Playing fiber “policy monopoly” is a risky game

Competition in international policymaking is unequivocally a good thing. When different jurisdictions adopt different policies, comparisons allow everyone to learn about what works best under different circumstances — which better informs future policy. If all jurisdictions adopted the same policies — in other words, a “policy monopoly” — then the policy environment would be as moribund as markets dominated by one-size-fits-all commercial monopolies. My academic research has been devoted to comparative analysis of different telecommunications policies in different parts of the world. However, some people with vested policy interests may feel threatened by such research. Over time, I have observed an interesting pattern in the way these people respond to findings that pose tough questions about the efficacy of their preferred position. They typically limit the range of comparisons they will accept as valid so as to make their own position appear like the only reasonable choice. They often include ad hominem attacks on the questioners, likely to limit the risk that those in their sphere of influence will be tempted to dabble in the critical literature. But, in the end, the truth tends to prevail, and the best-performing policies reveal themselves despite political pressure.

Hit or Myth?

The reaction of a former New Zealand Telecommunications Commissioner — a role similar to the Federal Communications Commission (FCC) Chairman — to my recent blog on the perils of government-funded fiber networks represents a typical response to uncomfortable questioning. Ignoring the substance of the offending material — in this instance, the effect that government subsidies and policies have on prices and network ownership incentives — the piece instead concentrated on peripheral matters. The former commissioner took exception to my observation that, four years after New Zealand’s subsidized fiber-to-the-home (FTTH) network was rolled out, uptake of the network stood at just 16.4 percent. This rate of uptake is sluggish compared to unsubsidized Reggefiber in the Netherlands, which achieved 30 percent uptake from day one in areas where it deploys fiber. The commissioner claimed that, in New Zealand, “demand for fib[er] services is greater than anticipated at this stage of the rollout.” His basis for comparison? The “pattern [in] other countries, such as Singapore, which have embarked on national fib[er] rollouts.” Supposedly, if it is okay for subsidized network operators in Singapore to take years longer than any unsubsidized private-sector operator anywhere else in the world to achieve an economically viable subscriber mass, then it’s okay for New Zealand. After all, the investment capital comes from taxpayers, not shareholders. Indeed, in getting to an arbitrary uptake level three years sooner than expected, New Zealand’s taxpayers have lost less than Singapore’s — backlogs, delays, and frustration aside. But does this constitute policy success?

Letting the data tell the story

The former commissioner also did not cite early uptake rates in South Korea — the country that inspired the recent bout of government-subsidized fiber networks. Korea shot to the top of the OECD broadband uptake rankings back in the days when any broadband uptake was considered a success. This elicits a question: Why are subsidized uptake rates in New Zealand and Singapore so low compared to when Korea was at a similar stage of deployment? But asking or answering this question would require opening another can of policy worms which the former commissioner would likely prefer remained closed: the role of structural separation of network and retail operations. This policy prevailed in Singapore and New Zealand, but not in South Korea, and it has played a significant role in delaying fiber uptake rates. Instead of addressing this fact, however, the commissioner stated the blindingly obvious: that the (subsidized) New Zealand fiber uptake rate would ultimately exceed the (unsubsidized) Reggefiber rate. But again, no mention of Korea. Following this was an ad hominem attack that misrepresented my position on the role of government funding in broadband networks. I have never held a “theoretical standpoint that state funding is unambiguously bad.” I have consistently advocated its use where private sector investment is uneconomic. My objection is to public investment that distorts prices and crowds out economically feasible private sector investment — the substantive argument of the blog that drew the commissioner’s response. The European Union imposes restrictions on government subsidies and investments to mitigate this risk. This is one reason why European fiber uptake, on average, lags Singapore, Japan and South Korea. However, infrastructure uptake metrics are not the only relevant measure of either technical performance or policy efficacy. Last-mile fiber connections carry Internet data over the shortest leg of its trip. Hence, it is far from clear that concentrating investment in last-mile fiber at the expense of investment in other parts of the network is the most efficient use of scarce investment resources. Neither is FTTH necessarily a “risk-free” or “future-proof” option. Technological innovations have provided increasingly more cost-effective ways to upgrade performance on existing copper and cable last-mile connections. Hence, measures capturing network performance over both the middle and last mile — such as Akamai’s quarterly average speed data — offer a much better means of comparing both providers and policies. Table 1 contains selected data from the latest Akamai report released in December.

Table 1: Selected Akamai Speed Data Q3, 2015

The argument that raw speed on fiber in the last mile is higher is technically correct, but this is a poor way of measuring end users’ effective Internet experience. Users’ actual experience of Internet speed is, for the most part, determined by what goes on in the middle mile and Internet trunk, where traffic congestion is a major consideration. It is impossible to determine conclusively if Akamai data showing “average connection speeds in New Zealand increased by 30 percent in 2015” is due to increased last-mile fiber connections or investments in the middle mile by all of fiber, copper and cable operators. But given technical considerations and the fact that New Zealand’s fiber uptake rate still sits below 10 percent of all broadband connections, it is hard to believe that Akamai’s observation is substantially due to a recent upswing in fiber connection rates alone — especially since the Akamai data pertain to the year ending September 2015 and do not include any increase in demand for fiber since that time.

Fact and counterfactual

Lastly, the former commissioner advances a counterfactual argument: What would have happened if New Zealand’s government had not subsidized the fiber network? In this view, the result would be that New Zealand would resemble the UK, the “laggard of Europe for fib[er] to the home deployment.” The policy counterfactual fits, as regulatory settings in New Zealand in 2008 — when the fiber subsidy was announced — were very similar to those in the UK. The incumbent operators (Telecom/Chorus and BT) were both privately owned with functionally separate network and retail arms. Whereas the New Zealand government went ahead with full structural separation and a heavily subsidized nationwide fiber network, the UK government left its policy settings largely unchanged. BT eschewed FTTH and proceeded with a mostly privately funded middle-mile investment strategy — hence the UK’s low fiber uptake rate. But how do end user experiences compare across the two countries? Akamai speed data is one proxy for this. As shown in Table 1, the UK may have negligible FTTH, but average Internet speeds there exceed those of Singapore. US average speeds are pretty much the same as Singapore’s, with DC’s rivaling Korea’s and Utah, Washington and Virginia all with faster averages than Japan. Furthermore, Singapore’s average speed fell in the last quarter, as uptake to the fiber network increased. As for the FTTH poster-child Korea? Average speeds there (and also in Japan, which has very high fiber uptake rates) have been consistently falling for over a year. In the UK and the US, however, average speeds are rising. If current trends continue, it won’t be long before the average speeds in all four regions are approximately equal.

The takeaway: Fiber-to-the-home may not be the optimal policy choice

On the basis of this evidence, the superiority of subsidized national FTTH networks looks somewhat shaky. It may be difficult or embarrassing for ardent FTTH advocates to acknowledge this. However, Australia offers an interesting policy comparison. In 2014, the incoming government ditched its predecessor’s expansive (and expensive) FTTH strategy in favor of a more nuanced fiber-to-the-node investment plan. Akamai expects that the rollout of FTTN equipment — which began in September 2015 — will lead to a rapid increase in average Australian Internet speeds. The report is curiously silent on expectations for New Zealand.

The post Playing fiber “policy monopoly” is a risky game appeared first on Tech Policy Daily.