4 May 2016

Borrowed knowledge - There is a problem in Indian research in the social sciences

Shubhashis Gangopadhyay 

In a recent meeting on setting a research agenda in an institution, there was a suggestion to find out how, and whether, the top-level bureaucracy, consisting of the national administrative services, affects governance. There was some opposition to this as it was felt that there are less than 5,000 people in the Indian Administrative Service at any point of time, and that it is highly unlikely that so few can affect the lives of a billion-plus people in a highly diverse society. The thing that struck me as odd was that the opposition did not have any evidence either way about it. It was also odd that a number of researchers in foreign universities have recently put out papers addressing this specific issue. These are researchers outside India, not all of them of Indian origin, who think it worthwhile to ask this question but our experts have already concluded that it cannot be an important issue to study.

This is a big problem in Indian research in the social sciences. There is no systematic generation of knowledge that can provide a basis for inputs in policymaking. Almost all funding in social science research goes towards the preparation of policy briefs but very little resource is put into developing the knowledge base on which such policy briefs are based. The government, through its ministries and departments, issues calls for advice on what it should do. Given that the government is always short on time, these outputs are expected in three to six months. It is obviously difficult for researchers in universities, who are also expected to teach, to come out with anything of value in such a short period of time. So whoever gets the funding, mainly so-called 'think tanks', put together something derived from what others have done in other countries and apply them to the Indian context. Most of these 'reports' are no more than literature surveys, with some quickly done empirical surveys, and a set of conclusions in keeping with policies implemented in one, two, or three other countries. Even when the reports go beyond this, they are mostly drawn from the knowledge about India generated by researchers outside India.

How does this affect policymaking? I give an example. Recently, the Telecom Regulatory Authority of India has come out with a recommendation for the reserve price of the 700 MHz spectrum auction. They believe that this should be at least four times the price received at the 1800 MHz auction. Let us see how this conclusion was reached. The point I want to make over here is not whether they are right or wrong; instead, I want to focus on the logic used to come to this conclusion and whether the methodology is acceptable or not.

At the outset, let me emphasize that Trai had looked into all aspects of what and how to auction spectrum bands in its April 23, 2012 recommendations. To arrive at these recommendations, they have dutifully listened to various stake-holders and considered all the issues regarding the auction of the 700 MHz band. However, while they have listened to all stakeholders and taken everyone on board, when it comes to taking a fundamental decision regarding the reserve price of the 700 MHz spectrum band, they have come up with a logic that is most difficult to comprehend.

At that time, 2100 and 2300 MHz bands (popularly referred to as 3G and 4G bands) were the only bands for which market prices were available in India (from the 2010 auctions). For the 700 MHz band, Trai had the option of extrapolating the reserve price from this reference, just like they did for other bands (800, 900, 1800 MHz). In 2012, Trai priced the 1800 MHz band approximately equal to the market price of the 2100 MHz band. The 800 and 900 MHz bands were considered twice as valuable and hence were priced at double that of the 2100 MHz band. However, for the 700 MHz band, Trai decided to look at other countries for references. They used six countries with auction price data for the 800 MHz band. Trai assumed that the current state of the technology, and the expectations from it, make the European 800 MHz band comparable to the Indian 700 MHz band (both likely to be used for 4G). Consequently, they decided to take the ratio of the auction prices of these two bands (800 and 1800 MHz) in other European countries. This would have done the trick as Trai had already extrapolated the 1800 MHz band reserve price for India from the 2100 MHz band market price. The data are given in Table 3.5 of paragraph 3.91 of the recommendations.

Trai lists the 1800 MHz auction prices for four countries - Germany, Italy, Portugal and Sweden - and the 800 MHz auction prices for six countries - France and Spain in addition to the four already mentioned. The auction prices are quoted in £/MHz/population. They have taken the (simple) average of the prices of the 1800 MHz bands and that of the 800 MHz bands. Then they used the ratio of the two averages and multiplied the 1800 MHz reserve price by this ratio to get the 700 MHz reserve price.

Using Trai's calculations, the 1800 MHz average price is 0.4809 while the 800 MHz average is 0.1727. The first average is 2.78 times that of the second average. First, it is not clear where they get the 800 MHz average. If one takes all the six countries, it works out to 0.4975; if one takes only the four countries where both bands have been auctioned, the 800 MHz average is 0.5000. This does not make a whole lot of difference because the ratio changes to 2.88 and 2.89, respectively. Both are higher than 2.78 but not as high as 4! So, why did Trai say that the 700 MHz band reserve price should be four times that of the 1800 MHz band price?

There is another issue with this calculation. If we are to look at the individual ratio of the two prices for each country, for Germany it is 28.5, followed by Italy at 3.1, Sweden at 1.8 and Portugal at 1.4. Clearly, Germany is way different from the other three countries. Also, working the ratio of averages by removing Germany results in a factor of 2.1. If four countries that have more reasons to be similar can be so different from one another, why is the Indian market assumed to be similar to the average of these four countries?

This is not a complaint about Trai but a commentary on the way decisions are made in India.With no knowledge base of our own, we have to look at where other countries are and hope that the one we like is the one nearest to India. Ideally, research on spectrum should have started receiving funds when the telecom industry started. Researchers would have asked and answered various questions on how the industry is growing and what are the future expectations from it. All of this would have been debated in seminars and conferences, as well as published in peer-reviewed journals. And, hopefully, we would have come up with a better logic for why the reserve price of the 700 MHz band should be four times that of the 1800 MHz band. We still may have got it wrong but at least it would not look so blatantly strange.

The author is Research Director, India Development Foundation 

No comments: