US Army let a matching algorithm assign officers to branches, not a committee

Every year thousands of new officers are assigned to Army branches — infantry, armor, cyber, and the rest — through a centralized matching process that decides how the Army uses its talent for decades.

The West Point mechanism used for the Class of 2020 made the Army's own policy aims harder to hit: cadets could game their preferences, and the result misaligned talent with the branches that the service most wanted to keep and staff.

Economists modelled it as a market design problem and found the Army's objectives uniquely pin down both which cadet gets which branch and what priority policy to use — a uniqueness result. They redesigned the mechanism and the priority policy together instead of patching one.

The repaired design, built on a cumulative-offer mechanism, was adopted for new classes at both West Point and ROTC, which assign around 70% of newly commissioned officers.

Fix the priority rule and the allocation rule together

An allocation rule without a coherent priority rule invites gaming and misalignment.

The Army's stated goals uniquely determine both the mechanism and the priorities, so you cannot fix one without the other.

A cumulative-offer mechanism reduces the payoff for cadets to misreport their true preferences.

Putting designers inside the institution made the redesign credible enough to actually be adopted.

Do not fix an allocation by tweaking the ranking; the ranking and the rule are one design.

The redesigned mechanism was adopted at both West Point and ROTC, and the project became a standard example of 'minimalist market design' — using a light academic touch to reform an existing institution rather than replacing it.