Should the UK really want to split the investment and retail arms of the banks? I'm really not sure.

The first pro argument is that, when people put their hard-earned cash in the bank for safe keeping, they tend to think that the bank keeps it safe in their vaults. Or at the very most, lends a bit of it to reputable local businesses. The reality is that the bank can take your money and do whatever they like with it – gambling it on the international markets if they want. Sure, when you turn up to draw money out again, they have to give you some – but they can do that from all the other people's deposits that are sloshing through over its counters. Unless, Northern Rock-style, people figure they've taken too many risks, and everyone wants their money back. So retail customers are being exploited to feed the bank's risk-taking.

Second, after the stock market crisis of 1929, America separated the retail and investment banks through the Glass-Steagall Act. From there on, we enjoyed almost 70 years of relatively stable financial markets. But after the 'big bang' deregulations in London and pressure on politicians in the US, Glass-Steagall was repealed in 1999. And look where we are now.

Third, if we are going to insulate retail customers from risky investment banking, the split has to be complete. If these functions remain part of the same institution, with just regulatory restraints in operation, the bank will undoubtedly find a way round these so-called Chinese Walls.

Fourth, many of our hybrid investment-and-retail banks are too big to be allowed to fail. If risky investment banks blow up, retail customers will lose their money and will have to be bailed out by the taxpayer. That is not a risk that taxpayers want to bear.

But there are many arguments on the other side, too.

First, the mechanics of this split are expensive. Systems will have to be re-jigged, branding changed and regulators and bank compliance officers hired. Even the Independent Banking Commission, which is proposing the reforms, says that will cost £10bn. Such is the way of these things that it will probably cost a lot more.

Second, it will put London at a disadvantage compared to other financial centres. Not only is there the extra structural and regulatory cost. In addition, UK banks' investment arms will no longer have access to all those funds that come through the retail sector; and the retail banks will no longer be able to give their savers the better deals that their lucrative investment activities once made possible.

Third, and more tellingly, the policy aims at the wrong target. It was not the investment banks that caused the banking crash. Arguably it was the inevitable consequence of a twenty-year binge of money and credit engineered by Western governments, and the inability of Western regulators and central banks to understand what was going on and act sensibly to deal with it. But even if you do blame the banks, which banks were the first to get into trouble? It was of course, the small former building societies like Northern Rock, which could not hack it as banks; and Halifax, which got swallowed by Bank of Scotland; and those like the Royal Bank of Scotland, which came a cropper because of injudiciously large takeover strategies rather than investment failures.

Fourth, the split would not obviate the need for future bail-outs; it would make them 100% certain. The whole purpose of the proposed split is to protect retail bank customers. The idea is that retail banks would be more conservatively managed and therefore less likely to fail, so there would be little need for any taxpayer bail-out. In fact, since the newly-split retail banks would be in no doubt that the government would be obliged to step in if things went wrong, they would have every incentive to act as riskily as possible, offering their customers the highest returns and the lowest costs they could squeeze out, so as to have an edge over their competitors.

So what should we make of all these arguments?

First, even if you believe the banks should be split, you have to concede that legislators will mess up the process. True to form, they will do it in some ham-fisted way that completely undermines successful banks and puts at risk the UK's financial services sector, with all the jobs, income and deficit-busting taxes that it generates.

Second, if people knew what the banks were doing with their money, would they still hand it over the counter for 'safe keeping'? Hardly. The real problem is that there is a lack of proper information in this market. Rather than split up the banks, we need the banks to tell their customers what the real deal is. If there is indeed a demand for low-risk but low-reward retail banking, the banks would then have every incentive to do their own ring fencing and provide what customers want. People would know that their money is not just put in the vaults, and they could choose exactly how much risk to accept. It seems like one of the first duties of a market regulator – to make sure that customers get the right information – but one that regulators have so far failed to deliver. Let us focus on that before we let ham-fisted politicians loose on our banks.

Third, our biggest problem is not the need for retail/investment separation, but the fact that we have too few banks and they are all too big to fail. Partly it has been the very burden of regulation that has driven all the mergers in the sector – banks these days need fleets of compliance officers to make sure all the boxes are ticked, and small banks can't afford that. And in other ways, public policy has promoted this giantism. But then, when one of these leviathans get into trouble, it's a major national (and indeed world) problem. We need to remove the bias towards large size. One useful policy, which could be implemented very quickly, would be to put more onerous reserves requirements on banks, the larger they are. Then banks would have an incentive to break themselves up, but in a way that made commercial sense, rather than how politicians might think it should be done.

Fourth, we need to recognise that separation of retail and investment banking is a second-best solution to problems that are of governments', rather than the banks' own making. It may be justified on the grounds that we have a banking oligopoly, and hence far too little competition, far too large banking institutions, far too little information to customers and far too much reliance on governments and regulators. We really need to attack those core problems, rather than treating the symptoms.