Mobile Edition
Print
Get the Mag
Weekly UpdatesC. How Much Should We Spend and Where?
A key problem with universal service is deciding what subsidies are necessary and how to distribute them. In principle, universal service subsidies are necessary when it is not economic for a firm to provide service. In that case, the ideal subsidy would equal the gap between the level of investment a firm would be willing to make and the investment required to provide service.
This cost-based approach has several problems. First, our regulatory history demonstrates that it is not possible to accurately calculate the true costs of providing service. (25) The task becomes more difficult when the provider has every incentive to make the cost of service appear high. Second, it becomes difficult to change once a firm is providing subsidized service. Potential new entrants would have to compete with a subsidized incumbent. Subsidies could be made available to those firms too, but that risks driving up the cost of the program.
Reverse auctions do not address the way in which universal service funds are collected. Instead, they focus on how those funds are distributed. When designed properly, auctions are a tool that can induce firms to reveal their best guess as to how much it would truly cost to serve an area. Next, different countries' experiences with reverse auctions are discussed.
III. GLOBAL EXPERIENCE WITH REVERSE AUCTIONS AND UNIVERSAL SERVICE
Subsidy auctions have been used elsewhere in the world with some success. This Section investigates auctions in Australia, Chile, Colombia, India, Nepal, and Peru. In a fair bidding process with multiple bidders, firms should bid the smallest subsidy necessary for them to provide service. Global experiences reveal that auctions are feasible and that the subsidies required are generally less than incumbents had previously led policymakers to believe.
A. Australia
In 2000, the Australian government decided to pilot the use of reverse auctions to distribute universal service subsidies in certain areas (see Figure 3). (26) Firms--both the incumbent and its competitors--were to bid for an $85 million subsidy to provide standard telephone service in 2003-2004. (27) This subsidy previously would have been available only to Telstra, the incumbent. (28) Bidding was to open in July 2001. (29)
As it turned out, none of Telstra's competitors bid to provide service in the pilot regions. (30) The Australian Department of Communications Information Technology and the Arts (DCITA) (31) reported that the competitors explained that the subsidy was too low for them to compete with Telstra given Telstra's existing installed capacity and information asymmetry. (32)
DCITA noted that while the results of the pilots were disappointing in that they did not lead to competitive entry, several factors contributed to the outcome, all of which may provide useful lessons. (33) First, the auctions took place at the beginning of a major downturn in telecommunications markets worldwide. (34) It is possible that firms were especially risk-averse during this time. Second, the auctions may have revealed that the existing subsidies were not excessive. (35) Finally, they highlight the need to consider carefully the role of the incumbent when designing these auctions. (36) As discussed below, India faced similar problems with respect to its incumbent provider.
The unique position of the incumbent raises the important question of identifying the goal of a reverse auction program. In Australia, the goal was to introduce competition. (37) As Australia's experience shows, however, introducing competition and reducing subsidies are not necessarily consistent, at least in the short run. (38) As the DCITA pointed out, encouraging competition may have required it to increase its spending on universal service. (39)
[FIGURE 3 OMITTED] (40)
B. Chile
Chile created its Fondo de Desarrollo de las Telecommunications (Telecommunications Development Fund) in 1994 to provide payphones in rural and low-income urban areas. (41) Regional and local governments submitted requests for payphones to the regulator, who then determined a maximum allowed subsidy to make the phone commercially viable. (42) Any firm could bid to provide the service and the winner received a nonexclusive thirty-year license. (43) The resulting average subsidy was US$3,600 per payphone, compared to the US$10,000-US$20,000 the government had paid previously. (44)
The average subsidy masks two other results that emerged from the bidding process. First, winning bids tended to be either very close to the maximum allowed subsidy or zero. (45) The dominant local firm bid 100% of the maximum subsidy in areas with no competitors which were close to its existing network, 90% of the maximum subsidy in areas with an emerging competitor which were close to its network, and zero in areas with strong competition. (46) Likewise, the satellite firm Global Village Telecom (GVT)--a Gilat Satellite Networks Ltd. subsidiary, which was a new entrant--bid 100% of the maximum in areas with no wireline network and did not bid elsewhere. (47)
Intelecon Research and Consultancy Ltd stated, "Chile's fund, which has been in place for four years, did not need to use subsidies at all in 656 of the villages it supplied with telephony, and managed to cover 77% of the designated villages with only 54% of the US$13.3 million of financing it had available." (48)
The second result was that bidding competition decreased steadily as the auctions proceeded. (49) Figure 4 shows that the average winning bid increased from 40% of the maximum subsidy during 1995-1996 to nearly 100% of the winning bid by 2000. (50) Bjorn Wellenius attributed this change to consolidation among telecommunications providers. (51) Other explanations, however, are also plausible.
