"Multidimensional Monotonicity and Economic Applications"

We characterize the extreme points of multidimensional monotone functions from [0,1]n to [0,1], as well as the extreme points of the set of one-dimensional marginals of these functions. These characterizations lead to new results for various mechanism design and information design problems, including public good provision with interdependent values; interim efficient bilateral trade mechanisms; mechanism (anti) equivalence; asymmetric reduced form auctions; and optimal private private information structure.

"Digital Ecosystems and Data Regulation"

This paper provides a framework in which a multiproduct ecosystem competes
with many single-product firms in both price and innovation. The ecosystem is
able to use data collected on one product to improve the quality of its other
products. The framework is used to evaluate the impact of three data policies on
pricing, innovation, and consumer welfare: restricting cross-product data usage
within the ecosystem, requiring the ecosystem to share its data with smaller
competitors, and establishing a data cooperative among small firms to facilitate
data sharing with each other.

"Monotonicity and Robust Implementation Under Forward-Inducing Reasoning"

In sequential games, the set of paths consistent with rationality and forward-induction reasoning may change non-monotonically when adding transparent restrictions on players’ beliefs. Yet, we prove that—in an incomplete-information environment—predictions become sharper when the restrictions only concern initial beliefs about types. Thus, strong rationalizability for games with payoff uncertainty characterizes the path predictions of forward-induction reasoning across all possible restrictions on players’ hierarchies of exogenous beliefs. With this, we can solve an open problem: The implementation of social choice functions through sequential mechanisms under forward-induction reasoning—which considerably expands the realm of implementable functions compared with simultaneous mechanisms (Müller, J. Econ. Theory 2016)—is indeed robust in the sense of Bergemann and Morris (Theor. Econ. 2009).

"Collusion without Patience (with Jacopo Bizzotto)"

Tacit collusion is typically associated with repeated interaction between patient firms. We show that it can also arise in a one-shot duopoly. In our model, capacity choices and outsourcing contracts with minimum order quantities create gaps in feasible outputs, allowing firms to commit not to produce intermediate quantities. This leads to three possible types of equilibria: collusive, partially competitive, and leadership equilibria. We then generalize the model to a broader setting with two-sided limited commitment and show that it yields the same set of equilibria.

“Social Learning with Endogenous Timing”

We show that social learning games with endogenously chosen timing of actions have qualitatively different symmetric equilibria reflecting different degrees of herding in the long-run outcome. With endogenous timing, we can distinguish between efficiency losses resulting from uninformed actions and losses from waiting. For large games, we show that the game has a fully informative equilibrium, where almost all the agents take the correct action, but this involves lengthy stochastic delays. We also identify Pareto-superior outcomes that combine fast partial revelation with a unanimous, but possibly ex post incorrect decision taken by almost all agents. We discuss the differences in the sets of equilibrium outcomes under continuous and discrete time modeling approaches.

"Keeping in the Dark with Hard Evidence"

We present a dynamic learning setting in which the periodic data observed
by the decision-maker is mediated by an agent. We study when, and to what
extent, this mediation can distort the decision-maker’s long-run learning, even
though the agent’s reports are restricted to consist of verifiable hard evidence
and must adhere to certain standards. We introduce the manipulation-proof
law of large numbers – that delivers a sharp dichotomy: when it holds, the
decision-maker’s learning is guaranteed in the long-run; when it fails, the scope
for manipulation is essentially unrestricted.

"Ranking Statistical Experiments via the Linear Convex Order and the Lorenz Zonoid: Economic Applications"

This paper introduces a novel ranking of statistical experiments, the Linear-
Blackwell (LB) order, equivalently characterized by (i) more dispersed posteriors
and likelihood ratios in the sense of the linear convex order, (ii) a larger Lorenz
zonoid—the set of statewise expectation profiles, and (iii) greater variability of the
posterior mean. We apply the LB order to compare experiments in binary-action
decision problems and in problems with quasi-concave payoffs, as analyzed by
Kolotilin, Corrao, and Wolitzky (2025). Furthermore, the LB order enables the
comparison of experiments in moral hazard problems, complementing the findings
in Holmström (1979) and Kim (1995). Finally, the LB order applies to the
comparison of experiments generating ex post signals in screening problems.

"Organizational Design in the Knowledge Economy"

We study a principal-agent model of spatial learning. The principal must make decisions across a continuum of problems, but lacks complete knowledge of how each problem maps to its solution. This mapping is represented by the realized path of a Markov process. The principal, facing a fixed number of agents, assigns one problem to each agent, who can privately exert effort to solve the assigned problem. We characterize the optimal assignment of problems to agents and the compensation scheme that robustly implements effort by all agents. The model highlights a trade-off that shapes organizational boundaries in a knowledge economy.

"Correlation Games"

We introduce a model of flexible information acquisition in two-person games. We focus on the incentives to acquire information that is positively or negatively correlated with that of the other player. The (sign and extent of) correlation can arbitrarily differ state-by-state. In states where a player views actions as complements, she seeks information that is most-positively correlated with the other player’s. In states where a player views actions as substitutes, she seeks information that is most-negatively correlated. These observations help characterize pure-strategy equilibria. In any game, however, there is a mixed-strategy equilibrium where the players’ information is conditionally independent.

 "The cost of capital and misallocation in the United States" (with Julian Kozlowski and Jeremy Majerovitz)

We develop a method to estimate the cost of capital using credit registry microdata, and apply it to study capital allocation efficiency in the United States. Our measure incorporates the contractual interest rate, expected default probability, recovery rate, and expectations of future rates. We estimate three distinct rates: (i) the lender’s discount rate, (ii) the firm’s cost of capital, and (iii) the social cost of capital. We derive a sufficient statistic for misallocation based on the first and second moments of the social cost of capital. Dispersion in this rate captures both heterogeneity in lender discounting and the presence of financial frictions. Normal times feature modest amounts of misallocation, corresponding to an output loss of 0.9%, but this increased to 1.8% during the 2020-21 period.

"Informed communication equilibrium"

We consider a privately informed sender selecting a mediated communication device to influence players' actions. In contrast to standard information design, there is no ex-ante commitment, and the communication device relies on information elicited from the informed parties. We define an informed communication equilibrium (ICE) as a perfect Bayesian equilibrium outcome of the informed mediated communication game. We show that the set of ICE is the subset of communication equilibria (CE) that yield sender payoff vectors bounded below by an equilibrium payoff vector of the silent game, under some consistent interim beliefs.  The principal's ex-ante optimal CE may not be an ICE. In sender-receiver games, the set of CE payoff vectors lies in the convex hull of the set of payoff vectors above silence. Hence, if the latter set is convex, as is the case when the sender has transparent motives, the sets of CE and ICE coincide.