Reverse auctions and universal telecommunications service: lessons from global experience.

At least three problems led to the failure of the first two auctions to create genuine competition for rural public service. First, the calculations for the benchmark subsidy were not plausibly based on accurate information or on the appropriate standard, which is the incremental cost of public telephone service. The cost data used for calculating these benchmarks were provided primarily by BSNL. While there were rigorous independent attempts to verify the information, BSNL's accounts are aggregated in a way that makes it impossible to separate costs for different operations, which in turn makes incremental cost calculations extremely difficult. (67)

Second, callers pay "access deficit charges" (ADCs), which are surcharges on telephone calls that, in theory, help pay for existing service in unprofitable areas. BSNL received nearly all of the ADC cross-subsidies. The incumbent has potential gains from manipulating how cost information is aggregated across service categories and across high-cost and low-cost areas because these data determine not only the benchmark subsidy for public telephones, but also the magnitude of the net deficit for all local access service. If some ambiguous cost elements are allocated to subsidized areas, the effect will be to increase both the public telephone subsidy and the ADC subsidy.

Third, bidding was open only to basic service operators already providing rural service in the area. BSNL, even though it historically had not served many villages, owned some facilities in these areas; however, few other firms had entered these markets, in part because they were opened only recently and in part because disputes about the terms and conditions of interconnection with BSNL remained unresolved.

The fact that the first two auctions covered VPTs in areas in which the incumbent operator had already built infrastructure gave the incumbent a distinct advantage and limited the ability of private operators to compete. Firms not yet operating could bid for the public telephone subsidy only if no other bids were received or if the bids by others exceeded the benchmark. By precluding firms that were not already present, the subsidy scheme did not encourage either entry or innovation in rural services.

The auction procedure that the DoT set up advantaged the incumbent while providing no incentive to improve efficiency. In particular, if only a single firm can qualify for the subsidy and if that firm is then reimbursed the difference between its own estimates of its revenues and costs, the subsidized firm has no incentive to reduce these costs unless it can do so in ways that can be hidden from the DoT. Moreover, with only one subsidized firm in the entire nation, even benchmark competition (whereby differences between monopolies in different areas are used to evaluate performance and adjust the subsidy) is impossible, while the subsidies themselves make it impossible for nonsubsidized firms to enter the market.

The subsidy scheme for encouraging investment in VPTs was only the first part of the reverse auction plan. The following three auctions were more successful, attracting additional firms and yielding better outcomes. While the incumbent won one of those three auctions and parts of the other two, private providers won parts of two auctions, and the subsidies in all three auctions were well below the benchmark amounts.

In September 2004, the government held an auction to provide a second VPT in 300 areas (called secondary switching areas, or SSAs) that already had one. The incumbent BSNL and Reliance Infocomm were the largest winners, and two carriers bid against each other in 115 out of the 300 SSAs. The total subsidy awarded was seventeen percent below the benchmark amount.

A fourth auction in November 2004 was for the obligation to provide VPTs in the remaining 67,000 villages without one. The incumbent BSNL won in all twelve service areas. It faced bidding competition in three service areas, and that competition reduced the total subsidy by fifteen to twenty percent.

A fifth auction for subsidies to install rural household phones was concluded in 2005 as a first step toward distributing funds for connecting individual households. This step is potentially far more important than the first. Many more telephone lines were at stake in devising a plan for implementing extensive residential access than for providing more public telephones. While even in the best of circumstances firms might not have found subsidies for a relatively small number of public telephones an attractive basis for entering rural areas, subsidies for a much larger number of residential lines clearly are more attractive.

Indeed, the 2005 auction generated more interest among private operators, and the bidding reduced subsidies by sixty to seventy-five percent of the benchmark. BSNL won subsidies for 1,267 Short Distance Charging Areas (SDCAs, the basic service unit identified for subsidies) while two private operators won subsidies for 418 SDCAs.

