{"id":"palomar/PALOMAR-2026-11-20-000012","url":"https://github.com/bocowgill/worker-compromise/tree/2c377b62b66f79ae9bbf127af1ec5ab0b2ea5468","json":"https://data.palomar-registry.org/entries/PALOMAR-2026-11-20-000012-v1.json","repo":"https://github.com/bocowgill/worker-compromise","tags":["econ.TH","math.PR","msc:91B68","msc:60C05"],"score":0,"title":"One extra worker rules out the square-root welfare compromise","trust":"high","domain":"econ.TH","source":"palomar","authors":["Bo Cowgill"],"reviews":0,"subject":"econ.TH","summary":"A clearinghouse observes a market with n firms and n+1 workers and can choose any stable assignment. Both sides draw independent rate-one exponential utilities for every firm-worker pair. Changing partners costs log(n)/(6 sqrt(n)). Write δ = 1/400000. As n grows, with probability tending to one, every stable assignment gives the matched workers a mean partner rank greater than n^(1/2+δ) and a mean utility less than (1/2−δ) log(n). Rank one is best. Consequently, even a clearinghouse that chooses separately for each objective cannot deliver mean partner ranks at the square-root benchmark on…","category":"research","comments":0,"disputed":false,"division":"econ","mentions":0,"verified":true,"discussed":0,"published":"2026-11-20","reactions":{},"formalized":true}