Export-led growth or growth-led exports? Western Europe in the "golden age".

• Author: Boltho, Andrea

1. Introduction

   One remarkable economic feature of the 20th century was the exceptional growth of Western Europe (and Japan) between the early 1950s and the early 1970s. This period has often been labelled the "Golden Age", in contrast to the mediocre growth performances of the decades that preceded and followed it. Most explanations of why such a growth explosion occurred stress the importance of catch-up, as Europe's relatively developed economies combined their surplus labour with technologies that had been pioneered in the United States in the first half of the century (Maddison, 1964; Kindleberger 1967, Abramovitz, 1989). This led to high investment rates as well as to rapid growth of both employment and private consumption. Additional inputs into the explanation have stressed the importance of a stable international environment and of rapid trade liberalization (Maddison, 1982), of domestic institutions supporting economic growth (Eichengreen, 2007) and even, if more tentatively, of the successful operation of demand management in smoothing economic cycles and inspiring confidence in continued expansion (Boltho, 1982).

   The econometric evidence in support of a catch-up explanation is robust. Limiting the sample to Western Europe and to the period 1950-73, even the simplest of specifications manages to "explain" nearly 70 per cent of the variance in the GDP per capita growth rates of 16 countries. For productivity growth, over the same period, the results are even stronger, with an adjusted /?2coefficient above 0.75 (Figure 1). Such a cross-section approach, however, does not, allow for individual economies' idiosyncratic behaviour. Less formalized investigations have thus emerged which have stressed other features specific to particular countries or periods. Thus, reasons for Britain's relatively slow growth have been attributed to a variety of reasons, going from too large a public sector (Bacon and Eltis, 1976), to endemic problems with industrial relations (Cairncross, 1992), to a long-run squeeze on profits induced by labour militancy (Glynn and Sutcliffe, 1972), or even to the operation of "stop-go" policies (Surrey, 1982); reasons for Germany's very rapid growth to the destruction wrought by World War II or to the reforms introduced during the reconstruction period which ushered in the Soliale Marktwirtschaft (Hennings, 1982; Abelshauser, 2004), etc. One particular approach, following Beckerman's pioneering work (Beckerman, 1962), has stressed the leading role of exports in stimulating economic growth. And it is true that export growth over the 1950-73 years was more rapid than that of any other final domestic demand component in 14 out of 16 West European countries. (1)

   The expression "export-led growth" has, thus, become very popular, suggesting, as it does, an almost painless road to prosperity which can be followed by developed and developing economies alike. Yet, despite its popularity, investigations as to whether export-led growth actually occurred in the advanced economies (as opposed to the developing ones) seem to be both relatively scarce and fairly inconclusive. This brief paper will consider whether evidence for the hypothesis can be found for Western Europe in the 1950s and 1960s. Section I looks at various possible approaches, Section II examines some tentative results. The Conclusions, unsurprisingly, attempt to conclude.

2. Approach

   Three major approaches have looked at the interactions between export and GDP growth. The oldest one goes back to Marshall who argued that for economic progress to occur a country needed, inter alia, 'access ... to markets in which it can sell those things of which it has a superfluity' and added: 'The importance of this last condition is often underrated; but it stands out prominently when we look at the history of new countries' (Marshall, 1956, p.556). What Marshall had in mind, of course, was the experience of the United States and of the British Dominions, areas blessed by the presence of vast natural resources, which benefited from the growth of European demand for food and raw materials in the 19th Century. This approach has been formalized in so-called "vent for surplus" models (Caves, 1965), but is one that is hardy applicable to the developed West European countries which are considered here.

