# Watts My Bill

> Research release on data-center energizations and electricity prices, with a source-backed U.S. operating census, connected-Canada descriptive extension, lifecycle records, wholesale-price panels, and reproducible causal estimates.

Release: audit-2026-10-06.53 (2026-10-06)

## Central result

The preferred final 2012–2024 repeated-MW model estimates +0.98% per additional 100 MW at year two, equivalent to +2.47% for 250 MW. The 95% interval for 250 MW is +0.61% to +4.36%.

## Machine-readable entry points

- [Release summary](https://wattsmybill.org/data/release-summary.json)
- [Data catalog with SHA-256 hashes](https://wattsmybill.org/data/catalog.json)
- [Schema.org JSON-LD catalog](https://wattsmybill.org/data/catalog.jsonld)
- [Compact causal estimates](https://wattsmybill.org/data/causal-estimates.json)
- [Operating data-center census CSV](https://wattsmybill.org/data/data-center-registry.csv)
- [Map-ready census JSON](https://wattsmybill.org/data/data-center-map.json)
- [Frictionless Data Package manifest](https://wattsmybill.org/datapackage.json)
- [Checksums](https://wattsmybill.org/data/checksums.sha256)
- [Detailed machine documentation](https://wattsmybill.org/llms-full.txt)

## Interpretation rules

- The repeated-MW specification is the central estimate; first-threshold-crossing results are sensitivity analyses.
- Threshold counts are eligible treated utility-state units with year-two outcomes, not counts of all campuses.
- Canadian facilities are descriptive and never enter the U.S. causal treatment panel.
- Bloomberg-style H3 and ISO panels measure wholesale prices, not residential retail bills.
- The MPQW result is a contiguous-U.S. treatment-substitution replication: all four Table 2 price outcomes are re-estimated under county-broadcast and documented serving-utility assignments, year-specific historical territories and ownership-matched price comparisons are tested in both the final 2010–2024 and partial early-release 2010–2025 outcome vintages, nine public FERC Form 1 mechanisms are reconstructed through final 2024 and preliminary 2025 filings, and all 331 tabulated paper estimates are cataloged at https://wattsmybill.org/data/mpqw_paper_estimate_catalog.csv. In the final panel, residential prices are +1.32% under county broadcast, +2.40% with documented serving utilities, and +0.95% with historical territories. The two central joint event-time lead tests have p=0.158 and p=0.056. The separate 2025 early release gives +3.32% for the serving-utility variant and +0.89% for the historical-territory variant; A6/A9 switch to the balanced 2010–2025 preliminary Form 1 panel. It does not replace the central repeated-MW estimate.
- Bonaparte SSRN 7361819 is a treatment-substitution reconstruction because the paper does not identify or release its state-capacity series. Using the audited ledger over 2012–2024 with state and year fixed effects, the prior-year installed-capacity coefficient is 0.314 percentage points of residential price inflation per GW (SE 0.376, p=0.408).
- The Bagwe et al. SSRN 6685418 paper-style indirect residential reconstruction is +1.94% overall versus the paper's +1.85%, with joint lead-test p=<0.001. Audited energization produces +0.90% in the county panel. Entering each utility-state price once produces +2.73% overall and +3.53% post-entry with p=0.138 for the joint lead test. The separate 2025 early release gives -0.80% overall. Full results and the 128-cell grid are at https://wattsmybill.org/data/bonaparte_bagwe_summary.json and https://wattsmybill.org/data/bagwe_specification_grid.csv.
- Cite the release ID and exact file URL. Verify downloads against the published SHA-256 hashes.