Needed, a national fibre optic network
Information and communication technology (ICT) has been shown to be a powerful facilitator for meeting the Millennium Development Goals. This supports the rationale for the roll-out of Internet access as an enabler of development. However, the provision of basic services using ICT is dependent on the availability of other complementary inputs. This means the decision on the level of a particular ICT service to be provided cannot be made without reference to the presence of those other inputs. The build-out of networks far in advance of the absorptive capacity of a region would therefore appear to be wasteful.
However, in an academic paper titled Universal Service Obligation in the Age of Broadband published in The Information Society, one of the authors establishes two criteria that could be used to support the build-out of networks in advance of the ability of target populations to use them. The first is “time to build”. If ICT infrastructure takes a long time to deploy, then the project needs to be initiated in anticipation of future absorptive capability. While wireless broadband is amenable to relatively rapid build-out, taking a fibre optic network from the last viable point (for example, the district headquarters) to the village can take two to three years.
The second is “technological discontinuity”. The capabilities of the technologies used for providing access develop in a discontinuous, step-wise manner. Each step represents discontinuous jumps in access speed per dollar of investment. At a certain point, rural areas must switch from lower to higher technologies due to the constraint of download speeds. The earlier they do so, the lower the total capital costs of connectivity.
While the optimization of an existing network requires little time to build and represents low technological discontinuity, the build-out of an optical fibre network involves high time to build as well as a high level of technological discontinuity. A fibre optic network is a “durable” solution in the sense that once deployed, it would take care of rural connectivity needs for many years to come, (although last-mile access would continue to be wireless). For these reasons, a case can be made for the government to get involved in the deployment of a national optical fibre network.
Who should pay for the fibre optic network? Recall that the fundamental driving force for government intervention stems from the role of connectivity as an enabler of development. Hence, the government’s financial obligation needs to be limited to the level of connectivity required to enable the provision of the requisite amount of developmental goods. The levels of provision of services implied by high-speed connectivity (especially through a fibre optic network) are so high that one may question the necessity of governments paying for the entire network. After all, these levels of development may not be a consumption norm or a systemic necessity in the society in question. As an example, governments do not normally place an emphasis on creating super-specialty hospitals in rural areas.
The degree of production and demand externalities that accrue after threshold penetration levels are reached may also not justify the expense. Therefore, the cost should be shared between the public and private sector with the public sector paying for the basic level of connectivity required to provide development inputs and to internalize the positive impact arising from demand and supply externalities. The calculation of this level would depend on commercial and technical factors, and may not be amenable to an elegant analytical solution. Yet, the point remains that the liability of the government needs to be limited.
However, the lack of a business case for deployment in the vast majority of target geographies, and the thorny business environment, make it difficult for the private sector to bear the initial investment. Therefore, the government should pay for the upfront costs of building the network and collect a revenue share for a specified number of years.
In parallel with the fibre optic project, a wireless network closely aligned with complementary inputs and the absorptive capacity of the target population should be rolled out to provide basic necessities immediately and prepare the population for the coming of the fibre optic network. Superimposing lofty urban wireless standards on these interim rural networks in the cause of parity would be profligate and short-sighted.
The option of using a build-own-operate-transfer model for building the national optical fibre network is not recommended on account of the operating challenges of rural networks, the low ability to pay, and the uncertainty regarding the availability of complementary inputs that would trigger a virtuous cycle of usage. The operating competence of the private sector should be leveraged by tendering projects for building and maintenance through a reverse auction process.
The national optical fibre network should be divided into a number of state-level projects in order to secure the buy-in of state governments, crucial for obtaining right-of-way permissions. Vertical integration of the private infrastructure operator and the service provider should be permitted in order to strengthen the business case and trigger operational efficiencies. A phase-wise roll-out should be planned: the universal service need not be a uniform service. First, the economically well-off subset of the 250,000 gram panchayats should be targeted. After demonstrating success in these clusters and incorporating lessons learnt, further roll-out should take place.
The aim is to create a humming optical fibre network, not a white elephant.
Rohit Prasad and V. Sridhar are, respectively, professors at MDI Gurgaon and IIIT Bangalore.
Comments are welcome at firstname.lastname@example.org