Part 1 is here, and Part 2 is here.
And here’s a brief recap of the story so far. Part 1 covers what industrial policy is, and why it is important. Part 2 helps us understand that India’s industrial policy, at least in terms of outcomes, has not produced great results.
Now let’s answer the question that really and truly matters: what should India’s industrial policy look like going forward?
We argue in this essay that we are at a turning point in development strategy. Strategies that worked well in the past are unlikely to do so in the decades ahead. In particular, the manufacturing- and export- based growth strategies that drove East Asia’s development miracles are no longer suited for today’s low-income countries; at the very least, they are insufficient. New technologies, the climate challenge, and the reconfiguration of globalization require a new approach for development emphasizing two critical areas: the green transition and labor-absorbing services. Unfortunately, policy makers do not have ready-made recipes or successful models to emulate. Confronting this challenge head-on will require therefore also building greater capacity to learn about new opportunities, constraints, and what works and doesn’t as governments experiment with new policies on a number of fronts.
https://drodrik.scholar.harvard.edu/sites/scholar.harvard.edu/files/dani-rodrik/files/a_new_growth_strategy_for_developing_nations.pdf
Dani Rodrik and Joseph Stiglitz have a new essay out on just this topic, called A New Growth Strategy for Developing Nations. And they begin by telling us that what worked in the past may not work in the future. And the reason for that in this instance, they say, is because of:
- New Technologies
- The Climate Challenge
- The Reconfiguration of Globalization
New Technologies
New technologies may well make you think immediately about AI and robotics, which we have covered yesterday and the day before. But the authors aren’t just talking about AI and robotics. They are trying to make a broader point:
The primary culprit was skill- and capital-biased technological changes in manufacturing. These changes increased labor productivity substantially in the advanced economies where innovations originate. But they also undercut the comparative advantage of low- income economies in traditionally labor-intensive manufacturing. The quality and technological standards set by leading firms in GVCs rendered labor-intensive production in export-oriented sectors even less viable.6 The result was that globally competitive formal manufacturing sectors in developing countries ceased being labor-absorbing sectors. They turned into “enclave” sectors where very few of these economies’ excess low-skilled labor could be employed. In countries where manufacturing output held its own, manufacturing employment shrank (as a share of total employment). In the few cases where manufacturing employment registered an increase, the rise was concentrated in small, informal, and low-productivity enterprises, while large, internationally-competitive firms generated little demand for labor.
https://drodrik.scholar.harvard.edu/sites/scholar.harvard.edu/files/dani-rodrik/files/a_new_growth_strategy_for_developing_nations.pdf
By the way, these trends have had somewhat surprising, but with the benefit of hindsight, entirely predictable consequences:
Mr Maskin’s theory relies on what he calls worker “matching”. Unskilled workers can be more productive when matched with skilled ones—that is, when they work together. Assigning a manager to a group of workers can do more for total output than just adding another worker. He places workers into four classes: skilled workers in rich countries (A); low-skilled workers in rich countries (B); high-skilled workers in poor countries (C); and low-skilled workers in poor countries (D). Crucially, he thinks low-skilled workers in rich countries (the Bs) are likely to be more productive than high-skilled workers in poor ones (the Cs).
Before the current wave of globalisation started in the 1980s, skilled and unskilled workers in developing countries—the Cs and Ds—worked together. Mr Maskin gives the example of a rural Indian man, fluent in English, who helped local farmers understand modern agricultural methods. Wage growth of high-skilled workers (Cs) was weak, because poor transport and communication links made it hard for them to work with skilled workers in rich countries. But low-skilled workers (Ds) did well: their interactions with the Cs boosted total output, which let them demand higher wages, so pushing down inequality.
The latest bout of globalisation has jumbled the pairings: high-skilled workers in poor countries can now work more easily with low-skilled workers in rich ones, leaving their poor neighbours in the lurch. Take “intermediate goods”, the semi-finished products that account for about two-thirds of world trade. The production processes outsourced by big companies in factories or call-centres are by rich-world standards unskilled. But when jobs are sent offshore, they are snapped up by C workers, the relatively skilled ones. According to research from Cornell University, the typical call-centre employee in India has a bachelor’s degree.
Globalisation in its latest guise means such workers come into more regular contact with less-skilled people in the rich world. The Anglophone Indian cited by Mr Maskin may go to work in an export factory where he meets tight deadlines laid down by its American owners. The Cs work with Bs and end up being more productive. The Ds are left by the wayside.
https://www.economist.com/finance-and-economics/2014/08/23/revisiting-ricardo
Question: Who do you think the D’s are likely to vote for in their countries the world over? Who do you think they have been voting for these last ten years or so?
Industrial policies have consequences, children.
Tread (trade?) carefully.
Hello, services.
And because of rapid advancements in technology and the resulting changes in patterns of globalization, the good ol’ industrial policies recommended and implemented in places ranging from America in the 1770’s to South Korea in the 1950’s no longer work. Also, I should note, because of climate change, as Rodrik and Stiglitz discuss in their paper.
And so, they say, services is where it’s at. It is the service sector that will have to absorb all the young people who are going to come of age in India oer the next two decades.
But there is (of course) a problem. The problem is that the kind of service sector growth you and tend to think about (IT/ITeS/BPO/KPO) has “limited potential to create large numbers of employment for the typically low-educated, low-skilled workforce of a developing economy. These new jobs, they say, are a “hodge-podge of largely self-proprietorships or micro/small firms, typically non-tradable and often informal”.
