Muldoon is dead, long live Muldoon (updated)

The New Zealand Herald give us an editorial on, what it claims is, At last, a telco victory for the consumer. Unfortunately the reasoning is a bit simplistic when it comes to the economics involved. There are a number of questions the editorial should have asked, which it didn't.

The editorial writes

The structure adopted by Communications and Information Technology Minister Stephen Joyce confirms as much by removing retail services from the orbit of the partnerships that will be charged with delivering broadband to three-quarters of New Zealanders' homes and workplaces within 10 years.

Now the writer of the editorial seems to think that the splitting of retail from the infrastructure is a good thing, and it may be. But to make this case an obvious question should have been asked: Why do firms vertically integrate? If splitting services up is a good idea then vertical integration has to be a bad idea but many firms do in fact vertically integrate, Why?

Let me offer three reasons. First, transaction costs. When transaction costs are low a market transaction may be preferable, that is, a firm will "buy" a component or service on the market, but when transaction costs are high then it may well be cheaper to "make" the component in house, that is a firm will vertically integrate with its supplier. For example, as the complexity of a component increases the transactions costs of negotiating a contract with a supplier of that input increases, which makes making the component more attractive. Secondly, relation specific investments. In a world of incomplete contracts ownership of assets affects the incentives that an agent faces. That is, an agent's incentives are different when he is part of an integrated firm or when he is an independent firm. The buyer of an input is may well invest more, assuming that such investments are non-contractible, if he has bought the supplier of the input than he would if the supplier is an separate firm. This is because the buyer can capture more of the returns to invest if he owns the supplier. So the more important it is to motivate investment by the buyer, the more likely is vertical integration. If the relation specific investment is made by the seller, then the more important the investment, the more likely it is that the seller will buy the buyer. An important version of the investment story is the Hold-up problem. If you make a sunk, non-recoverable, relationship specific investment then you could find yourself in trouble if the contract is renegotiated. The returns to your investment may be (mis)appropriated by the other party because your investment is only valuable within that investment and thus you have no alternative but to give up some, or most, the returns to your investment should renegotiation occur. One way you can stop renegotiation is to buy the other party. A third reason for integration is the problem of "double marginalisation". Think of two monopolists one of whom provides an input to the other. Each of the monopolists will apply the same marginalist rule, MR=MC, and hence you get two marginalisations in a row. This is inefficient. The monopoly actions by the supplier will affect the input choices of the buyer and the demand for the suppliers product falls. One way to deal with this problem is for the two firms to integrate which will result in an efficient choice of actions by the (old) supplier.

The point of all of this is that there are good economic reasons for vertical integration. What the government seems to be saying with it telco policy is that none of the above reasons apply in that case of broadband. This may be true, but we have seen no evidence to show it is, so we can reasonable be skeptical that the government is right. If it is wrong then its policy could have very bad outcomes for consumers, and this would not be a victory for the consumer.

The editorial goes on to say

Additionally, this country's unhappy telecommunications history suggests it would be unwise to leave investment in broadband infrastructure to the private sector.

The question here that the editorial writer should have asked is, Why do we have a unhappy telecommunications history? One reason is the actions of previous governments. In particular the actions of previous governments with respect to firms' property rights. Firms will not invest if they do not believe that they will capture he returns from heir investment and government actions are the major influence on the beliefs that firms hold. One historical example of where (local) government actions affected the investment plans of private firms involves the municipal take over of waterworks in the US.

This interesting case study is discussed in the Troesken and Geddes paper Municipalizing American Waterworks, 1897-1915. This paper provides support for a transaction cost interpretation of municipal acquisition of private waterworks. The incentive facing the local governments and the private companies are an important factor in the government take over.

The basic Troesken and Geddes story is that municipalities were unable to credibly precommit to not expropriating value from private water companies once (sunk) investments were made. This gave the private firms an incentive to reduce investment in water provision. This rational under investment was then used by local governments a pretext for municipalizing the private water companies.

Troesken and Geddes explain that their evidence supports the idea that because local governments held the assets of the private firms hostage, they could use their police powers to undermine the value of the private waterworks and thus acquire these firms at reduced rates. The water companies anticipated such opportunistic behaviour and rationally devised strategies to reduce its effects. For example, firms would demand contractual provisions limiting the power of local governments and reducing investments in idiosyncratic assets. Water companies would, for example, reduce the number of water mains they installed if they faced a high probability of future public takeover. Such underinvestment was then used by local authorities as justification for a takeover. Troesken and Geddes summarise the situation as

When Progressive Era reformers advocated municipalization of urban water systems, they often pointed to places like New Orleans and Duluth, where there was clear evidence that private water companies were providing inadequate service. While reformers were certainly justified in claiming that water companies in such places were failing to deliver adequate service, they did not appreciate the ultimate cause of that failure: the threat of municipalization itself, which was often accompanied by a host of opportunistic strategies. The fact that the threat of future municipalization discouraged private water companies from extending their distribution systems suggests failure stems not from private provision per se, but from a more fundamental contracting problem.

In New Zealand in recent times, the sorry story of government actions over the Telecom unbundling is another example of property rights being ignored and rational underinvestment being a possible result. In May 2006 secret plans by the government to take, without compensation, some of Telecom's property rights in its network were made public. This revelation caused a fall in the share price of Telecom which reduced shareholder wealth by around $3 billion in six weeks. Like the case of the waterworks, the Telecom case shows that government action can and do effect firms value and investment decisions. Rather than try to nationalise the broadband infrastructure on the grounds that, as the Herald puts it,

[...] it would be unwise to leave investment in broadband infrastructure to the private sector

the government could assist private investment by implementing measures that decrease the likelihood of opportunistic behaviour by the current and future governments. A reduction in regime uncertainty would be one of the cheapest and most effective ways the government has to increase investment in broadband. In fact it would be a good way to increase investment in general.

I have covered just a couple of problems with the government's plan, for more discussion of additional issues surrounding the plan see the posting by Matt Burgess here.

Update: Liberty Scott commented here, David Farrar comments here, Brad Taylor comments here and Not PC here.