pproaches 43103 F e b r u a r y 2 0 0 8 N o t e N u m b e r 1 8 Output-Based Aid in Mongolia Expanding telecommunications services to rural areas By Andrew Dymond, Sonja Oestmann, and Scott McConnell A ccess to telecommunications services has been ex- tremely limited in the remote and sparsely populated areas of Mongolia. Several factors have conspired against achieving universal access on a purely commercial basis--the country's vast and challenging geography, the no- madic lifestyle of the rural population, government ownership and incumbent control of the long-distance transmission net- work. As first steps in rolling out a universal access program, two pilot projects are bringing phone service to remote herder communities and both mobile phone and Internet services to rural villages. To support the delivery of these services, one- time subsidies were competitively awarded to licensed telecom- munications operators in Mongolia. These are output-based subsidies: the winning bidders take on the investment risks of expanding their networks, receiving subsidy payments largely only after meeting service targets. The operators are obligated to continue offering the services for the duration of their 5-year service agreements without further subsidy. Mongolia has the lowest population density in the world. Its population of 2.7 million is spread over a vast territory--featuring wide plains, high mountain ranges, and the Gobi desert. Its rural inhabitants, numbering just over 1 million, are spread very thinly. About a third of these live in 330 soum (district) cen- revenue. The mobile operators are growing much more ters. The rest are herders. Largely nomadic, they move rapidly than the partially privatized incumbent opera- with their herds between different summer and winter tor, Mongolia Telecom, which is facing market chal- locations. lenges in both voice and Internet sectors. The country's per capita GDP, though only around A third mobile operator won a license in 2006 and $950, has doubled over the past four years thanks started offering services the same year, and a fourth was mainly to rapid growth in mining and tourism and the licensed in 2007. The incumbent operator offers wireless impact on the rural economy of better winter weather local loop service in many localities and is expected to conditions. But significant urban-rural inequality has secure a fifth mobile license in 2008. The plans of the emerged, and many rural areas have seen their popula- mobile operators suggest that by the end of 2007 al- tion decline by 10 percent or more in the past few years. most 200 of the country's 330 soum centers will have at The telecommunications market was effectively liberalized by the early entry, in the 1990s, of two Andrew Dymond is the managing director, Sonja Oestmann private mobile operators. These operators--one using a senior partner, and Scott McConnell a senior rural the Global System for Mobile (GSM) standard and the telecommunications consultant with Intelecon Research & other Code Division Multiple Access (CDMA) technol- Consultancy Limited, Canada. ogy--account for more than 80 percent of all customer Special thanks to Tenzin Norbhu and Peter Silarszky, World connections and more than two-thirds of sector Bank task team leaders for the Mongolia OBA project. Supporting the delivery of basic services in developing countries pproaches least one mobile service and 150 of them will have two. community served, with per capita subsidies ranging Most will offer the possibility of broadband access. between $5.28 and $14.45, depending on location. This estimate was expected to provide the operator Toward universal access with a commercially sustainable rate of return after the one-time subsidy. The soum center wireless pilot was When Mongolia conceived its universal access program expected to require a total subsidy of $63,000­73,400, in 2004, public telecommunications service in most or $47.90­55.82 per inhabitant (table 2). The projec- rural areas was inadequate and unreliable. Service was tions for both pilots were within the range of subsidies available only at Mongolia Telecom offices in the soum granted in similar output-based subsidy competitions centers and only during business hours. Until recently in other countries (Chile, Peru, Uganda). there was no service at all outside the soum centers. Cost-benefit studies using consumer surplus cal- Setting targets culations showed that both pilots would yield positive economic returns on the subsidies. Economic analysis The universal access program set specific targets, such for the Internet service and school subsidies did not as having at least one mobile or wireless operator in show immediate economic returns, but it is very rare in each soum center, offering broadband wireless Inter- such cases to be able to demonstrate direct impacts in net service in some soum centers, and, having at least rural emerging markets. Instead, the investments repre- public access telephone service in the country's 1,500 sent "market seeding" to have economic benefits in the remote herder communities, or baghs. long run--with evidence of such benefits beginning to To finance network expansion into rural areas, the emerge in countries as diverse as Hungary, India, Peru, country's Communications Law of 2001 had called for and Uganda--often