Recently I’ve been suffering cognitive dissonance regarding the role of low interest rates in important policy matters.

On the one hand, prime minister-designate Justin Trudeau told us throughout the just-concluded election campaign that we should invest more in public infrastructure and do it right away while interest rates are low and borrowing to finance it won’t cost much. On the other hand, large parts of the policy community, though not necessarily Mr. Trudeau’s Liberals, have been telling us that because interest rates have been so low for so long, Canadians have borrowed excessively and invested too much in housing, which I like to think of as private infrastructure.

So, public infrastructure good? Private infrastructure bad?

The worry about growing private debt is that when interest rates start to rise again (assuming they ever do) consumers will have a hard time continuing to finance all the new debt they’ve contracted. The Economist recently spent the better part of a page on the supposed problem of our rising household debt.

OK. Got it. But isn’t public-sector borrowing subject to exactly the same vulnerability? What happens to all that new government debt if interest rates go up? Catastrophe wouldn’t necessarily be the result. But higher interest costs for government would be.

The Economist argues that although our household debt is at levels similar to where it was in the U.S. before its housing bubble burst in 2007-08, we don’t have nearly the proportion of sub-prime mortgages they did and our big banks, because there are so few of them, are much more careful about the housing loans they make. Even so, a run-up in interest rates might well cause a run-down in house prices, with consequent ripple effects.

By contrast, higher interest rates wouldn’t burst a bubble in infrastructure. Most infrastructure isn’t bought and sold on open markets. Most isn’t bought and sold at all. But a flood of new money into infrastructure might well drive up the cost of building it. We’ve been in an infrastructure boom in Montreal for the last several years now—much to the annoyance of anyone who tries to drive in the city. Piling on more demand will only worsen construction bottlenecks and push up prices. We’ll end up getting much less infrastructure per buck.

But cost doesn’t matter, people say. We need infrastructure, people say. That’s a little like saying cost doesn’t matter, we need stuff. Everybody needs stuff. But what stuff exactly? And how much stuff? And for what price? And can’t you reach a point where more stuff is too much stuff?

One reason people believe we do need infrastructure investment is that they look around and see our existing infrastructure crumbling. That’s clearly the case in my part of the country. But before we buy boatloads of new infrastructure we should think about why that is. It’s because governments have been chronically incapable of adequately maintaining existing infrastructure. Some glamorous new initiative is always coming along that’s much more tempting than boring old infrastructure maintenance and repair.

When I look around at all the private infrastructure in my neighbourhood (i.e., all the privately-owned housing), I don’t see the same problem. I see people replacing their roofs and gutters, re-paving their driveways, reinforcing their foundations, taking delivery of a new furnace, and so on. Sure, once in a while neighbours will tell you they’ve deferred maintenance because they’re between jobs or have had a tough year—though in such cases they usually do more maintenance themselves. But our governments have let things slide so often and for so long that now we’re apparently facing an infrastructure crisis.

If that’s true, do we really want to entrust these same governments with tens of billions of dollars to buy new infrastructure without reform (hoping they do better this time round)? Or should we look for completely new ways of acquiring the infrastructure, ways that create strong incentives, such as ownership rights for somebody—anybody—to maintain what they build?

Going ahead without changing our approach? I sense another bout of cognitive dissonance coming on.