Caste and entrepreneurship in India
The story of India's economic surge is dominated by two conflicting narratives.
The sceptics insist that growth has been largely jobless and deepened inequality in an already hierarchical society.
The optimistic refute this gloomy thesis and believe that the rising tide has lifted all boats.
As an example, they point to the emergence of a small but growing class of Dalit (formerly known as untouchables, the lowest in India's wretched caste hierarchy) millionaires.
So much so that Dalit activists like Chandra Bhan Prasad like to call it a "golden period" for Dalits where "material markers are replacing social markers".
But new research by Lakshmi Iyer, Tarun Khanna and Ashustosh Varshney paints a less rosy picture.
Delving into the relationship between caste and entrepreneurship, the researchers have found that scheduled castes and tribes, the most disadvantaged groups in Hinduism's hierarchy, owned very little businesses despite a decade of sprightly economic growth and a long history of affirmative action.
Mining information thrown up by the 2005 economic census covering more than 42 million enterprises, they found schedule castes owned only 9.8% of all enterprises in India in 2005, well below their 16.4% share of the total population.
The scheduled tribes owned only 3.7% of non-farm enterprises despite being 7.7% of the population.
However, ownership of business among OBC's - an acronym for Other Backward Castes or the "middle castes" who "neither suffering the extreme social and economic discrimination of the Scheduled Castes, nor enjoying the social privileges of the upper castes" - has grown.
OBCs comprise 41% of India's people. Their members owned 43.5% of all enterprises in 2005, and accounted for 40% of non-farm employment.
This is a remarkable achievement considering that affirmative action for this group was widely introduced only in the 1990s.
The pattern of dismally low ownership of businesses among the most disadvantaged groups, the researchers found, is not specific to any one region or state in India.
Even in states like Kerala, Andhra Pradesh, Tamil Nadu and Maharashtra that were among the first to have social movements to end caste discrimination, ownership of enterprises is low.
States with high population of the disadvantaged groups also show that they are under-represented in ownership of businesses.
The researchers say there could be a host of reasons - caste discrimination itself (members of other castes refuse to work with the lowest castes), lack of knowledge, illiteracy, and problems with securing finance.
"All these factors," they say, "can prevent scheduled castes from entering industries that have significant economies of scale."
Growth possibilities are limited by differences in the size of worker networks - scheduled caste owners find it easier to work with scheduled caste workers.
I asked Dr Varshney, who teaches at the US's Brown University, whether the findings really came as a surprise, given that deeper social changes in societies like India take a lot of time.
He said he wasn't.
"I should, however, add that the story of the rise of the Dalit millionaires is not small either. Though numerically insignificant, it is politically, economically and socially very significant," he told me.
Over time, he believes, the rise of Dalits "may well become comparable" to the rise of Nadars - a southern caste - in Tamil Nadu.
Until about 150 years ago Nadars - mostly "toddy tappers" - were condemned to a near untouchable status. Today, they are a leading business community in the state and are found in all classes.
I asked Dr Varshney whether increased representation in politics had anything to do with higher ownership of business for different caste groups? I cited the example of the increasingly influential OBC-led politics in the country.
"Whether that happens remains unclear. The correlation undoubtedly exists, but the causes are still to be sorted out," he said.
I also wondered why decades of affirmative action and more than two decades of economic liberalisation hadn't still unleashed entrepreneurial energies among the most disadvantaged.
"Such transformations can take a long time," he said. "The rise of the Nadars, for example, took nearly eight to 10 decades, depending on how one defines the rise."