You're wrong. Few reasons:
1. "A very large portion" - no, there are some motivated by different regulatory treatment for different transactions, but it's not "a very large portion." There's also nothing wrong with the regulatory-motivated trades generally done by banks. It's not the same as what this guy did which is clearly immoral/wrong - he was not doing something in an optimal way to reduce tax burden. He was doing things that caused governments to literally pay out new money - negative tax payments - possibly under fraudulent pretenses. As an analogy for what is appropriate - if you open a Corp in the US but decide to open an S Corp rather than a C Corp for better tax treatment - is that a reg loophole? No you're following the law to get better tax treatment.
2. Wide bid-offer isn't a rip off. Markets are generally competitive, corp clients can easily ask 3-4 banks for a price for the same product. What confuses novices is that in financial markets its typically easy to see the "mid" in addition to a bid-offer and then feel bad about profits. Would you feel equally guilty if you went to the store and saw the "mid" on a $1000 iPhone is $500 - and the bid/offer is 50%? Before you say "it's different!" - consider the complexity involved in arranging a profitable bank that has a competent sales team, trading desk, operations, compliance, capital, legal, etc. etc. - and also does profitable transactions. It's a moat - hard to replicate - and that's what leads to profits.
For the first point I agree. In this person's case it is more extreme. I don't know what capacity your involved in sell side trading but I'll share some cases I've seen in case anyone else reading is curious as well.
For rmb, krw, and a lot of Asian currencies in general there is an onshore and offshore market. A lot of the times both onshore and offshore books are managed by the same team. This creates two tradable currencies which should theoretically move in lockstep and some corporates have capacity to arbitrage/move currency across border. A common strategy is to simply bet that their spread converges when some technical factor like a large trade going through moves one market more than the other. Hardly nefarious but exists because of regulatory reasons and hardly rocket science. The problem happens when currency trends strongly in one direction in which case one book has massive losses and the bank ceo of that branch is not happy. Or when the onshore/offshore team starts trying to claim a stake of the profits because their resources/credit lines are being utilized. A lot of bickering and politics then ensues.
Another case, in China, regulation requires long usdcny fx forwards to be accompanied with an extra 300bp charge to stop cny depreciation. Option structure only require half the charge. A lot of the desk pnl is essentially made from selling synthetic forwards structured from calls.
Similarly a large part of the hybrid structured business in general is because some insurance/pension funds are only allowed to trade bonds but want fx/equity exposure.
For 2, from my experience it really is ripping off most of the time. The exception is when a client comes in with an order much larger than what the market can digest quickly. This usually requires a mixture of luck, skill, and balls to avoid it blowing up in your face and is one of the rare times I think traders really are doing their job. However these trades don't come too often
