Krugman, Paul (1991) Geography and trade, London: MIT Press/Leuven UP, p.142.
Economic geography is devoted to understand the location of production in space, in other words where things happen in relation to one another (p.1). Economists ought to remember that countries both occupy and exist in space (p.2). “To say anything useful or interesting about the location of economic activity in space, it is necessary to get away from the constant-returns, perfect-competition approach that still dominates most economy analysis”, and use such notions as increasing returns and imperfect competition (p.4).
Why space matters
Certainly the most striking feature of economic geography is concentration of population and production, the later indicating that such thing as increasing returns exist (otherwise why would producers bear the costs of crowding together?; p.5). In fact, trade does not only arises from comparative advantage but also from increasing returns which favour concentration of production even when confronted with a dispersed market (p.6).
First-comer advantage, sometimes acquired out of sheer luck, is by essence transitory but can often be locked in and some industries can as a result become heavily localized (p.10). The US manufacturing belt is certainly one of the best examples of such a process. By 1900 it contained 74% of the country’s industrial employment (64% in 1957). The North East region monopolized nearly all the footloose (i.e. not linked to natural resources) industries (p.12). Although the region’s dominance had been established when it was also the country’s main agricultural centre but it remained even after the initial advantage had shifted (p.13).
Of economies and scales
Obviously, each individual factory remained in the manufacturing belt because so were the other factories, its clients, competitors and providers. To understand the forces pushing factories to cluster together, one ought to consider the externalities coming from density (p.14).
“Given sufficiently strong economies of scale, each manufacturer wants to serve the national market from a single location. To minimize transportation costs, she chooses a location with large local demand. But the local demand will be strong precisely where the majority of the manufacturers choose to locate. Thus there is a circularity that then to keep a manufacturing belt in existence once it is established” (p.15).
In other words, producers have to choose between fixed and transportation costs (p.17). It is also remarkable that integration between two locations through lower transportation costs may allow their markets to reach a critical mass that would attract manufacturers and thus give them an initial advantage upon other smaller, less integrated markets. But noticeably, as transportation costs are also submitted to economies of scales (through indivisible investments such as railroads), hence integration is more likely to arise between larger markets (p.24). “Increasing returns and cumulative processes are pervasive and give an often decisive to historical accident” (p.25).
The system described above is a deeply conservative force, but things do change and when they do, it happens fast (p.26). For instance, the discovery of oil in California, fuelled an immigration that suddenly allowed the state to reach a critical mass starting a process of explosive growth (p.28). Change can be smoothen by workers’ and firms’ expectations (if it is clear that a region is going to grow, may as well move there early; p.29). There is thus a large share of self-fulfilling expectations in regional development: if information does not travel well enough a location may simply loose momentum (p.31). In the same way, pessimism can be self-justifying (p.33).
Why localization happens
Localization often depend on seemingly trivial historical accident. Following Alfred Marshall, the author identifies three distinct reasons for localization.
- Concentration: “allows a pooled market for workers with specialized skills; this pooled market benefits both workers and firms” (p.36).
- “Provision of nontraded inputs specific to an industry in greater variety and at lower cost” (division of labour).
- Information flows locally more easily than over greater distances, an industrial centre generates what we would now call technological spillovers” (p.37).
Labour market pooling
“The labour demand of the firms is both uncertain and imperfectly correlated” (p.38). Specialized workers will favour a location where several firms are susceptible to hire them so as to limit their risk of unemployment and avoid one employer’s monopsony over hiring (p.40). On the other hand, both firms and workers are susceptible to opt for a deconcentrated location to avoid competition on the job market (p.42) but labour pooling fails to be beneficial for both parties only in very extreme conditions (very low or very high wages, strict inelasticity of labour input, economies of scale so strong that there are very few firms; p.45).
“A localized industry can support more specialized local suppliers, which in turn maker that industry more efficient and reinforces the localization”. It is because of these economies of scale on intermediate inputs that large production centres have a key comparative advantage (p.46). These economies of scale arise when transport costs of the intermediary goods are significantly lower than those of the finished one.
In the USA, many industries – both cutting-edge and less innovative-intensive – are highly concentrated (p.58). “Localization within an industry tends to fade away, there are always new industries. […] Indeed surely there is a kind of product life cycle, in which emergent new industries initially flourish in localized industrial district, then disperse as they mature” (p.63).
Regions and nations
Nations matter because their governments’ policies may affect factor mobility (p.72). Krugman observes that European nations are less specialized than US regions. For instance, whereas heavy and light industries are present in both Germany and Italy, the Midwest has none of the latter and the South little of the former (p.77). So national policies do affect negatively localization (p.78).
Transports costs have not fallen as rapidly in Europe due to trade barriers. The EU might change that (p.79). In the US, a process of industrial delocalization is taking place but it may be compensated by a greater concentration in the service sector (p.81). National barriers to trade can prevent the constitution of a large enough market to allow specialization, economies of scale and concentration. In this case, a multi-core structure is likely to be created (p.87). On the other hand, early tariffs may protect a new manufacturing centre, which will later be able to create a core-periphery relation with the neighbouring countries (p.90).