Research            Lecture notes            Conference            Mini-courses

3rd Southeast Theory Festival
19–20 April 2024
Lecture Theatre B05, Chadwick Building, UCL
Organised by Deniz Kattwinkel (UCL) and Ludvig Sinander (Oxford)

This conference brings together a number of junior economic theorists in southeastern England.


The juniors:
Thomas Brzustowski (Essex)
Henrique Castro-Pires (Surrey)
Bruno Furtado (Royal Holloway)
Alkis Georgiadis-Harris (Warwick)
Duarte Gonçalves (UCL)
Deniz Kattwinkel (UCL)
Jan Knoepfle (Queen Mary)
Sophie Kreutzkamp (Oxford)
Krittanai Laohakunakorn (Surrey)
Paula Onuchic (Oxford)
Evgenii Safonov (Queen Mary)
Christopher Sandmann (LSE)
Ludvig Sinander (Oxford)
Yu Fu Wong (Warwick)


Schedule:

Friday 19 April 2024 Saturday 20 April 2024
09:30–11:00 Evgenii, Deniz
10:30–11:45     arrival 11:00–11:45     break
11:45–12:30 Yu Fu 11:45–12:30 Thomas
12:30–14:00 lunch 12:30–14:00 lunch
14:15–15:45 Sophie, Alkis         14:15– social / departure
15:45–16:30 break
16:30–18:00 Bruno, Krittanai
18:00–19:00 break
19:00–21:00 dinner


Talks:

Thomas Brzustowski
‘Induced higher-order uncertainty in mechanism design with limited commitment’ (with Alkis Georgiadis-Harris and Balázs Szentes)
AbstractWe highlight the role played by mechanisms that endogenously induce higher-order uncertainty (uncertainty about beliefs about beliefs…) in dynamic contracting problems characterized by limited commitment. We consider environments featuring a single agent, with a privately known and persistent payoff-relevant type, and a principal whose prior belief about the agent's type is common knowledge. The principal is committed to the terms of her chosen mechanism within each period, but not across periods. For any order k, we construct one such contracting problem such that there exists a dynamic allocation that cannot be implemented unless the principal uses a mechanism inducing uncertainty of order at least k. This finding challenges established “revelation principles” à la Bester and Strausz (2001, 2007) or Doval and Skreta (2022).

Bruno Furtado
‘Recovering garbled experiments (or Testable identification of finite mixture models)’
AbstractIn information economics, the analyst is often interested in uncovering the information structure available to an agent but can only observe a garbling thereof. This is the case, for example, in costly information acquisition models, where the signal-contingent action policy by the agent constitutes a garbling of their chosen information structure. This paper establishes sufficient conditions that allow a more informative experiment to be recovered from its garbling. Notably, these identification conditions focus solely on properties of the garbled experiment, freeing them from any constraints on the primitive experiment or the garbling matrix. This feature enables verification without direct knowledge or restrictive assumptions about unobservable model features. Building on this, I introduce a statistical test to assess whether these conditions are met and derive its asymptotic distribution. The test employs an extremum estimator with a strictly concave objective function, which makes simulating its asymptotic distribution computationally tractable. Additionally, a straightforward transformation of the same extremum estimator yields a consistent estimator of the more informative experiment and garbling matrix.

Alkis Georgiadis-Harris
‘Preparing to act’
AbstractThis paper develops a dynamic model of information acquisition in which the decision maker lacks control over the timing of her action. Dynamic experiments are constrained only in the quantity of information they can generate over any time interval. The optimal experiment concentrates resources in seeking a single piece of breakthrough evidence and obeys two key cost efficiency principles: breakthroughs keep shadow-prices constant and are in the direction where they are increasing. We further explore the dynamics of learning in two extreme environments: pure time pressure where there is a deterministic deadline; and constant timing-risk where there is a constant probability of acting in each instant. The presence of timing-risk forces the decision maker to generate contrarian breakthroughs which can significantly limit her learning, relative to the case where the decision time is known.

