Volume 4 NOV 2008
Feature articles When a firm is market-oriented
Accounting induced performance
anxiety
Understanding global imbalances
An interview with Robert E. Lucas Jnr.
The use and misuse of intelligent
systems in accounting
Alumni refresher lecture series
Real options analysis and
investment appraisal
New economic geography and
manufacturing
Print
this page
New economic geography and manufacturing
Understanding the existence of cities and regularities about the location of manufacturing activities within countries
by Russel Hillberry
Listen to lecture - August 2008 (MP3 10.0MB)
Revelations of night photos from space
The US space agency NASA publishes composite photos that show us night-time views of the earth from space.1 What is striking about these maps is that the casual observer is usually able to identify particular locations, with little more than the geographic scattering of light to help them along. For the purpose of this lecture we take light that is visible from space as a useful indicator of urbanisation.
This lecture discusses the New Economic Geography (NEG) literature in Economics. Among other things, the NEG literature is focused on explaining the pattern of light in these maps. Why do the (brightly lit) heavily populated urban areas and (dark) rural areas coexist? Why are these urban areas so often located near one another in economic space? These are not simply questions about why the world looks as it does at night. Rather, they are fundamental questions about human behaviour and the organisation of industrial societies.
A related topic, though one not as well captured by the night photos, is the extreme localisation of particular types of production activities. Why, for example, did Pittsburgh, Pennsylvania, become the home of several steel companies; while Detroit, Michigan, was the home of several auto companies? Why are most US band instruments produced in the small city of Elkhart, Indiana; while most carpet and flooring materials are produced in Dalton, Georgia? The recent revival of economic geography has focused on these topics.
Head and Mayer (2004) provide a guide to the characteristics of a standard NEG model. They note that an NEG model typically has five components. First, there are increasing returns to scale in production, and these are internal to the firm. Second, the existence of increasing returns generates market power, and this must be modelled appropriately. Third, there are costs of trading over geographic space. Fourth, firms have the ability to choose their locations. Fifth, there is endogenous location of demand, either through mobility of households, or via the effect on the demand for intermediate goods that occurs through the location decisions of downstream firms.
A model of this type has the following implications. Increasing returns to scale imply that production will occur in a limited number of locations, because spreading out production over multiple places would mean that scale economies would go unexploited. Because there are trade costs, firms choose to locate near the bulk of demand. Trade costs will also give consumers an incentive to locate near the firms, to avoid paying more for the goods.
Core and peripheral areas
This basic model thus generates co-location of firms and workers in a single location (i.e. a city). There are forces in the model that push some firms toward breaking away and locating in peripheral areas, but these are typically overcome by the benefits generated by increasing returns and trade costs for being in the geographic ‘core’. The core-periphery structure of equilibria in these models is the common outcome that makes the models useful and interesting.
Among the outcomes that emerge from the models is that in some equilibria, higher wages are sustained in the core rather than in the periphery. The core may also be able to support higher tax rates and more generous public services, than can be supported in the periphery. The advantages maintained by the core can sometimes persist against certain shocks (i.e. changes in trade costs). However, the effects of such shocks are non-linear. In some cases, such shocks will lead to rapid unravelling of the core.
These results offer us a way to understand not only the existence of cities, but also regularities about the location of manufacturing activities within countries. The example that is most often used in the literature is the ‘manufacturing belt’ in the northeast part of the US. From the late 1800s to the mid 1900s, US manufacturing prowess was unrivalled in the world. Its workers received very high wages by world standards, yet were often able to produce on a scale that made their goods seem cheap by world standards. During this period of US manufacturing dominance, much of its manufacturing activity was geographically clustered in a small section of the country – the Great Lakes region and the Northeast.2 Estimates reported in Krugman (1991) suggest that 70 per cent of US manufacturing activity occurred within this belt in 1900, and 64 per cent as late as 1957.
A map showing the distribution of night-time light in the late 1950s would have made clear to the naked eye the unusual levels of activity in this region. Much of the country’s population lived inside the manufacturing belt. Consistent with the models, manufacturing firms and the bulk of final demand were co-located. As some NEG models indicate, the manufacturing belt was able to sustain higher wages than the rest of the country. The region also sustained higher tax rates and better quality public services. For example, this region hosted the bulk of the country’s great cultural institutions, including the most prominent universities, museums and symphonies.
