CIS presentation for stock analysis for 20 years

Independent study on stock data of India

1. Fusion of Low Volatility & Momentum Investment Strategies Can we enjoy best of both worlds

2. Dr. Mayank Joshipura Presented by:- Vishesh Sahni Sap ID: 80101190610 Roll No.: G052 MBA-FT (core) CIS research under the guidance of

3. Agenda Factor Investing & Low Risk Anomaly 01 Objective of Study & Methodology 02 Results 03 Final Remarks & Scope for further studies 04

4. Factor Investing What drives stocks returns – CAPM is the oldest and most popular explanation It was believed that only two main drivers explain returns: Systematic risk and idiosyncratic risk Statistical Arbitrage Pricing Theory - there are other factors that can explain expected returns This led to start of factor investing. A factor is any characteristics of a security that helps in explaining the risk and return it provides Three main categories of factors : Macroeconomic Fundamental Most popular of the lot is factors based upon fundamental indicators. These include Value, Growth, Size, Momentum. Their main popularity came from establishment of the Fama French model. Various studies have proven that it is difficult for active fund managers to generate alpha and most of the time they fail to outperform the capitalization weighted benchmark. Studies also showed that when active fund managers align their portfolio towards well known factors like Size, Momentum etc, they tend to outperform the benchmark. ( Source : MSCI )

5. MSCI Fundamental Factors Definitions Fundamental Factors Factors What is it? Captured by? Value Excess returns of stocks that have low price related to intrinsic value Book to price, earnings to price, book value, cashflow Momentum Excess returns to stocks with stronger past performance Relative returns, historical alphas Low Volatility Excess returns to stocks with lower than average volatility/beta Standard deviations/downside Standard deviation Quality Excess returns to stocks that have low debt, stable earnings growth and other quality metrices ROE, dividend growth stability, financial leverage, accruals, cash flows.

6. Low Risk Anomaly Classical thought CAPM: Risk and Return have a linear relationship. This has been contested with development of Fama French Factor models which showed no relationship between beta and returns after controlling for size 87% 95% Developments and Studies MSCI launched minimum variance version and S&P launched low volatility index by 2010. NSE in 2012 launched a low volatility investment index which outperformed benchmark Nifty in absolute and risk adjusted basis. Only 1 ETF tracking this index is available in India. Clarke, Silva, and Thorley (2006), Blitz and Vliet (2007) and Zhang (2006) did major contributions to low risk anomaly literature. While Clarke uses a minimum variance portfolio, Blitz use ranking based portfolio with 3-5 years lookback What is it? It has been observed that low volatile stocks have higher returns than high volatility/risky stocks. Analytic Investors and Robeco launched funds between 2004-06 to benefit from this. Destruction of wealth in 2008 led to interest in low risk investment strategies. Behavioral Bias : Investor preference for lottery like payoff lead to riskier securities becoming overvalued and hence lower payoff. Market friction & constraints: Leverage constraints leads to investors moving towards high riskier securities rather than using leverage in low risk category. Performance Mandate: Investors with obligation to outperform benchmark show preference towards negative alpha and high beta stocks. Low Risk Anomaly persistence Low Risk Anomaly

7. The objective of this study is not to find out whether low risk anomaly is present in India or not as its existence has already been seen in markets across the globe. The aim of the study is to bring best of both worlds by forming high momentum and low volatility portfolios and to check if the returns are better than market-cap or other forms of portfolios on absolute and risk adjusted basis. Momentum effect has generally offered investors with highest Sharpe ratios compared to any other portfolio but momentum only portfolios have fallen strongly during worst times. This pushes away risk averse investors. Thus, by aiming to combine low volatility with high momentum we try to bring the best of both worlds. Objective of Study

8. Methodology For each quintile portfolio, the excess returns of the following month are calculated & exercise is repeated Time Series Formation We collect data from IIM A online library for Rf, size value & momentum factors Additional Data We remove stocks with 0 P/E as they aren’t trade enough and those with market cap less than 100 Cr because they might have liquidity problems Filtering At the end of every month, we sort data based upon momentum to create equal weighted quintile portfolios, each of which are further sorted on volatility to create equal weighted quintile portfolio. Thus we have total of 25 portfolios with combinations of momentum and volatility Portfolio Creation We keep only those stocks which have at least 12 months of stock returns data Data Availability Monthly stock data for all the listed companies collected from Prowess Database for period of 02/2000 to 12/2018 Data Collection Step 2 Step 5 Step 3 Step 4 Step 6 Step 1 Step 7 For the resultant time series, we run 1,3 & 4 factor regression with factor data as independent variables and portfolio excess returns as dependent variable to come up with values of Alpha and t-stat for each quintile portfolio. Results