Deniz Kattwinkel
‘Robust robustness’ (with Ian Ball)
AbstractTo find robust mechanisms, a standard approach is to seek maxmin optimality: the designer evaluates each alternative according to its worst-case expected utility over all priors in a chosen ambiguity set. We show that such mechanisms may perform poorly when the environment is slightly perturbed. We call an ambiguity set robust if each payoff guarantee over the ambiguity set extends continuously to priors that are sufficiently near the ambiguity set (in the weak topology). We show that continuous moment sets and Wasserstein balls are robust. However, support sets and total variation balls are not. Finally, we provide a behavioral characterization: an ambiguity set is robust if and only if the induced preferences satisfy a new continuity axiom that reflects the topology on the state space.

Sophie Kreutzkamp
‘Brevity’ (with Péter Eső and Dezső Szalay)
AbstractThis project studies incentives for private learning in a game of strategic information transmission. We show that information that is sufficiently costly to acquire can only be transmitted coarsely. We provide comparative statics results showing that a more nuanced categorization provides weaker incentives for information acquisition than a coarser one. Finally, we consider two welfare-improving extensions: information design and delegation.

Krittanai Laohakunakorn
‘Monopoly pricing with optimal information’ (with Guilherme Carmona)
AbstractWe analyze a monopoly pricing model where information about the buyer's valuation is endogenous. Before the seller sets a price, both the buyer and seller receive private signals that may be informative about the buyer's valuation. The joint distribution of these signals, as a function of the valuation, is optimally chosen by the players. In general, players have conflicting incentives over the provision of information. As a modelling device, we assume that an aggregation function determines the information structure from the choices of the players. We characterize the payoffs and prices that can arise in equilibrium for a natural class of aggregation functions. When both players are initially uninformed, equilibrium prices do not depend on the buyer's valuation. In contrast, if the buyer initially knows her own valuation, the price in every equilibrium is always equal to either the buyer's valuation or the uniform monopoly price.

Evgenii Safonov
‘Socially efficient approval mechanisms with signaling costs’ (with Inés Moreno de Barreda)
AbstractAn agent with a privately known continuous type applies for approval. While the agent always prefers approval over rejection, approving an agent with low type has social costs. The agent sends a report about her type that she can inflate by engaging in signaling costs that have the single-crossing property. We study approval mechanisms without transfers that maximize a social welfare function that takes into consideration both the approval decision and the signaling costs of the agent. We show that threshold approval rules, which are widespread in society, are never socially optimal. By introducing some randomness in the rule, we can reduce the signaling costs without substantially changing the screening of the rule. If we further assume that the marginal cost is strictly log-supermodular, an assumption satisfied by the quadratic-loss function, we show that the optimal approval mechanism induces an approval probability that is continuous in the agent's type. We provide necessary first-order conditions for the optimal rule, and illustrate them with an analysis of the case of quadratic-loss function and uniform distribution of the agents' types.

Yu Fu Wong
‘Dynamic monitoring design’
AbstractThis paper introduces flexible endogenous monitoring in dynamic moral hazard. A principal can commit to not only an employment plan but also the monitoring technology to incentivize dynamic effort from an agent. Optimal monitoring follows a Poisson process that produces rare informative signals, and the optimal employment plan features increasing entrenchment. To incentivize persistent effort, the Poisson monitoring takes the form of “bad news” that leads to immediate termination. Monitoring is non-stationary: the bad news becomes more precise and less frequent. When persistent effort is not required, the optimal incentive scheme features a trial period of non-stationary monitoring, and a combination of Poisson bad news that leads to termination and Poisson good news that leads to tenure.


Photograph from the conference dinner
Photograph after the conference (alt)

Thanks to UCL Economics for funding and hosting the conference.



2nd Southeast Theory Festival
21–22 April 2023
Large Lecture Room, Nuffield College, Oxford
Organised by Deniz Kattwinkel (UCL) and Ludvig Sinander (Oxford)

This conference brings together a number of junior economic theorists in southeastern England.


