Many people seem to think that finance is merely some sort of legalized gambling which makes some of the players rich and may be fun for some people but is totally useless to society, in the same way that a casino is useless.
Finance is similar to gambling and it does make some people rich but it is extremely useful to society. It's as if we invented a type of casino where merely by playing, people are doing useful work that makes the world a better place.
What is finance good for?
First, let's look at what finance can do:
Now, what is that good for? I'll go from least important to most.
Borrowing in particular makes our lives more pleasant by purchases expensive things in the middle of our lives instead of waiting until near the end.
A number of these are crucial for innovation to occur. For example, if you did not have investment, or if you could not diversify, or if you could not transfer risk, tech startups would rarely be born. If we used a method that was less good at predicting the future to allocate funding (for example, if committees of voters or of politicans allocated funding), then frequently the best startups would never be funded, as everyone has heard stories of awesome startups that most experts thought were silly ideas that would never work (until they proved the experts wrong by taking off anyways).
Most importantly, all of these lead to reduction of poverty. All of these things contribute to economic growth and economic growth is needed to give society the means to stamp out poverty worldwide (other things are needed too, but economic growth is one thing that is needed).
This doesn't mean that our financial system is perfect or that no regulation is needed. But don't throw the baby out with the bathwater; finance is not a senseless game that makes people rich, it is useful work that is making the world a better place.
This is a good idea but not because it would prevent bad things in itself. it's a good idea because it makes the system simpler and easier to understand. However, the real solution is braeking up Too Big To Fail banks.
There will always be financial institutions going broke. The best that we hope for is to attentuate the domino-like, indirect spread of the failure by making the system as loosely coupled as possible. An important part of keeping the system loosely coupled is not having single institutions that are "too big to fail". Such entities should be broken up -- not because they have done anything wrong, but because their size is suboptimal for society.
So, Glass-Stegal is not the answer. Breaking up Too Big To Fail is.
A pet peeve of mine: the media always says that 'the markets demand this or that'. Markets don't demand anything. Markets only predict.
This usage probably came about due to the technical economic term "demand" as in "supply and demand". But it is being misinterpreted as "demand" as in "I insist that you give this to me". These are not the same at all. For example, there is "demand" in the economic sense for Monty Python DVDs. But this does not mean that the consumers of those DVDs are "demanding" them in the sense that they will get angry and complain of injustice and perhaps sue or protest or riot if the DVDs become unavailable (well, some people would, but certainly not every consumer who is demanding the DVDs in the economic sense of the word).
This use of language where markets are said to 'demand' things causes various problems. First, it implies that maybe the 'demands' of 'the markets' could be taken into account when making decisions. What the market 'wants' is not a good way to determine which political decision should be made (because, again, markets don't 'want' anything. They only predict). Now, perhaps some market-like mechanism for decision-making could be imagined, but present-day financial markets are not it.
A second problem is that people get annoyed at 'the markets' for being so 'demanding'. Then they try to regulate the markets to make them less 'demanding'. You cannot make the markets less demanding, because they already aren't demanding anything at all (nor should they). The effect of most reforms intended to make markets less demanding is merely to make them less good at predicting things, which doesn't help anything.