Substack

Monday, February 6, 2023

Thoughts on India's structural transformation path

A new NBER working paper (here) looks at the nature of industrialisation in the US between 1880 and 1940, when the country transformed from a largely agrarian to an industrialised country.   
This short paper studies the co-movement of workers across sectors and space during a pivotal period of the US economy: its second wave of industrialization between 1880 and 1940. We show that industrialization was primarily a local phenomenon, with most sectoral reallocation happening not through long-distance moves towards industrial hubs but within counties. Moreover, within counties, the most significant sectoral shifts did not occur via the expansion of incumbent cities but rather through the birth of new cities and towns in the rural hinterland. Interestingly, the new urban structures had a much higher employment share in manufacturing than incumbent cities, which specialized more in providing non-tradable services. In other words, “factory towns” sprouting across Rural America were central to both US industrialization and urbanization.

The paper contrasts the contributions of spatial reallocation of workers from rural to industrialised counties with local transformation or the sectoral shifts due to within-county changes in industrial structure. 

In Panel B, we implement this decomposition for each decade between 1880 and 1940. Structural change was mainly about the transformation of local economies: the “Within” component, shown in orange, explains between 45%-85% of the decline in agricultural employment in each decade and 63% over the entire period. The “Across” component could reflect moves within the local labor market or long-distance migration across states. To quantify the importance of these different types of reallocations, we further decompose the across-county component into reallocations across counties within commuting zones, across commuting zones within states, and across states.

Panel B shows this decomposition of the across-county component in various hues of blue. Long-distance, cross-state moves from remote rural locations towards large industrial centers such as Cook County (Chicago) or New York County (Manhattan) played a minor role in the aggregate structural change. The within-state component accounts for at least 80% of aggregate structural change every year. Interestingly, the era of the Great Migration toward Northern States between 1910 and 1920 is the only decade for which cross-state migration accounts for a non-trivial part of sectoral reallocation...
The industrialization of Rural America (the union of counties with the highest agricultural employment shares that collectively accounted for 50% of total employment in 1880) occurred in two ways. First, workers moved to its industrialized cities... Second, within the hinterland, the importance of the agricultural sector also declined swiftly. In an accounting sense, the second channel is much more important for the industrialization of Rural America than the first since the hinterland accounted for a much larger share of Rural America’s population... In 1880, Rural America had only four incorporated cities identified in the US Census. By 1940, there were almost 250 cities, and their populationshare had risen by a factor of 50 from 0.3% in 1880 to 16.8%. Crucially, the entirety of this increase is accounted for by new cities sprouting in the hinterland: the incumbent “old” cities that already existed in 1880 only accounted for 0.5% of Rural America’s population in 1940... the gradual transformation of parts of the hinterland into towns and, from there, into incorporated cities... 

The newly founded cities were heavily specialized in manufacturing and corresponded precisely to the textbook idea of factory towns. While the manufacturing share also grew in old cities, by 1940, it was only half as large as in the new cities. Old cities continued to rely more on (consumer) services. However, despite the formation of new cities, the hinterland still accounted for more than 80% of the population in Rural America and contributed substantially to the fall in agricultural employment... Even outside the proliferating cities, Rural America urbanized: the hinterland’s share of urban workers increased from 4.9% to 17.6% between 1880 and 1940. Moreover, the rise in the hinterland’s manufacturing employment share was particularly pronounced among urban workers, highlighting the general trend of “densification” in Rural America: workers came together in and around factory towns long before their incorporation as cities.

Interestingly, this kind of sectoral shifts due to within-county/region changes has been the major driver of industrialisation in developing countries too. 

Changes in the local employment structure accounted for between 86-94% of the decline in agricultural employment shares in Brazil, Indonesia, India, and China. Interestingly, spatial reallocation plays the most negligible role in China, known for its stringent migration restrictions (e.g., “Hukou system”). 

The data above does not disaggregate the types and quality of non-agriculture within-county employment. More specifically, it does not make the distinction between rural non-farm informal employment and productive urban manufacturing and formal services employment. 

