House bill would require Pentagon to try commercial algorithms to calculate BAH

Rep. Don Bacon, R-Neb., introduced a bill in the House of Representatives on Aug. 18 that would require the Defense Department to launch a new pilot program to use industry-built machine learning and artificial intelligence algorithms to calculate the monthly rates of basic allowance for housing in certain places. 

Broadly, DOD sets military housing areas (MHAs) — or batches of zip codes around military installations — and assigns basic allowance for housing (BAH) rates for each of those areas. The department’s BAH program is a fundamental component of its pay packages for service members, and it’s ultimately designed to compensate for local median rental costs and average utilities costs for civilians with comparable incomes to each military pay grade in their specific duty locations via monthly payments. 

Officials look to private sector rental housing costs and associated data to compute BAH rates, which are updated annually.

Bacon’s new bill (the text of which hasn’t been released publicly but was shared with DefenseScoop by the lawmaker’s team this week) is titled the “Basic Allowance for Housing Calculation Improvement Act of 2023.”

The three-page legislation would mandate the secretary of defense — no later than Sept. 30, 2024 — to “seek to enter into an agreement with a covered entity pursuant to which the covered entity shall calculate, using industry standard machine learning and artificial intelligence algorithms, the monthly rates of BAH for not fewer than 15 MHAs,” through the proposed pilot program. 

A “covered entity” is defined in the bill as a “nationally recognized entity in the field of single-family housing that has data on local rental rates in real estate markets across” the U.S. 

DOD would have to submit a report evaluating that work to the House and Senate Armed Services Committees (HASC and SASC) within two years of the act’s passage. Beyond that time frame, the bill language does not include further details or directions for the pilot.

The legislation was referred to HASC, where Bacon serves as a member, following its introduction last week.

If the bill does pass, this wouldn’t be the first time AI would be examined by Pentagon officials as a tool to augment the BAH calculation process. A cohort of Air and Space Forces members completed a capstone project that assessed AI and machine learning models to optimize BAH determination procedures, last year, through the Department of the Air Force-Massachusetts Institute of Technology Artificial Intelligence Accelerator’s Phantom Fellowship Program. 

Members of Bacon’s team did not provide further information regarding what prompted the lawmaker to introduce the bill now. Congress is in recess until after Labor Day.