Linchen Zhang

University of California, San Diego · Economics · 2026–27 job market

Fields

Theory, Behavioral & Experimental

Job market paper

Recognizing Dominance through Local Improvement

Participants in strategy-proof mechanisms often fail to play their dominant strategies. Most existing simplicity concepts attribute these failures to the cost of contingent reasoning and lower the cost by comparing the worst case of the dominant strategy with the best case of an alternative. I argue that a participant can instead compare a strategy with its neighbors, such as the bids one step above and below, and move to a better neighbor until none remains. I formalize this idea with a neighbor relation on each strategy set and call a mechanism local improvement solvable if every such sequence of moves ends at a dominant strategy. Strategy-proofness is the criterion under the complete neighbor relation. Under their natural neighbor relations, the criterion holds in the sealed-bid second-price auction, the multi-unit Vickrey auction, and, in rich environments, serial dictatorship, top trading cycles, and deferred acceptance. In extensive-form mechanisms without duplicate dominant strategies, the criterion under one-situation neighbors is strict action-proofness, which yields a one-shot deviation principle for dominance. This criterion and obvious strategy-proofness classify the pick-a-number game, chess, and Sudoku in opposite ways. Among local improvement solvable mechanisms that implement the same outcomes, the model predicts more dominant-strategy play when the local comparison is salient.