We had discussed how risk and return are two sides of the same coin (read the series here). Also, we looked at the high price correlation of Reliance vs. the Nifty 50 index. The question yet to be answered is this: given a choice between owning Reliance and owning the NIFTYBEES ETF, what should an investor do?
Lets look at the pros of owning the NIFTYBEES:
- The NIFTYBEES represent the broader market. Investors get a diversified, market-weighted portfolio.
- 8.73% of the NIFTYBEES is, in fact, Reliance. So investors do get a slice of exposure to Reliance.
Now what would be the disadvantages of owning the NIFTYBEES vs. owning Reliance outright?
- If Reliance out-performs the index, then investors only get a small (8.73%) of the increase
- Investors may not want to buy the entire index and might prefer a concentrated portfolio of just resource stocks, of which Reliance is one.
To answer this question, lets turn our heads to two measures: alpha and beta (discussed here.) Reliance has an alpha of -0.0002946529 and a beta of 1.027928. What this shows is that Reliance actually underperformed the index (a –ve alpha) and it more or less tracked the index (a beta close to 1; confirmed by our correlation study.)
We are in the process of rolling out alpha and beta of individual stocks against the Nifty 50 index. Stay tuned!
Our previous discussion of correlation in the NSE looked at a years worth of data for the NIFTY 50 components to see how individual stocks correlated with the index. There are three ways to look at correlation:
- Highly correlated stocks can be substituted with each other. For example, if the price of stock A is highly correlated with the price of stock B (r approaching 1), then investors should be indifferent between owning A or B.
- Correlation can be used to expose relative value. For example, in the above example, if A pays more dividends than B, then owning A is better than owning B.
- Correlation as a trading tool. In the above example, say on a particular day A drops (or rises) more than B, then you can put on a trade betting on mean reversion – that ultimately A & B will start behaving similarly.
For example, lets have a look at RELIANCE over the NIFTY 50 index. I created a series of 10-day correlations (r)
and lastly for 2011:
It looks like RELIANCE is usually highly correlated to the NIFTY 50 and the range is somewhere between 0.7 and 1.0. Lets have a look at the histogram to get a better idea:
You need to ignore the deviations around stock splits and dividend ex-dates (for example, on 26-Nov-2009, RELIANCE issued a 1:1 bonus so the displacement that you see surrounding that date should be ignored) to truly appreciate what’s going here.
The charts show that there are significant number of instances when the correlation breaks down but it always moves back into the range. Looks like betting on convergence seems to be a no-brainer.
Have a trade idea? Let me know!
This is a good follow for the correlation article. Source: BW
Meet the worst nightmare of stock-pickers and the best friend of technical traders: Correlation.
This year has been unkind to stock-pickers as markets have tended to move in unison, with price moves increasingly driven by the ebb and flow of investors’ fears (risk-on/risk-off) over the economic environment. We setup a correlation matrix between the NIFTY50 stocks and the index itself to see how we faired this year.
The vertical axis here is the correlation between individual stocks (dots) with the index.
Now lets do a quick gut-check. The top 3 stocks with the highest correlations with the index were:
RELCAPITAL, ICICIBANK and STER. A quick look at their charts show that it is indeed so. The ones that were closer to zero are: GRASIM, SIEMENS and RANBAXY. They may have been highly volatile, but we are only looking at correlation for now.For stocks that went the other way, we have: HEROMOTOCO, BAJAJ-AUTO and HINDUNILVR.
There’s really no actionable information here, except that instead of tracking close to 30 stocks in the index, you are better off just buying the NIFTYBEES and saving yourself a lot of effort.
For the more inquisitive, the data can be found on google docs.