I develop a quantitative model of multi-plant oligopolists where each firm decides where to locate the set of plants and how to serve each market, taking into account cannibalization across its own plants as well as competition with others. In contrast to canonical trade models with multinational firms where neither spatial interdependency of decisions nor oligopoly is considered, I advance the existing research by allowing for interdependent entry, oligopolistic rivalry and variable markups. Despite having a high-dimensional discrete choice problem, I provide a toolkit to tractably estimate the model in a three-step procedure, leveraging the gravity-type regressions, the analytical expression for market price derived from the model, and the solution algorithm for a combinatorial problem when the location game is submodular. As an application I estimate the model for the cement industry in the US and Canada. Counterfactual experiments quantify firm-level responses to changes in environmental, trade and competition policies and highlight welfare implications of having multi-plant production.
Recent decades have witnessed the growing importance of trade in intermediate goods and pursuit of free trade agreements (FTAs). FTAs distort firms’ sourcing decisions internationally through preferential tariffs and preferential rules of origin (RoOs), a set of regional value content (RVC) criterion to ensure that goods are originated from member countries to be granted preferential treatment when exporting within the FTA. This paper unpacks the effects of FTAs and distinguishes the intermediate trade elasticities with respect to RoOs and tariffs focusing on the automobile industry. Decisions of how much car parts to acquire and which supplier location to select for each part are modeled for car assemblers. With the derived gravity form of equation, the estimation identifies significant diversion in sourcing induced by high RVC requirement, to an extent more than a direct tariff reduction. However, RVC below 60% tends not to be binding. The impacts of RoOs are also heterogeneous to different groups of firms, different degree of preferential treatment and different intermediate inputs. Interaction between the level of RVC restriction and firms' production inputs allocation leads to sophisticated but sensible patterns in trade flow changes among countries.
One of the central questions revolving multinational firm theory is whether firms choose to internalize or engage in arm’s length licensing when entering a foreign market. Decisions on firm boundaries become more complicated when firms carry more than one brand and multiple stages of operation before reaching consumers. This paper compares two approaches in the literature concerning a multinational firm’s organization of foreign operation, namely the transaction-cost theory and property-rights theory. The two differ in whether efficiency losses from incomplete contracts are endogenous and whether they are of the same nature even for integrated relationships. Two models yield opposite predictions which open the door for empirical investigation that I test using a unique brand-market level data in the beer industry. I empirically characterize the optimal allocation of ownership rights by brand quality, contractual environment, market size, competition structure, and the existed band portfolio of the licensor and licensee. Results support the property-right theory for beer companies. I also overcome the data limitation on licensing agreements and develop an innovative way to infer contractual relationship by combining brands’ ownership information at the national and global level.
We study whether consumer taste is biased towards beer brands originated from countries where the ancestors of consumers came. The empirical analysis takes advantage of a single Chicago-area grocery store chain, Dominick’s, having stores located at census tracts with different ethnic background composition. Combining the sales data at store-UPC level controlling for prices and brand offerings with 173 beer brands’ origin information scraped from one of the most visited beer online platform, we find significant home bias in consumption. The result sheds light on why this grocery chain has its market share plummeted after it was taken over by Safeway and replaced with house branded products.
This paper applies the gross exports accounting framework, initially proposed by Koopman, Wang and Wei (2014) and generalized in Wang, Wei and Zhu (2013), to 186 trading nations in 26 sectors from 1990 to 2011 based on the Eora MIRO database. Such decomposition provides insights on the intensity of cross-country production sharing across time, country and sector. There are five dimensions in determining a country’s position in global value chain after decomposing its gross exports, namely domestic value added to gross exports that are absorbed abroad (VAX), domestic value-added in final goods exports (DVA_FIN), domestic value-added in intermediate goods exports consumed by direct importing countries (DVA_INT), domestic value-added that are imported back home (VS1*) and vertical specialization (VS) that is the foreign value-added in exports. I link all these value-added terms with technological progress in different countries, expressed in the form of R&D expenditure, the number of patent applications, scientific and technical journal articles and researchers in R&D. The analysis considers the effects on value-added terms of technological progress in the domestic country itself regardless of the international transmission of technology from foreign countries. As technology become more advanced in a country, its DVA_FIN ratio falls, whereas DVA_INT and VS1* ratios rise, indicating the country is moving towards the upstream of production. Country’s VAX ratio decreases and VS share increases with technological progress. That is, the country is more vertically integrated. Meanwhile, I decompose countries’ overall effects on each value-added term to the within and between effects. Cross-country variation in aggregate term ratios due to different technology levels is to a large extent driven by the technology influence within industries instead of the composition of exports.
Available upon request.