Measured or assumed? Regulatory defaults, ownership, and the measurement content of the U.S. greenhouse gas inventory

Yichuan Wang , Suying Song

Municipal solid waste landfills are the third-largest anthropogenic source of methane in the United States, and the national inventory for this source is assembled almost entirely from facility self-reports filed under Subpart HH of the Greenhouse Gas Reporting Program (GHGRP). Airborne and satellite surveys estimate actual emissions at 1.4 to 2.7 times reported levels, but remote sensing alone cannot identify how individual filings or reporting methods contribute to this gap. We examine the construction of reported emissions under Subpart HH and distinguish facility-years according to the regulatory equation used. The analysis covers the full universe of Subpart HH filings—16,319 facility-years for 1,331 landfills from 2010 to 2023—linked to the GHGRP parent-company register and the EPA Landfill Methane Outreach Program database.
Use of the unmeasured branch varies systematically with ownership. It accounts for 13.7% of facility-years at landfills owned and operated by publicly listed waste firms, compared with 27.5% at municipally owned sites operated by a listed firm, 31.3% at private firms, and 33.2% at municipally owned and operated sites. We then use the staggered installation of gas collection systems at 119 landfills to estimate changes associated with crossing the equation boundary. Pre-treatment coefficients are generally indistinguishable from zero, and the two-way fixed-effects and Callaway-Sant'Anna estimates are similar (0.350 and 0.373). Installation is associated with a 34- to 38-percentage-point increase in implied mitigation and an approximately 45% decline in reported methane. Because installation changes both physical capture and the applicable reporting equation, public data do not permit those components to be separated. In contrast, switching from the regulatory default methane fraction to a site-specific value produces no detectable within-facility change. The 23-percentage-point cross-sectional difference between measured and default reporters therefore reflects selection. About one third of facilities also reverse their reporting of durable attributes such as truck scales at least once, whereas the gas-collection indicator is much more stable.
The regulatory partition is strongly reflected in the data. Under 40 CFR 98.343(c)(2), emissions for landfills without gas collection equal modeled generation adjusted by a table-based oxidation factor; 96.9% of facility-years in this branch correspond exactly to a permitted default value. The branch represents 25.0% of facility-years and 23.4% of reported methane tonnage. Prior research documents reliance on default oxidation factors among landfills without collection systems. We extend that evidence by quantifying the branch in tonnage, showing that the partition is nearly exhaustive, and documenting its stability over fourteen years. An extension to Subpart C stationary combustion, based on 274,858 fuel-year records at 7,479 facilities, indicates that the broader mechanism is not limited to landfills. Default-factor methods account for 92.9% of combustion facility-years and 64.2% of reported carbon dioxide; the tonnage share increased from 57.9% to 67.4% during the sample period, and tier downgrades outnumber upgrades by 1,488 to 974. Together, the findings show that regulatory equation assignment materially affects reported inventory totals and that reliance on default methods varies across ownership types and has increased in stationary combustion.