W Power 2024

It's Very Hard to Design the Right Auction For Resources

Arunava Sen and Kaushik Basu tell Forbes India that the ultimate motive of auctions should be to benefit customers

Published: Feb 14, 2013 10:50:12 AM IST
Updated: Feb 14, 2013 03:51:58 PM IST
It's Very Hard to Design the Right Auction For Resources
Image: Amit Verma
Left Kaushik Basu, chief economist at the World Bank and Arunava Sen, professor at the Indian Statistical Institute

India has struggled to understand why so many government decisions and policies regularly go wrong. The telecom spectrum allocation is a good example of this. Can we not design auctions more efficiently? What does the academic world have to say on how our government makes its policies? And if policy prescriptions are available then what explains government inability to incorporate them.

Forbes India quizzes Arunava Sen, professor at the Indian Statistical Institute, who specialises in auction design. Sen has been awarded the 2012 Infosys Prize for economics in December for his work, which highlighted how individuals holding information crucial to policy-making may benefit by misrepresenting it. Joining us in this discussion is Kaushik Basu, chief economist at the World Bank, who was not only a jury member for the prize but also the advisor to the finance ministry till July 2012.

Structuring of auctions for resources has been a burning topic in India. Can you share some insight into how a country like India could go about structuring such auctions?
Arunava Sen:
Auction is a very rich area for theoretical research and the kind of auctions you are referring to are among the most complex auctions. These are called combinatorial auctions where there are multiple goods to be sold, like the various circles in spectrum etc. And the crux of the problem is that various parties have private information about the values of these goods. It is a problem that is technically very difficult to solve.

The kinds of auctions that are being used are based on heuristics–it doesn’t have a good theoretical foundation. The design used in the last round for spectrum is fairly standard–one that is used all over the world.  Then there is the issue of what the objectives of the auctions should be–should they maximise expected revenue or should they be something else. But even if you decide the objective, to design the right auction is a very hard theoretical issue.

Kaushik Basu: 
I will give you an example. I had just joined the government when the 3G auction was being designed. We sat and calculated using the best information around, we decided that the spectrum should be sold at X amount. After the auction, we got three times that money. Now, had we sold it at our own estimate, we would have sold the spectrum at one-third the price. Maximising revenue is not always the government’s objective. But if you want to do that, a well designed auction can make a huge impact.  And we really learnt that through practical experience.

AS:  But there are few simple points like what is the role of the reserve price. There is an incredible amount of confusion on these issues. So in 3G auctions, the reserve price was rather low and the final price was much higher. Should you be pleased at it? Or should you say this is a disaster? In my opinion, it was good because you had price discovery. But in retrospect…you know it seems to me that what policymakers are interested in doing is to be very close to the final price otherwise governments face some heat that they are trying to undersell.

So how does one judge whether an auction was designed well or not?
KB:
The ultimate in some sense is to benefit the consumers. Just that if you leave a lot of money on the floor it is obvious that the consumers will benefit by that. And that is where the design comes in. So if you are lowering the price just so that the few people buying the spectrum will make a huge amount of profit, then it is not rightly done. But if the government can lower the price and ensure the benefits reach the consumers, then you can say the government has not only maximised the revenues but also distributed the benefits through the society.

Something similar afflicts the issue of land acquisition in India. Prof Sen, where do you think we might be missing on policy?
AS:
In my opinion, this is a problem appropriate for an auction rather than to set the compensation at four or seven times the market price. That may not be right. According to the standard notion of Pareto efficiency in cases like land acquisition the trade should take place when the valuation of the buyer is more than sum of the valuation of the individual plot holders. [A decision is Pareto efficient if you reach a point after which no one in the system can be made better off without making at least one individual worse off]
Now, the difficulty is that we don’t know these valuations. If you ask, those who want to sell may understate their valuation [an abnormally high valuation will keep buyers at bay] and those who don’t want to sell may overstate it.  

KB: But [after everyone has sold their land] if there is one person who is in the middle of the plot and just doesn’t want to sell, then you’ll never get a deal (even though you would have satisfied conditions of Pareto efficiency).

AS: But then you can sacrifice (notions of) Pareto improvement (and acquire land if, say, 90% of the people agree to sell).

Prof Basu, so why doesn’t this happen in actual policy-making? Why don’t we include such nuances?
KB:
It is very easy to begin to believe that I am an intelligent person with plenty of common sense and I can do it. So I feel that the first thought while structuring our policies is, “Why do I need these [economic principles]? It is just a lot of hocus-pocus.”  It gradually dawns on you. It is like engineering science. Today, no one would say I will build a bridge because I know roughly what to do. I feel this is beginning to happen in India now—just the sheer appreciation that you may be intelligent and have common sense but there is a problem, which goes beyond that. It is a kind of modesty you need in the policymaker to say, “Look I don’t know enough and so I need to bring in specialised knowledge.”

Prof Sen, just on that note have you been approached by the government?
AS:
Well I have been asked. [KB prompts AS and both laugh] Yes, once I was asked by Shah Rukh Khan, through an intermediary, to offer some advice on the first IPL auction. But returning to your question about why a government doesn’t do it…you know there is a technical reason and it leads to serious policy distortions. It is the inability of the government to commit. Auctions will not work if you cannot commit to an outcome. For example, in the last round of auctions, when the government found that it had not raised sufficient revenues, they auctioned it again. Now, if bidders realise that look I just have to wait for the reserve price to come down, then their behaviour would be completely different. Similarly, suppose some land acquisition auction runs into a hitch then you say, “Oh no, it did not work out and now we will have to do something else.” Once you do this, then all the attributes that good auctions disappear because you know that this is not the final outcome and we can go to court and negotiate etc.

KB: I feel in government there is a slow awareness that you need pre-commitments in certain things. That even if things are not ideal you will not change it or else you will damage future actions.

(This story appears in the 22 February, 2013 issue of Forbes India. To visit our Archives, click here.)

Post Your Comment
Required
Required, will not be published
All comments are moderated
  • A. N. Singh

    It will never be hard to design auction of resources if your team has actual domain knowledge. It is hard to find people with domain knowledge than to design an auction.

    on Feb 14, 2013