Woah there - the Treasury costing does not "conclude" or "predict" that student loan uptake will increase to 95% by 2008/09, it *assumes* student loan uptake by full-time students will rise to 95% by 2008/2009.
Treasury made their calculations based on those assumptions, which are just educated working estimates. They're not necessarily wrong and certainly not ill-informed - but they're just assumptions.
And those assumptions are most certainly debatable - while Treasury has models (fallible, but detailed and earnest models) to predict things like fees and student numbers, they don't have any statistical model or supernatural powers to predict how people will change their borrowing behaviour in reaction to the policy.
All they know is that if you gave them an interest-free loan, they'd take it. Me too. I would be more than happy to put money on Treasury's assumptions over Labour's assumptions, but I just want to make absolutely clear that assumptions and conclusions are two incredibly different things.
Having said that, I think it was particularly telling that Michael Cullen has been trying to argue that uptake won't dramatically decrease, when the real ripper in the costing is the assumption that voluntary repayments will cease altogether.
The loan debt was always going to go up, but it was expected to reach $13.5b by 2015 (my data here has been superseded, but the principle still stands), and stick around that point once it got there. But if voluntary repayments stopped, then the total student debt would rise faster for longer, and the point at which it stabilised would be much higher - meaning that the long-term cost of the no-interest policy would also be much higher.
(47% of all repayments are made directly by borrowers, most of this would be from voluntary repayments, though IRD doesn't have a figure.)
The cessation of voluntary repayments is the real boogeyman in the scheme, and it makes me greatly suspicious that Cullen doesn't take it head-on.
As for government debt part of the costing, it's been *way* overplayed. The government takes on debt to lend to student, but that debt is instantly offset by the debt that the student owes to the government (i.e. The government gains an asset). However, because it's interest-free, and because not all students pay back all of their loans, the real value decreases over time, which has a real impact on government's net worth, but only at a portion (~20%) of the face value of the debt. (Simplified: Just divide what National says by 5.)
In short, Treasury is only offering an opinion. A definitive opinion, but an opinion nonetheless. It's also an opinion that I agree with.