VoxEU Column Industrial organisation International trade

The world is not yet flat: How changes in transport costs and trade exposure shape geographical location patterns

Transport costs fell precipitously during the last century leading many observers to posit that the world has ‘become flat’. If this were true, the costs of transporting goods should no longer have much bearing on firms’ location choices and the spatial structure of economic activity. This column, using manufacturing data for Canada from 1990 to 2008, argues that despite a decline in geographical concentration of industries, location patterns still change with fluctuations in transport costs.

Transport and communication costs fell precipitously during the last century (see Glaeser and Kohlhase, 2004), leading many observers to posit that the world has ‘become flat’, that locational differences no longer matter (Cairncross, 1991). With the help of cheap transport and communication, business can be done almost anywhere, or so the story goes, leaving policymakers with the impression that we have entered ‘brave new frictionless’ world. But, if this were true, the costs of trading and transporting goods should no longer have much bearing on firms’ location choices and, thereby, the spatial structure of economic activity.

Changes in the costs of trading goods across space affect the concentration patterns, especially at short distances. This suggests that the micro-geographic structure of an economy—the clustering of firms in industries at a short distance—is influenced by changes in the trading environment

This debate about the magnitude of transport costs and their importance for economic outcomes may be viewed as a particular instance of the more general fallacy that consists of equating ‘low’ with ‘unimportant’.1 On two counts, the empirical evidence challenges this fallacy.

  • First, the tendency for economic activity to cluster in space is still strong. Many industries nowadays do exhibit strong location patterns, including for the entry of new firms that should face little locational constraints.

This runs counter to the view that firms are footloose. In a nutshell, the costs of trading goods still seem important, and location still seems to matter to a great extent (Rosenthal and Strange, 2003; Duranton and Overman, 2005; Ellison, Glaeser, and Kerr, 2010).

  • Second, and contrary to popular belief, it has been documented that trade costs—broadly defined—remain large. Anderson and van Wincoop (2004), for example, estimate that trade costs amount to a tariff equivalent of 117%, a large number by any metric.

Trade costs and clustering: Is there a link?

This is the question we ask in our recent work (Behrens et al. 2015). We use detailed microgeographic manufacturing data for Canada from 1990 to 2008 to document changes in patterns of geographical specialisation. We study how those changes are related to:

  • Changes in the costs of shipping goods across space,
  • Exposure to international trade, and
  • Shifting landscapes of access to domestic suppliers and clients.

We find strong evidence that location patterns change with fluctuations in all three aforementioned components. Thus, our results suggest the world is not yet flat. Transport costs continue to matter!

To establish that result, we first measure the geographical concentration of manufacturing industries in Canada and document changes in that concentration. We find that the extent of geographical concentration has declined from 1990 to 2008 (see also Behrens and Bougna, 2015). For the ten most concentrated industries in Canada, on average, about 35% of plants in 1990 were located less than 50 kilometers from each other, but by 2008 that figure had dropped to 21%. Although changes are less pronounced when concentration is measured in terms of either employment or sales—thus showing that clustering is partly driven by bigger and more productive firms—there is nevertheless a clear downward trend.

While these results suggest the world has become flatter, we show that this trend is associated with changes in the costs of trading goods across space. To illustrate this, Figures 1 and 2 below—which depict the coefficient estimates and 95% confidence bands for the impact of trucking costs and of the Asian share of imports on the measures of geographical concentration in Canada—show that the spatial effects of transport costs and Asian imports affect location patterns in a systematic way.2 As can be seen, the effects are especially strong at short geographical distances (less than 100 kilometer). In a nutshell, changes in the costs of trading goods across space affect the concentration patterns, especially at short distances. This suggests that the micro-geographic structure of an economy—the clustering of firms in industries at a short distance—is influenced by changes in the trading environment. Of course, the obvious next question is how strongly.

