The King of All Telecom Scams — Part 7

[Previously in the series: Part 1, Part 2, Part 3, Part 4, Part 5, Part 6.]

Radio spectrum is a very special kind of stuff. It was entirely worthless not too long ago. Now it is not only an extremely valuable resource but it is nearly impossible to properly price. The price continues to increase with advances in science and technology. There are two broad generalization I make. First, that the more advanced an economy is, the more important the use of radio spectrum becomes. Second, the more efficiently the radio spectrum is used, the more advanced the economy becomes. In short, a virtuous cycle exists between spectrum use and economic growth. Therefore mistakes in the use and allocation of spectrum can have far-reaching consequences with damages that are impossible to estimate.

“The time has come,” the Walrus said,
“To talk of many things:
Of shoes–and ships–and sealing-wax–
Of cabbages–and kings–
And why the sea is boiling hot–
And whether pigs have wings.” [1]

The time has come to talk of many kinds of stuff that the world is made of, the economist said.

There are physical goods (cars, pizzas), and there are intellectual goods (a novel, musical composition, the idea of double-entry bookkeeping). There are private goods, public goods and club goods. Private goods are rival in consumption (either you or I, but not both, can consumer the same piece of cake), while public goods are non-rivalrous. There are durable goods (ships) and consumables (shoes and sealing-wax.) Physical goods can be sold outright or rented out for specified periods of time. You can buy a car or rent one.

So what’s spectrum? It is not strictly not a public good since you cannot simultaneously use the same slice as I am using. It is not a private good since its use now does not mean that it cannot be used again later. It is neither a physical good but neither is it an intellectual good. It is a natural resource but unlike petroleum, air or water, the “stock” of spectrum is not limited since the stock grows with technological advances.

Spectrum is special in many ways and that make it very difficult to properly price. The practical way out of this is to rent it out for use for short periods of time. What’s short? A period during which the technology does not change radically, in my opinion. A generation, as is generally understood in the terms 2G, 3G, etc.

Note that the length of a generation depends on the rate of technological change. Thus a generation in the 1950 may have been 20 years, but now in 2010, it may be only five years. If you rent out spectrum for 20 years now, you may mis-price (most certainly an underestimate) it and those lucky enough to rent it at current prices will make windfall profits. This also implies that it should not be sold outright.[2]

OK, so spectrum should be rent out for short periods. But for how much? The answer to that is always straightforward: auction.

A word on auctions. At the bottom of it all, they are all auctions in a generalized sense. When you buy a tube of toothpaste, you are actually engaging in an auction. The store sets the price (ask) and if it is lower than what you would pay (bid), the trade takes place. In this case, there are tons of tubes of toothpaste being sold by various stores (auctioneers) and being bought by millions of buyers (bidders). So it does not look like an auction at Sotheby’s. But from a process point of view, they are all auctions.

What should be the asking price for the auction of spectrum? As we noted before (see part 6 of this series), much depends on the objective of the government. If the objective is (short-term) revenue maximization, it should sell spectrum in a competition for the market and guarantee that there will be little or no competition in the market.

Revenue maximization in the short term is a very short-sighted policy. It is also a policy that the government of India is particularly good at. Why do they do it? Because they (the politicians and bureaucrats) are there for maximizing their private wealth and they have to make the most of their short tenure. So if they make a percentage of the take, their incentive is to increase the sale price as much as they can.

Short-term revenue maximization is really a very bad idea from the point of view of the economy. The efficient use of spectrum is an indispensable input to nearly all production processes. If the cost of spectrum use is too high, then all these processes will suffer and eventually the economy will suffer.

Perhaps we will go into details some other time. But for now, a story. In some banana republic, telephones were considered a luxury. So the government priced telephone calls way way above cost. The marginal cost of a telephone call was only one penny but the government priced it at $1. The idea was revenue maximization.

Did people make lots of phone calls? No. To avoid a $1 phone call, they would take the bus for 20 cents and deliver the message. The bus trip was however subsidized from revenues earned from telephone use. The actual cost of the bus ride was 50 cents but subsidized to 20 cents. If you do the arithmetic, you will see how utterly stupid the whole scheme was.

To state the obvious, the stupidity compounded with cupidity of leaders of banana republics is the main reason for the country to be a banana republic. India is the world’s largest banana republic (thanks to Mr Nehru and his clan).

Anyway, let’s move on.

So if not revenue maximization, then what?

Let’s try this. Suppose spectrum was auctioned off to telecom companies (telcos, for short) for free. The winners would be telcos which provide specific services at the lowest prices.

To make that concrete, let’s put some numbers. TelcoA’s bid, “Will provide high speed internet services for Rs 100 per month” and TelcoB’s bid is “Will provide high speed internet services for Rs 120 per month.” TelcoA wins.

This is to give you a sense that there are different ways of doing this which make more sense in the long run. Of course there are ways to sabotage this sort of scheme as well but mechanisms can be designed to minimize corruption and collusion between firms.

I think this is all for now. I will have to continue this because we are certainly not done. We have to look at the distributional aspects of the various schemes. Let’s get to it later.


1. “The Walrus and The Carpenter” (from Through the Looking-Glass and What Alice Found There, 1872) is one of my favorite poems by Lewis Carroll.

2. The cautionary tale of Manhattan is worth mentioning here. It was sold to the Dutch in 1626 by “Native American Lenape people in exchange for trade goods worth 60 guilders, often said to be worth 24 dollars, though (by comparing the price of bread and other goods) actually amounts to around $1000 in modern currency.” [Wiki.]

3 thoughts on “The King of All Telecom Scams — Part 7

  1. Atanu Dey Thursday December 2, 2010 / 2:14 pm

    We should be careful when figuring out whether the Dutch made a good trade with the Lenape people in buying Manhattan.

    The principal amount of $1000 if compounded annually at a rate of 5 percent over 385 years works out to a pretty penny. At 5 percent annual rate of growth, the rate of doubling is 14 years.The money would have doubled (385 / 14 =) 27.5 times. The present value of that initial sale would be therefore (1000 * (2^27.5)) = 1.89812531 × 10^11, or around $190 billion.

    The land area of Manhattan is around 59.5 sq kms (645,834,624 sqft.) Thus $190 billion for the land (we are not counting the improvements on the land) works out to around $300 per sqft.

    So the Lenape got in present terms $300 per sqft of land they sold. I think they did well.


  2. SV Saturday December 4, 2010 / 1:04 pm

    I always had this hunch that selling the public goods for the highest bidder was a bad idea. High time our netas took an Economics 101 class. Or may be it should be made mandatory to be a representative


  3. Prakash Tuesday October 25, 2011 / 4:58 am

    It’s not normal of me to disagree with an economist, but I guess the devil is in the details. How the service to be provided will be specified is really the concern, right?

    We’d have to specify the service, the time taken for providing a requestor a service, its reliability, the customer service to be given and so on and a lot of such stuff. It would be like a repeat of the inspector raj.

    It seems easier to auction the spectrum and let the companies fight out on the services to be given, right?


Comments are closed.