The Financial Post published my response to the Montreal Economic Institute's idea that what we need is exactly three physical cellular networks, rather than four. I find the MEI is a reliable voice of pre-Internet ideas of telecommunications. The number of networks is largely irrelevant; what traffic they carry, and how they interconnect, is the supreme consideration. It is the apps, not the transport, that count. Policy needs to concern itself with less with the number of networks, than with the conditions under which they carry the apps.

Telecoms and broadcasting are being absorbed into the architecture of the Internet

When we speak of competition in telecommunications, what do we mean? Is it best understood as competition among networks, or is it best seen as competition among applications? Is the number of networks actually a useful proxy for the amount of competition?

Two visions compete: the silo model and the highway model. The silo model says, roughly, that competition consists of carriers attracting people to their respective networks, and that, ideally, they should not exchange traffic with one another except on such terms as are commercially negotiated, without government regulation. In this vision, competition among carriers is the way to help consumers.

By contrast, the highway model says that traffic should freely interchange among networks, and that this interchange should be subject to regulation if the carriers exert market power. In this vision, competition among applications is more relevant to the consumer interest than competition among physical networks. Thus, while competition among carriers is always relevant, useful, and important, the number of carriers is not the supreme consideration. The ability of the consumer to use large amounts of cheap bandwidth, and to reach the largest number of other people, is decisive. You just cannot pick a number between one and five and say: this must be the number of networks.

In their attack on government policy in wireless (Federal government should drop its seven-year crusade to force competition in telecom, May 6), Martin Masse and Paul Beaudry take issue with the doctrine that four competitors are needed. Three is enough, they say. I say that trying to fix the right number of networks is largely irrelevant — a pre-Internet idea. It is just so 1980.

Just as everyone’s cars should be able to get on any highway, so should everyone’s traffic be able to get on any network, for a reasonable price. And as large carriers exert market power, economic regulation is used to make sure that no access ramp is closed or subject to unwarranted discriminations against certain uses and users.

More than that, the prejudice against smaller carriers leasing parts of the access network of larger carriers is a relic of a pre-Internet theory of competition. Physical infrastructure at the scale of the backhoe limits the number of possible networks. However, infrastructure at the computer level has the same characteristics as the domain name business, where dozens upon dozens of competitors thrive.Thinking of competition as if it needed massive physical infrastructure is inappropriate. Yet we do it.

Telecom policy is in a time warp: it still thinks networks are exclusively governed by laws appropriate to massive physical plant, rather than computers. In consequence, the CRTC in its recent decision on wireless prevented the creation of virtual leased networks (called MVNOs) on the grounds that it would diminish the incentives of the larger carriers to invest in infrastructure.

Such a policy presumes that real competition consists of building out your networks in terms of physical plant rather than building customer-focused organizations based on the capacities of computers and personal service. Relevant competition, it assumes, is in infrastructure. That is akin to saying we need thicker steel in cars because it will increase investment.

“Convergence” is another distraction from essentials. Telecoms and broadcasting are not converging; they are being absorbed into the architecture of the Internet. Applications are what matters, just as cars are generally of more concern to us than highways. Nothing carriers can do can make them relevant to competition in applications. If they play in that space, all they do is favour some applications over others.

Ottawa may still be mired in the pursuit of a fourth national carrier, but for me this is not so much an error, as some think, as a pre-Internet idea of competition. A fourth carrier may be good or it may be bad, but is a distraction from the importance of applications riding freely over transport.

Timothy Denton is a former Commissioner of the CRTC and is Chairman of the Internet Society of Canada