   A modern reformulation of the Marshall thesis, still stressing the demand side, would argue that a favourable exchange rate, or an auspicious specialization in products in high demand on the world market, or geographic proximity to countries growing very rapidly, could boost export demand and profitability, as well as generating a trade surplus. This would relieve balance of payments constraints (still widespread in the Europe of the 1950s and 1960s) and allow imports of needed capital equipment. More importantly, buoyant exports would set in motion multiplier and accelerator effects at home which would raise demand and promote investment; economies of scale would be exploited, generating further gains in competitiveness, and thereby launching a virtuous circle. These growth promoting forces would be strengthened, in a world of fixed exchange rates as, by and large, was that of the "Golden Age", by the rising confidence which a favourable balance of payments position would inspire in economic agents. Nor would the monetary consequences of continuing external surpluses undo this mechanism since capital controls and sterilization policies would limit any rise in the money supply and hence in domestic inflation. (2)

   In an alternative view, however, it is the supply side that receives pride of place. Exports succeed in the world because domestic growth is rapid, new products are put on the market in response to buoyant internal demand, economies of scale are achieved and external competitiveness improves: thus, a country's exports grow rapidly because of 'the innovative ability and adaptive capacity of its manufacturers' (Kaldor, 1981, p.603); and/or because 'fast growing countries expand their share of world markets ... by expanding the range of goods that they produce' (Krugman, 1989, p.1039). Trade, according to these views, is not the primary engine of growth, but its "handmaiden" (Kravis, 1970). Growth is not "export-led". Instead it is exports that are "growth-led".3

   The most common approach to look for the presence or absence of export-led growth has been that of applying Granger-causality tests to time series. A large number of studies have applied this methodology to the data of many developing countries and of a few staple-exporting OECD countries such as Australia or Canada (see, for instance: Bodman, 1996). Overall results are mixed, with plenty of examples of countries whose experience fits the hypothesis and, equally, numerous examples of the opposite. A particularly thorough study sceptically concludes that extreme care should be exercised when interpreting much of the applied research on the export-led growth hypothesis' (Giles and Williams, 2000). Two meta-analyses (Mookerjee, 2006; Sannassee et al, 2014) surveying 76 and 447 studies respectively provide a broadly favourable response for developing countries--either manufactured exports or total exports tend to lead, more usually than not, to higher overall growth in emerging economies, particularly in those that are most developed.

   Surprisingly perhaps, there are fewer investigations of this kind for the industrialized West European countries considered in this paper and even here the few that are available have often explored the experience of slightly less advanced economies such as those of Greece, Ireland or Portugal (e.g: Panas and Vamvoukas, 2002; Fountas, 2000; Oxley, 1993), with causality running from exports to output in Ireland over the years 1981-94, but reverse causality, from output to exports, present in Greece (1948-97) and Portugal (over the very long 1871-1985 period).

   For the more developed countries the fairly generalized conclusion is that such tests have seldom succeeded in proving the hypothesis. Over the long 1950-85 or 195090 time spans two studies have found either no causal relationship at all going from exports to GDP for any of 13 industrialized countries of Western Europe (Afxentiou and Serletis, 1991), or only limited evidence for some, depending on specifications (Riezman and Whiteman, 1996), while weak evidence for export-led growth was found for Italy (1951-2004) (Pistoresi and Rinaldi, 2012). For more recent periods, the results obtained are less negative, but even here positive results in some cases are then contradicted by negative ones in others. For the years 1960-87, for instance, exports seem to lead GDP growth in Germany (Sharma et al, 1991) and manufacturing productivity growth in Germany and the UK (Marin, 1992); similarly, Belgium, Denmark, Italy, Spain and Sweden appear to fit the hypothesis in the period 1960-97 (Konya, 2006), Spain in the years 1959-99 (Balaguer and Cantavella-Jorda, 2001), Finland in 1965-85 (Pomponio, 1996), France and Germany in 1970-87 (Kugler, 1991) and Italy in the years 1975-97 (Yamada, 1998). At the same time, however, there are negative results for Switzerland and the UK (1960-97) (Konya, 2006) (confirmed by Kugler (1991) for 1970-87), Italy (1960-87) (Sharma et al, 1991) and again for 1960-92 (Anwer and Sampath, 2000), Austria (1965-85) (Kunst and Marin, 1989), Austria,

   Denmark, Germany, France, Italy and Norway (1965-85)...