Over the past 60 years, increasing European integration has brought peace and security, besides contributing to large social welfare gains (through lower prices and a larger variety of products). Still, national borders matter a lot within Europe and a vast literature has documented the large negative effect of national borders on trade – also known as the border effect (Nitsch 2000, Head and Mayer 2000, Anderson and van Wincoop 2003, de Serres et al. 2001, Chen 2004).
There are several reasons why national borders impact negatively on trade, including language and cultural differences as well as other trade frictions. However, while virtually all studies in the area have found that these factors are important, they are insufficient to fully explain the border effect (see e.g. Anderson and van Wincoop 2003). Poor international transport networks could be another factor hindering international trade. As more than 70% of merchandise trade among European countries is transported by road, the effect of poor international road links could be particularly harmful for inter-European trade.
The workhorse model to estimate the border effect on trade is the gravity equation, whereby trade between two locations is increasing in their GDP (or population) and decreasing in the great circle ('as the crow flies”) distance between the locations. The former is assumed to proxy potential trade opportunities, and the latter, the transport costs between the two locations.
One problem with this approach is that goods transported by road rarely fly like crows, but rather use existing road networks. Another problem is that the speed of transport is a key component of transport costs (as Benjamin Franklin wrote, time is money). If international road connections systematically have lower quality by being longer and slower than national ones, gravity estimates based on great circle rather than road distance can only explain part of the border effect.
To correctly infer the quality of the main national and international road connections, we downloaded road distances and travel times between 220 continental European cities – the ten most populous cities in each country – from Bing Maps Route Service (Braconier and Pisu 2013). The data for the 48 180 city pairs reveals that road distance exceeds the great circle distance by around 30%, on average (see Figure 1).
Figure 1. Intercity road and great circle distance in continental Europe
Source: Bing Maps Route Service.
More importantly, simple regression analysis suggests that national road connections are 10% shorter than international ones, given the size of cities and great circle distance between them. National road connections are more than 15% shorter than international ones for the Baltic countries, Italy, Greece and Bulgaria, but less than 5% shorter for Germany, France and Austria (see Figure 2). Austria is the only country for which the length difference between national and international road links is not significant at the 5% level. Thus, everything else equal, European road networks provide longer – and therefore more costly – connections between countries than within them.
Figure 2. Country-specific road-distance difference between national and international connections
Source: Braconier and Pisu (2013).
International road links are not only longer; travel speeds are also slower, further increasing international transport costs with respect to national ones. Travel speed (time) for national road links is around 5% higher (shorter) than for international ones, controlling for city populations and road distance. Figure 3 shows country-specific speed differences between domestic and international road trips. Speed differences are even more heterogeneous across countries than road-connection lengths. For the majority of countries the speed on national links is higher than on international ones. For Luxemburg and the Netherlands the opposite is true, probably because in these (small) countries – which are key European transit corridors – international transport will use motorways to a larger extent than in other countries.
Figure 3. Country-specific estimates of the speed difference between national and international connections
Source: Braconier and Pisu (2013).
Using the distance between the 48 180 city pairs, we computed population-weighted road distance and travel time within and between countries, capturing national and international transport costs respectively. These population-weighted distances take account of the relative importance of cities as arrival and destination points, thus gauging shipping costs more reliably than the simpler measures, such as the distance between capital cities (Head and Mayer 2010).
We then use these new road distance and travel time measures to estimate traditional gravity equations of bilateral trade flows. The estimates based on road distance yield border effects that are roughly 15% lower than estimates based on traditional great circle distances. Remaining border effects are largest for countries with longer international road connections in comparison with national ones, and for industries whose products are more costly to transport by road. Accounting for travel time lowers the border effect by an additional 10%.
From a policy perspective these findings suggest that improving international road connectivity can significantly enhance European trade and market integration. Enhancing the quality of international road connections to the same level of national ones could raise international trade by more than 15%, according to these estimates, yielding large welfare gains.
However, several caveats apply. First, the coincidence of national borders and geographic obstacles – for instance in terms of rivers and mountains – may render international links inherently more costly to build than domestic ones. Compensating fully for such disadvantages with more road transport investment is not socially efficient. Second, the results presented in this study indicate that the relative spending on cross-country links with respect to national one could be increased, but are mute on whether road investment is too low or too high. Finally, implementation and coordination problems and persistent optimism-bias in infrastructure cost-benefit analyses is considerable, and could be even bigger in international projects. This calls for careful evaluation of cross-European road projects to ensure that benefits outweigh building costs.
Authors's note: This column reflects the views of the authors writing in their personal capacity.
Anderson, J E, and E van Wincoop (2003), “Gravity with Gravitas: A Solution to the Border Puzzle”, The American Economic Review, Vol. 93(1), 170–192.
Braconier, H and M Pisu (2013), “Road connectivity and the border effect: Evidence from Europe”, OECD Economics Department Working Paper, No. 1073, OECD Publishing
Chen, N (2004), “Intra-national versus International Trade in the European Union: Why Do National Borders Matter?", Journal of International Economics, Vol. 63(1), 93–118.
De Serres, A, P Hoeller and C de la Maisonneuve (2001), "The Width of the Intra-european Economic Borders", OECD Economics Department Working Paper, No. 304, OECD Publishing.
Flyvbjerg, B (2009), “Survival of the Unfittest: Why the Worst Infrastructure Gets Built – and what We Can Do about it”, Oxford Review of Economic Policy, Vol. 25, No. 3.
Head, K and T Mayer (2000), “Non-Europe: The Magnitude and Causes of Market Fragmentation in the EU”, Weltwirtschaftliches Archiv, Vol. 136(2), 284–314.
Head, K and T Mayer (2010), “Illusory Border Effects : Distance Mismeasurement Inflates Estimates of Home Bias in Trade'”, in Brakman, S. and P. van Bergeijk (eds), The Gravity Model in International Trade: Advances and Applications, Cambridge University Press.
Nitsch, V (2000), “National Borders and International Trade: Evidence from the European Union”, The Canadian Journal of Economics / Revue canadienne d’Economique, Vol. 33(4), 1091–1105.