Why are the markets for many tradable goods segmented across national boundaries? Despite low levels of formal trade barriers and recent reductions in global communication and shipping costs, firms still command a disproportionately large market share in their home countries. Table 1 shows the extent of this ‘home market advantage’ in the automotive industry. The case of the German and French passenger car markets illustrates this phenomenon well: the home market shares of German and French brands are 55% and 52%, respectively. Cross-market shares are 9% and 19% only. What explains this discrepancy? These two neighbouring EU members have no formal trade barriers between them. Their demographics and levels of development are very similar. Could supply-related frictions account for the gaps, especially between other distant markets? Or is demand for product attributes sufficiently different across countries, with national producers catering to these heterogeneous preferences? Do consumers, after considering price, quality and other attributes, still have a preference for national brands? Addressing these questions would help one to understand the sources of market segmentation in similar oligopolistic durable goods, make predictions on the effect of new trade agreements (such as the TTIP, CETA, or TPP), and assess market entry strategies (acquiring existing home brands versus offering core brands of one’s own).
Table 1 Market shares by brand nationality
The automotive industry presents an ideal setting to address these questions. It makes up important shares of world trade and FDI, and receives above-average tariff protection. Prices of gas and space (for parking and avoiding congestion) show considerable variation across countries, most notably between North America and the rest of the world. The product is also seen as a source of national pride in many industrialised countries. For research purposes, detailed data are available on production locations, product attributes, and importantly on prices and quantities sold – as opposed to just having information on revenues or trade flows, which is typically the case in the bulk of the international trade literature (e.g. Eaton and Kortum 2002, Anderson and van Wincoop 2003). In recent work (Cosar et al. 2015), we construct a comprehensive dataset of model level sales, price and characteristics for the nine markets across three continents displayed in Table 1. We are able to match this data to model-level data on the assembly locations of automobiles, indicating the set of available sourcing locations for each model. Using the data in a rich model of demand and supply, we disentangle the supply- and demand-related drivers of market segmentation.
To distinguish demand-side and supply-side drivers of home market advantage, we estimate a structural model that extends the demand framework of Berry et al. (1995) to an international setting with a rich supply side. Consumers have heterogeneous preferences for horsepower to weight, size and fuel efficiency ratios. Importantly, these preferences are allowed to vary systematically across countries in order to capture differences in demand. In addition to differences in taste for observable characteristics, consumers may value brands differently across countries – with a potential preference for national brands. Once the demand-side parameters of the model are estimated, we recover marginal costs for each model assuming that producers set prices according to a Bertrand-Nash pricing equilibrium. We then estimate supply frictions assuming that firms endogenously choose to source cars for a particular market from a subset of their worldwide production locations to minimise costs. These frictions include the costs of shipping from the assembly location to the destination market and the costs of managing an assembly plant away from the headquarter country. Our model captures observed international trade flows in the global car market.
Results
We use our estimates to unpack the contributions of tariffs, trade/FDI costs, home preference, and taste heterogeneity to home market advantage. To quantify home market advantage, we calculate the difference in market share when a model is sold at home versus abroad. We find that, on average, a model's home market share is more than triple its share abroad – albeit from a small base, given that there are more than 200 models competing in most markets. Using our structural estimates, we then evaluate the contributions of potential drivers of the home market advantage by computing counterfactual prices and shares after removing various potential drivers of market segmentation and re-computing the home market advantage statistic. We find that home market advantage is most sensitive to the removal of home preference for domestic brands, declining by about 60%. In contrast, removing all supply-related frictions reduces the home market advantage by only 13%.
Years of market presence and dealer densities are important in brands’ market shares. Since, for historical reasons, these factors are correlated with home status, treating them as unobserved would bias the estimates of home preference. Our estimates of home preference already control for these factors, and the large preference for national brands remains strong. The median consumer’s willingness to pay for a home brand varies from $800 to $1,050 across countries, corresponding to a markup power around 4-5% (compared to average markups of around 15%). These findings point to home preference as an important feature of trade and multinational production that is underappreciated in the academic literature.
Implications
We believe these results have strong implications for companies, policymakers, and academics. For international firms, the strength of consumers’ home preferences represents an important consideration in determining where to enter a market and how to build a brand reputation. For example, we find that brand reputation is correlated with years since entry into a market, suggesting that long-running investments in brand reputation may be a dynamic driver of future sales. Furthermore, entering a foreign market through acquisition of a local brand can help to overcome the disadvantage of foreign firms. For policymakers, it is important to understand that market segmentation will not disappear immediately upon the conclusion of international agreements that reduce policy-based trade frictions. A significant portion of this segmentation is driven by differences in preference that are likely to persist following liberalisation. Finally, for academics, the strong importance of home preference identified by our analysis requires further investigation. Home preference may arise from familiarity on the part of consumers, nationalistic feelings, or the ability of home brands to provide unobserved characteristics (body styling, interior features such as cup holders) that better fit their home market. As each of these explanations has different implications for policy and firm strategies, achieving a better understanding of the origins of home preference is an important area for future research.
References
Anderson, J. E. and E. van Wincoop (2003), “Gravity with Gravitas: A Solution to the Border Puzzle", American Economic Review 93(1), 170-192.
Berry, S., J. Levinsohn and A. Pakes (1995), “Automobile Prices in Market Equilibrium", Econometrica 63(4), 841-890.
Cosar, K., P. Grieco, S. Li and F. Tintelnot (2015), “What Drives Home Market Advantage?”, NBER Working Paper No. 21583.
Eaton, J. and S. Kortum (2002), “Technology, Geography, and Trade", Econometrica 70(5), 1741-1779.
