Explorations in the New Economic Geography

Abstract: The point of departure of this thesis is the rapidly growing research field known as the New Economic Geography (NEG). It consists of four chapters, one introductory chapter and three self-contained theoretical studies. The chapters may be summarised as follows. The first chapter gives a brief account of the NEG. It explains the soaring interest in the location of economic activity, and it reviews two important strands of models in the literature. Some of the models’ weaknesses are pointed out, and the thesis’ aim and main results are presented. In the second chapter we examine under what circumstances a region in the technological lead may lose its front position due to agglomeration externalities. Firms agglomerate in one region due to increasing returns, transportation costs and input-output linkages among firms. This essay illustrates that the risk of locking in to an old technology is increasing in the benefits of agglomeration. In the de-industrialised region factor prices are lower and a new technology may be profitable to adopt in that region instead, inducing a change in the technological leadership. In the third chapter we challenge two features common to almost every NEG model: the absence of strategic interaction among firms and the catastrophic nature of agglomeration. We employ a homogeneous-good Cournot oligopoly model and analyse a two-stage location-quantity game with many firms and two regions. There are no linkages in the model; firms’ spatial competition for market shares drives all results. In a benchmark case with regions of equal size, we show that the firms will never agglomerate in the same location if transportation is costly between the regions. However, the set of location equilibria changes if a difference in market size is introduced. For high levels of trade costs firms locate in different regions. Lowering the trade costs beyond a critical level triggers a one-firm relocation to the larger region. The nature of agglomeration is gradual and trade costs have to be successively lowered for a full-scale agglomeration to take place. Finally, we also find that the market generates excess agglomeration, and that introducing imperfect consumer arbitrage affects the symmetry-breaking level of trade costs. In the fourth essay we analyse the political determination of transportation costs in an analytically solvable core-periphery model. In a benchmark case with certainty about where agglomeration will take place, we find that a majority of voters prefer low transport costs and the resulting equilibrium is an industrialised core and a de-industrialised periphery. Allowing for uncertainty we show that a high transport cost candidate, that guarantees the initial symmetric equilibrium, may defeat the concentrated equilibrium candidate. The reason is that the immobile factor of production does not know whether industry will end up in its home region when it votes, introducing a bias for status quo. This is so even if reform is efficiency-enhancing for society as a whole and needs no financing.