You're missing my point. Most of the benefits you observe — less traffic, better air, less congestion, etc — are benefits not of more people riding buses, but of fewer people driving cars. And it's true that if more people switched from driving cars to riding buses, we'd be better off. But the gain is from fewer people driving cars.
That's not a semantic disctinction: there are policy choices that flow from it. Subsidizing people to use transit is a horribly inefficient way to induce less car driving. Some of the people who benefit would have ridden the bus anyway. Many more will go on driving their cars, no matter how big the subsidy. And bus riding does not come without its own costs, in terms of road space occupied, fuel consumed, sprawl etc. Whereas charging car drivers to use the roads gives people a very direct incentive: drive less.
How they drive less is up to them, and should be: riding the bus is just one of many ways. Which is another point: subsidizing transit only encourages one particular way of reducing congestion.
The same might also be said of your other class of putative externality, the "helping people get around town" variety. In the main I'd argue that's a private benefit to the traveller, which it is possible to exclude him from if he does not pay. So the classic public good criteria do not hold. But in any event, even if there were an external benefit to mobility, that's not an argument for privileging transit over bikes, taxis, etc.. Indeed, by that logic, we should be subsidizing cars as well!

Not sure what you mean, but thanks for the style note.

I stand by the point: the social benefit from transit ridership is only an indirect one - ie, fewer people driving cars. But transit ridership *itself* has no social benefit, beyond that directly accruing to the rider (for which he pays, or should pay). And subsidizing transit ridership, to the extent it has any impact, only encourages people to overuse it.

Far better simply to address the problem directly. If people are driving more than is socially optimal, because the costs of driving are hidden from them, then make those costs known to them, through prices.

But it's not just distance travelled that matters. It's when and where. The gas tax is a rough proxy for distance (very rough, alas, since mileage varies so much from car to car and year to year). But you pay the same tax whether you're driving a deserted country road at midnight or the DVP at rush hour.

I'd still have a carbon tax, but for the specific task of capturing the cost of carbon emissions — which a road toll would be just as ill-suited for as gas taxes are for congestion. First rule of pricing externalities is to target them as precisely as possible.

But we don't want to encourage "transit ridership"! There's no external benefit to society from people riding the bus, per se. And indeed, subsidizing transit carries with it many of the same social costs as subsidized car use: sprawl, pollution, etc. It may be *less* costly in this regard than auto use, but subsidizing less wasteful waste is still subsidizing waste.

So even if there were much evidence that subsidizing transit encourages people to abandon their cars, which there isn't, the more direct and effective remedy for too much car use is to price the unpriced costs that are encouraging that specific overuse — not to suppress the costs of other activities.

And of course, as I say in the piece, transit — and the poor, as disproportionate users of transit — are among the big winners from pricing roads, since buses are stuck in traffic along with the cars. Far from subsidizing transit to get people out of their cars, pricing car/road use is the best way to get people to use transit. It isn't more transit that reduces congestion, but reduced congestion that leads to more transit.

Not dismissing telecommuting. Or carpooling or transit or any ot the rest. My point is only that people won't do any of these things without an incentive. Exhortation is not enough.

Agreed. It's the corollary of my don't-subsidize-it argument, but I can't make every point in every article.

I am not the editor of Maclean's, and never have been.

It could be justified as a "reasonable limit," based on the harm it would cause. But you'd have an easier time justifying a discriminatory marriage law than making it a crime, on the principle of "minimal impairment."

I, uh, I have no answer to that.

I think there's a difference between a duly elected government which had just had its Throne Speech endorsed by the House and an untested coalition of the kind I describe. I'm not saying she wouldn't have called upon them. I'm not even saying she shouldn't, necessarily. I'm just saying there's no assurance she would have. Which puts me in the same company as, among others, Michael Ignatieff. (Also Peter Hogg.)