If you still think that people are honest on Wall Street, you need to read the article on forensic finance in the latest issue of the Journal of Economic Perspectives. Jay Ritter documents several frauds that were so widespread that they were discovered by looking at aggregate data, thus instigating investigations.
Ritter documents four examples. The first the so-called late trading of mutual funds. Their price is set at the end of the trading day, but predictable market movements, say due to major announcements made while the market is closed, implies that their value continues changing. Yet some sold shares at old prices, because either they were in exchange allowed to invest in vehicles that were generating high fees, or employees where plainly enriching themselves.
The second example pertains to employee stock options backdating, whereby options where given with an exercise price set at the market price of the share, as required by law, but at the lowest price in recent trading. In other words, firms where pretending their filing was late, while it was really on time with an old price. Just from looking at abnormal stock returns around stock option grants, it became obvious that those returns where more than 3% lower on aggregate for firms that do not have a fixed stock option granting calendar. A major scandal ensued, and it uncovered that firms that were backdating lost 7% on the stock market. All this for allowing executives to gain on average half a million dollars annually per firm.
The third example is the so-called spinning of IPOs, that is, offering underpriced new shares to privileged people. As IPOs were generally oversubscribed (and could thus have been priced higher), bookrunners could choose whom to sell them. Of course, they chose those who could provide them with other favors, either separate deals or in the case of executives, loyalty. And said executives were fine with the underpricing that cost their firms.
The final example pertains to something that looked like fraud, but ended up being "just" incompetence. The Thomson Financial I/B/E/S database of analyst recommendations had at some times about 30% of the recommendations altered after the fact, putting them more in line with actual outcomes. While this looked like rewriting history to highlight the competence of these analysts, it turns out the coding and data entry policies were horrendous. Yet plenty of investors relied on this data. Such incompetence seem also present in rating agencies and elsewhere.
Some have called the increased government involvement in financial markets a step back. But looking as these generalized frauds or examples of incompetence, it seems regulation and consistent disclosure requirements that are also verified are necessary.