Daily Summary for 2026-11-11

daily

Transcript

Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.

Ruth: It's the eleventh of November, twenty twenty-six, and this is the day's research.

Warren: 29 new papers came out today.

Ruth: I'm Ruth.

Warren: And I'm Warren, here with Iris.

Ruth: We'll take the day in one pass, then pull out the papers we're staying with.

Warren: Start us off with Delegation without Priors.

The day's research: Ruth: We examine how principals make decisions when they only have private information about an agent.

Warren: That asymmetry is crucial for designing better contracts and systems because understanding it helps us.

Iris: The core idea is a rule to delegate authority optimally, minimizing regret even when circumstances are unknown.

Ruth: This optimal rule balances sticking to a standard default action versus deviating from it.

Warren: How much freedom the agent gets when they choose to deviate is another key element of that balance.

Iris: Hidden commitment power shows enforcement ability can be less than we think if information about commitment is hidden.

Ruth: This suggests only measures improving the worst-case scenario are truly valuable in this information gap.

Warren: We also have work testing if linear specifications accurately describe relational data, especially in dyadic regression models.

Iris: Certain bootstrap methods consistently approximate the limiting law even when parameters are estimated or the function is curved.

Ruth: This contrasts with foundational work on identifying discount factors and payoff functions in dynamic choice models.

Warren: That research uses assumptions about per-period payoffs to find those unknown economic parameters, which is different from linear specification testing.

Iris: There is also methodological work on where to collect data efficiently using Delaunay matching to determine the best next step for an intervention.

Ruth: This suggests geometric methods can significantly reduce worst-case bias when learning about a population's response to an intervention.

Warren: The work on interacting principal-agent games is significant because it tackles incentive compatibility fragility when multiple principals design systems simultaneously.

Iris: Exact equilibrium can be recovered even when allowing small expected incentive regret, as shown in Interacting Principal-Agent Games.

Ruth: This means in complex economic settings, we do not need perfect coordination for the system to work well if regret remains small.

Warren: That is supported by analysis of a one-member team-contest specialization within that general model.

Iris: This establishes existence of an exact Bayesian-Nash principals' equilibrium under conditions where feasibility varies continuously in Interacting Principal-Agent Games.

Ruth: Delegation without Priors states a principal finds a rule minimizing regret using private information from an agent.

Warren: Specification Testing for Dyadic Regression Models develops omnibus tests to check if linear specifications accurately describe conditional means in undirected dyadic data.

Iris: Hidden Commitment Power Is Powerless A principal's ability to renege on a contract is often limited by their private information about their commitment power.

Ruth: Robust Contracts with Exploration involves the principal contracting sequentially, allowing exploration before final commitment.

Warren: Double Machine Learning for Inequality Decompositions uses double machine learning to decompose outcome gaps using a reference outcome that depends on the choice of estimator.

Iris: A Note on Qualified Majority Voting Rules characterizes every qualified majority voting rule by examining properties like anonymity, responsiveness, and q-neutrality.

Ruth: Semiparametric Identification of the Discount Factor and Payoff Function in Dynamic Discrete Choice Models identifies those parameters using restrictions from economic theory.

Warren: Learning Where to Look Delaunay Matching for Discovery and Data Collection shows using a Delaunay triangulation minimizes worst-case bias when combining observational data with small intervention samples.

Iris: Interacting Principal-Agent Games In many economic environments, multiple principals design incentive systems that strategically interact with each other.

Ruth: Combining stated and revealed preferences proposes using stated choices to control for unobserved heterogeneity to recover causal effects from actual choices.

Warren: Continuous difference-in-differences with double/debiased machine learning extends difference-in-differences to continuous treatments by proposing estimators under a double/debiased machine learning framework.

Iris: An Algebraic Approach to Continuously Updating GMM Estimation and Inference provides an algebraic solution for the continuously updating GMM estimator, allowing for efficient computation of its minimum.

