The ethics of broadband

Most writing about broadband discusses technology or business issues. For a change, I would like to examine the ethical challenges facing us. My suggestion is that current regulatory and industry approaches are not achieving the best possible outcomes. There are three ways in which we can improve: measure the service given to those who are weakest in society; focus on radical increases in value; and pay most policy attention to those with “skin in the game”.

Service the weak
How should we judge the success of national broadband policies? At the moment we have bodies like the OECD compiling tables of speed and affordability by country. These offer some crude sense of progress. However, they are also very limited:

  • Large nations (e.g. the USA) will naturally have lower rank than outliers of an equivalent group that has been disaggregated (e.g. Luxembourg as part of the EU). Whole books are written about “tumbling down the league table”, with the panic based on nothing more than a statistical artefact.
  • Hard-to-service rural areas, or populations with low incomes, cost more to reach (all other things being equal). Thus we aren’t comparing apple to apples. That people in Luxembourg have great broadband is of no solace to those in Liverpool or Lombardy.
  • Even if we accept “speed” as being a good metric of end user value (it isn’t), then the statistics are highly skewed by a few gigabit users. Yet the last megabit per second creates negligible value compared to the first one, meaning these are over-valued.
  • Average speed measures mean that the distribution of service speed is ignored. I think you would agree that a country with universal 500Mbit/sec broadband is a lot more successful than one which delivers a gigabit to half the population, and nothing to everyone else.

I believe that we need better metrics to benchmark social progress against. This is not the place to offer a framework of moral philosophy, but a reasonable starting point for such issues is the work of John Rawls. If the dice were rolled and you were reassigned a random place in the economic pecking order, you’d want to be sure that a lowly position would not result in suffering.

I suggest a more appropriate metric of broadband progress is the actual service level delivered to (say) those performing janitorial work. Survey them, and report what a typical janitor is able to buy, and how long they have to work for it. If they can only afford mobile broadband, then we should focus our policy initiatives and public money on better and cheaper mobile broadband, and not fibre to rich people.

Get more for less
Affordable always-on broadband exists only because we all (statistically) share a common resource. If we each needed our own dedicated “lanes”, or a separate infrastructure for each application, then prices would rocket and progress would be restricted. This hurts the poor and weak the most.

Once we focus our attention on those at the bottom of society, then there is a moral imperative to deliver a lot more value for a lot fewer resources. That means further increasing the intensity with which we use resources, and eschewing policies that create idle resources or duplicate infrastructures.

What does this mean in concrete terms?

Firstly, regulate for delivered QoE (i.e. value produced), not speed (i.e. resources consumed). Data from Actual Experience suggests that speeds above 10Mbit/second have little or no quality improvement. (A client this week suggested typical customers can’t tell the difference between broadband services over 30Mbps.)

Secondly, move to a multi-service wholesale environment to increase sharing. Today we have competition among a limited number of physical access providers. In markets with unbundling we also have a “transferable monopoly” at the retail level, whereby ISPs compete to shackle you for a year or two. If I want to deploy a smart meter service, or a connected car, either I have to negotiate with a few giant oligopolistic providers (as in the USA), or face a highly fragmented market (as in the UK).

As a result we have duplicate infrastructures for some key applications and/or broadband services that aren’t being delivered at all (e.g. home call centre worker). Regulators should in future focus on enabling many concurrent services (and service providers) through richer wholesale products.

Finally, encourage (or at least don’t inhibit) better resource sharing and scheduling. For instance, “fast lanes” are a necessity to deliver better and cheaper services to users. Unfortunately (insane) policies like “no paid prioritisation” work to achieve the exact opposite, and thus I would judge them to be unethical. This rule stops a charity from buying quality-assured video sign language service for the deaf, for instance. Asking someone on disability benefits to buy a second line and service for one application is a financial non-starter.

Skin in the game
Anyone who has read Nassim Taleb’s works, or watched him eviscerate people on Twitter, will be familiar with his simple moral heuristic. Pay attention to whether someone has “skin in the game”, and thus will experience the (negative) consequences of whatever course of action they advocate, especially if it is disastrous. Pilots have skin in the game of the safety of their passengers, hence aviation is overall very safe. Software engineers of drone craft don’t have their skin in the game.

The GSM ecosystem has been a huge success at delivering to the poor globally. All operators all had skin in the game: they couldn’t profit by doing harm to users. Their views of the likely impact of a policy can be taken seriously. Indeed, any entrepreneur who is risking their capital has put their money where their mouth is.

In contrast, the policy advice of tenured academics, think tanks and lobbying sock-puppets should be treated as ethically suspect. They can (and do) campaign for actively harmful things, and can make a profit from doing it. They don’t carry direct personal consequences of any harm from their opinions and actions.

Consider the typical regulatory employee. How much “skin in the game” does she have? She probably has a stable well-paid job, and lives in the same place for a considerable period. She experiences no personal consequences if her policies make life hard for students or itinerant workers who move around, and for whom long contracts are not viable. Her “skin in the game” is small, and limited to her own social class as a broadband user. Unsurprisingly, broadband policy tends to serve the bureaucratic classes well, and doesn’t deliver such good results for the disadvantaged.

I believe policy-making needs to take more heed of how much skin in the game advocates of a course of action actually have. How can we do that? A good question, and I wish I had a good answer.

For the latest fresh thinking on telecommunications, please sign up for the free Geddes newsletter.