Where does broadband competition come from?

Blair Levin, who oversaw the development of a National Broadband Plan, recently spoke at length on broadband competition. We shared an excerpt of his remarks in What do we want broadband competition to accomplish? Today we share Blair’s second of three foundational questions:

In my experience, there are two related answers.(1) The first goes to the nature of the competitive enterprise and the second involves an economic equation.

What kind of enterprises are capable of intensifying competition? Certainly existing competitors can do so. In addition, we often speak of the need for new entrants. In my experience new entrants come in three varieties:

• Greenfield new entrants, constituting new ventures;
• Adjacent market entrants who bring asymmetric assets and interests into the market; and
• Resale entrants who depend on inputs sold on a wholesale basis, a strategy that can include regulated access to unbundled elements.

All of these enterprises, both existing and new, follow similar economic patterns.

First, intensified competition always requires a new capital allocation decision by one of those four kinds of enterprises.

Perhaps now my conservative friends are nodding and my progressive friends are getting nervous. But as a factual matter we should all agree that every time we have seen intensified competition, it follows a company shifting its capital allocation from one purpose to the purpose of providing or upgrading a communications service.

A second pattern is that the new capital allocation decision follows a change in the same formula.

Ask yourself, why don’t we have more intense competition now?

The reason we don’t have greater competition is that the new or incremental capital (C) and operating expenses (O) of a network capable of intensifying competition are greater than the risk adjusted (1-r) new or incremental revenues (R), plus the benefits to the system(2) (SB), plus the risk of lost revenues due to competition(3) (CL)

That is,

C + O > (1-r)R + SB+ (-CL)

If we want to intensify competition, we have to change that math, causing, where possible, cap ex, op ex and risk to go down and revenues, system benefits and competition to go up.

Third, historically, the biggest changes in the competitive landscape in communications result from changes in the formula which themselves result directly from changes in government policy.(4)

This is where progressives get interested and conservatives get nervous.

I won’t do the full history here, but a few examples of companies reallocating capital to intensify competition should suffice:

• Cable intensified competition with broadcast television when government rules lowered its cap ex and op ex through pole attachment rules and copyright rules that gave it access to programming.
• Direct broadcast satellite (DBS) intensified multichannel video programming distributor (MVPD) competition when the government lowered its op ex by granting non-discriminatory access to programming and the telcos did so as well when local franchising monopolies were prohibited and then, state franchising was adopted, lowering costs for the telcos.
• Wireless began competing with wireline voice when the government both enabled more wireless competition with the PCS spectrum auctions and lowered its op ex by reducing the terminating access charges wireless had been paying wired providers.
• Cable began competing with the telcos dial-up Internet service when faced with the loss of revenue due to intensified video competition from DBS.
• Google devoted more capital to its fiber project when cities expressed a willingness to reform construction-related and other regulations in ways that reduced cap ex, op ex and risk, and increased potential revenues. In turn, the telcos facing Google Fiber competition were able to take advantage of these same streamlined regulations and have devoted more capital to fiber deployment, causing cable to accelerate deployment of its next generation product.

These examples demonstrate how policy affects capital allocation and competition. They also suggest not all kinds of enterprises are equal in producing long-term competitive effects.

Assistant Attorney General Bill Baer cites online video distribution as “disruptive innovation.” He explains, “some innovation comes from incumbents smart and nimble enough to take advantage of these new opportunities. But new entrants deserve a lot of credit, too. Companies like Netflix and Amazon offer consumers flexibility and control; established players like CBS and HBO have been forced to respond.”

I agree about the value of disruptive, instead of traditional, competition. Indeed, after some period of time, markets tend to stabilize and it is difficult to affect the incentives of existing players without introducing a new competitor or better and/or cheaper technology substitute.(5)

To bring improvements in price and quality to such mature markets, disruptive competition has proven key. Indeed, that decision on wireless to wired terminating access that I noted above, and a similar decision for data that enabled inexpensive VOIP is the reason the discussion of pennies per minute long-distance charges in now an anachronism.(6)

But I disagree in nomenclature with Baer. Wireless, VoIP, Netflix, Amazon or other disruptors are not really new entrants. Rather, they are adjacent market entrants. They had assets and motives different than existing players. The experience of the last 20 years suggests that the asymmetry of those assets and motives, if unleashed in an adjacent market, leads to far greater disruptions than existing competitors or new entrants in a mature market are likely to cause.

Similarly Google Fiber could be seen as a new entrant but it had both existing network assets to lower its cost structure and motive to improve its search business revenues through better broadband performance.(7)

With Gig.U, we worked with some true new entrants but those efforts failed, and as we discuss in the handbook, efforts involving true new entrants have a higher likelihood of failure.(8) Reflecting those experiences, I would argue that regulators should be cautious about betting on a true new entrant but rather look to strategies that enable asymmetric, adjacent market entry.

