Research

Control-function estimators identify production elasticities from input variation that instruments predict and that the productivity control does not already absorb. We show that this residual variation can be arbitrarily small while point identification continues to hold, so that the elasticities are weakly identified in the sense of the GMM literature. The relevant object is an orthogonalized Jacobian, obtained by projecting the nonparametric productivity dynamics out of the second-stage Jacobian; its rank governs identification and the associated concentration matrix governs identification strength. Two structural channels drive it toward rank deficiency: because static inputs respond to productivity, quasi-differenced inputs are nearly collinear given lagged productivity, and proxy noise attenuates the identifying signal. Each channel has a limiting case in which identification fails outright, and near that boundary the continuously updated GMM estimator has a non-normal limiting distribution, so conventional Wald inference is invalid. We adapt identification-robust diagnostics, confidence sets, and specification tests to the control-function setting, and show in Chilean and U.S. data that identification strength varies enough across standard datasets to reverse conclusions about markups.

This paper studies identification and estimation of production-function parameters via the control-function approach when the control function is noisy, so invertibility fails and productivity is not recoverable from observables. We show that exact recovery of productivity is not needed: under conditional mean restrictions and a forecast-sufficiency condition, the production function is locally identified. We also propose a novel one-step estimator that avoids separate first-stage estimation and delivers a more parsimonious specification.

DSGE models with different structural parameters can be observationally equivalent for the moments used in limited-information estimation. This is consequential for policy: parameter values that generate the same moments can imply different welfare rankings of policy rules. We study the choice of an interest-rate rule from a statistical decision-theoretic perspective. We apply a quasi-Bayes decision rule that minimizes posterior expected loss. The rule averages welfare over the structural values the moments leave plausible. We illustrate the construction in a Monte Carlo design and in a medium-scale New-Keynesian DSGE model estimated by impulse-response matching.

This paper develops a dynamic general equilibrium model with heterogeneous firms, capital accumulation, and endogenous entry and exit. Aggregate outcomes depend on the entire distribution of firm states rather than on a representative firm. The framework extends the stylized entry–exit model of Hopenhayn (1992) by incorporating investment dynamics and capital adjustment costs, creating a richer firm block. This structure moves beyond static wedge approaches by providing a structural dynamic mechanism through which misallocation arises and by linking the analysis to micro-econometric estimates. This allows to quantify how frictions generate dispersion in marginal products and translate into aggregate TFP losses using empirical data.

We estimate total factor productivity (TFP) distributions across European countries using firm-level microdata from the Microdata Infrastructure (MDI), a secure platform developed by CompNet and supported by the IWH-CBPD. The MDI provides researchers and policymakers with remote access to confidential cross-country firm-level data, enabling rigorous comparative analysis of productivity, competitiveness, and firm dynamics. This rich microdata resource allows us to characterize the complete distribution of TFP across firms, regions, and sectors, moving beyond aggregate statistics to capture heterogeneity in firm performance. Our analysis reveals productivity growth dynamics, quantifies the role of factor misallocation in explaining cross-country income differences, and identifies systematic sources of variation in firm efficiency.

Despite the growing importance of digital trade in services, its measurement remains limited, particularly in developing countries. This study builds a database of over 2,600 websites from Argentina, Brazil, Chile, Colombia, Mexico, and Peru, analyzing their business models and functional attributes. Using clustering techniques, it identifies common patterns and typologies of online presence. Results reveal wide heterogeneity within sectors, with most sites generating value through direct sales or intermediation, and little variation across major regional markets. The findings provide insights to inform policies promoting digital trade in services.