Now imagine paying $18.3-billion for it.

That is the latest estimate of the amount needed to construct the high- speed rail link in Canada's most densely populated corridor - a project that at this point remains only a gleam in the eyes of its sponsors in the public and private sectors. And if the sheer size of the number is an eye- stretcher, so is the rate at which estimated costs have grown over the years:

. In 1989, Via Rail, which has been pushing the idea since 1981 in hopes of winning the right to operate the trains, estimated it would cost $2.4-billion (in 1989 dollars) to build a a high-speed line between Montreal and Toronto. . In 1990, Bombardier Inc., the leading candidate to build the trains, estimated the cost of a Quebec City to Toronto service using the same technology as France's famous train a grande vitesse (TGV) at $5.3- billion, of which about a third would come from government. The Montreal- Toronto portion was estimated to cost $4- billion. . In 1991, another Via report estimated it would cost $6.2-billion for a line that would stretch from Quebec City to Windsor. It pegged the taxpayer's contribution at $3.3-billion, just over 50 per cent of the total. . In 1991, a task force struck by the Ontario and Quebec provincial governments projected the cost of the same service, using Bombardier's technology, at $7.1-billion. Of this, at least 85 per cent would come from government. . Now, four years and $6- million later, the final report of the High- Speed Rail Feasibility Study, financed jointly by the two provinces and the federal government, puts the cost of a Quebec City-Windsor link at $18.3-billion over the next 10 years, including inflation and financing costs. About 75 per cent of this would have to come from government. The Montreal-Toronto portion alone would cost $10.7-billion.

These estimates, mind, come well before the first shovel hits the dirt - and the first strike delays construction. Yet far from deterring its supporters, each successive upward revision in the projected costs of high-speed rail only seems to spur the next study. Far from a setback, the latest report was greeted as a victory for Bombardier, since the study preferred its TGV technology to the slower - but much less expensive - alternative championed by its Swedish rival, Asea Brown Boveri Inc. Already there is talk of spending $50-million more on the detailed engineering surveys needed to proceed with the project, as if the $14- billion in public funding were already in the bag.

One reason why the project refuses to be derailed is because it appeals to so many current enthusiasms. It's high-tech. It's a job creator. It would save lives (by getting travellers out of their cars and off the highways). It's environmentally friendly (ditto). It's even a boon for national unity. All of which may be the case. But that still leaves the cost to be reckoned with, even setting aside the present financial condition of all three governments.

Every dollar spent promoting high-speed rail is a dollar that cannot be spent on other projects, each with its own claim to public funds. And the question each and every claimant to the public purse has to answer is: Why can't the users of this service pay for it themselves? Why shouldn't the providers of the service assume the risks - that is, the risks that potential users won't want to pay for it? Why should not the people who reap the benefits also bear the costs? If you don't ride the train, why should you pay for it?

The short answer is that the usual two parties in any transaction, buyer and seller, may not always be the only ones who get all of the benefits - or, indeed, pay all of the costs. The environmental case for subsidizing high-speed rail makes the strongest use of this principle. When people buy gas for their cars, the price they pay covers the oil companies' production costs, but not the costs imposed on others in the form of greenhouse gases and other air pollutants. Subsidies to mass transit might be justified on the grounds that, by inducing people to switch from driving hundreds of gas-guzzling cars to one relatively fuel- efficient train, the impact on the environment would be reduced. Even those who never take the train - most of us, in fact - would benefit from cleaner air.

But even a subsidy to more efficient forms of fuel consumption is still a subsidy to fuel consumption. If we want to reduce automobile emissions, better to increase the taxes at the pumps than subsidize trains. If, similarly, we wish to reduce congestion on the highways, better to charge tolls for road use: Rail use would be encouraged merely by removing the present subsidy to roads. As a rule of thumb, the instrument of corrective policy is best applied directly to the source of the problem.

Indeed, even if we wanted to support high-speed rail travel out of public funds, for any of the reasons listed above, that would be an argument for a direct subsidy to passengers, not for governments paying most of the capital costs up front. It is train-riding we might wish to encourage, not track-laying.