The idea that pension schemes might invest some of their money in infrastructure projects, such as building roads, railways, or schools, is not new.
It has become a bit more common in the past decade or so to invest in what we used to call public works.
In that time pension funds, which control billions pounds of money on behalf of their members, have been looking for new areas into which to invest.
The attractions of doing so are fairly obvious. You have a tangible asset - something you can see and feel.
And on the face of it, it should offer you some sort of fairly steady return, especially if it generates cash from paying customers. It might even be for the public good, which is always comforting.
Examples might be drivers paying road tolls on a motorway, train operating companies paying for the use of a new stretch of high-speed railway, or airlines paying for the use of a new airport or runway.
The pension fund or funds own the asset, charge for its use, and collect the money.
Such projects can be either for the benefit of a government, or just private sector companies.
Economically it should not be much different from owning some commercial real estate, like a shopping mall or a retail park, where the fund owns the stores and collects rent from the shop owners.
What could go wrong? Well, quite a lot of things.
Firstly, there is construction risk.
It is very hard to ensure that a large construction project is built to the original time and cost.
And operating a big project for many years also carries a risk.
What if people stop using it, or your customers go bust, or refurbishing it costs far more than you thought?
What would happen if you simply managed the asset badly?
Frankly, almost anything can happen over a 20-30 year investment horizon.
This is probably where the government hopes to step in, with some sort of guarantee, so that pension funds are happy to invest their money.
Another important risk, from a pension fund's point of view, is that such projects would be hard to sell.
In the investment jargon, they would be illiquid.
That would matter a lot if a pension fund wanted to sell its stake in a project that was not generating the expected return, or wanted to cash in a paper profit from a scheme that was doing very well.
PFI Mark 2?
There are no details available - the idea is, so far, just that: an idea. It has not yet reached the level of a plan, let alone a detailed one that any individual pension fund could consider in detail.
But if the government was able to eliminate some of the construction and operating risk then it might take off.
One question is why the government should even bother? Why not just sell inflation linked government bonds (gilts) to pension schemes and use the money on normal government capital spending?
Putting together some sort of mechanism to facilitate pension scheme investment in an infrastructure fund might let the government keep the cost off the government's balance sheet and stop it inflating the government's debt.
Won't the emerging idea look very much like the controversial PFI (Private Finance Initiative) schemes that have been used for more than a decade now to fund the building of schools and hospitals?
Yes, there are clear similarities.
But as evidence emerges that these are turning out to be very expensive for the government, and very profitable for some of the investors who put up the money, the government may just be looking for a cheaper version whose cost it can predict and control more closely.