Ruth: On the construction and representation of social welfare orders Satisfying Consequentialist Equity Axioms examines how to construct such orders using lexicographic preference relations.

Warren: Difference-in-differences with a mediator Causal mediation analysis is advanced within the difference-in-differences framework to identify natural indirect, direct, and total effects of a treatment.

Iris: Disappointment Aversion and Expectiles recasts disappointment aversion theory using a Savage framework and connects it to econometric models of asymmetric least squares estimation.

Ruth: Tracking the economy at high frequency develops a high-frequency economic index by integrating mixed-frequency official data to monitor the economy in real time.

Warren: Risk-insurance parity introduces risk-insurance parity, which associates different classes of insurance contracts with varying notions of risk aversion.

Iris: Uniform price auction with quantity constraints studies the equilibria of uniform price auctions where many asymmetric bidders have flat demands up to their quantity constraints.

Ruth: Random Collection uses an axiomatic approach to characterize various parametric models in choice situations where a decision maker selects a subset of options probabilistically.

Warren: Organizational Lifespan as Commitment identifies when an organization's lifespan commits future decision makers to continue or terminate based on investment rules.

Iris: Equilibrium Selection in Coordination Games with Planned Actions and Scouting studies stag hunt games where players screen for cooperation by observing noisy signals about their opponents' preparations.

Ruth: Ever since Ellsberg compares subjective expected utility theory with non-EUT models of ambiguity aversion using revealed preference methodology.

Warren: Where the Gains Come From The wedge between multiple-prior and uniform-prior FOSD-rationalizability is the empirical space where non-EUT models can outperform subjective EUT.

Iris: Work that Counts, and Work that Doesn't compares worker responses to wage changes in experiments versus actual work to reconcile differences in elasticity estimates.

Ruth: The Process and Dynamics of the Nobel Memorial Prize in Economics investigates the complex dynamics governing the awarding of the Nobel Memorial Prize by analyzing network data.

Warren: Jornada de trabalho, produtividade e informalidade calculates how much productivity gain is needed for different working hour options to keep output constant.

Iris: Wage-Setting Constraints and Firm Responses to Demand Shocks investigates how institutional wage-setting constraints influence firm responses to demand shocks across different countries.

Ruth: Identification-aware Markov chain Monte Carlo in Bayesian analysis develops MCMC samplers that combine local updates with moves across observationally equivalent parameter values for better identification.

Warren: Advancing AI Capabilities and Evolving Labor Outcomes analyzes how increasing AI capabilities are associated with changes in employment status and work hours across different occupations.

Lucky paper: 2609.35156: Ruth: So we're looking at Delegation without Priors today, which deals with how a principal makes decisions when they have private information about an agent.

Warren: It’s about designing a rule to minimize regret even when the actual circumstances are unknown, balancing whether to stick to a default action or deviate.

Iris: The setup involves modeling the principal's decision as a real number whose desirability depends on a one-dimensional payoff-relevant state that lies in a known interval.

Ruth: This study focuses on quadratic loss environments where both parties prefer higher actions as the state increases, but the agent’s preferred action is shifted upward relative to the principal's.

Warren: The optimal robust delegation rule has a structure with a common default action, a gap above it, and a discretionary interval with random boundaries.

Iris: This structure separates two margins: whether the decision departs from the default and how much discretion the agent receives conditional on doing so.

Ruth: The randomization of the discretionary interval’s floor and ceiling is what hedges the principal’s exposure across potential worst-case states.

Warren: The incentive logic exploits partial alignment where a default and gap create a hurdle for departing from the baseline action.

Iris: This mechanism uses increasing dispersion whenever it remains responsive to the agent’s information, stopping when further responsiveness requires too much additional dispersion.

Ruth: The deterministic outcome at the lowest state generates that common default and gap, while randomizing the boundaries of the higher interval reproduces both responsiveness and dispersion of an unrestricted optimum.

