Research

I study the welfare implications of technology-induced disintermediation using a production network model with endogenous intermediation, wholesaler entry and exit, and markups. Wholesalers economize on the search costs of forming buyer-supplier relationships, but as direct trade technology improves, demand for intermediation falls: marginal wholesalers exit, survivors gain market share, and markups rise. These higher markups distort relative input prices and misallocate resources, partially offsetting the gains from disintermediation. I test the model’s predictions using Turkish firm-to-firm transactions, exploiting the staggered rollout of fiber internet. Consistent with theory, provinces with faster fiber growth see less intermediated trade, fewer wholesalers, and higher wholesale markups. Calibrating the model to these responses, endogenous markup increases reduce welfare gains from fiber-induced disintermediation by 30%. The results demonstrate that technologies enabling firms to bypass intermediaries can generate unintended efficiency losses by consolidating wholesale market power, highlighting the potential role for complementary competition policy.

I study how reductions in internal trade costs, driven by the rollout of fiber internet across Turkish provinces, reshape product-market competition and aggregate efficiency. First, I document a pro-competitive effect: provinces experiencing faster fiber rollout saw a relative decline in aggregate manufacturing markup, with markedly larger declines in sectors that are more tradable. Second, I develop a spatial extension of Atkeson–Burstein that embeds sectoral heterogeneity in tradability and internal migration. The model quantifies welfare changes that operate through two opposing forces. On the one hand, falling internal trade costs reduce aggregate markups and within-province markup dispersion. On the other hand, uneven infrastructure development can widen markup dispersion across provinces, misallocating workers toward low-markup provinces and dampening the aggregate gains. The net effect hinges on the initial spatial distribution of markups and on each province’s comparative advantage in more versus less tradable sectors. These findings provide a rationale for place-based policy aimed at correcting spatial distortions due to product-market power. The potential for uneven infrastructure development to worsen aggregate efficiency provides a rationale for a more balanced approach to development within a country.

I study the optimal tariff when importers are granular. Optimal tariff arguments typically rely on atomistic domestic agents who fail to internalize the effect of their sourcing behavior on the world price. A large country then improves its terms of trade by setting a positive import tariff. In many settings, however, a small number of large importers dominate procurement. When a few large importers internalize terms-of-trade effects, the standard optimal tariff motive shrinks. Unlike granular seller power (Gaubert and Itskhoki, 2021), granular buyer power can eliminate the rationale for tariffs. This creates a complementarity between antitrust and trade policy: reducing importer concentration raises the optimal tariff by weakening private internalization of the terms-of-trade effect.