It is possible, for example, that the regulator-auctioned areas were expected to be more profitable initially. (52) In that case, firms would be willing to pay more and accept less to serve those areas and would demand higher payments for serving the less profitable areas that were auctioned later.
Another possibility is that each round of auctions provided the regulator with additional information about the true costs of providing service. The regulator could have used that information to better estimate the maximum subsidy necessary to provide service. If this occurred, one would expect bids to come close to the estimated maximum.
[FIGURE 4 OMITTED] (5)
C. Colombia
Colombia first used subsidy auctions in 1999 through its Compartel Program after a 1998 government report showed that few rural areas had telecommunications access. (54) Intelecon Research & Consultancy described the broad goals of the program:
GVT won the first auction in 1999 and provided 6,745 telephones and 670 Internet access points. (56) GVT received about US$32 million out of the US$71 million that had been available. (57)
The second auction was held in December 2000. (58) Only one firm bid for subsidies to install "21,500 residential lines and 61 community Internet centers by April 2002." (59) The Communications Ministry, however, declared the results of this auction invalid due to "various anomalies and omissions in the information supplied by [the sole bidding company]." (60)
The third auction occurred in November 2002 for the installation and operation of 500 telecenters for telephone and Internet service and also for building a 3,000-site fixed satellite network for rural areas over six years. (61) GVT won this contract after bidding for US$65 million in subsidies out of the US$100 million that had been available. (62) Intelecon reported that this network was operational by the fourth quarter of 2003. (63)
D. India
India's Universal Service Fund (USF) is intended to reimburse the net cost (costs minus revenues) of providing rural telecom service. (64) Because costs may vary across different types of service and different service segments, separate auctions determine the actual reimbursement to he awarded for each. When awarding licenses for cellular telephone service, the Department of Telecommunications (DOT) divided the country into twenty telecom "circles" (which loosely follow state boundaries). These circles were used as the basis for geographic reference in the rural subsidy auctions.
Telecommunications firms submitted bids to provide service. The firm that bid for the lowest subsidy, as long as the bid was no higher than a set benchmark, was eligible to be reimbursed for that amount from the fund. Benchmarks were set using information primarily from the incumbent, Bharat Sanchar Nigam Ltd. (BSNL). Any firm with a license to provide basic or cellular service in the relevant service area was eligible to bid. The winner received a subsidy for seven years, subject to review after three years.
India held several auctions, each for different types of telecommunications services. (65) The first, in March 2003, was to install village public telephones (VPTs) in 520,000 villages. The second, in September 2003, was to replace about 180,000 Multi Access Radio Relay-based VPTs. The third, held in September 2004, was to provide additional rural community phones in about 46,000 villages. The fourth, in November 2004, was to install VPTs in the 66,000 villages that had no public telephone facilities. The fifth, in March 2005, was to provide direct rural exchange lines in 227 regions. The most recent auction took place in April 2007 to provide mobile services.
The auctions yielded dramatically different results. The first two subsidy auctions, relating to Primary VPTs and replacing Multi Access Radio Relay-based VPTs, were disappointing. In nineteen of the twenty circles only one firm bid for the subsidies, the incumbent BSNL. Not surprisingly, given the thin market, BSNL bid exactly the benchmark amount, which was the maximum subsidy DoT was prepared to provide. By the final auction, however, some firms even bid negative amounts, demonstrating that they were willing to pay to provide service. (66)
FEED