Now this is a major problem, because as they themselves say earlier on in their essay, their entire attempt is focused on answering what they describe as their key strategic question:
Where will the better, more productive jobs come from?
The footnote associated with this question confidently asserts that “Our concern here is not with the overall quantity of jobs or with full-employment, but with the structure of jobs. Reasonable macroeconomic policies (along with “flexible” labor markets) will ensure full employment, but the resulting structure can be sub-optimal from a developmental standpoint and not growth-promoting. We focus, therefore, on the creation of more jobs in the more productive sectors of the economy as a vehicle of structural transformation and growth, analogous to the role that industrialization has played historically”.
But that, alas, doesn’t resonate with the “hodge-podge, non-tradable and often informal” service sector. And that is going to be a challenge along two fronts (at least).
First, this will have an impact on inequality. When the authors say that the best that can be hoped for is that some service sector firms will see an increase in productivity, but the vast majority will not, they are also saying that some workers will see a rapid increase in their wages, but others will not. Maskin’s model will continue to have legs in this brave new world of ours.
Second, expectations setting becomes even more important. Based on what we currently know, growth via even the most successful industrial policy won’t be as impressive as it was in the second half of the twentieth century. Short, brutal version: yes, India has missed the proverbial bus.
Even in the best-case scenario, a services-based model is unlikely to deliver growth rates approaching those experienced in East Asia. For one thing, increasing productivity in labor-absorbing services is likely to prove more difficult than in manufacturing, even if the strategies outlined previously prove successful and even if there remains a significant gap in productivity in the services sector between developed and developing country. This is because manufacturing technologies are more standardized and have proved easier to copy despite large differences in the context of developing countries. But there is another general-equilibrium reason as well. Under manufacturing-led growth strategies, a succession of export- oriented sectors – wigs, toys, garments, autos, steel – could take off one after the other without regard to domestic demand. By contrast, the expansion of non-tradable services – those that are most likely to absorb employment — is ultimately limited by the size of the home market. Individual service sectors cannot keep growing if other service sectors are not also expanding and increasing their productivity; the growth of retail, say, depends on the growth of personal services, hospitality, and the rest of the economy. Otherwise, the profitability of the more rapidly expanding services would soon collapse. This complementarity on the demand side necessitates balanced growth and lowers the ceiling on the potential growth rate of the economy.
https://drodrik.scholar.harvard.edu/sites/scholar.harvard.edu/files/dani-rodrik/files/a_new_growth_strategy_for_developing_nations.pdf
But Not Doing Anything Ain’t An Option
Why can’t the markets just “figure it all out”? Isn’t that what they’re supposed to be good for in the first place? Why bother with industrial policy at all?
Rodrik and Stiglitz recommend against letting markets do their thing, because there are things that stop the market from working the way it is supposed to. We call this market failure. And that is why, the authors say, all of the most rapidly growing economies of the past have relied heavily on industrial policies promoting productive diversification and the growth of new industries.
I actually prefer their argument in the very last paragraph of their essy, where they say that choosing not to have an industrial policy is also an industrial policy. You might well think that this is a most brilliant idea, but try telling that to the B and D workers in Maskin’s world.
So What Should We Be Doing?
The significant uncertainties in technological evolution, heterogeneities among production units, and the highly dynamic setting in these new areas require a new model of iterative, strategic collaboration between firms and government agencies (national and sub-national). The focus would be on experimentation and learning, with objectives, instruments, performance criteria and institutions developed and shaped over time. Government capacity would be accumulated in the process, rather than presumed as given.
https://drodrik.scholar.harvard.edu/sites/scholar.harvard.edu/files/dani-rodrik/files/a_new_growth_strategy_for_developing_nations.pdf
In plain simple English, this is what it means:
- “Significant uncertainties” means “We don’t really know what a new industrial policy will look like”.
- “Iterative and strategic collaboration” is academic speak for “We’ll figure it out as we go along”.
- “The focus would be on experimentation and learning, with objectives, instruments, performance criteria and institutions developed and shaped over time” means we are not joking about pt. 2
- “Government capacity would be accumulated in the process, rather than presumed as given” means “We know that governments don’t know anything about this either, but they’ll have to learn, same as the rest of us.”
So that paragraph is really saying this:
We don’t know what a new industrial policy will look like, but we will figure it out as we go along. We have to, we have no choice. And that includes you too, government!
Now, you may think that I’m being facetious, but I assure you that I am not. They’re quite right!
We really don’t know what the world is going to look like with the advent of robotics and AI. We really don’t know what impact climate change will have during and beyond the time-frame of us getting used to a world with AI. But employment markets are likely to be volatile, and there may well be job losses, particularly in the manufacturing sector. Generating twenty million jobs every year, year after year, for the next twenty years is going to be tough, but we have to do it. Otherwise, we are screwed. Manufacturing probably won’t do it for us, so we have to figure out how to make it work using a combination of manufacturing, export oriented services, and non-tradable services.
India’s challenge is to not just catch up with those who went before us, but to do so on a brand new map. Technology has torn up the old one, you see. We have to do it starting about five years ago, if not earlier. We have to do it with our current state of state capacity, and for a larger number of people than has ever been attempted before (including China from twenty five years ago). And we have to do it while contending with climate change, and in geopolitical scenarios that are crazier than they have been in a long, long time.
Let’s wish ourselves luck, because we’re going to need it.
EconForEverybody 