Invited participants:
Doruk Cetemen (City)
Duarte Gonçalves (UCL)
Deniz Kattwinkel (UCL)
Jan Knoepfle (Queen Mary)
Aditya Kuvalekar (Essex)
Paula Onuchic (Oxford)
Evgenii Safonov (Queen Mary)
Christopher Sandmann (LSE)
Ludvig Sinander (Oxford)


Local participants:
Dağhan Carlos Akkar,  Alae Baha,  Elizabeth Baldwin,  Péter Eső,  Carlos Gonzalez,  Ian Jewitt,  Hao Jiang,  Edwin Lock,  Stefania Merone,  Inés Moreno de Barreda,  Alex Teytelboym


Schedule:

Friday 21 April 2023 Saturday 22 April 2023
08:30–09:30 breakfast
09:30–11:00 Duarte, Paula
10:30–11:45     arrival 11:00–11:45     break
11:45–12:30 Christopher 11:45–12:30 Ludvig
12:30–14:00 lunch 12:30–14:00 lunch
14:15–15:45 Evgenii, Doruk         14:15– social / departure
15:45–16:30 break
16:30–18:00 Jan, Aditya
18:00–19:00 break
19:00–21:00 dinner


Talks:

Doruk Cetemen
‘Optimal project management’ (with Alessandro Bonatti and Juuso Toikka)

Duarte Gonçalves
‘Speed, accuracy, and complexity’ (with Nick Netzer)
AbstractFast choices are in some cases associated with greater choice accuracy, in others with less, a fact that has eluded existing theoretical models. We revisit a canonical model of optimal sequential sampling and find that across problems response time is quasiconcave in choice accuracy. In simple problems, decisions are fast and accurate, but while accuracy declines in problem complexity, expected response time is non-monotone. We show that this observation can reconcile conflicting existing evidence on the relationship between speed and choice accuracy. Although our results illustrate the limitations of using expected response time alone to identify problem complexity, we highlight how choice responses to incentive manipulation can be used to reveal problem complexity.

Jan Knoepfle
‘Should the timing of inspections be predictable?’ (with Ian Ball)
AbstractA principal hires an agent to work on a long-term project that culminates in a breakthrough or a breakdown. At each time, the agent privately chooses to work or shirk. Working increases the arrival rate of breakthroughs and decreases the arrival rate of breakdowns. To motivate the agent to work, the principal conducts costly inspections. She fires the agent if shirking is detected. We characterize the principal’s optimal inspection policy. Periodic inspections are optimal if work primarily speeds up breakthroughs. Random inspections are optimal if work primarily delays breakdowns. Crucially, the agent’s actions determine his risk-attitude over the timing of punishments.

Aditya Kuvalekar
‘Interdependence of information in games’ (with Deepal Basak and Joyee Deb)
AbstractWe study a class of N-person binary action games with incomplete information. Our goal is to understand how changes in the interdependence of information of agents affects the equilibrium outcomes. To this end, we propose some notions of interdependence of information for finite support distributions. The main results of the paper demonstrate equivalence of these stochastic orders with enlargement of equilibrium sets in a class of binary action games with either pure complementarities or pure substitutabilities. We further propose analogous orders for continuous random variables and show that in a class of binary action games, improvements in interdependence of information implies increases in the maximal coordination equilibria.

Paula Onuchic
‘Disclosure and incentives in teams’ (with João Ramos)
AbstractWe consider a team-production environment where all participants are motivated by career concerns, and where a team’s joint productive outcome may have different reputational implications for different team members. In this context, we characterize equilibrium disclosure of team-outcomes when team-disclosure choices aggregate individual decisions through some deliberation protocol. In contrast with individual disclosure problems, we show that equilibria often involve partial disclosure. Furthermore, we study the effort-incentive properties of equilibrium disclosure strategies implied by different deliberation protocols; and show that the partial disclosure of team outcomes may improve individuals’ incentives to contribute to the team. Finally, we study the design of deliberation protocols, and characterize productive environments where effort-incentives are maximized by unilateral decision protocols or more consensual deliberation procedures.

Evgenii Safonov
‘Slow and easy: A theory of browsing’
AbstractAn agent needs to choose the best alternative drawn randomly with replacement from a menu of unknown composition. The agent is boundedly rational and employs an automaton decision rule: she has finitely many memory states, and, in each, she can inquire about some attribute of the currently drawn alternative and transition (possibly stochastically) either to another state or to a decision. Defining the complexity of a decision rule by the number of transitions, I study the minimal complexity of a decision rule that allows the agent to choose the best alternative from any menu with probability arbitrarily close to one. Agents in my model differ in their languages—collections of binary attributes used to describe alternatives. My first result shows that the tight lower bound on complexity among all languages is 3·ceil(log2(m)), where m is the number of alternatives valued distinctly. My second result provides a linear upper bound. Finally, I call adaptive a language that facilitates additive utility representation with the smallest number of attributes. My third result shows that an adaptive language always admits the least complex decision rule that solves the choice problem. When (3/4)·2n < m ≤ 2n for a natural n, a language admits the least complex decision rule if and only if it is adaptive.