as a result of "seeding" by OBA or a universal service obligation fund. Later legislation other forms of assistance. stipulated that operators will contribute to the fund through a 2 percent levy on their taxable income start- ing in 2007. This fund was set up with the assistance of The bidding the World Bank through grants from the Public-Private Mongolia's Communications Regulatory Commission Infrastructure Advisory Facility (PPIAF) and the Global conducted separate minimum subsidy auctions for the Partnership on Output-Based Aid (GPOBA). The two pilots. Each request for proposals specified the World Bank also provided seed financing of $5.5 mil- maximum allowable subsidy and included a draft ser- lion for the initial universal access subsidy projects. vice agreement. These agreements set out the schedule of subsidy awards: 20 percent on contract signature, Designing pilots 60 percent on meeting progressive targets, and the last Two pilot projects, supported by a $259,400 GPOBA 20 percent after nine months of successful operation. grant, were designed as the first steps in rolling out the In addition, the agreements required the winning bid- universal access program: ders to provide a performance guarantee, to remain in Herder public access telephone network--to provide place until the 36th month of satisfactory operation. public access telephones for herder communities Bidders were required to submit evidence of cor- in 27 baghs, spread among 6 soums in 2 provinces porate and financial stability, experience in operating (aimags), Arkhangai and Bayankhongor. rural networks, and experience in operating in Mon- Soum center wireless network--to provide wireless golia. Eligible bidders would be evaluated first on the voice and data services to one soum center in technical and operating compliance of their bids and Tariat soum in Arkhangai, that had previously had only limited service. Besides providing mobile voice and Internet services to the public, including a Table 1. Expected subsidies for herder network pilot commercial cybercafe, the operator's responsibili- ties include providing subsidized Internet access to Total the school for three years, with a declining subsidy. Total population of the 6 targeted soums 24,824 Herder population 19,942 The expected subsidies Estimated subsidy per herder (US$) 7.69 The herder network pilot was estimated to require an Maximum allowable subsidy per herder (US$) 10.91 average one-time subsidy of $5,100­6,200 per herder Source: Intelecon. Supporting the delivery of basic services in developing countries pproaches Table 2. Expected subsidies for Tariat soum center wireless pilot, by component Estimated Maximum allowable Estimated subsidy Maximum subsidy Pilot component subsidy (US$) subsidy (US$) per person (US$) per person (US$) Digital voice telephony 31,000 37,500 23.57 28.52 Internet point of presence and ICT 23,000 26,000 17.49 19.77 Subsidy to operator 54,000 63,500 41.06 48.29 School Internet access support 9,000 9,900 6.84 7.53 Total subsidy 63,000 73,400 47.90 55.82 Source: Intelecon. then, in a second stage, on their required subsidy. The to be commercially less attractive, the competition, winning bidder would be selected from the compliant based on similar tender documents, again attracted bids solely on the basis of the lowest subsidy required. two bidders, including the winner of the first tender. Each tender attracted two bids from operators already This competition was won by the second (CDMA) mo- active in the target markets. Three of the four were in bile operator, which requested a $43,000 subsidy. substantial compliance. This subsidy, though much higher than that awarded for Tariat soum, was also significantly below expecta- Herder network competition tions. Both were won under highly competitive condi- The winning bid for the herder network pilot, submit- tions in which the winner made a low offer for a strategic ted by one of the country's two leading VSAT opera- reason: to gain market brand advantage and demon- tors with existing rural services, required a subsidy strate collaboration with the government's universal ac- of $180,480. This was $6,000 below the allowed cess program for rural communities. Such market-driven maximum. This bidder also offered to provide services "efficiency bonuses" are not uncommon in initial OBA beyond those required by including two additional competitions in the telecommunications sector. remote soum center service points. The second bid required the maximum allowable subsidy. To ensure year-round access to services by the Outputs so far--and lessons nomadic target group, bidders had been asked to The networks in both original pilots were implemented propose configurations for VSAT terminals that would in September­November 2006 and fully operational allow portability--a unique requirement, since VSAT before the 2006/07 winter season. Both operators met terminals are usually stationary. The winning solution targets for service availability and technical quality proposed installing several separate concreted metal ahead of schedule. The Chuluut soum center network pedestals in each community that would make it pos- has also been installed, with service formally inaugu- sible to move the VSAT terminals between summer and rated in June 2007. winter grazing areas. For the herder network