Comparing the performance of two ministers is always difficult because their circumstances are different. A finance minister coming in when the economy is on an upswing, for instance, always has it easier than one who inherits an economy in bad shape.

This doesn't apply so much to Dayanidhi Maran and his successor, A Raja, since both came in when the telecom sector was doing well. But what complicates things is that Raja has just come in; and any summing up of Maran's tenure has to factor in the way in which he dealt with the Tata-Sky DTH proposal, which was in competition with his brother's firm, and also the unfair manner in which his ministry dealt with the Tata group's telecom firm.

What this column concerns itself with, then, is just the public sector unit BSNL, which is the one area in which Raja's policy has been made public. In this, the contrast between the two ministers cannot be more stark, and to BSNL's detriment, Raja emerges the clear loser.

BSNL, readers would recall, had decided to float a tender for 45 million cellular phones two years ago, and when the bids were opened earlier this year, Ericsson won the bid, after quoting a price of $108 per line.

While BSNL was just getting ready to place the order for the much-needed lines, one of the bidders (Motorola) decided to go to court on the matter. Telecom circle talk is that Maran arm-twisted Motorola into withdrawing the case, paving the way for BSNL to go ahead and get the extra network it so desperately needs.

BSNL, which was adding more subscribers than Bharti in 2005 (see table), started slipping badly last year--the ultimate insult, in May, was Hutch beating it to even the number two slot.

Coming in after this, what's the first thing Raja does? He announces that he thinks the rates are too high and they need to be lowered to below $100 a line--since that's a saving of around Rs 1,500 crore (Rs 15 billion), it's difficult to fault Raja, and BSNL has also publicly toed this line.

But, wait a minute, how does Raja know what the correct price is? Apart from the fact that he's been in the ministry for just a few weeks, price fixing is very complex and varies in accordance with each network design--indeed, the price bids BSNL got range from around $175 a line to $108. And if the price has to be fixed based on what the minister thinks is right, why even have a tender process that took over two years?

Let's assume, though this looks a bit unlikely, given how competitive the bids were, the vendors do come around and reduce the price (without changing the specs!). If this gets done immediately, there may not be any great loss. But what if it takes 6-8 months (this may seem far-fetched, but remember the tender itself was put out in March 2006)? Before we do the maths, the assumptions need to be spelt out.

The average revenue BSNL gets per subscriber per month is around Rs 320 and, if it had enough equipment, presumably it would be able to maintain its 22-24 per cent market share average of 2005; so, if the industry is getting around 5 million new users each month, BSNL should be able to get 13.8 million users in a year.

In which case, by not having equipment, it loses Rs 5,299 crore (Rs 52.99 billion) a year, or Rs 441-442 crore (Rs 4.41-4.42 billion) a month, or an EDITDA of Rs 147 crore (Rs 1.47 billion)--a six-eight month delay then costs it Rs 883-1,177 crore (Rs 8.83-11.77 billion).

That is, a delay of 6-8 months virtually wipes out whatever gains the PSU can get by way of lowering project costs--in which case, even if Raja does manage to get his way, BSNL doesn't really gain, does it?

If the delays stretch more, and the price is reduced by less than the $8 Raja wants, the losses mount. The other cost we haven't spoken of yet is the value the market puts on these subscribers -- Vodafone paid Hutch around $700 a subscriber.

Even if you take a lower $600 for BSNL since Hutch ARPUs are higher, this means a capital loss of $690 mn a month or $4.14-5.50 bn in 6-8 months. As far as BSNL is concerned, the contest between the two ministers is a no-brainer.