Research

Digitalization, Factor-Biased Productivity, and Market Power

with Subal C. Kumbhakar

Request Paper Research Statement

Summary

Digital technologies may make labor more effective, raise productivity across inputs, or change firms’ ability to price above marginal cost. We develop a production-function framework that separates labor-augmenting productivity, Hicks-neutral productivity, and markups. Using Spanish manufacturing firms from 2000–2019, we find that lagged digitalization is positively associated with both productivity components in the perfect-competition benchmark, with stronger labor-augmenting gains among small firms. Allowing for imperfect competition leaves the labor-augmenting results broadly positive. The average markup response is small, but markups increase in industries with limited digital adoption and decrease in more digitally intensive industries. These findings show why evaluating digitalization requires distinguishing changes in how firms produce from changes in how they price.

Data and Method

The analysis uses the Encuesta Sobre Estrategias Empresariales and a digitalization index covering information technologies, digital human capital, automation, and stakeholder-facing technologies. A restricted translog model and the ratio of labor and materials first-order conditions identify labor augmentation; additional dynamic restrictions distinguish productivity and markup dynamics. Firm-specific price indices deflate output and materials, and a firm-block bootstrap repeats the estimation stages for inference.

Causal Forests versus Penalized Splines for Heterogeneous Treatment Effects

with Ivan Korolev and Nency Dhameja

Economics Letters, 268 (2026), 113205.

Journal Article Working Paper (PDF) Replication Code

Abstract

This paper compares causal forests with varying-coefficient penalized spline estimators for heterogeneous treatment effects. We conduct simulations with randomized treatment assignment that vary heterogeneity, predictor dimension, and discrete nuisance structure. Performance is measured by out-of-sample mean squared error. Forests perform best in Wager–Athey designs with sharp treatment-effect transitions. Penalized splines tend to dominate when heterogeneity is smooth, when the baseline mean and treatment effect differ in structure, and in mixed designs with continuous, binary, and group-level covariates. As group cardinality rises, forest performance deteriorates, while mgcv random-effect smooths handle group structure effectively and remain easy to implement.

The Scale-Technology Paradox: Technological Progress and Productivity Decline in Nigerian Banking

with Subal C. Kumbhakar

Economic Modelling (2026), 107859.

Journal Article Working Paper (PDF) Replication Code

Abstract

Nigeria’s banking sector has undergone substantial consolidation and technological modernization, yet bank productivity has continued to decline. This study examines this scale–technology paradox using annual data for 17 commercial banks over 2000–2023. We estimate a multi-output input distance function using a control-function approach to distinguish productivity change from technical change and examine the roles of scale economies and input costs. Bank productivity declined by 3.5% per year on average (3.1% at the median), with larger and internationally licensed banks experiencing steeper declines. Technical change was cost-reducing, while returns to scale averaged 1.46, indicating substantial unexploited scale economies. Lending was the main source of cost pressure. The findings show that technological progress and scale economies did not translate into productivity growth because scale economies remained underexploited and lending costs persisted. Strengthening credit information, collateral enforcement, bad-loan resolution, and lending efficiency is therefore central to translating technological progress and scale economies into sustained productivity growth.

Mediation Analysis in Production Functions with Unobserved Productivity

with Subal C. Kumbhakar

Under review, Economics Letters

Request Paper

Summary

Technology adoption can change output directly and indirectly through firms’ input choices. Conventional mediation methods can misattribute these channels when unobserved productivity affects adoption, labor demand, and output. We develop a sequential estimator that combines production-function estimation with a mediator regression controlling for recovered productivity. Under the model’s assumptions, the framework identifies direct, input-mediated, and total effects. Simulations show that omitting productivity from the mediator equation can substantially overstate the indirect effect. An application to Spanish manufacturing uses labor as the mediator and lagged digitalization as the treatment.

Regularization in Translog Production Models

Solo-authored

Flexible production models can be difficult to estimate because their interaction and curvature terms are highly correlated. I evaluate Lasso, Ridge, Elastic Net, principal component regression, and partial least squares by the accuracy and interpretability of output elasticities and related economic quantities, including substitution and returns to scale. The project asks when regularization stabilizes estimates without distorting the production relationships of interest.

Artificial Intelligence and Spatial Factor-Biased Productivity

with Subal C. Kumbhakar

I am extending the digitalization framework to study how AI adoption changes the relative productivity of inputs and how these changes spread across firms and regions. The project considers knowledge flows, worker mobility, and input-output linkages as potential channels of transmission. Planned simulations will assess how well the model distinguishes spillovers from common shocks.