I was just reading Andy Kessler's Wall Street Meat, and he notes that in 1975 commissions were 75 cents a share! By 1989 they were 5 or 6 cents a share, and today they are about 1 cent a share. If you bypass the sell side and use ECNs (Electronic Communication Networks, such as ARCA or Island), currently the commissions are around 0.2 cents a share. All this is for institutional trading, retail will always pay more.

I can't get any confirmation on that 1975 number, but if true that sure puts many anomalies in perspective. Given the average stock price has been $20 for the past half century, this means you had a 7.5% (1.5/20) hit for getting in and out of a position in 1975, but that's before the bid-ask spread, which would probably add another 2-5%. There aren't many opportunities that overcome those costs, like betting on horses the vig is too big.

The book (Wall Street Meat) is ok, a quick read with several funny anecdotes. His theme is that research is just fluff to generate interest in stocks, not very shocking. But it got me thinking. Who currently reads sell side research thinking it is giving them an edge? Mary Meeker, the internet hypster who actually made it through the crash without any repercussions (unlike Grubman, Quattrone, or Blodget), still works for Morgan Stanley doing research. I guess there really is no such thing as bad publicity.