Warren: The paper shows this construction preserves both players’ expected payoffs at every state, thus attaining the unrestricted robust guarantee.

Iris: It's interesting how randomization reproduces the behavior of a mechanism that isn't even fully specified beforehand.

Ruth: They find that the optimal random delegation rule has a common default action equal to b minus one over two e of one minus one over b times one over b, and it excludes all actions in the interval from a0 to two b minus a0 almost surely.

Warren: That default-and-gap structure suggests that investment is lumpy in scale, which the quantitative macro literature often captures through nonconvex adjustment costs.

Iris: The minimax regret achieved by this rule is Vd equal to a squared zero less the minimum of b squared and one over four, which is the optimum of the unrestricted stochastic mechanism design problem.

Ruth: The conclusion is that the optimal rule separates access to discretion from the scope of discretion using a gap and a higher interval with random boundaries.

Warren: The deterministic delegation forces an all-or-nothing choice between full discretion and a fixed decision, which is what they found in the analysis of this paper.

Iris: It also provides a principal–agent microfoundation for frictions where incentive considerations in delegated capital allocation can endogenously make intermediate investment scales unavailable.

Lucky paper: 2607.26366: Ruth: We're looking at Specification Testing for Dyadic Regression Models today, which is paper number two of this show.

Warren: This study develops tests for linear conditional-mean models with undirected dyadic data, focusing on how observations depend on shared nodes.

Iris: The paper establishes a uniform projection theorem that reduces the dyadic process to its latent first-order node projections, which is interesting because it simplifies the whole structure.

Ruth: So they're saying the second-order component of this process is negligible uniformly over the index set at the node level.

Iris: That means we can treat it like an ordinary empirical process over those latent nodes, which changes how we think about estimating things in this context.

Warren: The asymptotic behavior of that estimated residual-marked process has two distinct variance regimes depending on whether there is nondegenerate shared-node dependence or independent dyads.

Iris: Under nondegenerate shared-node dependence, it converges at the square root of n rate governed by first-order node projections, while under independent dyads, it converges at the faster square root of n n rate governed by dyad-level marks.

Ruth: That difference in convergence rates is what motivates their two different bootstrap procedures.

Iris: They propose a raw first-order node multiplier bootstrap and a covariance-corrected Gaussian bootstrap to handle this variance duplication when dyads are independent.

Warren: The raw node multiplier bootstrap works fine under nondegenerate shared-node dependence, but it double counts dyad-specific variation when the dyads are independent.

Iris: The exact covariance decomposition motivates the corrected Gaussian bootstrap because it removes that extra same-dyad contribution, which is essential under independent dyads.

Ruth: The empirical results show that the corrected Kolmogorov-Smirnov test provides the most stable size control while retaining substantial local power across different designs.

Warren: Even with this corrected KS test, an application to a law-firm network rejected additive linear and quadratic specifications but did not reject a richer specification containing an economically relevant interaction.

Iris: That suggests professional connections cannot be adequately described by an additive linear conditional mean, which is a big implication for modeling complex social structures.

Ruth: The comparison of bootstrap methods demonstrates the empirical importance of accounting for dyadic dependence because the naive dyad-level test overstates sampling variation when dependence is weak.

Warren: The corrected procedures are asymptotically equivalent under nondegenerate shared-node dependence, but they yield valid grid tests even when dyads are independent.

Iris: Their conclusion is that the corrected KS test delivers the best finite-sample performance across all designs considered while retaining substantial local power against both quadratic and interaction alternatives.

Ruth: The results illustrate the importance of accounting for shared-node dependence and how useful those tests are for detecting economically meaningful nonlinearities in dyadic conditional-mean models.

Warren: The technical foundations include Lemma one showing that the second-order component is constant time for each fixed function, and Lemma nine showing uniform recovery of node projections.

Iris: Theorem one confirms that the corrected tests satisfy asymptotic validity under both shared-node dependence and independent dyads, which is a strong result for robustness.