In other words, it was not migration to larger cities and towns, not even daily commutes to the nearest urban centres, but the creation of new urban localities centred on manufacturing that helped achieve the structural transformation in the US. 

This has important learnings for India, especially since it means taking jobs closer to the majority of population instead of making the majority of population migrate to where there are jobs. This assumes significance since one of the most important labour market frictions in India is the distance between labour  and jobs locations, and labour's higher marginal wage expectations for moving outside their commute zones. 

I believe that the biggest challenge for India in the years ahead is to manage the quality of its economic growth and create jobs to absorb its youthful population. On the former, I've blogged here and here about the need to ensure that economic growth is broad-based. It's on jobs that the structural transformation pathway is important. Specifically the nature and quality of the structural transformation. 

I'll describe this in terms of long and short routes to change. 

Notwithstanding India's 1987-2009 numbers in the table above, I'm inclined to argue that India's structural transformation and socio-economic mobility has followed what I will call the long longest route to change. Though farm productivity growth has lagged, thanks to a combination of horticulture and animal husbandry activities and migration, farm incomes have increased gradually. In the typical poor rural household (and they form the vast majority), the male migrates for work seasonally or all-year to the nearest metropolis or large city or another part of the country. They do mostly unskilled and semi-skilled jobs, especially in the construction or in urban informal services sectors, and mostly earn just slightly more than would have in the village. 

This kind of structural transformation and socio-economic mobility has its limits. Bar the tiny few who get into professional courses (this too may no longer apply to an Engineering degree from the local college) definitively exit poverty and those lucky fewer do well by entrepreneurship, the vast majority of rural poor migrant households remain stuck there. In any case, it does not create the broad base of consumption class required to sustain high growth rates for long periods. It's only in the degrees of poverty that their lives change.

A good measure of structural transformation and social mobility would be to get the trends on the proportion of rural poor who are qualitatively better off than their parents. Or that on those who have migrated into the lower middle class status. What's the trend on intergenerational mobility at different income quintiles? 

Instead the short route to change which the East Asian countries followed combined high intensity small-holder agriculture which sharply increased farm productivity and incomes, with exports-led manufacturing factories which created millions of jobs nearer the hinterlands. In these cases, rural incomes increased sharply resulting in the emergence of large rural consumption class, who provided the market for the emergence of new towns which in turn attracted manufacturing and other industries. These in turn further increased incomes, and the virtuous cycle went on. Local transformation trumped spatial reallocation. 

India needs to adopt something similar as its structural transformation pathway. It needs to spread industrialisation and productive economic activity beyond the current few urban cores to cover the hinterland areas. At least some of its currently numerous sub-optimally productive urban centres (largely parasitic economies dependent on government activities, like many state capital cities are) should embrace productive industrial activities, especially of the labour intensive kind.

I do not recollect coming across a study examining these trends in the Indian context. It would be good to have someone dive deep into this issue in the Indian context. It would require layering the trends in urbanisation with economic growth. What are the identifiable markers of local transformation? How many new towns/cities have emerged in the last two decades? What's the profile of jobs which have emerged in these new towns/cities and in the hinterland? How has inter-generational mobility changed, and to what extent?

In this context, I'll also draw attention to the much hyped point about India's entrepreneurship. As I have co-written here, most of the entrepreneurship that we associate with the poor in India is of the non-productive subsistence kind which arises out of their compulsion to earn something to survive. It's not the productive dynamic kind which arises out of an interest in creating something new and which builds companies and generates productive jobs. In an ideal world, such subsistence entrepreneurship should be replaced with wage earning jobs.  

The political leadership needs to formulate a narrative that prioritises local transformation and the bureaucracy needs to provide the policy enablers and instruments to catalyse and promote such growth. For this to happen, the academics, opinion makers and commentators should catalyse a credible and actionable discourse, trigger public debates, and influence important stakeholders to embrace this theory of change. 

No comments: