Entner: Special access — How government preference for some may mean higher prices for all

Wednesday, October 21st, 2015

Five years after the FCC called for data on the state of the special access marketplace from just a portion of the providers offering special access, the agency appears poised to modify contracts and embark on a new round of rate regulation based on market data from 2010 to 2012. This would not be a concern if, in fact, the market for special access services had stagnated in 2012 with prices and providers remaining constant; however, that is not the case. Why should we care if the FCC premises a new set of pricing regulations on outdated information? Let me explain.

Special access refers to the dedicated data connections that physically connect a business, an office park, a government building, a cell site, or other man-made structure to the larger public switched telephone network and the Internet. The number and kinds of companies offering special access services has increased substantially in the last few years. When Sprint issued RFPs for backhaul related to its Network Vision program, more than 20 providers responded. The pricing was so competitive – so low – that even at the most hard-to-reach sites where competitive pressure should be the smallest, at least one potential bidder decided not to submit a bid because they were unable to provide the service profitably.

There is much debate about the reality of the market for special access and market data is hard to get, but not impossible. Zayo is the largest stand-alone backhaul and special access provider in the country (it’s one of T-Mobile’s backhaul providers), and provides very timely pricing trend data. In its Q4 2015 Pricing Trends document, which Zayo publishes with its earnings release, we can follow the prices Zayo is charging in the market place. Traditional DS1 prices have declined from $1,147 per DS1 (also known as T1) per month to $783 per DS1 per month. For DS3s (also known as T3) prices have declined in the same time period from $4,081 per month to $3,269. Just to put this in context, a DS3 has 28-times the throughput of a DS1, for roughly four times the cost.

As part of the wireless industry’s drive to stay ahead of consumers’ appetite for high-capacity data services, building more backhaul has been essential. As Sprint has embarked on its Network Vision program, it has also revamped its backhaul provisioning. Gone are the days when Sprint was predominantly reliant on its direct competitors to provide backhaul; it now has a stable of thirty to forty alternative access providers, in addition to its own wireless backhaul in the 2.5 GHz range. T-Mobile has also been no slouch; almost all of its backhaul is now Ethernet fiber, which is part of the reason why its download speeds are so fast. The cost savings for both companies are substantial. In its Q4 2014 financial results, T-Mobile USA’s quarter over quarter cost of service was down 7.1% partially due to renegotiated backhaul contracts.

In addition to having a multiplicity of providers from which to obtain their special access lines, wireless companies continue to experiment with other solutions that can improve network performance and reduce cost. For example, companies are experimenting with self-backhaul where access and backhaul are part of the same system. This solution is intriguing but not currently viable in today’s spectrum environment due to the amount of spectrum needed to make it a reality. In the spectrum-constrained environment that characterizes the mobile market in the U.S., T-Mobile, AT&T and Verizon Wireless are all in the same boat. Because wireless self-backhaul could provide more options for carriers, this should be an additional reason for the government to redouble its spectrum clearing goals and aim for at least one gigahertz of spectrum for the wireless industry below 6 GHz.

It’s important to note that DS1 and DS3 lines are legacy products that are at the end of their technological life. As the industry has moved on to Ethernet connections, the number of DS3s sold by Zayo has declined from 3,569 in September 2013 to 2,772 in June 2015, a 23% drop in less than two years. For DS1s, that decline has been even more pronounced – from 3,569 to 2,772, a 38% decline from September 2013 to June 2015. It is perfectly understandable that the lack of new demand for DS1s and DS3s makes some providers hesitant about issuing new long-term contracts, as it would obligate them for many years to divert time and money to support a dying market. If we take Zayo’s data and project out the current decline rate then they will have stopped selling DS1s in three and a half years and DS3s in less than seven years. But these projections are deceiving, and likely too conservative, as declines are accelerating as the DS1/DS3 technology becomes increasingly obsolete.

Zayo’s data shows a massive shift to Ethernet connections, which are both faster and cheaper than DS1/DS3, and where the marketplace is essentially even as new entrants and incumbents are building capacity at the same time. Zayo’s data shows a steady increase in demand, as well as falling prices per unit. The way Zayo represents the data – grouped in 10 to 100 MB, 101 to 1000 MB, and above 1GB – shows that customers are buying larger and larger pipes for lower prices per unit, which is consistent with the commonly observed conditions in the marketplace. The market is so competitive that further industry consolidation among backhaul providers seems inevitable because so many competitors are barely, if at all, breaking even today.

Into this competitive dynamic steps the FCC, which seems to want to set prices. But nobody, including the FCC, knows how low special access prices can continue to fall which means the agency runs the real risk of setting rates at an artificially high level. In international markets where regulators set termination rates, everyone charges the government-set rate regardless of the actual cost, especially when the cost is lower than the government mandated rate. In the United States, where operators can freely negotiate call termination rates, per minute prices are $0.0007. In countries where the regulator sets the rate, the rate is higher than in the U.S. and Canada, as a 2012 OECD report shows. In the lowest-cost government controlled market, France, the termination rate was almost 20 times higher than in the U.S., with $0.0139, going up to $0.0878 in markets like Estonia, where the government-set rate is 125 times the rate U.S. operators negotiated with each other. Canadian operators that can also set prices freely went even further than their U.S. peers and eliminated termination rates entirely.

Mobile Termination Rates

Source: OECD Communications Outlook 2013

If the FCC were to set a price level – even if it is meant to be a price ceiling – the market would take it as a benchmark for what it should charge for special access charges. The shake out would continue with the weaker competitors selling to the lower-cost providers, but at a lower competitive intensity level and higher prices than if the FCC would have not intervened. The winners would be the special access providers that are able to offer service above the cost of the most competitive players, but below the government-set rate. It’s a classic rent-seeking scenario where marginal players ask for government intervention to safeguard their survival and increase their profits at the expense of end customers. The losers would be the end customers: businesses that have to pay the government-mandated rate when competition would have driven down prices below what the FCC deemed appropriate.

The impact of FCC intervention would be analogous to rent control in the housing market, which Economics Nobel Prize winner Gunnar Myrdal called “the worst example of poor planning by governments lacking courage and vision.”