Ruth: The empirical study on the law-firm network shows strong evidence against the additive linear specification, and the corrected bootstrap p-values reject the null hypothesis at a five percent significance level.

Warren: The disagreement in the last row of Table one between the naive and corrected procedures illustrates why a dyadic-dependence-robust specification test is needed.

Iris: They show that while raw node multiplier process converges to GR in scenario one and GD in scenario two, their covariance estimators converge to KG and Gamma G under those respective scenarios.

Ruth: It’s good they confirmed the corrected procedures are asymptotically equivalent under nondegenerate shared-node dependence, which is a key point for practical application.

Lucky paper: 2606.02769: Ruth: We're looking at Hidden Commitment Power Is Powerless from Yale University this time.

Warren: It deals with a principal who might renege if her default option is attractive, and that temptation measures her commitment power privately, which is a big asymmetry.

Iris: The Intuitive Criterion disciplines off-path beliefs by making every type of principal behave as if she has the least commitment power.

Ruth: So hidden commitment power is rendered powerless unless there's symmetry in information across all types.

Warren: That means only measures that improve the worst case are truly valuable for mitigating this pre-contracting information asymmetry, according to Hidden Commitment Power Is Powerless.

Iris: Theorem two establishes that in any perfect Bayesian equilibrium surviving the Intuitive Criterion, every type achieves the symmetric-information payoff corresponding to the lowest type.

Ruth: That means an issuer with a larger hidden endowment performs no better than one with a smaller endowment if penalties for reneging are negligible.

Warren: Agencies in those cases contract on terms dictated by the agency facing the mildest penalty under symmetric information, which means stronger legal exposure isn't rewarded if it remains hidden.

Iris: The intuition behind Theorem two rests on two pillars: first, a type can only earn more than this symmetric payoff if some type defaults along the equilibrium path.

Ruth: Second, the Intuitive Criterion rules out any default-separation pattern in that situation.

Warren: That leads us to think that disclosure improves welfare only by raising the worst case because it needs to induce a posterior belief that excludes the lowest type, as Corollary two states in Hidden Commitment Power Is Powerless.

Iris: The leading institutional example is credit rating, where agencies aggregate information and publish discrete grades anchored by the lowest type.

Ruth: Proposition three shows that an optimal signal exists and is monotone-partitional, which gives a contracting-theoretic rationale for the threshold structure we see in scales like S andP and Moody's ratings.

Warren: This structure implies every issuer within a band contracts on terms set by the floor of that band, rather than by their specific type.

Iris: When considering a designer who can credibly disclose information about the principal's type via a Bayes-plausible signal pi, every type's payoff under this posterior equals the symmetric-information payoff of the lowest type in that support.

Ruth: The designer maximizes their welfare by choosing a signal pi that maximizes Z delta theta V (supp mu) d pi(mu) minus C(pi), which confirms Corollary two again.

Warren: If the cost of a signal, C(pi), scales with the number of signal realizations in the support, this leads to a monotone-partitional optimal signal where each band is anchored by its lowest type.

Iris: This structure perfectly mirrors the discrete, ordered bands of credit rating agencies and shows that only measures improving the worst case are valuable in Hidden Commitment Power Is Powerless.

Ruth: So for designers of contracts or ratings, focusing on the floor of a band based on the worst-case scenario seems like a robust policy lesson from this work.

Warren: It's interesting how this translates directly into institutional practice when penalties for reneging are small, which is exactly what the authors explored in Hidden Commitment Power Is Powerless.

Iris: The paper suggests that credit rating systems where each issuer contracts based on the floor of her credit band is a policy lesson derived from this research.

Ruth: And it confirms that only measures improving the worst case are valuable when you're dealing with pre-contracting information asymmetry, which is a key finding from Hidden Commitment Power Is Powerless.

Warren: We see this idea reflected in how credit rating agencies anchor their discrete grades to the lowest type, and that structure is robust across refinements.

