Have technology and globalization kicked away the ladder of ‘easy’ development? Dani Rodrik thinks so

Dani Rodrik was in town his week, and I attended a brilliant presentation at ODI. Very exciting. He’s been one of my dani rodrikheroes ever since I joined the aid and development crowd in the late 90s, when he was one of the few high profile economists to be arguing against the liberalizing market-good/state-bad tide on trade, investment and just about everything else. Dani doggedly and brilliantly made the case for the role of the state in intelligent industrial policy. But now he’s feeling pessimistic about the future (one discussant described it as ‘like your local priest losing his faith’).

The gloom arises from his analysis of the causes and consequences of premature industrialization. I blogged about his paper on this a few months ago, but here are some additional thoughts that emerged in the discussion. He’s also happy for you to nick his powerpoint.

Dani identified two fundamental engines of growth. The first is a ‘neoclassical engine’, consisting of a slow accumulation of human capital (eg skills), institutions and other ‘fundamental capabilities’. The second, which he ascribed to Arthur Lewis, is driven by structural differences within national economies – islands of modern, high productivity industry in a sea of traditional low productivity. Countries go through a ‘structural transformation’ when an increasing amount of the economy moves from the traditional to the modern sector, with a resulting leap in productivity leading to the kinds of stellar growth that has characterized take-off countries over the last 60 years.

Manufacturing has been key to that second driver. It is technologically dynamic, with technologies spreading rapidly across the world, allowing poor countries to hitch a ride on stuff invented elsewhere. It has absorbed lots of unskilled Rodrik-figlabour (unlike mining, for example). And since manufactures are tradable, countries can specialize and produce loads of a particular kind of goods, without flooding the domestic market and driving down prices.

But that very dynamism has produced diminishing returns in terms of growth and (especially) jobs. Countries are hitting a peak of manufacturing jobs earlier and earlier in their development process (see graph). And it could get much worse – just imagine the impact if/when garments, the classic job-creating first rung on the industrialization ladder, shift to automated production in the same way as vehicle production.

At the same time, lower transport costs and globalization is making it harder and harder to pursue import substitution and use exports to boost the wider economy through local linkages – Ethiopia imports cardboard boxes from China in which it puts its cut flower exports. A return to import substitution would require much higher tariff barriers than in the past, to counter the lower transport costs and Asia’s hyper cheap production costs.

The result has been deindustrialization in much of Africa and Latin America, exacerbated by the recent commodity KATLboom (now ending). A byproduct has been rising inequality – manufacturing has become more like mining – a low employment/high wage enclave with few linkages to the rest of the economy.

There is very little sign that services or agriculture can substitute as long term drivers of labour-intensive growth – hence Dani’s pessimism. The easy road of industrial policy and manufacturing catch-up that has driven such spectacular gains, is coming to an end. It’s back to the long haul of neoclassical growth, unless something new turns up.

The panellists (Nick Lea, Dirk Willem te Velde and chair Stephanie Griffith Jones) went into straw clutchist mode, looking for possible rays of light to ease Dani’s gloom. These included:

The numbers may be better than we think: new calculations of GDP, especially in Africa, show a lot more economic activity than we realized, plus a lot of manufacturing may not be picked up in official stats.

Global Value Chains: intriguing – in one direction, companies are outsourcing more stages like design, which are then counted as services, whereas before they would have been part of manufacturing. On the other hand, companies are also taking some services in-house, such as after-sales support. Not clear which is greater, but it means that the division between manufacturing and services is a lot more blurred than it appears.

Fragile States:  Nick Lea saw possibilities of leaps in growth in the ‘non-linear, chaotic group at the bottom of the pool’

Climate Change: Stephany Griffith-Jones wondered if the new techs required for the green revolution could fill some of the gap.

Frugal Technology: she also wondered if the shift in design and innovation to countries such as India might partially reverse the trend to automation, and revive job creation.

Even the earlier peak in industrialization is still decades away for most of Sub-Saharan Africa – hitting a wall at a GDP per capita of $5,000 would be a nice problem to have for many governments (missed this in first draft – thanks to Dirk for pointing it out).

All great stuff, although despite the panelists efforts, I found Dani’s pessimism rather too convincing for comfort.

One minor whinge on the panel – 3 male speakers, and a woman chair (even if it is the wonderful Stephanie Griffith-Jones) does not pass the CGD test, and every single person called on in Q&A was (I think) male (including me, I know, my bad). Come on ODI, please get with the anti-Hoff programme!