9. Volatility Sorted Universe Return Analysis P1 (Low Volatility) P2 P3 P4 P5 (High Volatility) Simple Return 20.33% 25.38% 25.03% 26.09% 24.64% CAGR 19.74% 23.72% 22.00% 21.57% 18.02% Standard Dev. 20.97% 28.05% 31.88% 36.18% 40.37% Sharpe Ratio 0.97 0.90 0.79 0.72 0.61 Ex-Post Beta 0.65 0.90 1.02 1.15 1.28 Alpha 1 factor 4.48% (2.94) 3.61% (3.06) 0.25% (0.20) -1.93% (-1.19) -6.42% (-2.94) Alpha 3 factor 4.51% (2.90) 3.65% (3.03) -0.04% (-0.03) -2.25% (-1.36) -5.87% (-2.65) Alpha 4 factor 4.77% (3.01) 3.91% (3.20) 0.53% (0.42) -2.42% (-1.44) -6.80% (-3.05) Max Drawdown -55.40% -69.07% -74.63% -78.05% -80.74% Tracing Error 12.18% 5.47% 4.68% 7.70% 11.89% Skewness -0.19 0.01 0.29 0.38 0.50 Table 1: Results for quintile portfolio sorted based upon historical volatility 30% 15% 25% 10% 20% 30% 25% 25% 50% Risk Return relationship is initially positive but flatter than expected For P5, alpha is negative which gets reflected in drawdown as well. CAGR increases from P1 to P2 but then starts falling. Sharpe ratio is lowest for P5 showing higher volatility leading to lower risk weighted returns. All the alphas are significant expect for P3 and P4 Analysis

10. Momentum Sorted Universe Return Analysis P1 (Low Momentum) P2 P3 P4 P5 (High Momeentum ) Simple Return 13.30% 22.22% 25.34% 31.54% 33.43% CAGR 6.93% 18.22% 22.77% 30.88% 32.72% Standard Dev. 37.30% 33.31% 30.91% 29.28% 30.26% Sharpe Ratio 0.36 0.67 0.82 1.08 1.11 Ex-Post Beta 1.16 1.05 0.98 0.91 0.90 Alpha 1 factor -15.83% (-6.36) -4.17% (-2.51) 0.69% (0.56) 8.60% (4.60) 10.70% (3.65) Alpha 3 factor -15.98% (-6.29) -4.07% (-2.39) 0.65% (0.51) 8.53% (4.45) 10.88% (3.62) Alpha 4 factor -16.66% (-6.47) -4.64% (-2.70) 0.83% (-0.64) 9.03% (4.65) 11.45% (3.75) Max Drawdown -79.00% -71.90% -69.83% -68.94% -72.48% Tracing Error 0.11 0.06 0.05 0.08 0.11 Skewness 0.81 0.99 0.76 -0.14 -0.73 Table 2: Results for quintile portfolio sorted based upon historical momentum 30% 15% 25% 10% 20% 30% 25% 25% 50% Returns increase from P1 to P5 as we move from low to high momentum. Returns are maximum for P5 with CAGR of 32.72% while Standard Dev is maximum for P1 with 37.30% . No linear risk return relationship is visible Momentum effect is so strong that it leads to higher returns even while taking lower risk. Alphas for P1 and P2 are negative and significant while for P4 and P5 are positive and significant. Alpha is insignificant for P3. Analysis

11. 1 st Quintile Momentum Portfolio Return Analysis P1 (Low Volatility) P2 P3 P4 P5 (High Volatility) Simple Return 13.22% 15.66% 16.35% 15.20% 5.87% CAGR 10.39% 10.24% 8.87% 6.90% -3.45% Standard Dev. 26.75% 35.39% 40.04% 41.60% 45.26% Sharpe Ratio 0.49 0.44 0.41 0.37 0.13 Ex-Post Beta 0.80 1.08 1.21 1.27 1.35 Alpha 1 factor -7.01% (-2.82) -11.54% (-4.02) -14.22% (-4.12) -16.78% (-4.99) -28.17% (-6.45) Alpha 3 factor -7.47% (-2.94) -11.27% (-3.83) -15.47% (-4.42) -16.48% (-4.81) -27.90% (-6.27) Alpha 4 factor -8.28% (-3.22) -11.48% (-3.82) -16.86% (-4.79) -18.00% (-5.32) -27.32% (-6.03) Max Drawdown -66.61% -79.29% -80.64% -81.69% -89.98% Tracing Error 0.11 0.11 0.15 0.15 0.20 Skewness 0.41 0.76 0.81 0.88 0.74 Table 3: Results for quintile portfolio sorted based upon 1 st quintile momentum 30% 15% 25% 10% 20% 30% 25% 25% 50% Even if we pick low momentum bucket but we pick right stocks with low volatility, we would still have positive returns as shown by P1. P1 has the highest Sharpe ratio at 0.49 and CAGR of 10.39% while SD of 26.75% Low volatility is helping P1 to give positive returns even at low momentum The other extreme is P5 which is the worst selection out there. At negative CAGR and Sharpe ratio of 0.13 , such portfolio would only lead to wealth destruction. We would consider P5 as a speculative portfolio because no risk averse educated investor would choose such a combination. Analysis

