OECD Export Performance
How Important are Costs?

There is growing concern about the export performance of European industries in the face of increasingly fierce global competition. Yet, according to Wendy Carlin and John Van Reenen (both University College London and CEPR), the statistical basis of the relationship between relative costs and export-market shares is still poorly understood. Speaking at a CEPR lunchtime meeting on 21 January 1998, Carlin and Van Reenen posed two core questions: (1) How much do costs matter in explaining export performance in the 12 key industries in the 14 main OECD countries? (2) What determines the sensitivity with which exports react to cost changes?

The speakers argued that the answers to these questions had implications for the functioning of economic and monetary union (EMU) in Europe, and for the design and scope of industrial policies. With regard to the first question, Carlin and Van Reenen maintained that there are 'deep structural' factors that enable countries to maintain their share of export markets, and that these are ignored when attention is focused solely on labour costs. Recent work on economic growth had emphasised the importance of a range of structural and institutional factors. Similarly, long-run success in export performance is dependent on factors such as a country's education system and structure of corporate governance, which must be considered alongside conventional economic variables.

This was not to say that relative unit labour costs do not matter for exports. Over the longer run, a 10% improvement in costs – whether from wage restraint, productivity improvement or exchange-rate devaluation – leads to a 2-3% improvement in export-market share. Even when costs are netted out, however, some nations perform surprisingly well. West Germany, for example, managed to increase its market share slightly between 1970 and 1992, despite the fact that its costs rose nearly 1% per year faster than its competitors.

Clearly, technical progress, as embodied in new investment, also has an important role to play. Carlin and Van Reenen suggested, however, that there were two other factors that explained the ability of countries like Germany to overcome rising relative costs: high levels of schooling, and strong ownership concentration. Thus a more educated workforce appeared to enable faster improvements in the quality of manufactured products; the export performance of individual industries appeared to be boosted by improvements in efficiency in the broader business sector; and committed owners appeared to confer an advantage on exporters – for example, through facilitating long-term relationships with suppliers.

Turning to the second question, Carlin and Van Reenen noted that sensitivity of exports to costs differed in predictable ways across industries. For example, in high-tech industries the elasticity was much lower. Yet, despite the growth of such industries, costs actually had become more important for determining exports over time for most industries. This was probably because of the increased levels of global competition, since the industries with the biggest increases in cost sensitivity were also those which had faced the stiffest increases in global competition.

In conclusion, the speakers argued that underlying trends in export-market shares could be disruptive inside a monetary union. A member country with poor underlying export trends would find it necessary to achieve a lower growth rate of unit labour costs than its neighbours. In general, it appeared that in Europe the core EMU countries had less export sensitivity to costs than non-EMU countries, like the United Kingdom and Sweden. This helped provide an economic explanation for the reluctance of these countries to participate in monetary union, where they can no longer use what, for them, might be the potent weapon of exchange-rate changes.

Wendy Carlin, Andrew Glyn and John Van Reenen, 'Quantifying a Dangerous Obsession? Competitiveness and Export Performance in an OECD Panel of Industries', CEPR Discussion Paper No. 1628, April 1997