Unbundling can work to reduce prices but it discourages broad network upgrades. I think unbundling can be appropriate when the government finances the facility, as it did in the BTOP program, or when there are economic reasons that there is no appropriate way to make the economics work for providing an essential facility.(9) Some argued that we had reached that point in 2009 and bitterly criticized the National Broadband Plan for not recommending unbundling.(10) As of today, I think Google Fiber and other fiber efforts prove them wrong but we are still in the early innings. If those fiber efforts end before we reach bandwidth abundance in a critical mass of the country, then perhaps, the critics were right.(11)

In short, if we want intensified competition to deliver abundant bandwidth, we should be looking at how government affects that equation today, with particular attention to how it can incent adjacent market entry.(12)

1. I am consciously relying on my own experience rather than the uber text of competition, How Competitive Forces Shape Strategy, by Michael Porter, Harvard Business Review, March 1979, which lays out five forces that determine competition in a market. I would note that Porter’s work was brilliantly updated for the digital era in Unleashing the Killer App, (1998) by Chunka Mui and Larry Downes, which lays out how digitalization, globalization, and regulation/deregulation are overshadowing Porter’s five forces. My purpose here is not to fit what I have seen into either framework but to try to describe how policy has -- and could in the future -- intensified competition.
2. Benefits to the system refers to the benefits a service provider may obtain in markets outside of the area of the investment. For example, AT&T, by building out fiber in Raleigh, North Carolina, may derive some benefit in a market, such as Wilmington, North Carolina. In the experience of Gig.U, this is significant for Google but not significant for incumbent ISPs. Further, we could not see examples of where government policy could affect this factor. Nonetheless, it is a factor that is relevant to the formula for upgrades.
3. There are certainly other factors that affect the equation. For example, as the investments we seek are long-term, there is significant sensitivity to interest rates. Two that are not reflected in the equation but were significant in the Gig.U experience were entrepreneurial talent in network services and local leadership that could organize local resources to improve the economic opportunity. As to the first, it appears that the generation of entrepreneurial network talent that grew up at MCI and went on to start a number of CLECs and DLECs in the late 1990’s has largely left the sector, though a new generation is starting to emerge. As to the second, there has been a significant increase in local government interest and talent related to broadband networks, owing to a number of factors, including the sharing of lessons learned from the dozens of cities that have now successfully accelerated the deployment of next generation networks.
4. This is not always true. One counter-example would be Netflix, which transformed from a postal delivered service to a streaming service and an original programming service, thereby creating competition to MVPD. The critical change was the increase in broadband capacity and customers, making the streaming service viable. However, Netflix would not have made that transition if it were not for earlier government policies requiring interconnection, banning terminating access charges for data, and looking unfavorably upon blocking or throttling traffic. Government policy played a critical role but the timing was different from the examples cited. Going back even further, Netflix would probably not exist but for 17 USC 109, which codifies the first sale doctrine. If Netflix had had to ask Hollywood's permission first before buying and then lending out DVDs (or at least if first sale were not there as a backstop should negotiations fall through), the original business plan would have been unlikely to get off the ground.
5. For example, government policy did successfully enable new wireless new entrants into wireless through the 1994/5 PCS auction. In that case, the existing market penetration was low enough and the potential high enough to induce new entrants. Despite many efforts, subsequent auctions have not done so, as it is too difficult to dislodge existing efforts. T-Mobile has recently intensified competition, but only after it got a boost from a spectrum and financial payment from AT&T for the rejected merger. Adjacent market entry, through Wi-Fi, discussed later, is most likely to be the next disruptive competition.
6. The one exception is prison, where the FCC recently acted to lower rates. Without commenting on that decision or the unique market structure for prison phone services, it is worth noting that bandwidth abundance in prisons could also do a lot to increase communications, security, education and job training, while reducing the cost of prison operations and bringing the cost of voice services to where it is in the non-prison market. But that is a subject for another time.
7. In Porter’s model, this would be described as competition from both a buyer and supplier as Google is both a supplier to ISPs and a buyer from ISPs.
8. See Gig.U handbook, at page 25.
9. This is the heart of the economic inquiry in the FCC’s current review of the special access market. In that inquiry, the FCC has to make an assessment of, among other issues, under what circumstances is it economically feasible for a CLEC to be able to build its own last-mile fiber loops to a location, to what extent do lower wholesale rates provide negative incentives for a CLEC to construct its own fiber loops, and given that the ILEC, as the historical monopolist, likely has a first-mover advantage and thus a larger market share than the CLEC, how does that larger market share affect comparative costs between the ILEC and the later entrant? Those are issues far beyond the scope of this speech but is the subject of extensive economic analysis in the FCC docket.
10. Ending the Internet’s Trench Warfare March 20, 2010.
11. The equity research firm Bernstein, in its October 7, 2015 report on Google Fiber, suggested that an “aggressive expansion” of the project would reach 15-20 million homes in 6-8 years. If that were to occur, I believe it would drive a number of developments, including competitive responses and new products that would improve the economics of deployment throughout most of the rest of the country. But again, I could be wrong.
12. This is not the occasion for a full discussion of the FCC’s decision to pre-empt state laws restricting local broadband efforts except to note that the threat of competitive losses is, as demonstrated by the competitive response to Google and by our experiences with Gig.U, the single biggest driver of incumbents accelerating their deployment of next generation networks. Whether it is wise for cities to build their own networks is subject to a reasonable debate. (For such a debate listen. ) On the other hand, there shouldn’t really be a debate about whether a city having the ability to build its own increases the probability that the incumbent will act to make it unnecessary for a city to build its own. That is a factual question for which all the evidence is on the side arguing that just like any negotiation, more leverage increases the odds of a successful outcome. Which is why the National Broadband Plan favored pre-emption of such laws. See National Broadband Plan, recommendation 8.19.