serious, irrevocable loss in industrial capacity for the continent as a whole. During the last ®ve years, fragmentary data have been presented for a few countries (most notably Ghana) but these cover relatively short time periods and are not comprehensive. The World Bank's own Regional Programme on Enterprise Development (REPD) has in recent years been by far the largest research activity looking at industrial development in SSA in the context of adjustment with surveys of up to 200 manufacturing enterprises in four countries in SSA having already been undertaken. However, since the focus of the research has been on microeconomic issues and, as yet, only very limited time-series data are available, no ®rm conclusions can be reached on de-industrialization. This is nonetheless recognized as being a potentially key issue in some of the RPED countries, including Cameroon and Ghana and will therefore `be examined in future' (World Bank, 1996a, p. 45).
The paucity of data in support of the de-industrialization is a point taken up by the authors of the World Bank's high pro®le research report, Adjustment in Africa: Reform, Results and the Road Ahead (AIA) which was published in 1994. The Report argues that to make the case for de-industrialization in Africa, one would have to show that (a) there are signi®cant declines in industrial output, output shares of GDP and employment because of policy reforms, (b) the declines in output and employment are more than temporary outcomes of ecient adjustment, (c) the changes in the industrial sector are not shifting the economy toward greater eciency; and (d) the policy reforms are impeding long-run industrial growth and transformation by inhibiting the pace of investment and thus the shift to a higher growth trajectory' (AIA, p. 150). 1 Since the time-series data needed to test these propositions are not available and, in most countries, it is probably too soon to ascertain whether any contractions in the industrial sector are temporary', the Report argues that it is impossible to resolve the de-industrialization debate conclusively' (ibid.). Nonetheless, on the basis of the limited data that were assembled for the AIA Report, it is stated that there is no systematic evidence' for de-industrialization among the 29 countries that form the basis of the study. 2 By comparing rates of industrial and manufacturing growth in the pre-adjustment period (speci®ed as 1981±86) with those that prevailed during the adjustment period' (1987±91) that is scrutinized in detail, AIA concludes that countries that made large improvements in their macroeconomic policies had strong increases in the growth of industry and manufacturing Ð with median increases close to 6 percentage points between 1981±86 and 1987±91. The response was weaker in countries with small improvements in macroeconomic policies and weakest for countries with deterioration in policies. Only eight adjusting countries had contractions in their industrial sectors, and ®ve of those were from the group with worsening policies' (AIA, p. 149). In fact, from the data presented in Annex Table A.20 of the Report, industrial growth was negative in only three countries (Cameroon, Cote d'Ivoire, Rwanda) during 1987±91. As can be observed in Figure 1, it is the much higher median dierences in the industrial and manufacturing growth performances of 1 Point (c) is clearly mis-speci®ed because it is perfectly conceivable that a much smaller industrial sector could help to shift the economy toward greater eciency and point (d) is merely a repetition of point (b).