pilot, revenues stayed below expectations in the winter, mostly as a result of tech- Soum center wireless competition nical problems (now resolved) with the solar power The winning bidder for Tariat soum center was the supplies and herders' unusual migration patterns. After country's leading GSM mobile operator. The bid of- six months of operation, however, revenues were rising fered a full GSM base station for the voice services toward a more commercially sustainable level. A user and a Wi-Fi network for the Internet, both hubbed to survey showed that the project has cut by more than Ulaanbaatar by a VSAT link. Competition was intense, half the distance herders must travel to reach a phone resulting in a winning bid of $16,455, only about 25 and more than doubled the frequency with which they percent of the maximum allowable subsidy. The win- use the telephone service. ning bidder effectively asked for subsidy support only The Tariat soum wireless operator has reported that for the Internet and school service components. two-thirds or more of the community's households The resulting savings were reallocated to an addi- have become customers, in line with the forecasts of tional tender, one for similar services in Chuluut soum, the universal access demand survey. Moreover, ben- in the same province. Though Chuluut was considered eficiaries' monthly spending on calls has exceeded the Supporting the delivery of basic services in developing countries pproaches national average. A large share of regular customers are Skills transfer and national program ownership herders who live well outside the range of the mobile were at good levels at the critical times. base station and travel regularly with their handsets to Tenders were prepared in consultation with use the service. operators and stakeholders. Take-up and use of the Internet service have been Strong competition in the sector--together with lower than forecasted. But the first operator, while not the traditional importance of the countryside and making a profit from the service, is using the experience the project's least-subsidy design--has been a key to learn about Internet demand. Moreover, the opera- factor in overcoming the higher marginal cost of tor expects the business to be seasonal, with demand providing access and connections in remote areas. in some key months going beyond the one or two "an- The GPOBA subsidy enabled the program to be pi- chor customers," such as the country's leading rural loted before the operators started to contribute to bank. This experience suggests that the universal access the universal service obligation fund. That helped program, will need to be selective with respect to both demonstrate the fund's benefits for the operators, locality and pace of development. It also points to a increasing its acceptance. need for more training for teachers and students to help them benefit from Internet access. Next steps Key success factors In late 2007, the universal access program is already running a major competition to award subsidies for The pilots are considered to be a success: the bidding more mobile base stations to be installed in 2008 competitions resulted in fully compliant bids below in the remaining soums that would not otherwise the estimated maximum subsidies. The pilots have receive coverage. A similar competition for VSAT provided important knowledge about rural demand projects in the most remote areas will be announced for better communications services and allowed the in the first quarter of 2008. Development of Internet testing of technical and commercial solutions that can service provision will proceed more slowly, focusing be adapted to local conditions, including a nomadic on soums that have strong voice service markets, population. And the experience appears to have helped "Internet ready" schools, and the best chance of accelerate even commercial expansion by all the coun- sustaining commercial Internet services after the one- try's licensed operators. time subsidies. The program will eventually enable Several aspects of the Mongolian experience have broadband Internet to reach the entire country. led to this success--and also promise success for the But the pace of development must be realistic to entire universal access program: ensure that money is well invested, that risks are The Communications Regulatory Commission has minimized and benefits realized, and that the market been strongly committed to the program. is sustainably developed. About OBApproaches OBApproaches is a forum for discussing and dis- The case studies have been chosen and presented seminating recent experiences and innovations by the authors in agreement with the GPOBA for supporting the delivery of basic services to the management team, and are not to be attributed to poor. The series will focus on the provision of water, GPOBA's donors, the World Bank or any other af- energy, telecommunications, transport, health and filiated organizations. Nor do any of the conclusions education in developing countries, in particular represent official policy of the GPOBA, World Bank, through output, or performance,-based approaches. or the countries they represent. To find out more, visit www.gpoba.org The Global Partnership on Output-Based Aid Supporting the delivery of basic services in developing countries