Reaction to Susan Crawford's post on Telstra

Susan Crawford wrote a great entry on Telstra on her blog. I replied and changed my reply a bit. In short. There are parties trying to compete with Telstra in Australia. Telstra, the incumbent telco, is a very mean dog to fight with. They are as anti-competitive as it comes. Telstra's plans to roll out a new network, but threatens they will only do this if they are allowed a regulatory holiday (Just like Deutsche Telekom). The competing parties will still need access to Telstra's copper for the last couple of hundred meters.

I first thought that Telstra was going to do FTTH, but they are going the VDSL2+ route if I interpret this article by Stephen Bartholemeusz correctly, however the economics remain the same. Stephen has the right conclusions, except one: there is no reason to assume that wireless will ever be a real competitor to broadband. An addition it will be, but not a replacement. But he is right there is no reason to assume that Telstra's competitors will be able to build a viable network to compete wit Telstra's. This discussion is however not unique to Australia, but also appears in Germany, The Netherlands, the UK and everywhere else where VDSL2+ is rolled out. LLU just isn't possible with VDSL2+. You need to be 450meters away from the customer and that is just too much fiber for normal companies to pay.

The fundamental problem behind Net Neutrality, behind regulation etc. is: Who is going to pay for the (ftth-)network. The economics of a ftth-network are such that one set of fibers provides infinite bandwidth. Economics predicts in dynamic models that if two people invest in a large sunk costs, low marginal costs business that are indistinguishable from eachother, that the result will be a race to the bottom. The parties will sell at marginal costs and not at full costs. The result will be with fiber that one network is doomed to go bankrupt. However since the network can easily be bought and reused by someone else, the buyer can run it for minimal costs from bankruptcy. A good example was 360networks, who invested 875 million in trans-atlantic fiber. They went bankrupt and were bought for 18 million and have operational costs of 10 million. Now imagine how their competitors in transatlantic fiber felt. They must have been gutted, because all of the sudden somebody on the market had only paid 18 million and all of them paid 875 million. Well we know where that ended.

The trouble for Australia is, that it will cost a couple of billion to build a nation spanning fiber network. 3 billion for FTTN and 3-4 times as much probably for FTTH. With two networks chances are one of them will go bankrupt. If that network is bought by an outsider, that outsider will have a better base to compete on. (40x cheaper for the buyer of 360 networks) This will probably kill off the other competitor too in the "long run". Unless both parties agree to a cease fire, but that is a cartel and it would mean that the new entrant can reap enormous profits from the network.

Some might say: That is capitalism at its best and in the end the consumers will have two networks working at marginal costs. How great. They forget that somebody needs to invest in those networks. That somebody is quite often a pension fund. Those are the loosing parties. Many pensioners paid for the current bandwidth glut. Smart investors know that and therefore don't invest in ftth, because if you get a competitor, well it's mutually assured destruction.

So what to do. You can roll out a fiber network for 1500dollar/subscriber maximum. On top of that network you can build Wimax, UMTS, Wifi and anything else you like. All for 32 dollars a month/subscriber. That network pays itself back in ten years with 10% interest. But, there is an enormous but, it would mean governments accepting that layer one is not the layer there will be any competition at. It is just bad business to compete at layer one.

So regulate it and demand wholesale broadband access. Create chinese walls or even split it. It's cynical, but its true.