Most countries have far fewer operators than India does and rarely more than 6 in any service area. With a dozen or so operators in most service areas in India, the demand for spectrum far exceeds supply.  Auctions can have their problems - e.g. speculative bidding, possibly slower roll out in less lucrative markets etc. They may be unnecessary if all demand can be met easily. But how else do we allocate spectrum transparently in India? Alternatives to auctions seem arbitrary. Better auction design might be the better approach to avoid most difficulties.

The problem is that in India, it seems difficult to separate allocation from pricing. Elsewhere, if you can accommodate all demand you can give spectrum free or at some other price. In India, auction determines at least the upfront price when it determines the user of spectrum.

Sharing a part of revenue with the government, instead of paying upfront for spectrum seems attractive till you see how companies use freebies to beat the system. A company providing its subscribers free/cheap calls e.g. calls within the same network, or within closed user groups, still benefits since it attracts more subscribers, which, in India qualifies you for more spectrum.  It improves market valuations too.

Freebies wouldn't be objectionable for fixed line calls. But on mobiles, you are offering them without paying for spectrum, a valuable and scarce resource. I am no economist, but this is a perverse incentive if there was one. With over 90% of subscribers dependent exclusively on wireless, can we afford to give away spectrum free?

India's spectrum rules have several other perverse incentives. For example, basing allocation of spectrum on number of voice subscribers encourages companies to overstate subscribers to hog spectrum. Since only voice subscribers are counted, few take data seriously either.

India's spectrum management is a royal mess and manifestly unsustainable. However, opponents of market based approaches like auctions will need to offer workable solutions to current problems.

There is merit in cognitive radio-based shared spectrum approaches which offer higher efficiencies and other promising possibilities. However, with over 600 million connections (and possibly around 400 million unique users) paying barely 1c per minute, we cannot simply move over to the new method without evidence that the current experiments can be scaled to these kinds of numbers. Perhaps demanding that a larger part of spectrum be made available for such options might be the first step.

We simply, cannot run away from market based approaches to allocation and pricing of spectrum in India, at least for the dominant wireless market i.e. mobile telephony and data.