There's a great weakness in the (otherwise sound) argument that borrowing for infrastructure is a good thing, adding to demand in the short term and improving productivity and supply in the medium term. We should be doing a lot more of it, which would impose no unfair burden on our children.
That weakness has been exposed by the election campaign. Marion Terrill, of the Grattan Institute, has looked at the parties' promises to build new transport infrastructure and found that a lot of them would be a waste of money. If so, all bets are off.
The fancy arguments don't work if the projects are chosen for their ability to buy votes rather than for the value of their contribution to improving our transport.
On some measures, by the way, transport projects – including roads, rail, ports, public transport, airports and bicycle infrastructure – account for about 70 per cent of our total infrastructure effort. They leave out the national broadband network and state spending on water and power.
With the two big parties committed to returning the budget to surplus, to changing the mix of government spending in favour of capital rather than recurrent spending, and to using the expertise of Infrastructure Australia to improve the evaluation and selection of projects, you might expect – especially if you'd just arrived from Mars – to find all this influencing the promises they made at this election.
It will amaze you to learn that, come elections, the parties don't practise what they preach.
One of the biggest steps towards getting more economic benefit from our infrastructure spending was the establishment by Kevin Rudd in 2008 of the semi-independent Infrastructure Australia. The Coalition has fully supported it and various states have set up similar bodies.
As Terrill says, now we've got IA, the parties' selection of projects should be simple. It examines projects worth more than $100 million that are of national significance.
It assesses the "business case" – buzzword for a cost-benefit analysis – of such projects and sorts them into three categories.
Those that are fully assessed and pass the test are classed as a "project". Those that fail their full assessment – that are judged not worth doing – have their evaluation published on the IA website.
Those that are yet to be fully assessed are classed as an "initiative". A case in point is the Melbourne Metro Rail.
Terrill examined the Coalition's list of transport promises (worth $5.4 billion) the Greens' list ($6.5 billion) and Labor's ($6.7 billion).
The Greens have no projects that are IA-approved, and almost half the cost of their promises is for schemes that aren't even being assessed.
Only the tiniest part of Labor's spending is on IA-approved projects, with about a third going on schemes it isn't even looking at.
Of the three parties, the Coalition has the highest spending on IA-approved projects and the lowest on schemes it isn't assessing. Even so, most of its spending would go on initiatives still being assessed.
In other words, all three parties have selected their projects with little reference to the IA's evaluations. In fact, they make a mockery of the IA process and its efforts to ensure taxpayers get value for money, which they profess to support.
But why? In a word: politics. For a fuller answer, Terrill offers circumstantial evidence.
One clue is that both big parties have plenty of projects worth less than $100 million each, which thus can't be assessed by IA.
Whether it makes sense for the federal government to get involved with such small-beer projects when they can be handled by state and local government is a good question.
Another clue is that there are only six projects on the IA's list of things worth doing: the WestConnex motorway and the M4 motorway upgrade in NSW, the Ipswich motorway and the M1 Pacific motorway in Queensland, the Perth freight link in Western Australia and the Brisbane-to-Melbourne inland freight railway.
The Coalition has only four of these on its list of promises. Labor has just one. The Greens have none, having chosen to nominate one major public transport project in each capital city.
One of the six worthwhile projects – the Sydney M4 motorway upgrade – isn't supported any of the three.
Clue three is that Australia is a federation of states and territories. Clue four is that parties attain government by winning the most electorates.
Though it may be that some states don't have any projects of national economic significance while others have quite a few, it wouldn't be surprising to see projects being spread across the states in a way that roughly fits their shares of population.
Sorry, doesn't fit. Terrill divides the value of the three parties' promised projects between the states, ensuring no double-counting. Queensland gets the most and Victoria gets quite a bit more than the larger NSW. WA also gets a disproportionate share.
Actually, she says Queensland has been getting more than other states for a decade. Maybe it's the state where elections tend to be won or lost.
Whereas almost all of the projects IA approves are in capital cities – improving commuting and connections with ports and airports – the two main parties prefer to spend in the regions.
Perhaps the regions have more marginal seats (or National Party electorates needing to be squared away by the Coalition).
The Coalition is promising $185 million to duplicate the Princes Highway from Winchelsea to Colac in Victoria, which IA has found would yield benefits worth 8¢ for every $1 spent.
You don't have to be Einstein to conclude a lot of spending on capital works – federal and state – is used to buy votes, not to make worthwhile additions to our infrastructure than improve our productivity.
If so, such spending will leave a burden for our kids.