Iris: The model's core prediction is that each issuer borrows on terms warranted by the floor of her band, stemming from Theorem two in Hidden Commitment Power Is Powerless.

Ruth: It’s about how the structure of information and incentives dictates the actual contract terms, and that floor anchoring is what makes it work robustly.

Warren: That result seems very concrete when you think about real-world borrowing structures in finance, which is where this kind of modeling often applies.

Lucky paper: 2511.13433: Ruth: Alright, let's get into the details of Double Machine Learning for Inequality Decompositions. This paper is about taking the Kitagawa-Oaxaca-Blinder decomposition and reformulating it using potential outcomes and a doubly robust estimator based on double machine learning to handle common support issues.

Warren: That sounds dense. So, instead of just splitting the difference in means into explained and unexplained parts, they're using potential outcomes to highlight the role of the reference outcome.

Iris: Exactly. They break down the observed difference in means into three parts: delta r X which is explained by characteristics between groups, delta r one S which is an unexplained advantage for group one and minus delta r zero S which is an unexplained disadvantage for group zero.

Ruth: And the choice of that reference outcome really matters here. They look at classical choices like using the potential outcome of one group as a reference, or using an equilibrium reference outcome defined by a propensity score-weighted combination of both potential outcomes.

Warren: The equilibrium reference outcome is defined by Y sub two equals w X Y sub one plus (one minus w X) Y sub zero where w X is the proportion of individuals who share the same set of characteristics in each group.

Iris: That definition for w X, which is P of D equals one times x, is key because it defines that specific reference outcome.

Ruth: The identification strategy focuses on estimating that unexplained part delta r S, which they define as a difference between an observable mean and a counterfactual mean.

Warren: They consider several identification strategies for that, including outcome regression, Inverse Probability Weighting or IPW, and Augmented Inverse Probability Weighting or AIPW.

Iris: For the equilibrium reference outcome r equals two specifically, the identification strategy shows that Assumption two about common support isn't required to identify that unexplained part of the observed difference in outcomes.

Ruth: That's a big relaxation because it removes a key assumption for identifying delta r S when you use that equilibrium reference outcome.

Iris: But then they move on to the machine learning estimation part, where they use a double machine learning procedure based on Neyman orthogonality and cross-fitting to estimate the unexplained part of the decomposition.

Warren: That double machine learning method is shown to be robust to small mistakes in estimating nuisance parameters and achieves root n convergence rates for debiased estimators.

Ruth: And the paper shows that with the equilibrium reference outcome r equals two results are more stable and more precise compared to using reference outcomes r equals zero or r equals one.

Iris: The double machine learning method provides estimators robust to misspecification of regression models with valid inference, which is a big deal when you're dealing with messy data.

Ruth: Empirically, application examples suggest that the part of the difference in means unexplained by the set of characteristics is quite large when using this method compared to parametric approaches.

Warren: Practical considerations involve calibration and trimming, and they mention that calibration issues in machine learning models need attention because estimated probabilities might not match empirical probabilities.

Iris: Trimming is often recommended when propensity scores are close to zero or one for IPW estimators, but the paper says it's not required for the equilibrium reference outcome r equals two.

Ruth: The sensitivity of the decomposition based on Neumark’s reference outcome to including irrelevant explanatory variables is also highlighted as a point of reflection for applied researchers.

Iris: So, in conclusion, their main contribution is analyzing the unexplained component using this equilibrium reference outcome, demonstrating that it allows for a doubly robust estimator satisfying Neyman orthogonality.

Warren: That provides an alternative to extrapolation or trimming strategies by relaxing the common support condition and the need for correctly specified low-dimensional parametric models.

Iris: They conclude that applied researchers must carefully reflect on the relevance of the chosen variables, even when using these data-driven methods to analyze inequalities.

