Luke Maddock

Assistant Professor of Data Analytics, Denison University

Publications

  1. Maddock, L., & Pena, A.-A. (2026). Funding, Facilities, and the Face of Homelessness: Heterogeneous Impacts of Federal Grants on Sheltered and Unsheltered Counts. Southern Economic Journal. DOI

  2. Maddock, L., Nelson, F., Altringer, L., & McKee, S. C. (2026). Estimating the Value of Ecosystem Services: A Machine Learning Approach for Missouri Wetlands. Ecosystem Services, 79, 101849. DOI

  3. Petach, L., Maddock, L., & Pena, A.-A. (2026). Housing Affordability and Homelessness: Evidence from Commuting Zones. Economic Development Quarterly.

Working Papers

  1. Maddock, L. Voucher Household Mobility and Rental Market Dynamics: Evidence from Source of Income Protection Laws. View PDF

    Abstract

    I study how source-of-income (SOI) protections affect where voucher households live and rents in the segments where vouchers transact. I assemble an annual panel of incorporated places with at least 65,000 residents from 2011–2019 and exploit staggered adoption across 57 jurisdictions (first effective dates 2013–2018). Identification follows Callaway & Sant'Anna with doubly-robust adjustment using 2012 baseline covariates and never-adopters as the counterfactual. Across four measures of voucher household geographic concentration, event-study profiles show flat pre-trends and no detectable post-adoption change. By contrast, lower-tier rents rise after adoption: the 25th-percentile contract rent increases by 0.049 log points (about 5%), the 75th percentile is unchanged, and the stock share affordable at less than 30% of HAMFI falls by 1.8 percentage points. Effects are larger under stronger laws and in tighter rental markets. Results are robust to a synthetic staggered DiD estimator and to allowing one year of anticipation. The pattern is consistent with pass-through of compliance and screening costs in voucher-relevant segments without re-sorting of voucher households within places.

  2. Maddock, L., McKee, S., & Altringer, L. Total Economic Valuation of Inland Wetlands in the United States: A Machine Learning Approach.

    Abstract

    Wetlands provide critical ecosystem services including flood regulation, water purification, and carbon storage, yet over one-third of global wetland area has been lost since 1970, while states lack consistent, spatially explicit valuations to inform conservation decisions. This paper develops the first national-scale framework for valuing inland wetland ecosystem services across all fifty states using machine-learning-based benefit transfer. We implement a three-stage methodology: stratified random sampling of National Wetlands Inventory polygons by EPA Level III ecoregion, wetland type, and size class; extraction of ecological and socioeconomic covariates from wetland footprints and 50-km buffers; and training supervised models on the Ecosystem Service Valuation Database to predict per-hectare values while avoiding double-counting. Our framework generates the first fifty-state estimates of wetland ecosystem value at consistent resolution, with improved predictive accuracy and better capture of local heterogeneity than traditional transfer methods.

  3. Gifford, T., & Maddock, L. The Mental and Physical Health Consequences of Legalized Sports Betting in the United States.

    Abstract

    Following the Supreme Court's 2018 decision striking down PASPA, mobile sports betting has rapidly expanded across the United States. This paper examines the causal effects of mobile sports betting legalization on physical and mental health outcomes using the staggered rollout of state legislation between 2018–2022. We employ the Callaway and Sant'Anna (2021) difference-in-differences estimator with data from the Behavioral Risk Factor Surveillance System and American Community Survey. Preliminary findings suggest associations between legalization and several health outcomes, including increased tobacco use and worsened mental health indicators, though effects attenuate in population-wide averages that include non-gamblers.

  4. Maddock, L. Counting the Uncounted: Machine Learning Approaches to Imputing Unsheltered Homelessness During the COVID-19 Pandemic.

Works in Progress

  • Maddock, L., & Cicero, V. Consequences of Controversial Referee Decisions: Evidence from the First and Second Tier Brazilian Football Leagues.
  • Petach, L., Maddock, L., & Pena, A.-A. Homelessness and Local Business Outcomes.

Technical Reports

  • McKee, S. C., Blanchard, M. D., Maddock, L., et al. (2026). Economic Valuation and Natural Resource Damage Assessment Methods at Myakka River State Park, Florida. U.S. Army Engineer Research and Development Center (ERDC). DOI
  • McKee, S., Blanchard, M., Miller, R., et al. Quantifying Wild Pig Damage Reduction Using a Before-After-Control-Impact Design at USACE Richard K. Yancey, Louisiana. U.S. Army ERDC. DOI
  • Maddock, L. The State of Homelessness in Colorado. CSU Regional Economic Development Institute (REDI). Report

Presentations

2025 — Southern Economic Association (Tampa, FL); North American Regional Science Council (Denver, CO); Western States Graduate Workshop (Fort Collins, CO); Wildlife Damage Management Conference (virtual) ; Eastern Economic Association (New York City, NY)

2024 — Southern Economic Association (Washington, DC); Eastern Economic Association (Boston, MA); Western States Graduate Workshop (Salt Lake City, UT)

2023 — Midwestern Economics Association (Cleveland, OH); Western Economics Association; Western States Graduate Workshop (Fort Collins, CO)—