Let me start with an anecdote. Two days ago the rainy season began in Kenya (almost a month late) and this morning the roads in the Kariobangi industrial cluster became very muddy. A truck got stuck in one of the roads and it took almost an hour to get it out –all businesses in that road were damaged from a very simple problem.
An entrepreneur came up to me and asked “why did the Italians construct tarmac roads in the Korogocho slum and not the Kariobangi industrial cluster? Nobody cares about small industries in this country”. He was referring to a slum upgrading program financed by the Italian Cooperation to improve the roads in the nearby informal settlement of Korogocho, which was not extended to the Kariobangi Light Industries; he was evidently not happy about it.
I could not come up with a good answer at that moment, but his question reminded me of an excellent paper by Robert Wade, professor of political economy and development at the London School of Economics, which discusses the role of industrial policy in Asia and how donors completely neglected it in Africa.
I came to Addis in the summer of 2004 with Joe Stiglitz, to participate in what is called a Policy Dialogue. We talked to several groups. At the meeting with aid representatives working in Ethiopia—about 20 of them—we invited each to identify the priorities of their agencies in Ethiopia. They identified more or less the same ones, virtually all in ‘governance’ or the ‘social sectors’ like primary health and primary education. They made virtually no reference to investment in the ‘hardware’ of productive capacities. The closest any of them got was ‘rural roads’, mentioned by just one representative. No-one mentioned aid for agriculture or irrigation or manufacturing or services. When they talked of ‘improving governance’ they were not referring to improving the governance of the processes through which capital accumulation, technological progress, and diversification of production can be directly accelerated, but governance in broader terms related to ‘rule of law’, ‘property rights’, ‘anti-corruption’ and the like. They talked, implicitly, of ‘market-enhancing governance’, not ‘growth-enhancing governance’, as though the former equated to the latter.
… The agendas of the aid donors Joe Stiglitz and I talked to in Addis in 2004 could scarcely be more different from the agenda of America’s large-scale aid for Japan, South Korea and Taiwan in the postwar decades. The comparison is important because the Americans were very serious about accelerating economic development in Northeast Asia. They saw the region as the front-line in the battle against global communism and they wanted economically prospering, militarily strong, and politically stable capitalist allies on the front line. So what did the Americans direct their aid to, when they were deadly serious about accelerating development (as distinct from opening up the economies and claiming that opening was the best route to development)? US aid was targeted explicitly at the development of productive capacities, including lots of assistance for new basic industries (in metals, chemicals, petrochemicals), infrastructure like roads, electricity and water supply, technology institutes, and agriculture (irrigation, improved crop varieties, crop research institutes). US aid even supported large-scale, semi-expropriative land reforms in all three countries, in order to make the rural sector politically stable as well as to raise the productivity of the land.
Yet for the past 25 years the consensus in the donor community has been, implicitly, that the development of productive capacities will result from free markets, provided free market policies are complemented with the development of ‘market-enhancing governance’ capabilities, like those noted above, as well as by investment in primary education, primary health care and the like. The comparison makes one wonder whether the donor community has a collective interest in accelerating development as distinct from reducing poverty and opening up developing economies—which is not the same thing.
Ungated version of the paper here