The Blue Grid Files
Read 4 of 12

What the auditor counted

Published 9 October 2026

The CAG report described an audit conducted from January to September 2010 covering 2003-04 through 2009-10. Its scope included implementation of the Unified Access licensing policy and allocation of 2G spectrum to new and existing operators. It used the records made available by the telecom department, related files seized by the CBI in October 2009, material supplied by the Finance Ministry and public corporate records used to examine eligibility. That source description is important. The audit was not simply a calculation made from press reports, nor was it a criminal trial with witnesses cross-examined on every allegation. The report said it incorporated departmental and Finance Ministry replies to observations communicated in July and September 2010 and discussions at a further meeting on 4 October 2010. Its stated questions were whether the policy had been implemented efficiently, whether licences and spectrum had been allocated fairly, transparently and efficiently, and whether revenue potential had been optimally managed. Those are connected but not identical tests. A procedure can be criticised for unfair access even where an exact counterfactual price cannot be fixed, and a revenue estimate does not itself identify a criminal agreement between named accused. (CAG Report No. 19 of 2010; CAG Report No. 19 of 2010 (NDTV-hosted full copy))

The report's organisation reflected those questions. Chapter 3 examined policy implementation, Chapter 4 the procedures for licences and allocation, and Chapter 5 financial impact. The audit's ending was also more qualified about amount than a single headline can convey. It maintained that loss had occurred while expressly saying the amount could be debated. That position did not make the alternative models a statistical confidence interval, and it did not amount to the later criminal court's finding of a proved offence. The report demanded responsibility and accountability for the lapses it identified and called for review of departmental systems. It should therefore be described both as a challenge to financial management and as a detailed criticism of administrative process. The later trial and appeal address a different evidentiary question, while the audit remains a dated institutional assessment of the records it examined. Its scope, replies, method and concluding qualification should travel with its figures. (CAG Report No. 19 of 2010; CAG Report No. 19 of 2010 (NDTV-hosted full copy))

The CAG did not present its largest figure as an audited bank transfer or a recovered bribe. It explicitly described the exercise as a presumptive valuation. The executive summary explained that scarcity, competition, business plans, the number of operators and sector growth could affect the market value of spectrum at a given time. It acknowledged that the assumptions underlying economic models could be questioned and disputed, and that there was no foolproof market-discovery mechanism available retrospectively. Chapter 5 therefore used indicators from the available record to estimate what a different allocation process might have realised. The distinction between an actual receipt and a counterfactual receipt is the core of the number. The government did receive entry fees; the report compared those receipts with prices inferred from alternative indicators. That is an argument about the model and its assumptions, not an admission that pricing was fair, and not a demonstration that an identified person pocketed the full estimated difference. (CAG Report No. 19 of 2010; CAG Report No. 19 of 2010, chapter 5)

The headline maximum of Rs 1,76,645 crore was the combined total of several categories under one valuation method, not the price of 122 licences alone. The report's summary table attributed Rs 1,02,498 crore to new licences, Rs 37,154 crore to dual-technology approvals and Rs 36,993 crore to spectrum beyond the contracted quantity. The alternatives in the same table produced combined totals of Rs 67,364 crore using the S Tel indicator, Rs 69,626 crore using the Unitech equity indicator and Rs 57,666 crore using the Swan equity indicator. Keeping these rows intact prevents a frequent denominator error: taking a total that includes dual technology and additional spectrum, then describing it as the loss exclusively caused by the 122 new licences. It also prevents a subtler error in which the smallest and largest totals are treated as a statistical confidence interval. They were outcomes of distinct valuation indicators with different assumptions, not endpoints generated by a single probability model. The right label is the CAG's range of presumptive estimates across methods. The Rs 1.76 lakh crore shorthand is acceptable only when the account immediately explains the model, the included categories and the fact that it was the upper estimate. (CAG Report No. 19 of 2010; CAG Report No. 19 of 2010, chapter 5)

The S Tel calculation was based on an offer, not an auction that had already collected the money. The CAG recorded that S Tel first offered an additional revenue share of Rs 6,000 crore for a pan-India licence and later raised the proposal to Rs 13,752 crore over ten years for 6.2 MHz of GSM spectrum. It also recorded willingness to increase the bid if there were a counterbid or auction. The report discounted future receipts to arrive at an indicative present value. For the new licences and dual-technology approvals together, it calculated Rs 65,909 crore against actual receipts of Rs 12,386 crore, producing a difference of Rs 53,523 crore for those two categories. This is not the combined Rs 67,364 crore total in the report's final table, which also includes beyond-contracted spectrum. The distinction between the two totals is important enough to spell out. The CAG used it as an indicator of market perception rather than claiming that the government had forgone a completed sale. (CAG Report No. 19 of 2010, chapter 5)

The department told the audit that S Tel's offer contained conditions unacceptable to the government and that the company later withdrew it in the Supreme Court. The CAG rejected the department's inference that withdrawal demonstrated no loss. It pointed out that the withdrawal occurred in March 2010, more than two years after the contested allocation, when rivals had received licences, established infrastructure and begun services. The audit argued that the competitive conditions had changed, including the presence of 13 or 14 operators in some areas. This exchange defines the real dispute more accurately than saying merely that one side claimed a large loss and the other claimed none. The department challenged the offer's usability and subsequent withdrawal; the audit defended its relevance as a contemporary signal of willingness to pay before those later changes. (CAG Report No. 19 of 2010, chapter 5)

The report's largest estimate used prices discovered in the later 3G auction as an indicator of the value of the 2G allocations. The CAG cited TRAI's 2010 discussion of services that had developed beyond basic 2G and its recommendation to use 3G prices as the current price for the 1800 MHz band. The report emphasised scarcity and demand as well as technical capabilities. It calculated an indicative value of Rs 1,11,512 crore for the new licences against Rs 9,014 crore received, and Rs 40,526 crore for dual technology against Rs 3,372 crore received. The difference for those two categories was Rs 1,39,652 crore; adding the beyond-contracted spectrum category produced the headline maximum. These details clarify both the strength and the limitation of the method. The audit did not infer its upper estimate from a 2008 auction of the same bundle. It transported a later price benchmark across technologies and dates, defended that benchmark with regulator reasoning, and treated the result as presumptive. (CAG Report No. 19 of 2010, chapter 5)

The equity-based estimates rested on another inference: that foreign investors' willingness to inject large sums into new licensees indicated the value of access to licences and spectrum. The CAG reasoned that inexperienced entrants had attracted investment before establishing a substantial operating position, making spectrum access a central source of their valuation. It calculated an indicative pan-India licence value between Rs 7,758 crore and Rs 9,100 crore, compared with the Rs 1,658 crore entry fee. But its chapter also recorded Unitech's contrary explanation that effort and expenditure had already gone into the business and that company value was not spectrum alone. Company equity, fresh capital, transferred shares, licence rights and operating assets are different objects. The Supreme Court criticised the large gains reflected in these transactions in the constitutional proceeding; the criminal court subsequently rejected the prosecution's theory of criminality on the trial record. The High Court's 2024 order identified the treatment of these transactions as one issue requiring deeper appraisal. (CAG Report No. 19 of 2010, chapter 5; Supreme Court judgment, 2 February 2012; Delhi High Court leave order, 22 March 2024)