Lucky paper: 2509.19823: Ruth: So, we're looking at A Note on Qualified Majority Voting Rules today. This paper characterizes every qualified majority voting rule in environments with just two alternatives through three axioms: anonymity, responsiveness, and q-neutrality.

Warren: That sounds like it gets into how a parliament decides whether to keep or replace the status quo when they only have two options.

Iris: The introduction sets up this problem by saying that if enough legislators strictly support one alternative, they are certain about it.

Ruth: But if neither side gathers a large number of strict supporters, they aren't very certain about which is the best option.

Warren: This reasoning suggests that the neutrality axiom gets weakened when you only have two alternatives because people aren't sure.

Iris: The paper suggests weakening q-neutrality so that a full preference reversal triggers a swap if either side has a sufficiently high number of strict supporters.

Ruth: So, the core result is that for any quota q greater than half the voters, the rule is anonymous, responsive, and q-neutral if and only if it's a qualified majority voting rule with that quota.

Warren: That’s quite specific about when this equivalence holds true for larger quotas.

Iris: The proof shows this equivalence by proving that the set of such rules equals the intersection of rules satisfying anonymity, responsiveness, and q-neutrality when q is greater than half the voters.

Ruth: If you look at smaller quotas, like q less than or equal to half, there actually isn't a rule that is anonymous and responsive and q-neutral.

Warren: That distinction based on the degree of neutrality really separates these qualified majority voting rules from other types of rules out there.

Iris: The paper points out that for larger quotas, specifically when q is greater than half the voters, it's exactly equivalent to being a qualified majority voting rule with that quota.

Ruth: This characterization provides a description of a common family of voting rules rather than characterizing every single type of rule.

Warren: That seems like a practical tool for understanding how these decision rules actually behave in practice.

Iris: The paper notes that this characterization is intuitive, which helps explain why it captures this common family of voting rules.

Ruth: Let's look at the environment defined as a four-tuple: N, A, R, and sigma. N is the set of voters and A is the set of two alternatives.

Warren: And R is the set of all preference profiles where for every voter you have to define whether they weakly prefer one alternative to another.

Iris: You also have Pib if a voter strictly prefers one alternative over another, and Aib if they are indifferent between them.

Ruth: The voting rule sigma takes a preference profile and associates it with some alternative, which is what we're trying to understand.

Warren: Anonymity means the rule doesn't care who the voters are when making a decision, formally stated as being independent of voters’ identities.

Iris: Responsiveness means if you add one voter who strictly prefers the winning alternative while keeping everyone else fixed, it doesn't change what wins.

Ruth: Q-neutrality is defined so that a rule is independent of the labels exactly when some alternative is strictly top-ranked by at least q voters.

Warren: That definition links the neutrality directly to a specific number of voters topping an alternative.

Iris: The theorem states that for any quota q greater than N over two, the rule satisfies these three axioms if and only if it's a qualified majority voting rule with that quota.

Ruth: So, for large enough quotas, we know exactly what kind of voting rules we are dealing with based on these properties.

Warren: That’s a very concrete result connecting the abstract axioms back to a specific class of rules.

Iris: The paper proves this equivalence by showing that the set of rules satisfying all three axioms when q is greater than half equals the set defined by qualified majority voting rules with quota q.

Ruth: Conversely, if q is less than or equal to half the voters, there isn't actually a rule that can be anonymous and responsive and q-neutral.

Warren: That means for smaller quotas, we lose that specific combination of properties entirely.

Iris: The paper uses this distinction based on the degree of neutrality to rule out seldom used voting rules where some alternative wins only if it is strictly top-ranked by at least q voters when q is less than or equal to half.

Ruth: Overall, this note gives a characterization of every qualified majority voting rule that we haven't characterized before.

Warren: That provides a fairly intuitive description for a very common family of voting rules in decision-making scenarios with two choices.

Iris: It essentially provides an intuitive description of how these specific rules work when you have this constraint on the number of voters and alternatives.

More episodes

← Home