When testing models of firm conduct, weak instruments can threaten inference. We develop a framework to understand instrument strength, and a diagnostic to measure it. We show its empirical relevance in an application to testing models of vertical conduct econometricsociety.org/publications/q…

How can we test for validity of conditional moment restrictions when the model identification is compromised? We propose a specification test that is asymptotically standard normal and robust to first-order local identification failure econometricsociety.org/publications/q…

Quantitative Economics Volume 15, Issue 2 (May 2024) is now online
econometricsociety.org/publications/q…

Can political clientelism be detected? Theory shows and empirical evidence from India confirms that voters support local incumbents only in response to private welfare benefits but not public goods delivered. econometricsociety.org/publications/q…

The paper proposes a test for deciding whether the high-dimensional betas in a continuous-time latent factor model differ at two time points. Empirically, the test rejects strongly the constant-beta hypothesis for asset prices on days of FOMC announcements econometricsociety.org/publications/q…

Unveiling risks in non-sharp identified sets: This paper reveals that non-sharp identified sets, contrary to the intuition, could be misleading. It provides a framework to address discrepancies for more robust empirical analysis econometricsociety.org/publications/q…

The editorial team at Quantitative Economics is excited to announce that Bernard Salanié will be the next QE Editor. Bernard will bring his deep knowledge and insight to the journal econometricsociety.org/publications/n…

We quantify the relative contribution of two types of skill-biased technological change to the increase in US inequality since 1980. Higher returns to skill are the main driver while higher dispersion of firm productivity plays a smaller role Manuel Macera econometricsociety.org/publications/q…

We study asymptotically exact inference for randomized experiments with multiple treatments in “matched tuples” designs. We introduce novel standard errors and argue that commonly recommended approaches based on strata fixed effects can be invalid econometricsociety.org/publications/q…

Trade policy impacts trade in a highly non-linear way: Tariffs' marginal impact on trade? Strong for low & medium barriers, weak for high ones—especially with substantial non-tariff barriers. Crucial for trade policy evaluation! Katharina Erhardt econometricsociety.org/publications/q…

Clustering reveals grouped heterogeneity in panel models, but misclassification is possible. Which units are classified reliably? Which group assignments are suspect? We quantify sampling uncertainty jointly for all units. Okui, Ryo 奥井亮 econometricsociety.org/publications/q…

Quantitative Economics Volume 15, Issue 1 (January 2024) is now online
econometricsociety.org/publications/q…

Exploring the dynamics of panic buying, we show how increased shopping costs in emergencies cause societal inefficiencies, even with rational consumers. We quantify its welfare cost and propose policies to mitigate such crises. 野田 俊也（Shunya Noda） Kazuhiro Teramoto econometricsociety.org/publications/q…

Trust is key for economic exchange. We propose a novel way of experimentally measuring trust in institutions & exploit exogenous variation to test if the institution’s fulfilled promise raises trust and changes financial behavior @spenczynski Maria Isabel Santana econometricsociety.org/publications/q…

Read more about upcoming editorial board changes econometricsociety.org/publications/n…

Quantitative Economics Volume 14, Issue 4 (November 2023) is now online
econometricsociety.org/publications/q…

Standard robust inference in linear regressions may not be accurate in small samples. As an alternative, we develop a permutation test for subvectors that is exact under an independence assumption and asymptotically valid under weaker conditions econometricsociety.org/publications/q…

Forming predictive models of data generating processes is difficult and biases the kinds of beliefs people form. Experimental evidence explains how this occurs and allows us to identify which notions of complexity predict errors in belief formation econometricsociety.org/publications/q…

Misspecification under multicollinearity can cause biased predictions and prevent simulation-based algorithms from converging. We show that neural networks are robust to multicollinearity by solving an optimal fiscal policy problem with multiple maturities econometricsociety.org/publications/q…

During the Barcelona Summer meetings EEA The Econometric Society we organized a roundtable on 'Data sharing when collection is costly' with fantastic panelists (@guido_imbens Marie Connolly, José G. Montalvo, and Sule Alan). Watch the video here 👇 youtu.be/Pl3TLHViTOM