I'm not clear what's supposed to be secret about this, but I suppose it makes for a better headline.
The banks are said to be keen provided someone else takes the risk. That's hardly a surprise. Anyone ought to be keen on a business deal in which someone else takes the risk and leaves you only the rewards.
Government probably can't take the risk. The risk is - in effect - the RAB charge that David Willetts is already paying and there will only be scope here through something that amounts to regulatory arbitrage - in other words to a different basis of calculation that comes out with a different £ figure for the same risk.
Institutions can, but why should they? The banks will want to do business with elite institutions because their students get the best jobs, but elite institutions (assuming the AAB plan goes into effect) will have uncapped access to the state-funded loans, for free. At best, if they share my view about the legal risks of the AAB plan, private loans will give elite institutions that wish to grow their FTUG numbers a way to get uncapped numbers for all qualified entrants.
I suspect the set of elite institutions that wish to grow FTUG numbers and agree with Andrew Fisher is a small one.
There may be scope for non-elite London institutions to be involved, simply because their graduates tend to earn better at least at the start of their careers (due to London Weighting), and some of these may wish to expand - in fact many of them certainly do. Will the banks and the institutions be able to agree on a value for what the risk is in these cases? We have good data at the institutional level six months after graduation, but not thereafter. My instinct is that it will be difficult to get enough longitudinal data determine the level of risk at all accurately (there is a certain brute unknowability about the future). Banks may find the residual risk well within their tolerance, since lending money is what banks do for a living, but the institutions may not.
There may be a middle way where the Government (in effect) bribes institutions to accept the risk - the bribe may be less than the RAB charge would have been. Thus overall numbers can increase as less risk-averse institutions expand. Provided the bank-funded loans have exactly the same terms and conditions as the Government-funded ones (and why wouldn't they, if the institutions are paying the banks for the extra risk this involves?), then the distributional issues are resolvable.