The video of the event is also up on the ODI website.

Subscribe to our Newsletter

You can unsubscribe at any time by clicking the link in the footer of our emails. For information about our privacy practices, please see our .

We use MailChimp as our marketing platform. By subscribing, you acknowledge that your information will be transferred to MailChimp for processing. Learn more about MailChimp's privacy practices here.


14 Responses to “Have technology and globalization kicked away the ladder of ‘easy’ development? Dani Rodrik thinks so”
  1. Tom Berliner

    Its a shame that this phenomenon that Roddrik highlights so well (early deindustrialization/declining opportunities for jobs in manufacturing) isn’t getting more coverage in policy circles, press and academia.As a result, we’re nowhere near agreeing on steps forward (as Rodrik admitted in the talk). We need to develop alternative solutions, and fast. There are some proposed solutions in this ODI blog here:


  2. Allan Moolman

    Maybe the paradigm ‘jobs=growth=development’ is wrong? How is it even remotely possible the number of jobs be increased in a world where efficiency (lower labour inputs for everything) dominates? Future economies will be low labour economies, our technological progress makes that inevitable.

  3. P Baker

    I really like Dani Rodrik’s work, especially his book The Globalization Paradox.
    However, like all economists, this present analysis lacks any environmental or scientific underpinning.

    Hall & Klitgaard’s book “Energy and the Wealth of Nations: Understanding the Biophysical Economy” shows clearly (to me at least) that labour productivity increases in direct proportion to the amount of energy used per worker hour.

    The missing factor in all economists’ understanding is that it was abundant cheap energy that transformed UK, US and now China, but in ways that we now realize are completely unsustainable. China doubled its GDP 12 times faster than 19th cent UK for 100 times the number of people – it did that through prodigious use of coal and oil, pushed through by an authoritarian regime.

    As a physicist Duncan, I think you should read that book which is a model of clarity and reason and then explain how important it is to your readers.

  4. Owen Barder

    I find it very hard to be depressed about a world of rising productivity. If incomes are rising faster than jobs because of technological change and falling costs, the big picture is surely that this is welcome. (We get more incomes – and higher consumption – without having to work for it).

    Of course in real life it isn’t as simple as my previous paragraph makes it sound: a shift from labour-intensive production will effect distribution of income (and power!) and so we have to do a better job than we have in the last few decades of redistributing those gains more evenly, both nationally and internationally. But overall I think that is a nice problem to have – I’d rather confront that challenge than +not+ see gains from rising productivity and technological change. If robots can do most of the work in the future, so much he better for all of us.

    • Duncan Green

      I remember as a schoolkid being fascinated by two imminent utopias – free energy through nuclear fusion and the forthcoming age of leisure, as predicted by Michael Harrington – the first is still ‘just a few decades ahead’, and the second didn’t quite work out like that did it? As you say, power and politics got in the way. They will determine whether a high productivity economy with no jobs more closely resembles Equatorial Guinea or a Harrington-esque Utopia.

      • Owen Barder

        Duncan – I am not convinced that we will enjoy/endure an “age of leisure” because there is always work to be done. We may choose to enjoy more leisure time as we get richer because leisure is a “luxury good” in the sense that our consumption of it increases more than proportionately as our income reasons, though someone needs to tell that to the Americans.

        Going back to Rodrik’s point: we can either try to resist these changes, by pushing economies towards more labour intensive (ie lower productivity) configurations, or we can embrace the change, and concentrate on how to deal with the power and politics required to share those (overall very welcome) benefits more fairly, locally and globally. I think the first option is neither possible nor desirable.

        • Duncan Green

          Dani didn’t go into this so much, but the same thought did cross my mind. During the structural transformation period, I think the East Asians pretty clearly won the argument on government intervention, industrial policy etc, codified in the work of Dani, Ha-Joon Chang, Alice Amsden, Robert Wade and others. Governments were able to accelerate transformation through a wide variety of policies – some picking winners (South Korea), others focussing more on the enabling environment (Taiwan). But the policy prescriptions may be different if we are now back to the heavy lifting neoclassical version of growth – one path might be the one you outline, but that worries me. Without some kind of predistribution to ensure a wider group of people are driving growth, political power is highly likely to follow the concentration of wealth produced by enclaves of tech-driven growth, so the redistribution you suggest (whether of money or time) is unlikely to happen. You can’t delink the economy and politics.

