not confirmed additions.

PH location DB supplies current 20th Congress names through qualified geographic context and the strict local-map screening rules. Uncertain locations stay blank. Names describe geographic coverage, not sponsorship.

Source: local HB 10858 Volume I-C text extraction and FY2027 NEP project extract. HB amounts are allocations attached to extracted candidate rows. The corrected native-text parser reconciles the printed program totals; still check each project against the original bill page; matching does not verify project scope, chainage, direction, package, or completion status.

Potential removals · domestic NEP-only line items

1,002 NEP lines have no selected or plausible HGAB counterpart, carrying ₱64.417B in NEP allocations. The reverse title screen holds out another 252 lines with compatible counterparts scoring at least 83%. These are candidates for line-item removal, pending scope verification. A lower allocation on a retained project is not a removal. Renaming and segmentation may conceal a counterpart; absence from this matching result does not certify deletion.

Allocation adjustments · retained domestic projects

These selected NEP/HGAB counterparts remain in both stages with different allocations. Positive changes are increases; negative changes are decreases. Nonexact title counterparts remain provisional. These lines are excluded from the insertion and removal shortlists.

Projects by region

Project counts by region for the selected fiscal year. This is not adjusted for population or need.

Rainwater Collector System (RCS): district allocation

Evaluated separately from category rankings. The key question is whether the same allocation across districts matches rainfall exposure, local demand, and project capacity.

INSPIRE Building: single-project comparison

Kept separate from program-category totals. This compares individual NEP entries while excluding records whose category or title mentions roads or bridges.

Site accessibility question

The listed INSPIRE site is in Laoag, at the northern end of Luzon. If the facility is meant to serve users from across the country, test whether that location is accessible to its intended users. Batangas is a possible comparison site: the Philippine Ports Authority describes Batangas Port as providing access to the Visayas region. Compare travel time, expected users, transport links, land and operating costs before concluding which site would serve the purpose better. PPA on Batangas Port access.

Other priority programs by district

District rankings for multi-purpose buildings, access roads and bridges, and flood control. INSPIRE is shown separately above as an individual project; adjacent water supply categories remain in the name-shift watchlist.

District allocations

How to read the pattern: compare office totals and project counts, then verify project titles and scope.

Exact ₱5M roads/bridges and flood-control allocations

District offices with one or more project entries budgeted at exactly ₱5,000,000 in the access roads/bridges category or flood mitigation categories. Counts are entries, not distinct projects verified against source documents.

Methods: hypergeometric overlap test and region-stratified randomization test. The hypergeometric test treats the 199 district offices as equally likely to receive each program’s ₱5M entry, fixes the observed counts (96 road/bridge offices and 45 flood offices), and calculates how often random independent placement would produce 35 or more shared offices. The region-stratified test is a second check: it keeps the observed road/bridge offices fixed and repeatedly reassigns the 45 flood-office flags among offices in the same region, preserving each region’s flood count. Its p-value is the share of 100,000 reshuffles with overlap at least as large as observed. Small p-values mean this degree of overlap is rare under these chance models. They do not prove deliberate allocation or a “consolation” motive. The analysis is exploratory because ₱5M and these programs were selected after noticing the pattern; the models also do not adjust for project need, workload, or project mix.

District offices with entries in both programs

Water-sector name-shift watchlist

Includes projects under the shared PAP2 water-sector umbrella plus district project titles containing rainwater, water collector, catchment, cistern, water supply, water system, or water storage terms outside that umbrella. Keyword matches are review candidates, not confirmed reclassifications.

District Engineering Office allocation ranking

Ranked by the selected measure in the selected year. In this single-year dataset, allocation and share have the same order because each share uses the same annual total. The concentration figures do not adjust for district population, project needs, scope, or cost.

Repeated amounts across offices

Groups require at least three offices within the same category. Rainwater Collector System is assessed in its own section above; other category labels can still cover different project sizes.

Largest statistical outliers

The current score compares each amount with the median for its broad category. A high score can reflect a genuinely larger project, not an inflated unit cost. Use this list to locate records that need scope checks.

Supporting methodology · NEP baseline screens

These explanations describe the NEP baseline screens on the DPWH patterns page. They are distinct from the HGAB insertion and removal comparison above.

