I'm not sure that opacity is necessary to the banking industry. Unless you think something like fractional reserve banking depends on opacity - which I don't. Under fractional reserve banking, there is more money listed in bank accounts than actual money (for certain meanings of actual money), but every moderately sophisticated participant knows what's going on. If you are financially sophisticated enough that you are bothered by fractional reserve banking, then you ought to agree the opacity isn't an accurate description. And if you aren't that sophisticated (bounded rationality and such), then fractional reserve banking is irrelevant to you.
This article makes the claim that financial systems are opaque in order to help us overcome our risk aversion:
Opacity is not something that can be reformed away, because it is essential to banks’ economic function of mobilizing the risk-bearing capacity of people who, if fully informed, wouldn’t bear the risk. Societies that lack opaque, faintly fraudulent, financial systems fail to develop and prosper. Insufficient economic risks are taken to sustain growth and development. You can have opacity and an industrial economy, or you can have transparency and herd goats.
I don't entirely agree. There's also a risk-pooling aspect (similar to insurance) and a middleman research value-add (individuals don't have to study thousands of possible investments). Still, it's an interesting idea. What other areas of the economy could make a similar argument?