Christopher Sandmann
‘Market structure and adverse selection’ (with Dakang Huang)
AbstractThis paper adopts a unified perspective on multi-contracting in competitive markets plagued by adverse selection. We subsume the two polar cases of exclusive and nonexclusive competition by introducing the concept of a market structure, i.e., a trading rule that specifies the subset of sellers with whom buyers can jointly trade. The existing literature shows that the market structure matters greatly in shaping competitive allocations, allowing for either separating allocations (as shown by Rothschild–Stiglitz) or layered pooling (Jaynes–Hellwig–Glosten) allocations. We prove the existence of intermediate ‘Pooling + Separating’ equilibria that allow for simultaneous pooling and low-risk buyer separation. Crucially, those allocations alleviate at the same time the concern of excessive rationing under separation of and cross-subsidies paid by low-risk buyers. They oftentimes Pareto dominate the Rothschild–Stiglitz separating allocation. Our analysis singles out the ‘1+1’ market structure where sellers are separated into two subgroups so that buyers can trade with at most one seller from each subgroup. Any ‘Pooling + Separating’ allocation is an equilibrium here. Finally, we prove that ‘Pooling + Separating’ allocations satisfy a notion of stability that we call serendipitous-aftermarket-proofness.

Ludvig Sinander
‘Optimism and overconfidence’
AbstractI revisit the standard moral-hazard model, in which an agent’s preference over contracts is rooted in costly effort choice. I characterise the behavioural content of the model in terms of empirically testable axioms, and show that the model’s parameters are identified. I propose general behavioural definitions of relative (over)confidence and optimism, and characterise these in terms of the parameters of the moral-hazard model. My formal results are rooted in a simple but powerful insight: that the moral-hazard model is closely related to the well-known ‘variational’ model of choice under uncertainty.


Map (click to enlarge):



Photograph from the conference dinner

Thanks to Nuffield College for funding and hosting the conference.



1st Southeast Theory Festival
1–2 April 2022
Clay Room, Nuffield College, Oxford
Organised by Deniz Kattwinkel (UCL) and Ludvig Sinander (Oxford)

This conference brings together a number of junior economic theorists in southeastern England.


Speakers and talks:

Duarte Gonçalves (UCL)
The dynamics of social instability (with César Barilla)
AbstractWe study a model in which two groups with opposing interests try to alter a status quo through instability-generating actions. We show that even if these actions are costly and do not lead to any short-term average utility gains, they can be used to secure longer-term durable changes. In equilibrium, the level of instability generated by a group decreases when the status quo favors it more. Equilibrium always exhibits stable states in which the status quo persists indefinitely. There is long-run path-dependency and inequity: although the process of social instability eventually leads to a stable state, it will typically select the least favorable one for the initially disadvantaged group.

Deniz Kattwinkel (UCL)
Optimal decision mechanisms for juries: Acquitting the guilty (with Alexander Winter)
AbstractA group of privately informed agents must choose between two alternatives. How should a principal design the decision rule if agents are known to be biased in favor of one of the options? We address this question within the classical Condorcet jury setting. Applications include the optimal decision mechanisms for boards of directors, political committees, and trial juries. The optimal mechanism is a voting mechanism. In the terminology of the court example, when jurors are more eager to convict than the lawmaker (principal), then the optimal mechanism convicts the defendant if and only if neither too many nor to few jurors vote to convict. This kind of mechanism accords with a judicial procedure from ancient Jewish law.

