I have watched the work of Jishnu Das and his co-authors on the geography of economics research with great interest. What is going on here? I have four conjectures.
A supply argument based on researchers: If there are more American researchers in the world, then they are likely to write more papers about US data as they take interest in the US. By this argument, we will get more research on India when the number of economists in India per million matches that of the US.
Economics is an observational science where the vital raw material is data. A supply argument concerns datasets: In the India of old, there were no good quality datasets. NSSO and ASI have problems of low quality survey administration. The NFHS is well measured, but it's not panel data.
Only in the last 20 years have well administered panel datasets come along. These are the consequence of the maturation of the economy, through which firms and government agencies are capturing more information, mandating disclosure, etc. This includes CMIE data about firms (now widely used) and households (example), the NSE data (example), the NSDL data (used by Tarun Ramadorai), the home loans data (used by Tarun Ramadorai), the KGFS data (used by Renuka Sane and Susan Thomas), etc.
I may conjecture that until a country hits middle income, it is stuck with datasets like NSSO or ASI, which hold back the emergence of good quality economics. From this point of view, the outlook for research on India is very good. Vast amounts of high quality data is now being produced by the private and public sectors, and this is increasingly getting to researchers.
The third aspect is the selection process operated by editors. The papers which are addressing an interesting question in the eyes of an editor in the US are very different from the papers that are interesting questions as seen here. For research on India to get published in the US, it has to be interesting in the context of the conversation that is taking place among editors and their colleagues in the US.
For an analogy, we who live in India know that many NRIs lack common sense on India. Their sense of what looms large and what is trivial is often out of place. When the journals are based on editors and referees who have a low connection to India, their notions of what's an interesting question are frequently odd.
Jishnu and Quy-Toan say that in 20 years, the top five economics journals published 39 papers on India. I am always on the lookout for knowledge and insight into India, and I don't think these 39 papers have contributed much to my understanding. Editors and referees also don't catch many blunders on datasets and data handling when it comes to India, which further reduces the value of these papers.
The fourth aspect is the distinction between first order and second order work. Datasets in the US like the NLSY (say) are well established and very well studied. Barring outstanding breakthroughs of creative work, most of the first order work using this dataset has been done. The average researcher is doing incremental stuff, and the average editor and referee is used to plodding along with incremental stuff. Diminishing returns have set in, and they're polishing the apple.
In contrast, all the datasets about India, mentioned above, are only now coming into the hands of researchers. There are opportunities for doing the first order stuff where the space is being mapped out for the first time. We are Johannes Kepler with Tycho Brahe's dataset, and we should think like that. At the same time, projects that are imbued with ambition and curiosity generally run afoul of the selection process operated by editors in the US, who are more comfortable with with Kepler fitting epicycles. All too often, I see researchers defining their projects through the question `what do I have to do to get past the editors and referees'. Very often, these filters are burned so deep inside the skull that the victims have stopped noticing them. When the editors and referees are not interested in India and lack intuition in it, this strategy is going to yield papers that don't illuminate.
There are two ways of doing research connected with India. The first is to look at our backyard, understand what are the interesting questions, and go after them. The second is to look at the top journals, and think of what one could give the editors that happens to exploit an identification opportunity found in the Indian data. Many papers that do the latter have a superficial engagement with the data and suffer from blunders in implementation that aren't visible to the top journal editors and referees. Even when well executed, I find myself unexcited about many such papers.
The incentives of the profession are strongly wired towards publishing in the top journals. Very few people have the ambition and curiosity to care more about the phenomena of interest and less about the editors and referees. This generates systematic under-performance by the economics profession, even though a major constraint (datasets) is now being eased.
The number of economists
A supply argument based on researchers: If there are more American researchers in the world, then they are likely to write more papers about US data as they take interest in the US. By this argument, we will get more research on India when the number of economists in India per million matches that of the US.
The raw material of data
Economics is an observational science where the vital raw material is data. A supply argument concerns datasets: In the India of old, there were no good quality datasets. NSSO and ASI have problems of low quality survey administration. The NFHS is well measured, but it's not panel data.
