Free public transport in Melbourne?

Frank Fisher is reported in the Sunday Age (here) as advocating free public transport in Melbourne. This means public transport will be 'unpriced' not 'free' since we still as a community have to pay for it. With this qualification, is the 'free' transport proposal sensible? I think not.

The core idea is that private car use creates congestion, pollution, greenhouse gas and accident externalities so people use their cars excessively from the view of the social advantage. Implicit too (I guess, this is not sketched out) is that it is expensive to price away these externalities so a 'second-best' option is to divert private car users to public transport by subsidising it. Since cross-price elasticities of demand for public transport by private car users are low these subsidies must be hefty and, given substantiual scale economies in public transport marginal costs are low. Hence it might be just be best to zero price public transport.

There are many issues here:

(i) Public transport is over-priced at present since, even with subsidies, its pricing reflects average not much lower marginal costs. Gans et al argue this point well.
(ii) Congestion in the CBD and along major motorways can be cost-efficiently priced by a combination of a cordon pricing scheme and curbside pricing of major arterials. See my paper with Andrew Hawkins here.
(iii) Making public transport 'free' would subtantially increase its use (The Age suggests a 30% increase) but much of this would come from additional travel not from people abandoning car use. Moreover there are also additional large costs of purchasing additional rolling stock to meet increased demands. The reason as stated, cross-price elasticities are low. Making public transport free would encourage socially-excessive total travel.
(iv) Meeting all public transport costs from the public purse would leave diminished incentives for private operators to improve service that match demands. The difficult issues are in Melbourne's periphery where demands for travel are low and journeys are often cross-town. The best way to address such demands is to provide private incentives for innovatory transport solutions. Such incentives will be reduced if public transport was 'free'.
(v) Pollution, greenhouse and accident externalities can be internalised through petrol excises and policing of roads. These policies are already in place.
(vi) Melbourne's traffic problems are location-specific. Thus the inner city is well serviced, the periphery less so. Providing 'free' transport to all would redistribute incomes away from low public transport using periphery regions towards intensive (and affluent) inner city users. In short it would redistribute transport resources from those poorly-serviced toward those well-serviced. This type of income transfer would be generally regressive.
(v) Generally the Fisher approach is based on the false premise that private travel by car is evil. It is not. It is socially-harmful only if private motorists do not impose all the costs they impose. Road and other types of pricing policies can internalise these costs.

Melbournians would be worse-off with 'free' public transport. We need a better public transport system which has greater flexibility and which offers improved service frequencies to reflect demands. Private sector incentives that reward operators who provide needed transport services will do this. We need to price key roads and public transport at social marginal cost but going further and providing transport for 'free' is ill-advised.

Update: After writing the above I see Joshua Gans broadly endorses the proposal but assumes comprehensive electronic road pricing in conjunction with unpriced public transport.