As Notorious B.I.G. put it, Mo Money Mo Engineers. I'm just finishing a book that has a long section on this exact topic and I'll karma whore with a link to this excellent paper with handy charts:
"Skill Biased Financial Development: Education, Wages and Occupations in the U.S. Financial Sector." Scroll down to Figure 6 in particular, which compares the salaries of people with graduate degrees when they go into finance or engineering. They were roughly equitable for decades and then split in the late 80s. Currently, controlled for education level, people with grad degrees make 110K in finance vs 80K for engineers. Even the tight tech job market of the dot-com years didn't close the gap much. The paper also has good stuff on job risk and other factors.
After talking with plenty of people in both sectors, I was surprised the bankers are quick to admit that it's a "quants beget quants" situation. The more sophisticated analysis and arcane devices they come up with, the greater the demand for math and similar hard science PhDs to figure them out, improve them, and keep up with the competition. (MIT and other schools are now offering Financial Engineering, of course.) The managers of the banking companies have even less understanding of these things than the average Fortune 500 tech CEO has about the cutting edge engineering going on in the company he runs. They can only go on what the models and results say, which, considering how unregulated these markets are, means they are basically testing powerful new drugs on the general population.
All of this research hasn't made me think more regulation is the solution, however, at least not how people take that to mean. Monitoring and making data public is essential. Even the banks would appreciate an outside agency that could provide perspective on an industry that is still basically a black box even to the participants. (Current derivative market: $1.4 quadrillion. Enjoy.) That's really what is lacking, sunlight and the context to understand what is really happening system-wide. Goldman and others figure out ways to shave points here and there, but they're making money on concessions, not on the movie tickets.
Getting vaguely back on topic, the conclusion of the book I'm working on is that this is a cultural problem, not just a money problem. Risk aversion is at a peak after building for decades and the relative sure thing of a good salary on Wall St is beating out the chance to change the world or become the next [insert tech billionaire here] in Silicon Valley. This isn't about the current downturn, it's been going on for decades. The American appetite for risk is disappearing. In the same industry, this is why we see half of Valley startups being launched by immigrants these days. Not just because they are qualified, which of course they are, but because they have the entrepreneurial spirit.
Two of the book's co-authors are VC and angel veterans (Max Levchin, Peter Thiel) and they argue strongly that this is having serious effects even among those staying in tech. Instead of working on a big idea for years, some of the best and brightest are cashing out in various ways. Selling their IP, going with a quick-to-market improvement instead of the Big Idea, etc. It's hard to criticize someone for making a buck, but cumulatively it's a slow-moving disaster for innovation -- both by having less in tech and by having more on Wall St.