The Herald’s transport correspondent Linton Besser reinforces standard confused thinking about motorway tolls in yesterday’s edition. He reports that the NSW Government’s ‘Cashback’ scheme, whereby private motorists can claim reimbursement for the tolls they pay on the M4 and M5, has cost well over half a billion dollars since it was introduced in 1997; and argues that this ‘is in sharp contrast with the cries for transport relief from residents of Liberal seats on the North Shore and in the north-west, who endure soaring road tolls and long-haul bus routes’.
Experts say that money could have been better spent delivering transport solutions for gridlocked commuter routes across Sydney, such as the Spit Bridge, which failed on Monday night, stranding thousands of motorists. The Government has yet to act to fix this bottleneck after dumping its election promise to widen the bridge with two extra lanes in May last year. The Roads Minister, Eric Roozendaal, said at the time that a $56 million cost blow-out had made the Spit project, which straddles the blue-ribbon Liberal seats of North Shore and Manly, an impossibility. It means that while the Government has allocated $102 million for Cashback in the coming year’s budget, there is still no plan to fix the Spit Bridge.
However, he could have devoted several more paragraphs to listing other things the money could have been spent on, in the transport, school, hospital, community services and criminal justice systems. Public finance issues are prone to a tendency for arbitrary coupling of allegedly competing prioities — telecommunications infrastructure versus reforestration, electricity infrastructure versus a new metro, and so on — and this is just another example. In fact, each project needs to be assessed according to its own balance of costs and benfits. Obviously, resources and government budgets are not infinite, but it’s the price system and the interest rate that should ultimately determine the parameters of those costs and benfits in individuial cases.
The construction of a motorway is justified if the benefit, in terms of travelling time saved, exceeds the cost of acquiring the land, building and maintaining the facility, and compensating those affected by externalities like noise and fumes. But, up to a point, a motorway is a public good, and the marginal cost of using it is close to zero — indeed it may well be negative if the alternative is that drivers use back streets, clogging them up, and making the local road residents miserable. Therefore, anything that deters people from using it is socially wasteful, and — ironically — more so in Western Sydney, where it easier to substitute to alternative routes. The trick, as with other public goods like golf, radio broadcasts and computer software, is to find a way to cover the fixed costs with a lump sum fee, while keeping the ‘per-unit’ price as low as possible. One way to do this is would be to raise car registration fees by the amount necessary to pay for the motorway. But this would obviously discriminate against people who use the motorways rarely, so a better scheme would be a quarterly permit for each vehicle, allowing unlimited use of specified motorways.
The current scheme toll system merely allows the operators to exploit their monopoly power to extract the most consumer surplus from the customer; and when that’s the sole criterions it’s bad welfare economics, irrespective of whether the proceeds fund infrastucture or go into the pockets of Maquarie Bank shareholders.
When motorways become congested it’s a different story: obviously the marginal cost becomes significant, and a higher per-unit fee is appropriate. But this argument applies to any road — the only reason to single out motorways in this respect is that it’s generally easier to put a toll-booth on a motorway than on a street. But thanks to satellite technology, it seems that the dawn of congestion taxes is coming, so (allowing for some moronic resistance) we may soon see a more systematic application of the principle; that is, charges for both motorways and other roads — presumably in whole zones — to reduce congestion. That will speed up the shift to public transport, especially if we lower the price of bus and tram tickets, as Joshua Gans urged a while back. (Andrew Leigh also has some practical suggestions about this last year.) It will also provide some revenue for the Spit Bridge expansion and some of those urgent bus and rail projects, and perhaps we’ll hear fewer irrational, revenue-focused arguments for motorway tolls.
OK I’ll be the mug. Why are golf and computer software public goods? Both are certainly non-rival (although golf not completely so because you can only fit a certain number of golfers on a course at a given time), but neither is really non-excludable. Golf course proprietors don’t seem to have any difficulty excluding non-members, and software manufacturers have found ways (albeit imperfectly and with difficulty) to inhibit illegal copying. There must be some definitional aspect that I don’t understand. Care to explain?
Both excludable and non-excludable goods are public goods in the sense that there is a straightforward case for public provision: without it, there will be under-supply in the former case, and under-consumption in the latter. I was taught to refer to goods that are both non-rival and non-excludable as pure public goods. If you reserve the term public good for goods that have both attributes, you have the problem of what to call goods that are excludable but non-rival. Some textbooks use the term ‘club goods’, but nobody would have understood what I meant by that. In any case, excludability is a matter of degree, since fences and tollgates are costly.
None of this makes any difference to my argument.