I am an M.A. Economics candidate and a Certified Residential Appraiser currently conducting a national-scale valuation of land occupied by U.S. golf courses. Rather than valuing them as going-concern businesses, my thesis focuses on estimating the "opportunity cost" of the land based on a Residential vs. Agricultural Highest & Best Use (HBU) counterfactual with regards to applying a hypothetical of a change of the zoning laws.
The Methodology & Data Strategy:
My first phase involved establishing a total land area (acreage) for all registered courses. I utilized a hybrid spatial model in R combining Tigris (Census) and OSM polygons, which successfully captured ~62% of the total known courses. For the remaining 38%, I utilized Multiple Imputation by Chained Equations (M.I.C.E.), applying my limited graduate-level econometrics to generate a statistically valid national estimation.
The Master Golf Course List:
To establish the universe for this study, I compiled a comprehensive list of U.S. golf courses using various public and online directories, including data sourced from the National Golf Foundation. The raw data required significant cleaning, including:
For the valuation phase, I am currently leveraging:
The "Hawaii Problem" & Zoning Distortions:
In my home market of Hawaii, I’ve observed significant tax distortions where commercial golf operations sit on P-2 Preservation zoning. This results in near-zero tax liability, which I argue violates the spirit of "preservation" while effectively subsidizing a private business. I am also investigating other policy "traps," such as the New Jersey "Cemetery" loophole, where land is misclassified to avoid development-level taxation.
I am looking for input from this community on:
I have not yet fully compiled my initial findings, but when I do have it ready I will post it on my academic website: manolakis.work. I would appreciate any "ground-truth" insights you can share from your respective markets.
The Methodology & Data Strategy:
My first phase involved establishing a total land area (acreage) for all registered courses. I utilized a hybrid spatial model in R combining Tigris (Census) and OSM polygons, which successfully captured ~62% of the total known courses. For the remaining 38%, I utilized Multiple Imputation by Chained Equations (M.I.C.E.), applying my limited graduate-level econometrics to generate a statistically valid national estimation.
The Master Golf Course List:
To establish the universe for this study, I compiled a comprehensive list of U.S. golf courses using various public and online directories, including data sourced from the National Golf Foundation. The raw data required significant cleaning, including:
- Parsing: Using regular expressions to break messy, single-string data into discrete columns for address, city, state, and course type.
- Manual Correction: Implementing a manual correction loop for records that failed automatic parsing to ensure data integrity.
- Geocoding: Ensuring each course was associated with specific latitude and longitude coordinates to allow for later spatial joins.
For the valuation phase, I am currently leveraging:
- Urban/Suburban: FHFA Residential Land Value data and House Price Index.
- Rural: USDA/NASS Agricultural land values.
The "Hawaii Problem" & Zoning Distortions:
In my home market of Hawaii, I’ve observed significant tax distortions where commercial golf operations sit on P-2 Preservation zoning. This results in near-zero tax liability, which I argue violates the spirit of "preservation" while effectively subsidizing a private business. I am also investigating other policy "traps," such as the New Jersey "Cemetery" loophole, where land is misclassified to avoid development-level taxation.
I am looking for input from this community on:
- Data Proxies: I am currently using FHFA land residuals for urban sites at scale. Are there better public proxies you’ve found reliable for land values across different MSAs?
- Regional Loopholes: Beyond P-2 or "Cemetery" designations, are there specific regional zoning or tax codes you've seen "trap" golf course land in its current use despite a significantly higher alternative value?
- Acreage Estimation: Have you found Tigris or OpenStreetMap polygons to be a reliable source for large-scale land area estimation in your own work?
I have not yet fully compiled my initial findings, but when I do have it ready I will post it on my academic website: manolakis.work. I would appreciate any "ground-truth" insights you can share from your respective markets.
