| The economic activity of many industries is geographically clustered. Over time, some industrial clusters remain in their initial location, some move to a new location, and some break apart. This dissertation considers the U.S. automotive industry, which is seemingly teetering between these three options.;The agglomeration of the automotive industry in the upper Midwest of the U.S. is one of the most prominent and persistent industrial clusters. Historically, automotive production in the U.S. was dominated by the "big three" domestic manufacturers (Ford, G.M., Chrysler) and the clear center of the industry was the upper Midwest. However, in the last 30 years, many foreign-owned manufacturers (e.g., Honda, BMW) have opened plants in the deep South, hundreds of miles from the industry center in Michigan. These assemblers are attracted to the South by the lower rates of unionization and lower labor costs in those states.;I quantify three major costs that drive (upstream) supplier plant location in this industry utilizing data on vertical contracting and supplier plant locations. In particular, I estimate the magnitude of location-specific fixed costs, labor costs, and distance costs (e.g., transport and just-in-time related costs) for automotive supplier firms. In chapter 1, I estimate the relative contribution of labor costs and distance costs to the variable cost of producing and delivering a given part. In chapter 2, I simulate to what extent supplier plant locations are driven by fixed costs, labor costs, and distance costs. In chapter 3, I utilize panel data on where supplier firms open and close plants to determine how their plant opening and closing decisions have been impacted by the shift in the composition and location of automotive production in the U.S.;As a framework for the analysis, I develop a model that includes both the supplier plant location decision and the assembler procurement decision (which determines supplier profit). In the first stage of the model, supplier firms choose locations. In the second stage, they bid on contracts from assemblers in a competitive auction. Using the efficiency of assembler procurement (i.e., the low-cost supplier gets the contract) and data on which supplier firm supplies which part to which assembler, I estimate supplier variable costs. Using the "location stage" of the game and the estimated variable costs, I simulate equilibrium supplier plant locations to understand the impact of costs on plant location.;The variable cost estimates indicate that being closer to the assembler or in a right-to-work/low unionization rate state decreases supplier costs. For a foreign-owned supplier firm having a plant in a right-to-work state is equivalent to locating 2,800 miles closer to the assembler (for a typical part), while for a domestic supplier firm these costs are equivalent to locating 1,400 miles closer.;To assess the relative importance of fixed costs and variable costs in determining plant location, I simulate supplier locations if fixed costs in all locations were equal. In this simulation, the geographic mean of foreign-owned supplier plants would be 450 miles south (farther from the Midwest) of its actual location in 2007. Further, I find that when foreign-owned supplier firms locate plants solely on the basis of distance costs, the geographic mean of their locations is in the Midwest, suggesting that the main driver of the automotive supplier industry towards the South is labor costs.;In chapter 3, I extend the two-period model to include the dynamic decision of supplier firms to open and close plants to maximize their profits. I estimate the reduced form policy functions of this model to see how supplier firms respond in their opening and closing behavior to the shifts in downstream demand. I find that while the least profitable firms are more likely to close down a large share of the firm's plants, the expected number of plant closures at the firm level does not decline with firm profitability. Furthermore, the most profitable firms open more facilities both gross and on net. Also, conditional on closing or opening a plant, a firm is more likely to open or close that facility in the location that maximizes variable profits. These results are consistent with plant relocation, where firms shift the locations of their facilities to maximize profits, as being an important driver of plant openings and closings in this industry. |