An important feature of economic activity within the US manufacturing belt was the importance of intermediate goods trade. Detroit produced autos, so auto parts were produced in nearby areas of Michigan, Ohio and Indiana. NEG models have been adapted to include intermediate goods. A typical result from thinking through this exercise is that the inclusion of intermediate goods trade magnifies many of the results from the basic model – the core pays an even larger wage premium, even larger gaps between core and periphery taxes can be supported, but that the unravelling of the core can be even more disruptive.
Intermediate goods trade can also be understood as a reason for the high levels of localisation that are sometimes observed. Note the co-location of auto parts producers and auto producers in the region surrounding Michigan. The existence of auto producers in Michigan (along with trade costs) made it too costly for auto parts producers to locate anywhere else. Once the auto parts producers are located in the region, the auto producers have an additional reason to stay in this location. It is in this way that intermediate goods trade magnifies the results of the basic model. A higher wage gap between core and periphery can be maintained, as can higher tax rates. However, if a shock is large enough to induce firms to leave (i.e. low wage manufacturing in China, lower costs of trading auto parts over distance), the implications for the core are more severe.
As an aside, let me note that I believe that many of the current political and economic tensions in the US have to do with the unravelling of the core that was once the manufacturing belt. Lower trade costs, rising incomes in the rest of the world, the emergence of low-wage manufacturing in Asia – all these could result in the unravelling of the belt. Indeed, manufacturing activity is moving offshore, but it is also moving to other parts of the US. States that were in the core are finding it difficult to maintain more generous public spending and welfare provisions to which their citizens had become accustomed. Wages in much of the old manufacturing belt are stagnant or falling, especially when compared to the rest of the country. These tensions have to do with the unravelling of the core that was the US manufacturing belt, which is a large enough phenomena to affect political and economic choices at the national level.
China and India
The causal empirical lessons we learned from the US manufacturing belt are also useful for helping us think about China and India. Rapid growth in Chinese manufacturing is often attributed to very low wages there. While that has certainly been important, another part of the story may well be that China is in a very good position to fully exploit any scale economies that exist, be they external or internal. Not only does China have a vast domestic market, but because of lowered tariffs, it also has access to a vast world market. Thus, it seems likely that China is able to reap all the benefits of scale that the US once did, and more.
Given that this is the case, we might also expect to see certain features of the US manufacturing landscape appear in China. Just as cities in the US became so specialised that they became identified with particular products, we are seeing the same in China. For example, Qioatau, in Wenzhou province, is known as the button city, for it produces 60 per cent of the world’s buttons, and 80 per cent of the buttons produced in China. Other products, including lampshades and badminton racquets, seem to be localising in much the same way.
As in the US, it also appears that Chinese manufacturing is beginning to co-locate within particular geographic regions. At the moment, there seem to be about three main agglomerations (around Shenzhen, Shanghai and Beijing) that contain much of the activity. While these cities are distant from one another, it is notable that they are all located in China’s east – close to much of China’s population and the global marketplace.
As in China, the NEG models have policy lessons for India. Compared with China, India remains relatively rural and poor. There do not appear to be the same vast agglomerations of manufacturing activities – for which at least two policies are partially responsible. First, India has had an explicit policy of retaining some sectors for small firms. This policy limits the ability of Indian firms to achieve scale economies. Fortunately, this is rapidly being changed.
A second policy constraint that limits agglomeration is a number of interstate trade barriers that limit trade within India. Faced with such restrictions, it can be more appropriate to produce in multiple states than to produce in a core region and export to the periphery.
Despite these tensions, there remain great prospects for further development in India and China. China seems to have already begun exploiting the scale economies that can be achieved in large-scale agglomerations of manufacturing activity. It may well be that a Chinese core will remain competitive even after Chinese wages rise. One big threat to such a core will be India, which also has the potential to be a core manufacturing centre that serves the world.
References
Head, Keith, and Thierry Mayer. (2004) “The Empirics of Agglomeration and Trade,” Handbook of Regional and Urban Economics: Cities and Geography, Volume 4, Amsterdam: North-Holland, 2004.
Krugman, Paul. (1991) Geography and Trade, MIT Press, Cambridge.
1 http://nssdc.gsfc.nasa.gov/planetary/image/earth_night.jpg
2 Incidentally, much of Canada’s manufacturing activity lies just over the border from this region.
