indexes like the dow and sp500 are constantly revised. poor performers are replaced with better ones. there is an incredible bias. i understand the usefulness of indexes but how can they be relied upon historically?
(1) The DJIA isn't revised constantly. (2) Revision happens more often with broader-based indices. (3) You should welcome revisions because it brings about trading opportunities in. (4) Remember, an index is just a number. Don't get hung up on what it "should" be.
i was thinking along those lines, pekelo... a thread on firewall problems or a $20 dell coupon can get 100s of replies, and yet serious trading questions can go unanswered. this is a trading forum right? are these questions too difficult?
dividend, I do all kinds of index work. 'Nothing wrong with adding/removing components. If no upkeep is done, then an index will just whither away and die. The idea of having a "publicly available" index with as long a track record requires just that...periodic maintenance otherwise people would then be complaining that they don't have an index with a longer life span than 10-20 years. Whether the added/deleted components were the "right" ones, well that can be argued to infinity but keep in mind that many stocks actually perform better *after* being deleted to an index and worse after their addition. Look at the Merrill Lynch "HOLDERS". Many of the original companies on those baskets are no longer traded and the indexes are no longer representative/tradable compared to when they originated.
Exactly. And that's what the OP complained about, again correctly, that there is no historical validity comparing the Dow of today to the Dow of the 60s.... You are talking about index maintenance, which is fine up to a point, and we are talking about reliance on any index as a historical measuring standard. We simply can not rely on them, but Wall Street would like you to believe that stocks are such a good performers over the long run, and they use a fake measuring stick to prove their data. I actually have a better idea of measuring stockperformance over the long run. Instead of a bucketcase of always good performers, at the start of every decade or 5 years we could pick 100-500 stocks randomly and see them through the 5-10 years how they performed. When the period is over, we would pick another randomly choosen group. In this way true performance would be measured.
If you're really so concerned about comparing today's index levels to those of the 1930's, 60's or 90's; compare P/E's, dividend yields, book values and other "fundamental" measures. I'd rather get that $20 Dell coupon instead.