When FCC Chairman Tom Wheeler said back in February 2015 that his agency was not interested in regulating the rates charged by Internet service providers, he neglected to offer any caveats. And for good reason: The more caveats, the less powerful his message of “forbearance.”
What he meant to say was, “No rate regulation for residential broadband offerings. Business broadband is fair game.” But that doesn’t have as nice a ring to it.
The FCC Chairman revealed his intention when he publicly voiced his opposition to a bill that would neuter the agency’s ratemaking authority. Clearly, the FCC wants to preserve its right to engage in ex post rate regulation of residential broadband offerings.
The FCC is even more clear about its intentions for business broadband. The agency already embraced rate regulation for wholesale rates charged to resellers of business broadband. In August 2015, the FCC imposed an “Interim Rule” that compelled telcos (but not cable operators) to make IP-based broadband connections (running over fiber connections) available at “reasonably comparable rates, terms, and conditions” to resellers if the telco decommissions its TDM-based services (running over copper connections) to a building.
The FCC is now doubling down by considering a proposal from the resellers’ trade association that would bring rate regulation to retail rates in its “special access” proceeding; a telco’s Ethernet services offered over new fiber networks would be subject for the first time to pre-approved price caps. So much for forbearance. [Update: Yesterday, the FCC issued a news release announcing a forthcoming NPRM that would solicit comments on how to structure price regulation of business broadband.]
Basic economics suggest that, by truncating returns on investment, price controls undermine a network operator’s incentive to invest. But by how much? A new paper I released last week on behalf of USTelecom models how fiber deployment to business districts likely will unfold with and without rate regulation.
To model the impact of retail or wholesale price regulation, one first needs to model the “baseline” level of investment—that is, predict how many buildings in a given city not currently lit for fiber would be wired over the coming years if the FCC does nothing.
This is where things get tricky. The cost to wire an unlit building in a given city depends on the distance of that building to the telco’s existing fiber footprint. To understand the distribution of distances (and associated connection costs) for unlit buildings, we selected a city (Charlotte, North Carolina) that was representative of the population of cities in terms of number of total establishments, the number of large establishments, and large establishments per square mile.
Our investment model was aimed at simulating AT&T’s fiber-expansion plan in Charlotte. Although the same model could be used to simulate the likely deployment patterns of competitors, who collectively serve over 267,000 buildings with fiber across the nation, the greatest investment impact of price regulation aimed at incumbent telcos will be (drum roll please) on incumbent telcos.
Today, AT&T’s fiber network reaches roughly ten percent of all buildings with over 20 employees in Charlotte. Although we cannot see the layout of AT&T’s fiber network, using the location of AT&T’s lit buildings, we can construct the most efficient network connecting those buildings using a minimum-spanning-tree algorithm. Next, we calculated the distance from the existing network to the closest unlit buildings, which can be translated—with estimates of underground and aerial fiber costs per mile—into the building’s fiber-connectivity costs.
We combined this building-specific cost data with an estimate of telecom revenues for the building. Unlit buildings that generated positive cash flows net of capital expenses in 18 months were assumed to be lit, which expanded the existing network and reduced the distance to serve the marginally unprofitable (and unlit) buildings. We allowed this iterative process to bring more and more buildings onto the network. By the time the process was complete, AT&T’s fiber network covered 20 percent of all buildings in Charlotte. This is not to say that 20 percent is the long-term maximum; a future technology shock could reduce AT&T’s costs, making marginal buildings profitable.
To gauge the impact of price regulation, we started the simulation over from scratch, this time reducing the expected revenue in each unlit building by 30 percent. As it turns out, the economic literature shows that prior unbundling efforts and price-cap rules have been associated with price declines of this magnitude. Which regulation ultimately causes Ethernet prices to fall is irrelevant; the point is that some unlit buildings that would otherwise have been served are no longer profitable to serve.
The results were sobering: Rather than expand to 20 percent penetration, AT&T would halt its investment in Charlotte at 14 percent. Put differently, 324 buildings in Charlotte that would have been lit under the baseline would no longer be profitable to serve under rate regulation. Because it costs on average $80,000 to wire a building, this means that AT&T will invest roughly $26 million less in Charlotte due to price regulation (a 55 percent decline in investment).
We extrapolated these results to the population to determine what the investment impacts would look like assuming the Charlotte experience was representative. If there is no rate regulation, incumbent telcos would light up nearly 122,000 buildings nationwide, representing $9.9 billion in capital expenditures and 4,900 new fiber route miles. Price regulation would cut projected investment (again by 55 percent) to an estimated $4.4 billion, providing fiber to only 55,100 buildings with 2,200 new fiber route miles.
Policymakers with a fixation on prices might not be moved by investment declines in the abstract. To put a human touch on these declines, we used fiber-specific employment multipliers from the literature. Because every $1 million in fiber investment supports 20 full-time jobs through the multiplier effect and another 20 jobs via the spillover effect, removing slightly over $1 billion per year (equal to $5.5 billion reduction spread over five years) means 43,000 fewer jobs.
The purpose of the exercise is to show policymakers the tradeoffs associated with intervening in today’s competitive broadband markets. While the FCC has a role in promoting the public interest, it also has a mandate to promote broadband deployment. Rate regulation is anathema to that mandate.
[[This piece originally appeared in Bloomberg BNA on April 22, 2016. It has been reposted here with Bloomberg BNA’s permission]]