Industrial policy is not dead: Shoemaking in Ethiopia

1000 or more factories are producing leather shoes in Addis Ababa. Most of them employ only 10 workers or fewer, but several factories have hundreds of workers.

In the early 2000s, China-made leather shoes flooded into the Ethiopian market plunging the local industry into a slump. Remarkably, however, the industry soon resumed vigorous growth, not only taking the market back but even finding its way into the international market.

While the majority of Ethiopian factories sell their product to domestic markets, some are exporting shoes in bulk to Italy and other developed countries as well as neighboring African countries.

I had lunch with Kei Otsuka this week. He and coauthors have taken their study of industrial clusters in East Asia to Sub-Saharan Africa. And they are hopeful about what they see. The above quote comes from a paper on Ethiopia.

What happened in Ethiopia?

A major finding is that the growth ofthis industry was driven initially by the massive entry of new enterprises established byformer employees of the existing shoe factories but more recently by the growth in enterprise sizes due to improvements in the quality of products, marketing, and management.

Such improvements were first made by highly educated entrepreneurs and subsequently followed by other enterprises. While the followers have grown in size, the leading enterprises have grown faster.

Such a development pattern appears similar to the experience of successful cluster-based industrial development in China,Taiwan, and Japan.

So what’s the secret to African industrial development?

An industry ceases to grow when the profitability of producing low-quality products falls as their market supply increases relative to marke tdemand. This is typically the case if the increase in market supply is not accompanied by improvements in product quality. By contrast, those clusters where product quality was successfully improved have continued to grow

Two things, they say, are crucial here: the absorption of technological and management knowledge from abroad, and better marketing.

This leads to at least three roles for government. First, support for investment in the managerial skills. Basically, more MBAs for Africa. Second, incentives for industrial clusters (which promote knowledge spillovers)  through infrastructure, industrial zone, and marketplace construction. Third, more access to credit for innovative enterprises for growth.

They have other examples, like the tailoring sector in Kenya. A book is due later this year, on the similarities between Asian and African industrial development.

4 thoughts on “Industrial policy is not dead: Shoemaking in Ethiopia

  1. Who is authoring the forthcoming book on the similarities b/w Asian & African industrial development? Kei Otsuka?

  2. >> “more MBAs for Africa.”

    How do you make sure that people don’t just get paper degrees for paper degrees subsidies?

    Also, in the Eastern block there was plenty of education but outcomes were not particularly stellar. Dumping education (incentives issues aside) on a population subjected to stupid policy is not productive.

    >> “Third, more access to credit for innovative enterprises for growth.”

    But you don’t know which are the “innovative enterprises” and you probably don’t want a bureaucrat picking.

    This might sound like a stupid, principles of macro critique, but if A is low, dumping K and H in an economy doesn’t help. And if A is high, K and H accumulate on themselves.

  3. I’ll have to read the article sometime this weekend, but the post it self kind of confuses me. First, why are we’re comparing Asia and Africa here? The lessons learned from this Ethiopian case seems like standard development advice, mostly simply broken down as “improve two of your basic inputs to production, namely here labor and capital, and increase access to credit.” The East Asian case is a very particular kind of industrial development, where a centralized bureaucratic state, fairly insulated from popular politics, was able to pick “industrial winners,” usually in heavy industries, and then make them national champions through various policies — subsidies, protections from outside competitors, and export policies that helped increase domestic companies’ competitiveness (many of those policies wouldn’t be stood for in the WTO now, and especially not by countries who aren’t critical to a Cold War fight, or something analogous in the 21st century). The key here is the strength of the centralized state, and its ability to shrewdly pick winners, develop those sectors, and then wean the firms off of state support. It’s a particularly unique feature of political intervention not seen in Latin America (where the aristocracy, not the meritocratic, go into the state, and then usually concoct a slew of terrible programs) or Africa (where unstable politics forces leaders to invest heavily in the military, and aim for much shorter term goals).

    Second, it seems like the focus here is on African economic clusters, which have been much written about by Michael Porter, and introduced as far back as Alfred Marshall. What’s the new thing we’re learning here that is being driven out of a fairly strange comparison between East Asia and Africa? East Asia has economic clusters, but it’s not a unique feature to the success stories there. If the comparison to East Asia only goes as far as “develop labor and capital” and “have economic clusters,” couldn’t we then make the same comparison from Africa to Britain and America? There are some lessons there, but I’m not sure how much more insight we get from what is standard development orthodoxy.

  4. Agent Continuum,

    The big problem with your macro critique is that nobody in mainstream macro really knows what A is, where its comes from, and how its development is related to K and H. Perhaps it suggests that the framework is not really very useful at all for an analysis of economic development.

    But anyway, to stay with you, isn’t “the absorption of technological and management knowledge from abroad, and better marketing” exactly about improving A? Holding A constant because a way to accomplish this is by having better educated managers – i.e. improving H – is not at all an effective critique of Blattman’s point above, but rather a showcasing of the severe limitations of the analytical framework.