Aditya Kuvalekar (Essex)
#Protests (with Deepal Basak and Joyee Deb)

Paula Onuchic (Oxford)
Information acquisition and disclosure by advisors with hidden motives
AbstractA sender acquires a signal about an object's quality and commits to a rule to disclose its realizations to a receiver, who then chooses to buy the object or to keep an outside option of privately known value. Optimal disclosure rules typically conceal negative signal realizations when the object's sale is relatively profitable to the sender and positive signal realizations when the sale is relatively less profitable. Using such disclosure rules, the advisor is able to steer sales from lower- to higher-profitability objects. I show that, despite this strategic concealment of some signal realizations, the receiver may prefer to be informed by a non-transparent sender, because the sender's hidden motives produce an additional incentive to invest in acquiring a precise signal of the object's quality. I use my model to evaluate policies commonly proposed in the context of financial advisors, such as mandatory disclosure of commissions, or commission caps.

João Ramos (Queen Mary)
Optimal political career dynamics (with Avidit Acharya and Elliot Lipnowski)
AbstractWe examine politicians' career dynamics generated by political accountability, characterizing voter-optimal equilibrium play under repeated moral hazard. When moral hazard binds, equilibrium play is non-stationary: Re-election prospects improve with good performance and deteriorate with bad. First-term politicians are among the most electorally vulnerable and the hardest-working, and effort and electoral vulnerability both tend to decline with tenure. These dynamics enable a detailed analysis of limited voter commitment, voluntary retirement from politics, and adverse selection with politicians' ability and effort being complementary. Our analysis highlights how a politician's career is shaped by voters' evolving “goodwill” toward her.

Christopher Sandmann (LSE)
Oligopolistic nonlinear pricing: A random search model
Abstract(early and preliminary) This paper proposes a new framework for oligopolistic nonlinear pricing: I incorporate the standard monopolistic screening problem in a dynamic random search model with explicit search cost. Firms' market power derives from search frictions and horizontal differentiation. I show that under private values firms' quality provision attains the first-best level, i.e., were one to maximize buyer's utility minus cost. This confirms and generalizes findings in the celebrated paper due to Rochet and Stole (2002) to a dynamic environment and a broader class of preferences and distributions. In contrast, first-best efficiency fails to obtain when there are common values. As a novel result I report that under adverse selection quality provision is less than the first-best whereas quality provision exceeds the first-best under advantageous selection for all but the buyers of least and greatest valuation. This both confirms and disproves the intuition derived from Rothschild and Stiglitz (1976) where there are two types only: There are downward distortions, but they are largest for intermediate types, not at the bottom. Also, unlike in Rotschild–Stiglitz, profit is not constant (i.e. zero) across types. Instead, under adverse (advantageous) selection firms make less (more) profit on high valuation buyers.

Ludvig Sinander (Oxford)
The converse envelope theorem
AbstractI prove an envelope theorem with a converse: the envelope formula is equivalent to a first-order condition. Like Milgrom and Segal's (2002) envelope theorem, my result requires no structure on the choice set. I use the converse envelope theorem to extend to general outcomes and preferences the canonical result in mechanism design that any increasing allocation is implementable, and apply this to selling information.

Akhil Vohra (Cambridge)
Bayesian persuasion: A reduced form approach (with Juuso Toikka and Rakesh Vohra)
AbstractWe illustrate the usefulness of reduced form representations of Bayesian persuasion problems with two applications. In one, the reduced form representation yields a simple algorithm to determine the support of the optimal solution. If n is the number of actions, the worst-case complexity of the algorithm is O(n^3 log n). By comparison, the worst-case complexity of solving the obedience formulation as a linear program is O(n^6). In the second application, the reduced form approach leads to a simple greedy algorithm to determine the maximum value a sender can achieve in any cheap talk equilibrium.


Local participants:
Dağhan Carlos Akkar,  Jean-Paul Carvalho,  Michael Eldar,  Péter Eső,  Barton Lee,  Meg Meyer,  Inés Moreno de Barreda,  Dan Quigley,  Aidan Smith



Schedule:

Friday 1 April 2022 Saturday 2 April 2022
08:30–09:30 breakfast
09:30–11:00 Christopher, Akhil
10:30–11:45     arrival 11:00–11:45     break
11:45–12:30 Aditya 11:45–12:30 Ludvig
12:30–14:00 lunch 12:30–14:00 lunch
14:15–15:45 João, Duarte         14:15–open social / departure
15:45–16:30 break
16:30–18:00 Paula, Deniz
18:00–19:00 break
19:00–21:00 dinner        


Map (click to enlarge):



Thanks to Nuffield College for funding and hosting the conference.