Is faster trade more trade? Evidence from EU enlargement

Cecília Hornok 09 July 2011



Costs of international trade are still large and come in a variety of flavours. Total trade costs for a developed country are around 170% of the traded value (Anderson and Van Wincoop 2004), of which only a minor share are due to trade-policy instruments like tariffs or quotas. The focus of trade economists and policymakers has shifted recently towards the indirect costs of trade, such as those related to the quality of the transport infrastructure or the regulatory and administrative environment. A curious feature of some of these trade barriers is that they are better described by the amount of time lost in the trading process than by pecuniary costs.

Time as a barrier

There is at least one reason why time barriers to trade might have recently become more painful to trading firms. International production is increasingly fragmented (Feenstra and Hanson 1996, Hummels et al. 2001), and when stages of production are located across the border, it is essential that trade linking stages is synchronised in a timely manner. Any delay in deliveries can potentially hold up the whole production process and cause serious losses for the producer. If a firm wants to be flexible in responding to changing consumer tastes, it will only source its intermediates from – or locate stages of production to places where – fast trade is possible. It may do so even at the expense of paying higher wages and prices (Evans and Harrigan 2005). Consequently, a sufficiently large fall in the time required to trade with firms in a low-wage country can lead to the redirection of trading relationships, reorganisation of international logistics systems, or complete relocation of plants.

Faster trade in an expanded EU

In a recent study (Hornok 2011), I find evidence that the cost of time in trade is large and better “timeliness” can have strong positive trade effects. My evidence is gathered from the enlargement of the EU in 2004 and its effect on trade with the eight Central and Eastern European countries (Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovakia, and Slovenia). The decline in time costs comes from the elimination of the time-consuming border and customs procedures for these countries once they enter the enlarged EU. In these trading relationships free trade in terms of direct trade policy tools had already been established for most manufactures years before 2004. Hence, the episode of EU enlargement serves as a reasonably good case for an experiment to study the effects of other trade barriers, such as time.

Trade within the EU accelerated for the eight new members after 2004. But how much of it was because of falling trade costs and – more specifically – falling time costs? To infer total trade costs from trade flow data, I use a theory-based bilateral trade cost index following Novy (2010) and Head and Ries (2001). The index is built on the ratio of cross-border trade to domestic trade between two countries. This means that it captures trade costs in a broad sense, i.e. it accounts for all factors that make cross-border trade fall short of domestic trade. Figure 1 shows that there was a substantial decline in the bilateral trade cost index after 2004 for country pairs with new members (“new with new” and “old with new”). In contrast, trade costs were relatively stable within the pre-enlargement EU (“old with old”).

Figure 1. Bilateral trade cost index

Note: Simple averages of trade cost indices across country pairs and 19 two-digit NACE manufacturing industries, excluding food and energy. Logarithmic scale, normalised to 2003.

Isolating the effect of time

To capture the time-related elements of this decline, I use a triple-differences estimation strategy. If the declining time costs have played an important role, the fall in the trade cost index for country pairs involving a new member would need to be stronger than the change in trade costs between country pairs of the pre-enlargement EU, particularly for time-sensitive industries. Comparing the differences-in-differences estimates for time-sensitive and for non-sensitive industries can be indicative of the importance of time costs in the total enlargement-induced trade cost decline.

But how to distinguish industries in which the time cost matters a lot from those where it matters less? One possibility is to borrow the results of Hummels (2001), who estimates by industry the premium firms in the US are willing to pay for air (instead of sea) transportation to save time. Time-sensitive industries turn out to be those at the higher end of the industry classification, i.e. machinery and transport equipment. According to evidence on six developed EU economies in Breda et al. (2008), these industries are also among the ones with the highest import content of export, a statistic widely used for measuring international production fragmentation.

Indeed, the differences-in-differences estimate is twice as large for the time-sensitive industries than for the time-insensitive group of industries. Trade costs fell by 2.4% in ad valorem terms for time-insensitive industries and by 5.0% for time-sensitive ones. These figures translate into a 17-percentage-point greater growth of foreign trade in time-sensitive industries than in other industries.

How is the estimated improvement in timeliness related to the abolition of border and customs control? I use country-pair-specific information on the enlargement-induced decline in the border waiting time and check whether the additional trade cost decline for time-sensitive industries is larger for country pairs where border waiting time fell more. The decline in border waiting time is determined by data on the level of pre-enlargement waiting time (Figure 2) together with the assumption that waiting time was completely eliminated with EU enlargement.

Figure 2. Waiting time at border crossings before EU enlargement

Note: Data source is International Road Union. Simple averages of waiting hours at different crossing points and years in 2000-2002.

My results support the hypothesis that a larger decline in the border waiting time comes with significantly larger decline in the trade cost index of time-sensitive industries. Numerically, I estimate that each additional hour of waiting time adds 0.8 percentage points to trade costs in proportion to the traded value in time-sensitive industries. In terms of trade creation, this amounts to about 5% more foreign trade for an hour’s reduction in waiting time. At the same time, the decline in the border waiting time has practically no effect in industries that are not sensitive to time.

The measured effect is no doubt large for a single hour improvement in the trading time. We should bear in mind, however, that it may not reflect only the direct costs of waiting (product deterioration, e.g.), but it also may involve the cost of uncertainty about the delivery time, a cost that can get large if trading is frequent and production stages are fragmented. Moreover, border waiting time can also proxy other time-intensive trade barriers that declined with EU enlargement such as other administrative or informational barriers.

All in all, the message is clear. The case of EU enlargement suggests that time plays an important role in foreign trade and removing time-intensive trade barriers pays off. The lesson especially applies to industries that are more sensitive to timeliness, which appear to be the more technologically advanced industries in general.


Anderson, JE and E van Wincoop (2004), “Trade Costs”, Journal of Economic Literature, 42:691-751.

Breda, E, R Zappariello, and R Zizza (2008), “Vertical Specialization in Europe: Evidence from the Import Content of Exports”, Banca D'Italia Working Papers No. 682.

Evans, C and J Harrigan (2005), “Distance, Time, and Specialization: Lean Retailing in General Equilibrium”, American Economic Review, 95:292-313.

Feenstra, RC and GH Hanson (1996), “Globalisation, Outsourcing, and Wage Inequality”, American Economic Review: Papers & Proceedings, 86(2):240-245.

Head, K and J Ries (2001), “Increasing Returns Versus National Product Differentiation as an Explanation for the Pattern of US-Canada Trade”, American Economic Review, 91:858-976.

Hornok, C (2011), “Need for Speed: Is Faster Trade in the EU Trade-creating?”, CEPR Discussion Paper 8451.

Hummels, D (2001), “Time as a Trade Barrier”, GTAP Working Paper No. 18.

Hummels, D, J Ishii, K-M Yi (2001) “The Nature and Growth of Vertical Specialization in World Trade”, Journal of International Economics, 54:75-96.

Novy, D (2010), “Gravity Redux: Measuring International Trade Costs with Panel Data”, Working paper, University of Warwick.



Topics:  EU policies International trade

Tags:  free trade, protectionism, Trade barriers

Postdoc Researcher, Kiel Institute for the World Economy


CEPR Policy Research