In 2007, the government conducted two auctions for mobile service in eighty-one "clusters" that include 250,000 villages. (68) The first auction was for the right to build infrastructure that could be used by other firms to provide service. (69) BSNL won eighty percent of the $570 million to build this wholesale infrastructure. (70) Although BSNL dominated the winning bids, bidding competition reduced the subsidy to thirty percent below the benchmark. (71)

The second mobile auction in 2007 was to provide service over this "passive" network. Bidding was so intense that in many cases the winning bid was either zero or negative, meaning that the operator was willing to pay the government for the right to provide service. (72) The Economist noted:

Unfortunately, it is not quite as easy to interpret these results as The Economist would suggest. These auction results demonstrate strongly that competition for subsidies can bring down the subsidy. Because these appeared to be bids to operate on a network being built by someone else, it is unclear why subsidies would be offered in the first place. The government of India apparently decided to separate ownership and operation of the network from service provision. The wisdom of such structural separation is heavily debated and centers on whether consumers are ultimately better off when firms compete by investing in facilities or by offering service over the same facilities. Mandatory sharing of network facilities is likely to lead to more intensive use of those facilities, but can also reduce the incentive to invest in the network itself.

In this case, we do not know what the bidding might have revealed if firms had bid simply to provide service at the lowest cost.

E. Nepal

In 2000, the Nepalese government decided to use a reverse auction process to provide telecommunications service to the 534 village development committees (VDCs--the second-smallest administrative units in Nepal) that had no such access. (74) Firms were to bid for a one-time subsidy and a ten-year renewable license with a five-year exclusivity guarantee. (75) In exchange, they were to provide two public access lines in each VDC. (76) Unlike most reverse auctions, in Nepal, the maximum available subsidy was not made public. (77)

Two firms bid in September 2000, but a "security situation" caused the winning firm to back out of its agreement. (78) The regulator, the Nepal Telecommunications Authority, attempted the auction again in 2003 with more success. (79)

Two firms bid in the 2003 auction, and the winning bidder asked for approximately US$11.9 million to do the project. (80) The winner appeared to be on track to meet its first three rollout agreements by the end of 2004. (81) The company notes that after rolling out service to more than 500 villages in 2004, it now serves "over 1,800 sites" and plans to expand service into western Nepal. (82)

F. Peru

Peru conducted reverse auctions from 1999 to 2001 for service in areas the regulator determined unprofitable. These included rural towns as defined by the National Institute of Statistics and Data Processing, district capitals with 3,000 inhabitants, areas without basic telecommunications services, sparsely populated areas, isolated villages, and poor areas. (83) The Organismo Supervisor de Inversion Privada de Telecomunicaciones (OSIPTEL) plan was to first auction subsidies for payphones, followed by Internet access, and finally subscriber-fixed telephony. (84)

For the first auction, firms bid for the twenty-year non-exclusive licenses to provide service in six regions of the country. (85) Winning firms were required to install at least one public payphone in each rural locality and public Internet access in each district capital. (86) The regulator had allocated US$150 million for the project, paid for by a one percent tax on all telecommunications revenue. (87) The bidding process reduced the total allocated to US$50 million. Winning firms used a range of wireless technologies, including Very Small Aperture Terminals and wireless local loops. (88)

The number of telephones and payphones per capita increased substantially following the auction process. While the auction seems to have effectively reduced the subsidy granted for providing these rural services, several factors make it difficult to truly evaluate the program's effectiveness.

First, countries around the world began liberalizing their telecommunications sectors in the 1990s, leading to rapid increases in investment. (89) An increase in Peru, therefore, cannot simply be attributed to one policy intervention absent a well-designed test of its effectiveness. Second, some winning firms did not meet their rollout obligations. (90) Assuming corruption was not a factor, a "winner's curse" might have left firms unable to provide service profitably. That is, the winning firms may have underestimated the costs of meeting the obligations and bid too little. (91) Finally, winning firms were given spectrum rights to provide service. The true subsidy, therefore, includes not just the US$50 million granted to the winning firms, but also the opportunity cost of these spectrum rights.

Print Get the Mag Weekly Updates