Some time ago, a book I read stated that the very nature of capitalism was to allow firms to die without any consequences to the system. The author wondered why the same thing isn't possible with banking institutions. That author was Nassim Nicolas Taleb and the book was titled Black Swan. (However, I wouldn't really recommend the book) Although at first I thought the author was right, after considering it I realized that the banks are not at all like any other firm.
If tomorrow Google were to file for bankruptcy, the consequences would be a lack of a good search engine and thousands of Google employees left without a job (and you wouldn't be reading this!). Otherwise, no harm would be done to the ordinary person. Now imagine what would happen if one of the major banks were to collapse: millions of people would suddenly be worried of being left with little or no money (although a country guarantees approximately 100,000 euros of deposits by each person in each bank). In the end, people would get their savings back (expect the persons which had amassed millions who will only get 100,000), however, the whole procedure would be traumatic to the ones who were unlucky enough to get caught in it. Nevertheless, such a procedure is preferred to the cost of pumping a bank with new money every now and then just to sustain it.
Thus, the question becomes: When should banking institutions be rescued? The answer is simple, although it may be tricky to define at times: When they are solvent, largely problem-free and have faced trouble only as a result of an unpredictable event to which they bore minimal or no responsibility. This, for example, would be the case with most EU banks which suffered great losses from the Greek haircut (although there were other banks which had overextended credit over the years).
If a banking institution is thought to be worthy of rescue then the course of action should be by directly providing liquidity to them. This would mean that governments, Central Banks or (preferably) an institution like the ESM, should provide liquidity to the banks; with the latter issuing new shares with voting rights, committing to repurchase them at an amount obtained after accounting for a fixed interest rate, and agreeing that the institution's participation in the bank's equity capital would be reduced each year, thus making sure that the stream of payments from the bank to the state would be continuous. Even if government (or Central Bank) intervention in banking institutions is not considered a good thing, it is much better than merely purchasing preferred shares, as the very recent Barclay's example indicates. By agreeing to pay interest on the amount given to them and repurchasing those shares at fixed intervals, it is assured that the taxpayers' money is not just thrown down a bottomless pit, making the banks more careful in their future investments and gradually reducing government intervention within them.
Following this, the distressed bank should implement some austerity measures, either by cutting down salaries or other benefits or, better, by reducing its activities both at domestic as well as at an international level. Needless to say, the bank's heads at the time of government intervention should be replaced. The bank could spin-off some of its operations if they are profitable, however, this should be done in a way that the new company is an independent one and not with the bank still controlling a substantial interest in it. Likewise, the spin-off should not be sold to another bank as this would increase the other bank's risk of becoming too-big-to-fail.
In the case of the bank becoming insolvent, problem-ridden with issues which are not due to unexpected events (e.g. management issues, poor lending decisions, etc) then it should be let collapse just like any other ordinary firm. If the bank has some operations which are profitable and well-functioning, then these should be spun-off, or sold to another company, even though selling them to another bank should be avoided for reasons of the other bank becoming extremely large. What the reader may inquire is what if the second bank is already too big? What difference would it make if the profitable operation is sold to them? First of all, a new spin-off will require more employees (think of new management, trainees, secretaries, even cleaners) than if it is sold to another company where existing economies of scale would take over. This would have a positive effect on unemployment especially since other bank's employees are expected to be laid off as a result of the collapse. Then, even the acquiring bank is already too big, in what scenario would the government (or another institution) be spending less for stabilization: saving a 100 billion bank or a 105 billion bank? And remember that's billion...
Under the current scheme, most banks are either classified as too big to fail or too little to have any exposure. Take the UK for example where only 5 independent large banks exist, with the government having interest in two of them. What should have been done was to reduce the level of activities of the Royal Bank of Scotland and Lloyds Banking Group to levels where such intervention would be easier in the future, also allowing the creation of more banking institutions, which would promote a competitive environment instead of an oligopolistic one.