Finding, counting, or determining the existence of Nash equilibria, where players must play optimally given each others' actions, are known to be computational intractable problems. We ask whether weakening optimality to the requirement that each player merely avoid worst responses – arguably the weakest meaningful rationality criterion – yields tractable solution concepts. We show that it does not: any solution concept with this minimal guarantee is “as intractable” as pure Nash equilibrium. In general games, determining the existence of no-worst-response action profiles is NP-complete, finding one is NP-hard, and counting them is #P-complete. In potential games, where existence is guaranteed, the search problem is PLS-complete. Computational intractability therefore stems not only from the requirement of optimality, but also from the requirement of a minimal rationality guarantee for each player. Moreover, relaxing the latter requirement gives rise to a tractability trade-off between the strength of individual rationality guarantees and the fraction of players satisfying them.

Bidders in combinatorial auctions face significant challenges when describing their preferences to an auctioneer. Classical work on preference elicitation focuses on query-based techniques inspired from proper learning—often via proxies interfacing between bidders and an auction mechanism—to incrementally learn bidder preferences as needed to compute efficient allocations. Although such elicitation mechanisms enjoy theoretical query efficiency, the amount of communication required may still be too cognitively taxing in practice. We propose a family of efficient LLM-based proxy designs for eliciting preferences from bidders using natural language. Our proposed mechanism combines LLM pipelines and DNF-proper-learning techniques to quickly approximate preferences when communication is limited. To validate our approach, we create a testing sandbox for elicitation mechanisms that communicate in natural language. In our experiments, our most promising LLM proxy design reaches approximately efficient outcomes with five times fewer queries than classical proper learning based elicitation mechanisms.

We explore stability and fairness considerations in decentralized networked markets with bilateral contracts, building on the trading networks framework [Hatfield et al., 2013]. In our trading network game, we show that a well-defined subset of Nash equilibria can be supported as competitive equilibria. Considering an offer-based trading dynamic as well as a stochastic price clock market, we prove new convergence results to Nash equilibrium and competitive equilibrium, providing a rationale for stability properties in decentralized, dynamic trading networks. Turning to the tension between fairness and (core) stability, we prove several negative results: inessential agents always receive zero utility in any core outcome, and even essential agents can get zero utility in all core outcomes.

In markets with budget-constrained buyers, competitive equilibria need not be efficient in the utilitarian sense or maximize the seller's revenue. We consider a setting with multiple divisible goods. Competitive equilibrium outcomes, and only those, are constrained utilitarian efficient, a notion of utilitarian efficiency that respects buyers' demands and budgets. Our main contribution establishes that when buyers have linear valuations, competitive equilibrium prices are unique and revenue-optimal for a zero-cost seller.

This paper develops algorithms to solve strong-substitutes product-mix auctions: it finds competitive equilibrium prices and quantities for agents who use this auction’s bidding language to truthfully express their strong-substitutes preferences over an arbitrary number of goods, each of which is available in multiple discrete units. Our use of the bidding language and the information it provides contrasts with existing algorithms that rely on access to a valuation or demand oracle. We compute market-clearing prices using algorithms that apply existing submodular minimization methods. Allocating the supply among the bidders at these prices then requires solving a novel constrained matching problem. Our algorithm iteratively simplifies the allocation problem, perturbing bids and prices in a way that resolves tie-breaking choices created by bids that can be accepted on more than one good. We provide practical running time bounds on both price finding and allocation and illustrate experimentally that our allocation mechanism is practical. Funding: E. Baldwin and P. Klemperer were supported by the Economic and Social Research Council [Grant ES/L003058/1]. P. W. Goldberg and E. Lock were supported by a JP Morgan faculty fellowship during the work on the final version of the paper. Supplemental Material: The online companion is available at https://doi.org/10.1287/moor.2019.0248 .

This letter provides an overview of our recent work on COVID-19 testing mechanisms that appeared at EC'23. Large-scale testing is crucial in pandemics but resources are often prohibitively constrained. We study a scenario in which a population under lockdown utilizes a limited budget of tests to allow healthy individuals to resume in-person activities. Our work explores the optimal allocation of pooled tests in populations that are heterogeneous with respect to individual infection probabilities and utilities that materialize if included in a negative test (and being permitted to resume in-person activities). Non-overlapping allocations of tests, where no individual in the population is included in more than one pooled test, are both conceptually and logistically simpler to implement. We show that the welfare gain from overlapping testing over non-overlapping testing is bounded. Moreover, we design a heuristic mechanism for finding test allocations that is fast and empirically near-optimal. We also implement our mechanism in practice and provide experimental evidence on the benefits of utility-weighted pooled testing in a real-world setting. Our randomized trial at a higher education research institute in Mexico suggests that performance and mental health outcomes of participants under our testing mechanism are no worse than under the counterfactual of full access for individuals without testing.

