Empirical testing of the CAPM on the JSE Mike Ward, Chris Muller Gordon Institute of Business Science University of Pretoria NERSA Conference August 2012
An economic return on the RAB? Regulatory Asset Base Shareholder Capital Debt Capital The cost of equity “The CAPM” Re = Rf + β.MRP The cost of debt
The Capital Asset Pricing Model Beta = 1.0 Return Rf = 7% MarketRiskPremium = 5% High beta shares are more risky, so give better returns 0.8 Rf = 11% Risk (beta)
Betting Against Beta, Andrea Frazzini and Lasse H. Pedersen, Oct 2011 Prior Research Data: All US Shares
Betting Against Beta, Andrea Frazzini and Lasse H. Pedersen, Oct 2011 Data: 18 International Markets
Fama and French (2004) estimated betas for every share on the NYSE, AMEX and NASDAQ from 1923 – 2003 using 2-5 years prior data and compared with their return over the next 12 months:
The largest 600 US shares over the period 1963 – 2006 placed into 10 portfolios ito beta.
Prior research on the JSE Strugnell, Gilbert & Kruger (2011) IAJ – “Beta has no predictive power for returns on the JSE” – Data from 1994 – 2007 – Included too many small shares van Rensburg & Robertson (2003) IAJ – “If anything, beta is inversely related to returns!” – Data from 1990 – 2000 – Included too many small shares
Rational for research The CAPM is a pillar of financial theory: – taught on all finance courses – found in all finance text books – used regularly in the financial services industry – Markowitz, Miller & Sharpe shared a Nobel prize We have 25 years of JSE data – 1985 to 2011 We can improve on the methodology
Methodology Select the largest 160 companies in Dec 1984 Estimate betas using prior years return data – OLS beta 60 monthly data points – Dimson Multiple regression (+1,0,-1,-2,-3,-4) Rank betas Construct 5 equal weighted portfolios of 32 shares Measure portfolio return over the next 3 months Repeat for next quarter
99% of JSE’s market capitalisation
Presentation of findings We track the daily value of each portfolio (quintile) We re-balance each portfolio quarterly – We retain the value of the portfolio – Equally weight – We ignore transaction costs We graph the results We benchmark against the ALSI total return index We plot a price relative versus the J203
Results
OLS Betas - monthly
OLS Betas - weekly
Dimson Betas - monthly
Dimson Betas - weekly
Volatility - Daily
Summary of Results Annualised returns for equal weighted portfolio quintiles over the period 31Dec Dec2011 Risk Measure Number of Obs ALSI Index R203 Highest Beta Quintile Quintile 2 Quintile 3 Quintile 4 Lowest Beta Quintile OLS Monthly Beta6015.7%7.7%12.1%18.1%21.6%20.4% OLS Weekly Beta %4.6%15.4%20.4%19.9%19.3% Dimson Monthly Beta6015.7%7.9%16.3%19.3%19.0%17.4% Dimson Weekly Beta %5.7%15.3%19.0%19.5%20.8% Volatility Daily6015.7%9.7%13.5%17.7%20.8%18.2% Average annualised Return15.7%7.1%14.5%18.9%20.2%19.2%
Conclusion: High risk (beta) = Low return Ben Graham once argued that: "Beta is a more or less useful measure of past price fluctuations of common stocks. What bothers me is that authorities now equate the beta idea with the concept of risk.
Questions… For those interested: The full paper will be published in the forthcoming: – Investment Analyst Journal –