Broadband access to every citizen is a key pillar of Digital India. The issue is more significant for the rural population.

The BharatNet project has a vision to establish a scalable network by 2017 towards providing affordable broadband connectivity of 2 Mbps to 20 Mbps to all rural households and institutions.

This project has evolved from the earlier National Optical Fibre Network (NOFN) project of providing 100 Mbps to all gram panchayats (GPs). This project was approved in 2012 but made slow progress. One of the main concerns of this was the slow implementation by three central public-sector undertakings (CPSUs)—Bharat Sanchar Nigam Ltd (BSNL), Power Grid Corp. of India Ltd (PGCIL) and Railtel Corp. of India Ltd)—diffused control and the relation between Bharat Broadband Network Ltd, the company managing the project) and the CPSUs, and lack of appropriate ownership by BBNL.

A committee constituted in January reviewed the earlier project and proposed a modified project called BharatNet, in a report submitted on 31 March. The revision included almost three times the fibre length, a mix of infrastructure technologies, at twice the capex ( 73,000 crore) but with a significantly reduced opex, so that it works out to be a lower total project lifecycle cost.

The content provisioning for the end user is envisaged by private service providers (PSPs), with competition, to be selected through an auction route. This is also the reason for the reduced opex, where it has been assumed that 75% of the costs would be recovered through auction revenue.

This renders the cost comparison between BharatNet and NOFN unfair since NOFN had not incorporated any revenue model.

While costs and revenue can be debated, the real issue is that quality broadband to every citizen is essential infrastructure. The main challenge is getting it done through a sustainable framework.

Starting with the end user and the PSPs, affordable connectivity would require regulation on pricing and ensuring availability of certain essential content, including those provided by the government.

A welcome idea of BharatNet is that states are to be given the choice of having BBNL drive the implementation or drive it themselves (through an special purpose vehicle with BBNL). There are two issues of implementation.

The first is whether the backbone network, as envisaged in BharatNet, should be implemented by BBNL/states as EPC (Engineering, Procurement and Construction) or BOOT (Build-Own-Operate-Transfer). The unbundling for this decision could be at a district level, which forms the basic unit of technological integrity.

A second related one would be, if BOOT, then who would engage the PSPs, BBNL or the BOOT operator? To avoid conflict of interest, it should be ensured that the BOOT operator does not also become one of the PSPs.

On the first issue, considering both the criteria of speed and quality of implementation, with a 2017 target, our recommendation is to go for BOOT in those districts where we can get operators of reasonable standing, and within an acceptable VGF (viable gap funding) annuity.

The VGF annuity would be discovered by bidding. The acceptable level could be 40% of the capex cost or such that at least 50% of the districts get an operator. Each district would go to one BOOT operator who could, of course, bid for and get more than one district.

The possibility of getting such operators would increase if the second issue of engaging the PSPs is bundled, rather than done by BBNL/states, since the BOOT operator then has a greater commercial incentive and can package offers for the PSPs. Since the BOOT operator would be in a monopoly position, fibre offer rate must be capped by regulation, at least for the regulated end services.

Those districts, which do not get taken in the first round, need to be bid out again after the first round of districts start putting out content and generating revenues. The perceived risk by bidders would be lower here. If even in the second round the bidders do not come, then the EPC route could be chosen.

In terms of other infrastructure sectors, an unbundling between the hard infrastructure and service provisioning has helped. Roads, ports, airports, and rural telecom are examples. It has enabled a level playing field and avoided conflict of interest. Further, BOOT models in infrastructure creation has seen good project management resulting in quality assets, with ownership for maintenance and operations. Issues have arisen where the revenue model has not worked out, with government reluctance to provide more. Hence, EPC gets considered as an option.

The key issue is with governmental ownership. Can we create the same quality of assets and maintenance, at costs comparable to what needs to be shared with the BOOT operator?

We conclude by saying that BOOT is a preferred option, with PSP engagement being part of it. A regulator to control end user pricing and fibre pricing between BOOT operator and PSPs, both in a limited context, and for licensing and service quality oversight would be essential. And finally, prior homework in generating appropriate contract documents would be critical.

Prof. G. Raghuram specializes in infrastructure, transport systems, and logistics and supply-chain management. He conducts research on the railway, port, shipping, aviation and road sectors.

Prof. Rekha Jain focuses on policy and regulation in telecom sector and IT. In telecom, her interest areas are Internet governance, spectrum auctions and net neutrality.