Figure 1. Coefficient estimates for the impacts of trucking costs on the geographical concentration of manufacturing industries in Canada, 1990-2008

Figure 2. Coefficient estimates for the impacts of the Asian share of imports on the geographical concentration of manufacturing industries in Canada, 1990-2008

We find that if imports from low cost Asian countries had remained at their 1992 levels in Canada, the geographical concentration of industries would have, on average, decreased by 60% less (or about 9 percentage points). The corresponding figure for changes in trucking costs is 20% (or 5 percentage points). Had trucking costs not decreased between 1992 and 2008, we would have observed more geographical dispersion. Table 1 summarises our results and reveals that changes in transport costs, in import exposure, and in access to clients and suppliers all have a sizable effect on changes in location patterns.

Table 1. Counterfactual changes in geographic concentration holding measures of trade costs constant

With this evidence in hand, the key message of our findings is that changes in the geographical concentration of industries due to changes in transport costs, in international trade exposure, and in access to clients and suppliers are all large. The world may be getting flatter, but we are not yet quite there. The lesson for policy makers is that small changes in trade costs—stemming from trade agreements—or in transport costs—due to infrastructure projects—or in any other policy that changes the costs of trading goods across nations and regions, still impact the economic geography of industries. They may impact them especially strongly in a world where firms compete globally and where any locational advantage drives profit margins—and thus locational incentives—to a sizable extent.

References

Anderson, James E., and Eric van Wincoop. 2004. "Trade Costs." Journal of Economic Literature, 42(3): 691-751.

Behrens, Kristian, and Théophile Bougna. 2015. “An anatomy of the geographical concentration of Canadian manufacturing industries.’’ Regional Science and Urban Economics 51: 47-69.

Behrens, Kristian, Théophile Bougna, and W. Mark Brown. 2015. "The world is not yet flat: Transport costs matter!" CEPR Discussion Paper 10356, Centre for Economic Policy Research, London, UK.

Brown, W Mark and William Anderson. 2015. "How thick is the border: the relative cost of Canadian domestic and cross-border truck-borne trade, 2004–2009." Journal of Transport Geography, 42: 10-21.

Cairncross, Frances. 2001. The Death of Distance: How the Communications Revolution Is Changing our Lives, Cambridge, MA: Harvard Business Review Press.

Duranton, Gilles, and Henry G. Overman. "Testing for localization using micro-geographic data." The Review of Economic Studies 72, no. 4 (2005): 1077-1106.

Ellison, Glenn, Edward L. Glaeser, and William R. Kerr. "What Causes Industry Agglomeration? Evidence from Coagglomeration Patterns." American Economic Review, 100: 1195-1213.

Glaeser, Edward L., and Janet E. Kohlhase. 2004. "Cities, regions and the decline of transport costs." Papers in Regional Science, 83(1): 197-228.

Rosenthal, Stuart S., and William C. Strange. 2004 "Evidence on the nature and sources of agglomeration economies." Handbook of regional and urban economics 4: 2119-2171.

Footnotes

1 This is highly reminiscent of the ‘kaleidoscopic comparative advantage’ debate in international trade. As Jagdish Bhagwati neatly put it: “[...] I was arguing that we now had “kaleidoscopic” comparative advantage – what we call in economic jargon, “knife-edge” specialisation – so that specialisation would shift among countries with small changes in cost conditions. The factors that had produced this situation were several, e.g. interest rates were less unequal across countries with integrated capital markets; technology used by multinationals located in different countries became more available across nations; the spread of technical education also meant that many in India and China read the same textbooks as Americans and Europeans; and so on. So, with kaleidoscopic (or “thin” or “knife-edge”) comparative advantage in many activities, we were now confronted with volatility in, not the end of, comparative advantage.” (Jagdish Bhagwati, “Why the world is not flat”, 2010)

2 We also document that the access to suppliers and clients—constructed using plant-level distances and national input-output shares— is strongly associated with changes in the industries’ geographical concentration. See Behrens, Bougna and Brown (2015) for additional details.

2,205 Reads