Alaskan Hunting License Lotteries are Flexible and Approximately Efficient

37 Pages Posted: 25 Apr 2019 Last revised: 10 Feb 2020

See all articles by Nick Arnosti

Nick Arnosti

Columbia Business School - Decisions, Risk, and Operations Division

Tim Randolph

Columbia University - Department of Computer Science

Date Written: April 11, 2019

Abstract

We analyze the k-ticket lottery, which is used to allocate hunting permits in the state of Alaska. Each participant is given k tickets to distribute among lotteries for different types of items. Participants who win multiple items receive their favorite, and new winners are drawn from the lotteries with unclaimed items.

When supply is scarce, equilibrium outcomes of the k-ticket lottery approximate a competitive equilibrium from equal incomes (CEEI), which is Pareto efficient. When supply is moderate, k-ticket lotteries exhibit two sources of inefficiency. First, some agents may benefit from trading probability shares. Second, outcomes may be ``wasteful": agents may receive nothing even if acceptable items remain unallocated. We bound both sources of inefficiency, and show that each is eliminated by a suitable choice of k: trades are never beneficial when k = 1, and waste is eliminated as k approaches infinity.

The wastefulness of the k-ticket lottery has some benefits: agents with strong preferences may prefer k-ticket lottery outcomes to those of any nonwasteful envy-free mechanism. These agents prefer small values of k, while agents with weak preferences prefer large values of k. Together, these results suggest that the k-ticket lottery performs well under most circumstances, and may be suitable for other settings where items are rationed.

Keywords: matching, market design, allocation without money, competitive equilibrium from equal incomes

JEL Classification: D45, D47, C78

Suggested Citation

Arnosti, Nick and Randolph, Tim, Alaskan Hunting License Lotteries are Flexible and Approximately Efficient (April 11, 2019). Available at SSRN: https://ssrn.com/abstract=3370605 or http://dx.doi.org/10.2139/ssrn.3370605

Nick Arnosti (Contact Author)

Columbia Business School - Decisions, Risk, and Operations Division ( email )

3022 Broadway
New York, NY 10027
United States

Tim Randolph

Columbia University - Department of Computer Science ( email )

New York, NY 10027
United States

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
115
Abstract Views
841
Rank
436,579
PlumX Metrics