The problem of bank profits

Not, I hasten to add, the problem where they're supporting their profit margin through counterfeiting, which is the essential problem with the "bad loans" currently causing such a mess with the global credit system.

The problem is that there are three ways for a bank to make a profit.

1. Transaction fees
2. Financial services
3. Return on investments

Transaction fees are effectively a regressive private tax, which to the extent that they generate profit for the bank function as a brake on the general economy. I can't see any strong argument that these should be, in a well regulated banking environment, permitted at all.

There's a weak argument that something like Interac transactions are best handled by a common infrastructure, and that the agency maintaining this infrastructure must survive somehow, and so charge fees.
I don't think this argument is an obvious support for per transaction fees on private citizens; it might be an argument for charging banks for their use of the infrastructure, and seeing which banks attract the most customers based on their degree of electronic banking offerings, or it might be an argument for a regulated not-for-profit monopoly providing electronic banking infrastructure to all, but as an argument for a regressive private tax, I think it's much too weak.

Financial services are, if adequately regulated (=no doing risk transfer in favour of you or your buddies, or in return for favours), conceptionally unobjectionable; you're selling specialized expertise, something in principle no different from a dentist or a roofing contractor.
In practise, two problems arise; it's a lot easier to tell, in the case of an immediate service (like getting a filling or a new roof) if the specialized expertise did a good job than in the case of a long-term financial service, for one, and for two, there's a misalighnment of incentives; the financial services guys get paid now, you get your investment returns later. The reputation feedback that helps people select a dentist or a roofer doesn't apply, because the results haven't happened yet. (And the best you have to go on is a very long term version of what the results used to be, based on the results people are having now; that's not a good indicator of what the different people who run the financial service now are likely to do.)

So the problem here is to tie the bank's profits to the client's profits; the simple way to do that is to forbid fee-for-service models and insist on fixed-rate -- 2%, say -- agency fees based on profits received by the client, due after the profits have been received.

Three, though, return on investments, is the seriously difficult one.

On the one hand, individual people certainly don't want the bank taking chances with their money, or at least chances that they didn't specifically consent to have the bank take. Nor do the officers of the bank want to lose money; that makes them look inept and drives away customers to other, more able or more careful banks. This makes well-run banks extremely conservative investors; they aren't going to take any chances on what they invest in, so they're only going to invest in kinds of business that have made a profit before.

On the other hand, there's obvious social benefit to innovation; at the present time, there's a pressing need to innovate so we can replace the fossil carbon energy economy with something else, but in general, technical innovation is a very large net social good, and requires funding. The venture capital model wants a high rate of return, to balance their high rate of risk; this is OK for economic areas that are already started and have an obvious possibility of high return; it doesn't work well for somebody who has come up with a small good idea, one that applies on the scale of a small business of, say, fewer than 20 employees.

The question then becomes, how does one get a bank to invest some money in something they don't know will work? There's good historical reason for being confident that, as a sort of statistical generalization, this is a good idea; the bank will, or at least can, make money, and the support for innovation will have general social and economic benefit. If banks are going to make substantial profits (which they certainly want to do), this is the area where policy would want to push them into making those profits.

How, though, continues to not be at all obvious.