Infrastructure-based Competition in NGA: The Best Way Forward?
According to EC’s recommendation, member states have to encourage infrastructure-based competition as the best and fastest way for broadband development [1]. At first, the arguments behind infrastructure-based competition policies seem solid. Competition does good to the market: provides efficiency incentives to operators, reduces prices, increase penetration, etc. This seems to cheer everyone: 1) Customers are happy for receiving cheaper service, 2) incumbents are happy because alternative networks start spending money in infrastructure, and gradually stop cherry picking over their infrastructure, 3) operators see higher penetration rates as a way to increase their market potential and revenue streams. In all this ideal setting, regulators’ overall efficiency, as measured by EC, increases.
However, broadband development and infrastructure-based competition come with some significant costs. Social benefits from higher broadband penetration (stemming from infrastructure competition) are offset by a) operation inefficiencies of duplicating/redundant infrastructures and b) the cost of laying out these infrastructures.
So, here’s the billion dollars question: Do the benefits outweigh the costs? Does network competition truely advances the market (and at what extent) or do competitors simply eat out each other’s profits. Infrastructure competition can lead to price wars (when you compete at the conduit level, what else there is to compete over) which then makes the business case of the networks worse. Intuitely, this means that only those with the deeper pockets are going to survive. International experience provides the evidence: Telco’s bankruptcies and consolidations imply that indeed firms’ benefits (producer surplus) is negative.
For that, telcos seem to initially choose unserved geographical markets to deploy their networks and avoid direct competition, as Benoit points out , but this will not necessarily last for ever .
Thus, we are faced with a serious potential deficit: Socially sub-optimal over-spending in duplicating, parallel infrastructures. Would it be possible that infrastructure-based competition does not serve social welfare in the long-run? Should we look beyond short-term gains of increased broadband penetration?
Maybe, it should be worth revisiting the current policies aiming at quick broadband penetration, especially in these times where huge piles of money are being prepared to be invested in physical infrastructure. Public, publicly-owned or public-private physical infrastructure operated under open access rules is a serious alternative to the current policy guidelines and it must be re-assessed by regional/national regulators.
—
[1] EC also suggests that in the regions that experience a market failure the Government must intervene and subsidize, finance and/or facilitate physical infrastructure on some sort of public ownership terms (wtih the aid of EU funds). The three most evident examples of EU funding in regional/municipal physical infrastructures are Sweden’s in the late 90’s early 00’s, Ireland’s in 2003-2006 and Greece’s in the period 2005-2008. In all three cases, municipal metro networks were built and FTTB connections to public buildings (education, health, public administration etc) were constructed.
Fiber is cheap, ducts are expensive. Therefore, it makes sense to share ducts in most cases but the case for sharing fiber is weaker. This is certainly the case in the so-called vertical outside plant. In buildings, the situation is a bit different where multiple fibers can be difficult to install for each apartment and landlords are often hesitant to let multiple providers install infrastructure (they open the door only once, if at all…)
From a competitive point of view, infrastructure based competition, rather than service based competition, also stimulates innovation. This we know from the DSL world where LLU is gaining ground in favor of bitstream access. Reason being that competition at the lowest possible layers allows more degrees of freedom to differentiate. And also avoids a situation where we actually create a monopoly on the lower layers.
So infrastructure competition is in principle a better choice. But sometimes such competition is not technically feasible (e.g., if there are constraints in ducts or sewers and multiple fiber cannot be installed) or economically feasible (e.g., in some rural areas where the business case for FTTH can be rather weak). This is where bitstream access/wholesale/open networks (whatever you want to call it, it’s all the same L2/L3 wholesale offering in the end) makes sense. But don’t impose it, if it happens let it be.
Thomas