          • Owen Barder

            Duncan I’m not opposed to activist industrial policy per se; but I think governments should think very carefully before trying to push their economy towards a luddite nirvana of high employment, low-productivity growth.

            I don’t agree with Dani’s basic worry that we are back to “neoclassical” growth. The type of growth that Arthur Lewis was describing comes about when a complex adaptive system goes through a transition to self-organising complexity (though of course Lewis didn’t describe it this way) and I see no reason to think that other countries cannot do this too.

            I agree that predistribution is in principle preferable to ex post redistribution – but I reckon it is likely to work better if governments try to increase skills to enable people to participate in a high productivity labour market rather than trying to predistribute by reducing the skills needed to benefit from growth by pushing productivity down.

    • P Baker

      Robots – yes wonderful, but a huge problem with this: as everything becomes more automated and optimized, the supply chains for our vital needs get ever more complex, in fact ‘chain’ is a misleading word now, web is better.

      The exponentially increasing number of components (computers, sensors, robots, satellites, etc.) each have their own chain. The total complexity will defy analysis. Increased efficiency always means reduced resilience, i.e. more susceptible to disruption.

      In an increasingly chaotic world, are you okay that most of your food will be produced by robots? Nothing can go wrong?

      • Owen Barder

        P Baker – We are indeed increasingly being supplied by a complex web of people, organisations and countries. And complex systems are difficult to analyse (they do not however “defy” analysis, as our friends in the natural sciences have discovered – though my fellow economists seem to be reluctant to adopt those tools).

        But complex systems can be resilient and self healing – more so than simple systems. (Think of the internet, for example, which by design routes around problems.)

        Complex systems are not the same as chaotic systems (if you are using “complex” and “chaotic” in their technical meaning).

        I watch my partner’s father bringing in the harvest on a combine harvester, doing work which would have taken a dozen men before. Am I OK with that? Sure I am. Lots can go wrong: but nothing we can’t fix.

        • P Baker

          Owen – Comparing our food system with the Internet is a false comparison – the latter was designed to be a resilient and redundant system by the US military that ran it initially. The former has evolved and now captured by a limited number of companies; robotics will further increase their control.

          Yes, resilience could be increasingly built into the food system, but in a free market, where short term profit is the goal, it won’t happen, any more than it has for the financial system (which is still not fixed by the way).

          When I mentioned ‘defies analysis’, of course it can be analysed to death, but it’s a complex system whose behaviour cannot be predicted, especially to black and grey swan events. Has anyone even tried to do this?

          ‘Nothing we can’t fix’ doesn’t really reassure me. The problem these days is that things don’t seem to get fixed – whether it’s the climate, the financial crisis, the Middle East (the whole of it), Greece, Calais and so on.

  5. Owen Barder

    P Baker – The internet wasn’t “designed”: it evolved, as do all complex adaptive systems. Nor was ARPANET, the precursor, designed. Individual parts of it were designed, and some of them (notably TCP/IP) were designed to bring about resilience in the system as a whole. Complex adaptive systems can evolve with resilience (and many do: think of most ecosystems). Normally the key to resilience is loose coupling, which can be a feature of evolved or of designed systems. If we want the food system to be more resilient, we should think about the system dynamics, not the degree of mechanisation.

  6. P Baker

    The differences between our food system and the Internet could not be more stark – the latter is modern and had a huge input from government to build in resilience. Quite the reverse of our food system, which is massively dysfunctional as the new Lloyd’s report states:

    “The continued globalisation of modern food networks is introducing an unprecedented level of complexity to the global food system, bringing both significant benefits and systemic risks. Disruptions at any one point in the system would be likely to reverberate throughout the food supply chain. Volatile food prices and increasing political instability are likely to magnify the impacts of food production shocks, causing a cascade of economic, social and political impacts across the globe. [Lloyd’s Emerging Risk Report – 2015 Innovation Series June 2015 30pp.]

    To fix that is going to require major change; ‘thinking about system dynamics’ just won’t cut it. Govt intervention will just deal with the effects and not the causes. Mechanization cannot be achieved by smallholders – we see this very clearly in the coffee system, where mechanization in Brazil, plus rents from deforestation drive them out of business. What will happen to them?