We prove that the classic problem of finding a competitive equilibrium in an exchange economy with indivisible goods, money, and unit-demand agents is PPAD-complete. In this "housing market", agents have preferences over the house and amount of money they end up with, but can experience income effects. Our results contrast with the existence of polynomial-time algorithms for related problems: Top Trading Cycles for the "housing exchange" problem in which there are no transfers and the Hungarian algorithm for the "housing assignment" problem in which agents' utilities are linear in money. Along the way, we prove that the Rainbow-KKM problem, a total search problem based on a generalization by Gale of the Knaster-Kuratowski-Mazurkiewicz lemma, is PPAD-complete. Our reductions also imply bounds on the query complexity of finding competitive equilibrium.

Product-mix auctions are sealed-bid mechanisms for trading multiple divisible or indivisible units of multiple differentiated goods. They implement competitive-equilibrium allocations when these exist, based on the bids that participants make in a simple geometric language. All concave substitutes (respectively, strong-substitutes) valuations can be uniquely represented, and no other valuations can be represented, by bids in the corresponding version of this language. This provides new characterisations of ordinary substitutes, and of strong substitutes, when goods are indivisible. We discuss implementation of the auctions, and extensions and variants of the language, e.g., allowing for budget constraints.

Markets with multiple divisible goods have been studied widely from the perspective of revenue and welfare. In general, it is well known that envy-free revenue-maximal outcomes can result in lower welfare than competitive equilibrium outcomes. We study a market in which buyers have quasilinear utilities with linear substitutes valuations and budget constraints, and the seller must find prices and an envy-free allocation that maximise revenue or welfare. Our setup mirrors markets such as ad auctions and auctions for the exchange of financial assets. We prove that the unique competitive equilibrium prices are also envy-free revenue-maximal. This coincidence of maximal revenue and welfare is surprising and breaks down even when buyers have piecewise-linear valuations. We present a novel characterisation of the set of"feasible"prices at which demand does not exceed supply, show that this set has an elementwise minimal price vector, and demonstrate that these prices maximise revenue and welfare. The proof also implies an algorithm for finding this unique price vector.

This work sets out a testing and containment framework developed for reopening universities in Mexico following the lockdown due to COVID-19. We treat diagnostic testing as a resource allocation problem and develop a testing allocation mechanism and practical web application to assist educational institutions in making the most of limited testing resources. In addition to the technical results and tools, we also provide a reflection on our current experience of running a pilot of our framework within the Instituto Tecnológico y de Estudios Superiores de Monterrey (ITESM), a leading private university in Mexico, as well as on our broader experience bridging research with academic policy in the Mexican context.

1. Sovereign debt vulnerabilities are high and rising. Debt-to-GDP ratios in many countries stand at record highs—with the COVID-19 pandemic being latest factor sharply contributing to debt accumulation across the globe. While central bank actions in Advanced Economies have managed to keep financing costs low for countries borrowing at safe interest rates, this holds less true for Emerging/Developing Economies issuing riskier debt.

This article addresses the computational challenges of learning strong substitutes demand when given access to a demand (or valuation) oracle. Strong substitutes demand generalises the well-studied gross substitutes demand to a multi-unit setting. Recent work by Baldwin and Klemperer shows that any such demand can be expressed in a natural way as a finite list of weighted bid vectors. A simplified version of this bidding language has been used by the Bank of England. Assuming access to a demand oracle, we provide an algorithm that computes the unique list of weighted bid vectors corresponding to a bidder’s demand preferences. In the special case where their demand can be expressed using positive bids only, we have an efficient algorithm that learns this list in linear time. We also show super-polynomial lower bounds on the query complexity of computing the list of bids in the general case where bids may be positive and negative. Our algorithms constitute the first systematic approach for bidders to construct a bid list corresponding to non-trivial demand, allowing them to participate in “product-mix” auctions.

We propose a novel testing and containment strategy in order to contain the spread of SARS-CoV2 while permitting large parts of the population to resume social and economic activity. Our approach recognises the fact that testing capacities are severely constrained in many countries. In this setting, we show that finding the best way to utilise this limited number of tests during a pandemic can be formulated concisely as an allocation problem. Our problem formulation takes into account the heterogeneity of the population and uses pooled testing to identify and isolate individuals while prioritising key workers and individuals with a higher risk of spreading the disease. In order to demonstrate the efficacy of our testing and containment mechanism, we perform simulations using a network-based SIR model. Our simulations indicate that applying our mechanism on a population of $100,000$ individuals with only $16$ tests per day reduces the peak number of infected individuals by approximately $20\%$, when compared to the scenario where no intervention is implemented.

Consider a vertex colouring game played on a simple graph with $k$ permissible colours. Two players, a maker and a breaker, take turns to colour an uncoloured vertex such that adjacent vertices receive different colours. The game ends once the graph is fully coloured, in which case the maker wins, or the graph can no longer be fully coloured, in which case the breaker wins. In the game $g_B$, the breaker makes the first move. Our main focus is on the class of $g_B$-perfect graphs: graphs such that for every induced subgraph $H$, the game $g_B$ played on $H$ admits a winning strategy for the maker with only $\omega(H)$ colours, where $\omega(H)$ denotes the clique number of $H$. Complementing analogous results for other variations of the game, we characterise $g_B$-perfect graphs in two ways, by forbidden induced subgraphs and by explicit structural descriptions. We also present a clique module decomposition, which may be of independent interest, that allows us to efficiently recognise $g_B$-perfect graphs.