Some time ago in Jamie White's book, "Bad Thoughts" I read some evidence that stock traders perform no better than chance. In Sunday's "Observer" Daniel Kahneman, who is a professor of psychology at Princeton, and got a "Nobel" award for his work on economic science in 2002, confirmed this. His take is quite interesting.
In the first place he points out that good performance is actually impossible if the economic theory underpinning the market is correct. Market theory assumes that rational people pursue their economic interest under conditions of good information. It is true that not everyone will have good information, so they might make mistakes. But professional stock traders do have good information. They put time and diligent effort into reading the relevant accounts and macroeconomic forecasts; they find out about the "quality" of the managers of companies; they monitor changes in structure and trends and compare all of that with other firms. So we can assume that they are informed; that they intend to make money; and that they take decisions about buying and selling, with a view to maximising their profit.
That being the case, it follows logically that there should be no market. They would all look at the same informaton; evaluate it; and come to the same conclusion about each stock. If trader X wishes to sell stock A it is because he thinks that stock is going to fall in value. If he thinks that, and if there is objectivity to that conclusion, then nobody will buy it: because they think it is going to fall too.
But a market exists. So is it then that some traders are more skillful than others? It is perfectly possible that some people are just better at analysing the same information than others, and so they will make better decisions. In that case there is room for disagreement, and room for a market. But if that is the case it should quickly become apparent: and again the market should disappear because everybody with any sense will follow the decisions of the most skillful. If Trader A is the best then if he wants to buy stock you hold, that is sufficient reason not to sell it: if he wants to sell his own holding, that is sufficient reason not to buy it. Once again the market cannot exist.
Financial companies are not transparent: they like to keep their business quite secret, and "commercial in confidence" is a ridiculously strong argument in many arenas: that being so it is possible that there are real skills, but the results of each trader are hidden: in that case there is also room for a market because they cannot just follow the best. They don't know who that is.
So although it is impossible in theory the market is not perfect and there is scope for skill. Whether their is any actual skill is perfectly testable: and it has been tested. There is not.
Kahneman first reports a study by Terry Odean, who is a finance professor at University of California, Berkeley. Odean had an unusually large sample to work with in social science terms. He got the trading records of 10,000 accounts of individual investors for a period of 7 years. There were details of 163,000 transactions. Odean then analysed every instance where a trader sold one stock and quickly bought another. Clearly any trader who does that believes that the stock sold will not perform as well as the stock bought: that is why he does this. So Odean looked at the actual outcome performance of both stocks in the year after the trades were made. After taking the costs of the transactions into account, the stocks sold did better than the stocks bought by an average 3.2% in the year.
Odean and a collaborator called Barber then followed this up with a further study which showed that the more active a trader, the poorer the results. This confirms that the decisions are poor when the trader trades: and that doing nothing at all is the better strategy.
On the face of it there is no skill: but there might be some other explanation, and so in order to come to that conclusion it is helpful to do another kind of study: and that has been done too.
There is an old adage: "the more I practice, the luckier I get". It applies to games of skill which are seen by those who do not play them as if they were games of chance. The point is that if there is a skill, then those who have that skill will perform consistently better than those who do not have it: they may lose from time to time, but if skill is in play, then over time the more skillful will win and the less skillful will lose. Chess is a game of skill: the amateur with no skill will not beat a grandmaster over a series of games: he will hardly ever beat him at all. Poker is also a game of skill and again the person who has that skill will win over a series. Bridge is a game of part skill and part chance. A pair with good skills will beat a pair with poor skills over a series too: but not by so much. Roulette is a game of chance: everybody is just as likely to win as everybody else, and despite the search for a "system" if the wheel is honest there is no difference over time.
If the traders have skills then it can be assumed that they are not equal (for reasons already given, and on the basis of common sense as well): if follows that there should be consistent rankings if you compare a group of traders with each other over time. This can be measured using rank correlations. Kahneman reports that he had he opportunity to do this using the investment outcomes of 25 "wealth advisers" for each of 8 consecutive years. These advisers achieved a "score" each year based on those outcomes and that score determined their bonus. So this was an ideal set of data to rank using correlation coefficients, and that is what he did. the average of the 28 correlations so produced was 0.01. Correlations of that sort range between 1 (perfect correlation) and -1 (inverse correlation): 0.01 means no correlation at all. The results are random.
This is in line with other research which has looked at the performance of Mutual funds over a period of 50 years. Apparently 2 out of every 3 underperform the market in any year: and the year on year correlations are also very small indeed. according to Kahneman
What this means is that there is no skill at all: to put it another way, if you can get a job in this field you cannot fail. You are as lucky as anyone else. And since they all fail to perform some of the time, they discount that failure and focus on the years when by chance you make lots of money.
What about all that work they do analysing information, though? Well I invite you to visit a bookie: nobody works harder at gathering information than some of the people who bet on horses. Bookies don't go bust, though.
What is obscene is the rewards these people get for gambling and getting the results you would achieve by chance. And these people think of themselves as skilled and important and "worth it".
ETA: I found this little gem when reading around this subject and I think it confirms all I have said above: but framed in quite a different way.
www.swing-trade-stocks.com/trading-business.html 