Potential repeated line items: FY2027 NEP → FY2026 GAA, then FY2025 GAA

Multi-purpose buildings, roads, bridges, and flood-control projects. The name screen found 283 FY2027 NEP lines with a 90% or higher title match in the FY2026 enacted GAA, and 454 with a 90% or higher title match in the FY2025 enacted GAA. Bridge candidates require overlapping plus-form chainage. Road matches are excluded when both stated ranges are disjoint; missing chainage is flagged for review. Different package, phase, stage, lot or segment labels require scope review. They may identify legitimate separate work or changes in how the same work is described; overlapping chainage still signals shared extent. Explicit opposing directions are excluded. For all categories, verify location, scope, phase, and completion status before treating a row as the same project. Explicit opposing directions are excluded as location conflicts. Amounts are in pesos. Title similarity is a screening lead, not proof that two appropriations cover the same scope.

Projects by region

Project counts by region for the selected fiscal year. This is not adjusted for population or need.

Rainwater Collector System (RCS): district allocation

Evaluated separately from category rankings. The key question is whether the same allocation across districts matches rainfall exposure, local demand, and project capacity.

Rainfall is not uniform across districts. This budget extract has no rainfall, collection capacity, or project scope fields, so equal amounts are a clear allocation pattern but cannot establish whether the distribution matches need. A defensible comparison needs rainfall normals joined to the implementing districts.

The office name may not define the exact project boundary; verify district service areas before matching climate data.

INSPIRE Building: single-project comparison

Kept separate from program-category totals. This compares individual NEP entries while excluding records whose category or title mentions roads or bridges.

The listed INSPIRE site is in Laoag, at the northern end of Luzon. If the facility is meant to serve users from across the country, test whether that location is accessible to its intended users. Batangas is a possible comparison site: the Philippine Ports Authority describes Batangas Port as providing access to the Visayas region. Compare travel time, expected users, transport links, land and operating costs before concluding which site would serve the purpose better. PPA on Batangas Port access.

Other priority programs by district

District rankings for multi-purpose buildings, access roads and bridges, and flood control. INSPIRE is shown separately above as an individual project; adjacent water supply categories remain in the name-shift watchlist.

How to read the pattern: compare office totals and project counts, then verify project titles and scope.

Exact ₱5M roads/bridges and flood-control allocations

District offices with one or more project entries budgeted at exactly ₱5,000,000 in the access roads/bridges category or flood mitigation categories. Counts are entries, not distinct projects verified against source documents.

Methods: hypergeometric overlap test and region-stratified randomization test. The hypergeometric test treats the 199 district offices as equally likely to receive each program’s ₱5M entry, fixes the observed counts (96 road/bridge offices and 45 flood offices), and calculates how often random independent placement would produce 35 or more shared offices. The region-stratified test is a second check: it keeps the observed road/bridge offices fixed and repeatedly reassigns the 45 flood-office flags among offices in the same region, preserving each region’s flood count. Its p-value is the share of 100,000 reshuffles with overlap at least as large as observed. Small p-values mean this degree of overlap is rare under these chance models. They do not prove deliberate allocation or a “consolation” motive. The analysis is exploratory because ₱5M and these programs were selected after noticing the pattern; the models also do not adjust for project need, workload, or project mix.

Water-sector name-shift watchlist

Includes projects under the shared PAP2 water-sector umbrella plus district project titles containing rainwater, water collector, catchment, cistern, water supply, water system, or water storage terms outside that umbrella. Keyword matches are review candidates, not confirmed reclassifications.

District Engineering Office allocation ranking

Ranked by the selected measure in the selected year. In this single-year dataset, allocation and share have the same order because each share uses the same annual total. The concentration figures do not adjust for district population, project needs, scope, or cost.

Repeated amounts across offices

Groups require at least three offices within the same category. Rainwater Collector System is assessed in its own section above; other category labels can still cover different project sizes.

Largest statistical outliers

The current score compares each amount with the median for its broad category. A high score can reflect a genuinely larger project, not an inflated unit cost. Use this list to locate records that need scope checks.

Most frequent exact amounts

Across all categories. Rounded figures recur often in budget proposals and are not suspicious by themselves.

Foreign assisted projects · allocation adjustments and reconciliation

The printed FAP schedule contains 29 lines, ₱44,749,011,000. All are excluded from the domestic additions shortlist. This excludes 4 unmatched FAP lines carrying ₱806,910,000; it does not establish that those projects are absent from NEP. Loan identity, counterpart funding and project scope need separate verification.

25 FAP lines with selected NEP counterparts

25 selected counterparts are compared below. Unpaired lines have no assumed zero NEP baseline. Fuzzy counterparts remain provisional. PDF page numbers use the parser’s physical-page convention.