Adrian Kulesza

Working Papers

An ‘Austrian’ Model of International Specialization
Pol Antràs and Adrian Kulesza
March 2026
Abstract: We develop a general equilibrium model of international trade in which the temporal structure of production is a key determinant of comparative advantage. Building on Böhm-Bawerk’s theory of capital, the model formalizes the idea that production processes with longer average periods of production (APPs) entail higher financing costs due to the time lag between input payments and revenue realization. We embed this insight into a multi-sector Ricardian framework with endogenous interest rates. Under autarky, countries with more patient consumers or more developed financial markets exhibit lower equilibrium interest rates and higher wage rates. With international trade, these countries typically gain a comparative advantage in sectors with longer APPs, though the model can also generate multiple equilibria and unconventional specialization patterns. We extend the framework to include trade costs (inclusive of shipment delays), global value chains, and international capital-market integration. Empirically, we present evidence showing that countries with more developed financial systems export disproportionately more in sectors with longer APPs, even after controlling for standard neoclassical and institutional determinants of comparative advantage.

Combinatorial Discrete Choice with Deep Reinforcement Learning
Adrian Kulesza
November 2024
Abstract: I utilize new computational methods to study how economic agents solve combinatorial optimization (CO) problems, where an optimal solution is selected from a large, discrete set. In structural trade/spatial models CO problems are ubiquitous from firms choosing where to source inputs and open plants to social planners choosing where to allocate infrastructure and enact policies. Normally, these problems are approached using either heuristics or specialized algorithms informed by model assumptions and precise parameterizations.CO problems are ubiquitous in structural trade and spatial models, from firms deciding where to source inputs and open plants to social planners determining where to allocate infrastructure and enact policies. I use a machine learning model to approximate policy functions that learn to solve CO problems through repeated interaction with a simulated economic environment. I benchmark this approach to existing algorithms for several CO problems from trade, often yielding either optimal or superior policies with competitive computational times. I then demonstrate how this method can be applied to models with rich interdependencies, for which current methods do not work. I estimate a model of export market entry with complementarity in fixed costs and substitutability through increasing marginal costs.