Rather than ranting about Indian corporates and politicians, in this column I will take a dig at my own profession (of brokerage analysts) and highlight the damage that my peers might be causing to investors’ portfolios by using erroneously low assumptions regarding the cost of equity in India.
Step 1: estimating the Equity Risk Premium (ERP) for India
As every business school graduate has been taught, the Capital Asset Pricing Model (CAPM) says that the “Cost of Equity = Risk free rate + (beta x ERP)” where ERP is the extra return that stocks have to offer relative to Government bonds to compensate for the higher risk of investing in stocks.
In the Indian context, using historical data over 1997-2009, my colleagues estimate the Equity Risk Premium (ERP) to be 8.7%. However, academics claim that historical data usually overestimates ERP partly because it does not account for survival bias in the data and partly because it does not take into account the maturing of an economy. Hence India’s ERP going forward is likely to be lower than 8.7%.
In the US, the median ERP used by analysts is 5% for local investments. Given our stage of development, India’s ERP should obviously be higher than the US. The question is “how much higher?”.
Moody’s rating on the Indian government’s long term bonds is “Ba2” which translates into a default spread of 300 bps over US Treasuries. Moreover, since according to Professor Aswath Damodaran (an authority on the subject), equities are 1.5x riskier than bonds (implied by the standard deviation in returns from equity and bonds), India’s ERP from a global investor’s perspective is likely to be around 9.5% i.e. 5%+(300 bps x 1.5).
This, however, does not take into account that an average Indian investor has fewer investment opportunities than a global investor and hence cannot expect the same ERP. Assuming that a local investor has access to only half the investment opportunities of a global investor, his ERP will be 7.25% i.e. 5%+(150 bps x 1.5). Since 80% of India’s market cap is held by local investors, we prefer to use 7.25% as our estimate of ERP.
Step 2: Moving from ERP to the cost of equity (COE)
We calculate beta from first principles (i.e. regressing individual stock returns vs market returns) for each of the BSE 100 stocks. Weighting each of these beta estimates by the market cap of the stocks gives us a beta of 1.1 for the BSE 100. If we then assume that the risk free rate is 7.4% (i.e. the yield on 10 year Government bond), the cost of equity for the BSE 100 turns out to be 15.5% i.e. 7.4% + (1.1 x 7.25%)
How far our estimates from our rivals’?
Based on our discussions with other sell-side analysts and based on a recent poll of Indian analysts done by Pablo Fernandez of IESE Business School in Madrid, we find that most brokerage analysts in India use an ERP of around 6%. This produces a cost of equity of around 11-13%. Our “reverse DCF” exercise on the Sensex also suggests that the market is factoring in a cost of equity of 13.3%.
Both of these COE estimates are 200bps lower than our COE estimate (which is grounded in a sensible ERP estimate of 7.25%). Interestingly, Fernandez’s research also suggests that Indian corporates use ERPs in the range of 6.6%-9%. This produces estimates of cost of equity of around 15-17% i.e. not a million miles from our 15.5% estimate.
Why does this matter?
If investors are systematically underestimating COE for Indian stocks, they will be overvaluing riskier stocks and overallocating capital to these stocks at the expense of lower risk stocks. To understand how easily this could happen, let’s take a world where an investor desires a return of 12% and has a choice of only two stocks to invest in. Stock X’s true COE is 15% but brokerage analysts say that it is 13%. For stock Y the true COE is 11% but brokerage analysts say it is 9%. With the 12% return in mind and using the brokerage analysts’ COEs as a guide, the investor allocates 75% of his portfolio to stock X and the rest to Y (if CAPM works, the weighted average return of such an allocation should be 12%). However, had he known the true COEs the investor’s allocation would have been only 25% in stock A. Using incorrect COEs therefore erroneously pushes the portfolio allocation towards the higher risk stock.
September 09, 2010
Dear Saurabh,
Wouldn't it be incorrect if we try to estimate a COE of whole Indian Market rather than a specific firm. As most of the cost to equity arises from the firm specific risk rather than the market risk, which is assumed to be diversifies under the basic assumption of marginal investor in CAPM model also the ERP will take care of the market risk for India.
By calculating an CoE for market as a hole is like building a biased opinion on how the analysts should assign cost to all the companies listed or unlisted.
You have taken country default risk to be 50% arguing less availability of invest able instruments but in fact unavailability of investment avenues would lead to less diversification and higher degree of risk.