Anyone can make a mistake – I've made plenty – but totally botching an alleged "core competency" four times in a row at the cost of many billions of dollars, well, that takes a special talent. Collectively, that's a talent the traffic forecasting trade has in spades – and lots of other earth-moving equipment.

It's a talent that should bring the ambulance chasers out soon enough with a class action or two on behalf of Brisconnections investors who have done their dough thanks to the flights of fancy that were sold as traffic forecasts for the Sunshine State's Airport Link tunnel.

Good luck finding anyone with money to sue – I can't imagine Arup, the consultants getting the credit for the failed forecasts, having enough to be worth pursuing after the legal fees. Or maybe the receivers will have a go, following the example of Korda Mentha having a crack at Aecom, the mob who did the incorrect forecasts for Brisbane's failed Clem7 tunnel.

If it was just Brisconnections, I could have some sympathy for Arup, as Brisbane drivers are indeed strange and, with the benefit of hindsight, unpredictable beasts. Aside from being tightwads, unwilling to spend a few dollars to save up to 20 minutes in the traffic jams above-ground, it seems Brisvegans are simply scared of the tunnel – it's different, it's long, it's lonely down there and what if I take the wrong exit, could I end up in Tasmania?

(That's in keeping with Queenslanders being the nation's worst drivers – the most likely to tailgate, the most likely to sit in the fast lane causing tailgating, the most likely to display aggro behind the wheel; the mob that needed to be taught how to "zipper", what the rest of the country calls "merging". I can safely say this because I am a loyal Queenslander, albeit now a Sydneysider, and we can criticise our own. Anything said by cockroaches and Mexicans (NSW league fans and Victorians) has no merit or meaning – they're just jealous and/or ignorant, particularly the ones with the funny football.)

When Airport Link opened – and was free – there was a brief rush of novelty-item interest, with more than 80,000 people a day checking it out and apparently being terrified by the experience. As Matt O'Sullivan reported, Arup had predicted that Airport Link would attract about 135,000 vehicles a day just a month after it opened in July, rising to 291,000 vehicles by 2026, but usage has actually been falling. In December, an average of just 47,000 vehicles a day used the toll road.

But it's not just Brisconnections and it's not just Arup. There's also Sydney's Lane Cove tunnel – fail. And Sydney's Cross City tunnel – fail. And Brisbane's Clem7. And while it was people rather than cars, the Sydney Airport rail link was a similar fail. The forecast traffic numbers on which the financial models depended all turned out to be useless. Chook gizzards under a full moon would be as accurate.

While traffic forecasting for some toll roads has worked well enough, the percentage of hits to misses could make this game as much luck as skill. It would be tempting to believe all these big insolvencies had been the fault of firms one bead short of an abacus, but maybe the game is just too difficult to pretend to play it with certainty.

The reality is that forecasting is always a risky business. For all the attempts to de-risk it and the attendant pseudo-science, predicting how people will react to various situations and opportunities is as much guesswork as craft.

The Reserve Bank, our best economic forecaster, has been honest enough to research and publish that you still wouldn't want to bet your house on the best in this particular field. The bad odds are demonstrated by the ability of most new products to rapidly fail in the marketplace – the people who made them really couldn't know for sure if the people they made them for would want to buy them.

When the failure is a K-tel doggy poop scoop, it doesn't matter. When it's the latest in a string of public-private partnership investments in infrastructure, and one that cost several billion dollars and might have blown half of it, that matters quite a lot – and not just because of the investors and bankers done over.

Governments that can't afford to build the infrastructure their cities need thought they could solve the problem by outsourcing it to private investors. The end users would pay the investors for it and everyone would be happy – except perhaps the end users who find the tolls expensive.

And sometimes that is just what happens. More often now, another set of infrastructure investors buy the failed projects from the receivers at an enormous discount and do quite nicely out of other people's mistakes. The end users still find the tolls expensive. But the risk premium investors now require to build new projects in light of their miserable tunnel experiences means the toll charges would have to be sky-high to start with – so high that few people would be prepared to pay them and it won't happen.

Enter the pleas for a government guarantee, a variation on agrarian socialism, the ability to privatise profits and socialise losses. It's a wrong and dangerous thing and we shouldn't go there.

Cities sometimes need infrastructure that, of itself, can't generate a healthy private-sector rate of return. Yet the total return to society of that loss-making infrastructure can be positive – the savings from bottlenecks cleared, productivity improved, safety enhanced.

Brisbane needs the Airport Link tunnel and the rest of its shiny new roads. Good transport systems are necessary to make a place attractive to live and to invest in. The benefit of having them will become more obvious in the years ahead. The next wave of Mexicans who can put up with Melbourne tunnels (they close a lane when it gets busy!) will love the Airport Link. Right now, though, the traffic figures show Airport Link is more than Brisbane needs. So was Sydney's Harbour Bridge back in the 1930s.

Which gets to the essence of the problem: Australian governments can borrow more cheaply than the private sector to invest in infrastructure. The federal government in particular can borrow extremely cheaply on very long terms. We need the infrastructure and we can afford to service those cheap long-term loans of a couple of percentage points when they're spent on projects that have long-term total returns in double digits.

It's what the Gang of Five present and former Reserve Bank board members are, to a greater or lesser extent, recommending we do. But we have managed to scare ourselves rigid over the mention of the words "government debt", so we won't. The nation as a whole has become sillier than Brisbane drivers – especially those with the funny football.

Disclosure: One of my many mistakes was having the family super fund take a small punt on Brisconnections after it had tumbled – a little gamble on a restructuring deal and Queenslanders suddenly realising what a great toll road it is. Such losses happen in portfolios. If the Reds beat the Waratahs this weekend, it will still be all right.

Michael Pascoe is a BusinessDay contributing editor.