Only in the last 20 years have well administered panel datasets come along. These are the consequence of the maturation of the economy, through which firms and government agencies are capturing more information, mandating disclosure, etc. This includes CMIE data about firms (now widely used) and households (example), the NSE data (example), the NSDL data (used by Tarun Ramadorai), the home loans data (used by Tarun Ramadorai), the KGFS data (used by Renuka Sane and Susan Thomas), etc.
I may conjecture that until a country hits middle income, it is stuck with datasets like NSSO or ASI, which hold back the emergence of good quality economics. From this point of view, the outlook for research on India is very good. Vast amounts of high quality data is now being produced by the private and public sectors, and this is increasingly getting to researchers.
What the American journals find interesting
The third aspect is the selection process operated by editors. The papers which are addressing an interesting question in the eyes of an editor in the US are very different from the papers that are interesting questions as seen here. For research on India to get published in the US, it has to be interesting in the context of the conversation that is taking place among editors and their colleagues in the US.
For an analogy, we who live in India know that many NRIs lack common sense on India. Their sense of what looms large and what is trivial is often out of place. When the journals are based on editors and referees who have a low connection to India, their notions of what's an interesting question are frequently odd.
Jishnu and Quy-Toan say that in 20 years, the top five economics journals published 39 papers on India. I am always on the lookout for knowledge and insight into India, and I don't think these 39 papers have contributed much to my understanding. Editors and referees also don't catch many blunders on datasets and data handling when it comes to India, which further reduces the value of these papers.
First order versus second order work
The fourth aspect is the distinction between first order and second order work. Datasets in the US like the NLSY (say) are well established and very well studied. Barring outstanding breakthroughs of creative work, most of the first order work using this dataset has been done. The average researcher is doing incremental stuff, and the average editor and referee is used to plodding along with incremental stuff. Diminishing returns have set in, and they're polishing the apple.
In contrast, all the datasets about India, mentioned above, are only now coming into the hands of researchers. There are opportunities for doing the first order stuff where the space is being mapped out for the first time. We are Johannes Kepler with Tycho Brahe's dataset, and we should think like that. At the same time, projects that are imbued with ambition and curiosity generally run afoul of the selection process operated by editors in the US, who are more comfortable with with Kepler fitting epicycles. All too often, I see researchers defining their projects through the question `what do I have to do to get past the editors and referees'. Very often, these filters are burned so deep inside the skull that the victims have stopped noticing them. When the editors and referees are not interested in India and lack intuition in it, this strategy is going to yield papers that don't illuminate.
Conclusion
There are two ways of doing research connected with India. The first is to look at our backyard, understand what are the interesting questions, and go after them. The second is to look at the top journals, and think of what one could give the editors that happens to exploit an identification opportunity found in the Indian data. Many papers that do the latter have a superficial engagement with the data and suffer from blunders in implementation that aren't visible to the top journal editors and referees. Even when well executed, I find myself unexcited about many such papers.
The incentives of the profession are strongly wired towards publishing in the top journals. Very few people have the ambition and curiosity to care more about the phenomena of interest and less about the editors and referees. This generates systematic under-performance by the economics profession, even though a major constraint (datasets) is now being eased.
"For an analogy, we who live in India know that many NRIs lack common sense on India. Their sense of what looms large and what is trivial is often out of place..."
ReplyDelete:)
Was it then a good move to appoint an NRI as the RBI governor? With his recent political utterings he has proved that maybe just maybe career bureaucrats or those rooted in India are better to get the job done.
The above point aside, as a layman with deep interest in economics, I find research on India coming out of the US to be fun to read yet sometimes quirky and not important to our concerns. But more interest in the Indian economy is always welcome. I can find research and data on China much more easily than India.
I question the quality of education and the PhDs that come out of our universities. There is no inquisitiveness. Just a desire to finish the work and get a job. The system emphasizes on policing against cheating rather than openness and creativity. I don't expect good economic research to come out of our country for a long time unless there some well-resourced, visionary sponsor comes up. The government and it's agencies are definitely not going to be such sponsors.
Just to add to your apt observations, most of the top journals don't care about the phenomenon of interest even if it is not related to India. I find "i scratch your back, and you scratch mine" being the gameplay. Hence, the repetition of even words at time let alone trying to solve a problem at hand.
ReplyDelete