View PDF

Climate change will increase the frequency and severity of natural disasters. This paper examines the effects of such increases on the spatial organization of firms. Using data on the global car industry and an event-study design, I document that nearby floods significantly reduce assembly plant production, with partial reallocation to unaffected plants within the firm. I develop a novel, quantitative, multiregion model in which firms choose their plant locations and capacities to maximize expected profits amidst weather disruption risk. The model captures firms' incentives to diversify capacity across locations and hedge against potential local disruptions. I estimate the model for the automotive industry and use it to compute plant location and capacity choices under different probabilities of weather disruptions according to possible climate change scenarios. With heightened risks, firms build additional, smaller plants with larger spare capacities. This spatial reorganization entails productivity losses, resulting in higher consumer prices.

Joint work with Gaurav Khanna, Nicolas Morales and Nitya Pandalai-Nayar

View PDF

We characterize how firms structure supply chains under climate risk. Using new data on the universe of firm-to-firm transactions from an Indian state, we show that firms diversify sourcing locations, and suppliers exposed to climate risk charge lower prices. Our event-study analysis finds that firms with suppliers in flood-affected districts experience a decline in inputs lasting two months, followed by a return to original suppliers. We develop a general equilibrium model of firm input sourcing under climate risk. Firms diversify identical inputs from suppliers across space, trading off the probability of a climate shock against higher input costs. We quantify the model using data on 271 Indian districts, showing real wages vary across space and are correlated with geography and productivity. Wages are inversely correlated with sourcing risk, giving rise to a cost minimization-resilience tradeoff. Supply chain diversification unambiguously reduces real wage volatility, but ambiguously affects their levels, as diversification may come with higher input costs. While diversification helps mitigate climate risk, it exacerbates the distributional effects of climate change by reducing wages in regions prone to more frequent shocks.

Joint work with Alonso Alfaro-Ureña, Sebastian Fanelli and Eduardo Morales.

View PDF

We estimate a model of firm export dynamics featuring cross-country complementarities. The firm decides where to export by solving a dynamic combinatorial discrete choice problem, for which we develop a solution algorithm that overcomes the computational challenges inherent to the large dimensionality of its state space and choice set. According to our estimated model, firms enjoy cost reductions when exporting to countries geographically or linguistically close to each other, or that share deep trade agreements. Countries, especially small ones, sharing these traits with attractive destinations receive significantly more exports than in the absence of complementarities.

Joint work with Benny Kleinman

Reject and Resubmit, Econometrica

View PDF

We explore how the labor share relates to the price of materials in the economy. Under conditions of imperfect competition and complementarity between materials and primary inputs, a higher price of materials lowers the labor share and raises the profit share of income, without requiring a change in markups. We show that fluctuations in materials prices align with aggregate trends in the U.S. labor share, including a sharp decline during the 2000s commodity boom and stabilization in recent years; provide causal evidence for this effect across industries and across local labor markets; and quantify the importance of our mechanism using an input-output general equilibrium model of the U.S. economy. We conclude that absent our mechanism, the labor share's decline would have been smaller and smoother. We extend our framework to a multi-country setting and demonstrate, theoretically and empirically, how shocks to global commodities yield heterogeneous changes in the labor share across countries.

Joint work with Paula Bustos, Joan Monras and Jacopo Ponticelli.

View PDF

Labor-saving technologies in agriculture can foster structural transformation by releasing workers who find jobs in manufacturing. The traditional view is that factor reallocation towards manufacturing generates innovation and productivity growth. We document, instead, that regions more exposed to a large and exogenous increase in agricultural productivity in Brazil industrialized but experienced lower manufacturing productivity growth. Workers released from agriculture were mostly unskilled and primarily moved to the least skill-intensive manufacturing industries. This paper explores the various mechanisms that can account for the observed manufacturing productivity decline. Changes in worker composition and lower incentives to innovate within manufacturing play prominent roles.

Joint work with Eugenia Menaguale, Eduardo Morales and Alejandro Sabal.

There are significant differences in the set of products multinational firms sell in distinct countries. These differences may be driven by variation in consumer tastes or supply costs and may be magnified by cannibalization forces that push firms to reduce the set of products available in each market. In this paper, we use data on the global car industry with information by firm, country, and car model on production, sales, and prices. We use this data to estimate a quantitative model in which car manufacturers decide where to produce and sell the models in their portfolio. Cannibalization forces imply our model exhibits substitutabilities in the firm’s decision to sell different models in the same destination market; transport costs increasing in distance imply our model displays complementarities in the firm decision to produce and sell the same model in geographically close markets. We introduce a new algorithm to solve high-dimensional combinatorial discrete-choice problems that exhibit complementarities and substitutabilities across choices. Using this model, we determine the role demand and supply factors play in determining the observed differences in the set of products each firm sells in each country. We also study the effects of country-specific consumption or production subsidies on the global production location, products offered in each market, and prices. Our counterfactuals highlight the importance of taking interdependencies into account and show that, e.g., local consumption subsidies lead to changes in foreign outcomes through the firms' endogenous reallocation of production.