Graduation Year

2026

Document Type

Campus Only Senior Thesis

Degree Name

Bachelor of Arts

Department

Economics

Reader 1

Nicholas Kacher

Reader 2

Nayana Bose

Abstract

The Inflation Reduction Act (IRA) is the largest U.S. climate investment to date, yet little is known about how its dollars are distributed across states. This thesis assembles a new 50-state dataset of IRA allocations per capita by combining (i) U.S. Department of Energy (DOE) Home Energy Rebates, (ii) U.S. Environmental Protection Agency (EPA) Climate Pollution Reduction Grants (CPRG) implementation awards, and (iii) U.S. Department of Agriculture/Natural Resources Conservation Service (USDA/NRCS) conservation obligations. I relate log dollars per person to proxies for environmental burden and state context—log CO₂ emissions per capita, PM₂.₅ concentrations, log median household income, log gross state product per capita, unemployment rate, share of clean-energy jobs—and a legislative control dummy (Democratic=1, Republican=0), estimating cross-sectional OLS with HC3 standard errors. I also estimate the model separately by funding channels to assess whether the relationships differ by program mechanism—formula (DOE), competitive (EPA), and apportionment/sign-up (USDA/NRCS)—and to identify which channels drive the aggregate pattern. In the baseline models, in which I run specifications with and without legislative control, two patterns are robust. First, states with higher CO₂ per capita receive more IRA dollars per person; in my later regressions, this association is strongest in the USDA/NRCS channel and smaller but present in DOE. Second, higher unemployment is linked to less funding per person. This is evident in the aggregate regressions, and within DOE and USDA, consistent with capacity frictions that may exist in application or administration. PM₂.₅ has no stable relationship with allocations, and the partisan dummy is positive but imprecise once other factors are included. EPA/CPRG coefficients are broadly imprecise, reflecting a winners-only sample (N = 29) and competitive selection. Overall, results indicate partial alignment of funds with carbon intensity, alongside a shortfall for states with weaker labor markets. Policy implications include strengthening technical assistance, simplifying applications, and making need-weighting explicit in formulas and rankings. Limitations include using state-level data (within-state heterogeneity is not accounted for), using cross-sectional data (not seeing the real-world benefit—cause and effect—of these allocations), and EPA selection style (not just amount) not modeled. These limitations point to future work using finer geographies, using panel-data and expanded variables (e.g., CO₂ emissions reduced), and further models that dive deeper into how and which states are selected to win aid.

Download

This thesis is restricted to the Claremont Colleges current faculty, students, and staff.

Share

COinS