Essays on International Cartels and International Trade

The first chapter of my dissertation evaluates the effect of international cartels on the volume of bilateral trade by relying on the gravity model of trade and the Poisson Pseudo Maximum Likelihood (PPML) estimator. Using a rich cartel data set that contains information on 170 discovered and prosecuted international cartels as well as the most disaggregated bilateral trade data for all OECD countries for the period from 1988 until 2012, I estimate the effect of each cartel on trade individually and find that the impact of cartels on international trade flows is quite heterogeneous across and within sectors. About 50% of the cartels in my sample exert a statistically significant effect on international trade with approximately half of these effects being positive. This suggests that the existence of a cartel between trade partners could actually facilitate trade in some cases. One possible explanation of this rather unorthodox finding is that cartel participants may need to engage in cross-hauling of goods in order to sustain collusion. On the other hand, the negative and statistically significant effects confirm the conventional wisdom that international cartels represent major obstacles to trade. A close examination of the potential determinants of these differing effects provides further insight into the workings of contemporary international cartels.

The second essay (with Constantinos Syropoulos and Yoto Yotov) constructs a symmetric, multimarket duopoly model to study the consequences of economic integration for collusive discipline, trade flows, global profits, and welfare. Two features of the model stand out: Firms interact repeatedly in quantities in each other’s home markets (intra-regional trade) as well as in third-country markets (inter-regional trade). Moreover, firms pool their incentive constraints across all markets to sustain collusion. When the no deviation constraint is active, national markets are strategically linked and thus intra-regional and inter-regional trade cost reductions affect output deliveries and welfare levels in all countries. In addition to unveiling the dependence of collusive stability and cartel discipline on trade costs, our analysis sheds light on the welfare consequences of economic integration. When intra-regional trade costs are sufficiently low, preferential trade liberalization triggers trade diversion and reduces welfare in third-country markets. Exactly the opposite is true for intermediate trade cost levels. However, when intra-regional trade costs are initially high, the direction of change in the volume of inter-regional trade and the outside-country welfare depends on the extent to which firms value future profits. Our analysis also sheds light on the consequences of inter-regional trade cost reductions for intra-regional trade and welfare.

The final chapter develops a duopoly model of multi-product firms that interact repeatedly in segmented, national markets to study their ability to collude and the consequences of collusive conduct for the volume of trade and efficiency. Each firm produces two goods and faces constant marginal costs of production. Importantly, each firm has a competitive advantage in the production of a single good (captured by a certain difference in marginal costs) and national markets are separated by trade costs. Because firms do not internalize the pecuniary price externalities they generate for each other under Cournot-Nash interactions, they may produce and export both goods under competition. In contrast, under collusion, because the cartel rationalizes production and trade, it may choose to shut down production of the inefficiently produced good and opt for trade. Thus, provided trade costs are relatively low, it is possible for collusion to induce higher volumes of trade as compared to Cournot competition. By balancing efficiency gains against the costs of international exchange appropriately, collusion may welfare-dominate competition regardless of whether trade costs take the form of tariffs or transportation costs. Our analysis reveals that, in the case of transportation costs, there always exists a set of marginal cost differences and trade cost values that render collusion superior to Cournot competition. However, in the case of tariffs, this possibility arises only if goods are sufficiently distant substitutes for each other. In other words, in this case, competition is welfare superior when international cartels produce relatively homogeneous goods. We also analyze the dependence of the sustainability of collusion on trade and marginal costs of production, and find that trade liberalization can enhance cartel stability, thereby sustaining the improvements in efficiency noted above.

Many thanks to Delina’s Committee: Committee Co-Chair: Constantinos Syropoulos, Professor of Economics, Trustee Professor of International Economics Committee Co-Chair: Yoto Yotov, Associate Professor of Economics Committee Member: Vibhas Madan, Professor of Economics, Director, School of Economics Committee Member: Matthew Weinberg, Assistant Professor of Economics Committee Member: Irina Murtazashvili, Assistant Professor of Economics