Contrarians would like to test it at the bottom of a bear market or top of a bull market.
Insider trading is one way to generate higher-than-normal returns. The most egregious forms are illegal, or at the least, require statutory public notification. However, some investing methods based on insider knowledge are legal.
Anybody who buys into the sector where he works (not necessarily his own employer) does so because he can make accurate judgements about the sector. “Insiders” know about hiring-compensation patterns, changes in credit cycles, marketshare, and so on. They don’t need insider financial data.
Peter Lynch extended the logic. He suggested consumers can beat financial analysts by investing in industries catering specifically to them. If you prefer shopping from a specific supermarket, or like a new brand of chips, shoes, and so on, check out the investment possibilities.
Another insider scenario is not discussed much because it involves brokerages and can lead into grey areas. Front-running is illegal. But brokerages can develop trading strategies by marrying public data on trading volumes and delivery ratios, to their proprietary data. They can mine their own books to profile their clients by trading styles, success rates and categories such as retail, institutional, hedge fund and conservative, and develop insights.
The data-mining is legal, though if it's done improperly, it could raise concerns about privacy or client confidentiality. It's only if the brokerage makes proprietary trades or recommendations on the basis of what it learns that it may be sailing close to impropriety. Even then, it is not necessarily doing something illegal, or transgressing fiduciary boundaries. But brokerages are always very discreet about this in order to avoid legal issues.
This sort of data-mining can be described as a variation on “crowd-sourcing” to borrow a term from internet jargon. While it isn't possible to do it precisely without access to granular trading data, anyone can use public information on the Internet and Web 2.0 to “crowd-source”.
Also Read
Internet makes it possible to easily track recommendations from institutions, and from popular TV programs. Such expert opinions can be aggregated and correlated to actual trade data. If X recommends a stock and that triggers a jump in volumes, we know X has an effect on sentiment.
Crowd-sourcing methods that track Web 2.0 may create a paradigm shift in our understanding of sentiment. One simple possibility is to track and aggregate twitter feeds using specific keywords about markets. This offers insights into retail attitude. Obviously any such systems must be tweaked in terms of parameters, and back-tested to see if they may give decent results.
Another method of crowd sourcing could be to follow market-related Facebook updates, and feeds from popular blogs and forums where investors post. This is more difficult to parse without human intervention because the information is less structured than in 140-character tweets.
To do any of this efficiently needs a lot of computing power, programming skills and a fast Internet connection. However, the required resources are not prohibitively expensive and can be bought “off-the-shelf” by individuals. So crowd-sourcing programs can be set up and run by retail investors.
The value of such a program for the retail player depends not only on strike rate, but also on persistence of opportunities. If the identified opportunities are momentary and yield small returns, only institutions with vast resources can profit. For a retail investor, trends discovered by crowd-sourcing must last at least an hour, preferably longer than one session, to be exploitable. Also, returns must be higher.
There is some evidence to suggest that opportunities from crowd-sourcing do persist long enough and have large enough payoffs to be useful for retail players. But to distinguish between strong, persistent sentiments and momentary, weaker sentiments is difficult. Also, as more sophisticated traders get into crowd-sourcing, it is likely that signals and opportunities will get shorter in duration and smaller in magnitude.
A lot of short-term trading is built around judgements calls about a market or specific stocks being either overbought or oversold. The crowd-sourcing possibility adds a new dimension to the whole exercise. It is a fascinating new area for traders and investors.
Institutional players have always held a massive edge in quantitative analysis because they have better equipment and better access to information. The institutional edge certainly persists but the Net and Web 2.0 reduces the gap.
Unfortunately, Web 2.0 hasn't been around long enough to build models that can be back-tested across a full boom-bust investment cycle. Contrarians especially would like to see how sentiment signals respond duringa crash, at the bottom of a bear market and at a bullmarket top.