12. 5 th Quintile Momentum Portfolio Return Analysis P1 (Low Volatility) P2 P3 P4 P5 (High Volatility) Simple Return 26.45% 35.11% 31.15% 35.44% 34.04% CAGR 26.35% 35.81% 29.21% 33.44% 29.84% Standard Dev. 23.34% 28.80% 32.33% 34.78% 37.44% Sharpe Ratio 1.13 1.22 0.96 1.02 0.91 Ex-Post Beta 0.62 0.82 0.94 1.01 1.07 Alpha 1 factor 10.84% (3.12) 14.59% (4.05) 7.37% (2.08) 9.97% (2.56) 7.09% (1.57) Alpha 3 factor 11.79% (3.33) 14.34% (3.92) 7.24% (2.00) 10.71% (2.69) 6.97% (1.51) Alpha 4 factor 13.11% (3.67) 15.66% (4.24) 8.27% (2.25) 10.31% (2.54) 6.37% (1.35) Max Drawdown -60.41% -67.91% -73.95% -77.73% -79.08% Tracing Error 0.18 0.15 0.13 0.15 0.17 Skewness (0.66) (0.46) (0.48) (0.56) (0.49) Table 4: Results for quintile portfolio sorted based upon 5 th quintile momentum 30% 15% 25% 10% 20% 30% 25% 25% 50% Momentum benefit is visible as returns are significantly higher than 1 st Quintile portfolio Momentum effect is so strong that even if we pick the riskiest one (P5) within high momentum bucket, it even can lift returns for that also. Visible by Sharpe ratio of 0.91. If we compare Sharpe ratio of P5 with the pure volatility sorted universe, this is performing better than almost all volatility sorted portfolios. The best combination is high momentum and low risk which is P1 and P2. With significant and large positive alpha this is the best portfolio choice. We would consider these conservative portfolios. On a risk adjusted basis, Sharpe ratio of P1 and P2 at 1.13 and 1.22 is higher than that of P5 in Momentum sorted universe. Hence, this is the best choice to get highest risk adjusted returns. Analysis

13. Other Momentum Sorted Portfolios 30% 15% 25% 10% 20% 30% 25% 25% 50% In 4 th Quintile portfolio, we start seeing a large rise in returns overall when compared to lower momentum buckets. Momentum effect starts to overpower the volatility effect, returns increase from P1 to P4 but fall for P5 which is the most risky portfolio. Sharpe ratios are significantly higher for P1 to P3 Alpha for P4 and P5 are not significant Analysis In 2 nd Quintile portfolio, the volatility effect is so strong that even though momentum has increased than 1 st Quintile but the benefits are not enough to compensate for higher volatility A positive but flattish risk return relationship was seen All the factor alphas were insignificant for P1,P2 and P3 but were significant and negative for P4 and P5 which proves our point that higher volatility ( Higher risk ) doesn’t lead to higher returns. In 3 rd Quintile portfolio, there is no linear risk return relationship. Even the Sharpe ratios don’t follow any linear trend from P1 to P5. Analysis

14. Final Remarks & Further Scope This is the most desirable combination. Even in high momentum, high volatility, the momentum effect is so strong that it even lifts the most risky stocks when selected with high momentum attribute. High Momentum & Low Volatility Portfolio This is the worst portfolio with negative CAGR. Momentum effect is so strong that it dominates even if low extreme negative momentum is taken with low volatility it would not allow stock to give substantial return. ideally within low volatility we should avoid stocks which have negative momentum. Low Momentum & High Volatility Portfolio On a risk adjusted basis, high momentum and low volatility portfolio gives superior returns than purely high momentum or purely low volatility portfolios. Thus, we can enjoy the best of both world using this strategy. Superior returns We can combine factors like profitability, size, growth or value with volatility to come up with other combinations of portfolio that increases the benefits of factor investing. We can also use market cap weighted portfolios instead of equally weighted portfolios. Scope for further studies

15. THANK YOU This would not be possible without the guidance of Dr. Mayank Joshipura