Chris Dillow has been arguing in favour of a state bank. Recently I finished reading "The Truth About Markets" by John Kay, in which Kay describes the fiasco of the development of British Advanced Gas Cooled nuclear reactors.
These reactors were announced in 1965 by the Labour Minister of Power Fred Lee, who promised that we would "hit the jackpot" in exports of AGRs.
The AGRs were sold for £1.9 billion when British Energy was privatised in 1996, along with a promise by the government to underwrite any future clean-up costs. The total cost of developing the AGRs was around £50 billion at 1996 prices.
On the face of it this sort of thing makes it look like it is a bad idea to put the state in charge of allocating investment. Kay's point is more subtle though, he explains that what markets possess and the state does not, and what makes markets such effective wealth creation systems, is the disciplined pluralism of the market.
In the market new ideas are implemented all the time. Most fail. Companies either ditch the bad ideas or go bankrupt. The problem with the AGRs was that the government wanted the electricity industry to speak with "one voice", and the civil servants who made such bad decisions were never properly held to account. There was no plurality of ideas and no discipline when things didn't work.
The problem then is not that necessarily with the state funding things, but rather with the state funding things without its decisions being subject to any kind of disciplined pluralism.
Back to Dillow's point: can it really be so hard for the government to create an independent state bank? It could be run in a similar fashion to the BBC, with an independent income guaranteed from some kind of levy.
As Dillow points out, the future is inherently unknowable. Professional VCs expect only around 1 in 8 of their investments to work out. Small loans made in sufficiently large quantities may generate enough benefits to outweigh the losses.