Subject New Markets Tax Credit (NMTC) · place-based policy · rural investment
The rural mobilization gap in a U.S. place-based tax credit
Does the rural private-leverage gap persist when rural and urban projects are compared within the same Community Development Entity?
- Status
- Job market paper · August 2026
- Role
- Sole author; Katia Antunes and Elizaveta Gonchar are acknowledged contributors
- Authors
- Ian Helfrich
- Method
- Community Development Entity fixed effects · order-invariant decomposition · bunching test
- What this record does not establish
- The design is descriptive and noncausal, does not identify why intermediaries differ, and leaves the mechanism unresolved.
Public artifactsJob market paper
Abstract
Summary
A descriptive, noncausal decomposition of the rural-versus-urban leverage gap in the U.S. New Markets Tax Credit using Community Development Entity fixed effects. The raw private-leverage gap of −0.262 falls to a mean within-CDE estimate of −0.047 (95 percent CI: −0.245 to +0.152); the median estimate is −0.001, and neither margin estimate is statistically distinguishable from zero. In the base order-invariant decomposition, most explained movement is associated with observed CDE composition: −0.185, or 86 percent. The design does not identify why CDEs differ, and the mechanism remains unresolved.
Question
Does the rural private-leverage gap persist within the same Community Development Entity?
Choose a comparison. Within CDE compares rural and urban projects handled by the same Community Development Entity. The confidence interval and decomposition are reported separately.
Raw rural gap: −0.262.
| Raw gap | −0.262 |
|---|---|
| Mean within-CDE estimate | −0.047; exact 95% CI [−0.245, +0.152] |
| Median within-CDE estimate | −0.001 |
| Base decomposition | −0.185, or 86% of explained movement, is associated with observed CDE composition |
The analysis is descriptive and noncausal. It does not identify why CDEs differ, and the mechanism remains unresolved.
Methods and notes
Ian Helfrich originated the project and is its sole author. Contributors: Katia Antunes and Elizaveta Gonchar. This contributor acknowledgment does not assign or imply specific roles.
Read the August 2026 working paper (PDF) · Inspect the public repository and evidence files
The paper makes three contributions. First, it constructs a project-level panel from 19,907 QLICI transactions covering 8,024 projects with origination years 2001–2022, linked to issuing CDEs and receiving QALICBs, with rural/non-metro tract flags from OMB and ERS definitions. Second, it reports a descriptive, noncausal leverage-gap decomposition with CDE fixed effects: most of the explained movement is associated with observed CDE composition. The median estimate is near zero and neither margin estimate is statistically distinguishable from zero; the mean interval is wide, so the design does not establish that a within-CDE component is absent or identify why CDEs differ. Third, it tests for bunching at the 20 percent non-metro administrative target using realized cumulative-deployment data. No excess mass is detected around the target; allocation-agreement compliance is unobserved, so the test does not establish whether the target binds through those agreements. A subgroup scan across twenty-four cells leaves one exception after a multiple-comparison correction: rural commercial real-estate rehabilitation at −0.439, reported as exploratory because that dimension was added after an earlier draft flagged real estate.
The results motivate evaluating intermediary-allocation levers alongside project-level interventions. Because the design is descriptive and the mechanism remains unresolved, it does not establish a causal or lower-cost policy fix.
BibTeX
Citation
@misc{helfrich2026rural,
author = {Helfrich, Ian},
title = {The rural mobilization gap in a U.S. place-based tax credit},
year = {2026},
note = {Job market paper, August 2026},
url = {https://ihelfrich.github.io/research/ian-helfrich-nmtc-working-paper.pdf}
}