Download presentation
Presentation is loading. Please wait.
Published byLeon Newman Modified over 9 years ago
1
Asset Allocation – What have you done for me lately? Tom Robinson, PhD, CFA, CFP, CAIA Managing Director, Education Tom.robinson@cfainstitute.org http://www.linkedin.com/in/trrphd
2
Agenda The Investment Management Process Asset Allocation in Practice: How Important Is It? Asset Allocation in Practice: The Financial Crisis
3
Investment Management Process – Potential Steps Decision to Invest Strategic Asset Allocation Tactical Asset Allocation Market Timing Rebalancing Investment or Security Selection
4
Asset Allocation Asset Allocation in Practice – How Important is it?
5
Relative Importance How important is strategic asset allocation versus other parts of the process? Determinants of Portfolio Performance
6
Actual Return 8.3% [Portfolio Weights X Portfolio Returns] Asset Allocation Return 5.2% [Portfolio Weights X Benchmark Returns] Selection Return 9.4% [Benchmark Weights X Portfolio Returns] Benchmark Return 6.4% [Benchmark Weights X Benchmark Returns Classic Brinson Attribution
7
Brinson Hood & Beebower 1986 (FAJ) Variation of ReturnsAverage ReturnsMinimum ReturnMaximum Return Strategic Asset Allocation93.6%10.11%9.47%10.57% Tactical Asset Allocation1.7%-0.66%-2.68%0.25% Investment Selection4.2%-0.36%-2.90%3.60% Interaction & Other0.5%-0.08%-1.17%2.57% 100.0%9.01% 91 Large Corporate Pension Plans 40 quarters of data 1974-1983
8
Ibbotson & Kaplan 2000 (FAJ) How much of the variability of returns across time is explained by strategic asset allocation (policy)? How much of the variation in returns among funds is explained by differences in strategic asset allocation? What portion of return level is explained by strategic asset allocation? There are really three Questions
9
Ibbotson & Kaplan 2000 (FAJ) Variability ExplainedBrinson 1986Brinson 1991Mutual FundsPension Funds Mean93.60%91.50%81.40%88.00% MedianNA 87.60%90.70% Question 1 – Time Series Analysis ” …high r-squared result primarily from funds participation in capital markets in general.” Question 2 Cross Sectional Analysis of 94 balanced mutual funds 1988 – 1998 100% Question 3 Impact of strategic asset allocation on average returns for balanced mutual funds and pension funds (active management did not contribute on average) 40%
10
Xiong, Ibbotson, Idzorek and Chen 2010 (FAJ) Variation in a portfolio’s returns from: –Cash or risk-free asset –The market portfolio (average in this case) –Strategic Asset Allocation Policy –Active Management Tactical Asset Allocation Security Selection Fees
11
Xiong, Ibbotson, Idzorek and Chen 2010 (FAJ) U.S Equity FundsBalanced FundsInternational Funds Market Movement83%88%74% Strategic Asset Allocation18%20%19% Active Management15%10%26% Interaction Effect-16%-18%-19% 100% Sample: 5,628 mutual Funds, up to 11 asset Classes; Time Series Analysis Variability of returns explainded.
12
Xiong, Ibbotson, Idzorek and Chen 2010 (FAJ) Sample: 5,628 mutual Funds, up to 11 asset Classes; Time Series Analysis after removing market returns. Variability of returns explained. U.S Equity FundsBalanced FundsInternational Funds Strategic Asset Allocation48%36%49% Active Management41%39%45% Interaction Effect11%25%6% 100%
13
Xiong, Ibbotson, Idzorek and Chen 2010 (FAJ) Monthly cross-sectional analysis. 120 regressions to explain variation among funds as opposed to over time. Also effectively removes market return as it is the same for all funds. On average strategic asset allocation explained 40% of variability.
14
Ibbotson 2010 (FAJ) Asset allocation provides passive return (beta). Remainder of return is active (alpha). On average alpha sums to zero across all portfolios. Asset allocation, on average, determines 100% of returns before costs. After controlling for interaction effects – at the firm level: –About 75% of return variation comes from market movement –Remaining variation split about evenly between asset allocation and active management.
15
Observations Beta versus Alpha Passive versus Active Average alpha/returns from active management should be zero. Importance of asset allocation
16
Asset Allocation Asset Allocation in Practice: The Financial Crisis
17
Modern Portfolio Theory & The Crisis Efficient Frontier versus Efficient Markets Assumptions –Returns are normally distributed. –Risk can be measured by standard deviation. –Investors are risk averse and rational. –Investments are infinitely divisible and priced in line with risk levels. Limitations –Investors are not always rational and markets are not particularly efficient at pricing – fear, panic and greed also drive markets and as we saw in this crisis markets can seize up. –Asset Allocation Inputs Expected Returns: Returns are not quite normally distributed and extreme events occur much more frequently than normal. Expected Variability (SD): Variability is variable! Expected Correlations: Correlations are not stable, they change over time and are asymmetrical, particularly in periods of market stress like the recent crisis.
18
Black Monday and Black Swans Bogle (2008) 19 October 1987 Still a record daily decline – US Stock Market dropped 22.6% in one day Risk versus uncertainty: risk is a measurable quantity for which probabilities and distributions are known whereas uncertainty is immeasurable.
19
Three Sigma or More Events Black Monday is still a record: Next two largest daily declines were 28 October 1929 at 13.5% and 29 October 1929 at 11.7%. Largest daily gain in DJIA was 15.3% on 15 March 1933. 20,839 DJIA daily returns from 1 October 1928 (when it began using 30 stocks) through 23 September 2011. Mean daily return 0.025%, Standard deviation 1.16%. Three sigma or greater events expected for a normal distribution would be about 0.27% or about 56 events Actual number was 364 or six and a half more times than would be expected for a normal distribution. (203 negative events and 161 positive). There is also a clustering of these events in short time periods. Seven three sigma days occurred from 4 August 2011 to 23 September 2011. 25 such days occurred from 15 September 2008 to 1 December 2008.
20
Large Cap Stock Return Distribution Fat-tailed (Excess kurtosis = 2.95), Negatively Skewed (Skewness = -.91)
21
30 Year US Government Bonds Return Distribution Fat-tailed (Excess kurtosis = 3.22, Positively Skewed (Skewness =.51)
22
22 How rare are black swans, really? “Once-in-a-century crisis” Q: What is the probability of a 50% drawdown in stocks during a 100-year period? Assume: 7.5% annual return 18% standard deviation A: 90% (50% decline almost certain!) (50% decline almost certain!) — Alan Greenspan Source: Zhou, Guofu, and Yingzi Zhu. 2010. “Is the Recent Financial Crisis Really a ‘Once-in-a-Century’ Event?” Financial Analysts Journal, vol. 66, no. 1(January/February). Image source: sxc.hu
23
Asset ClassPeriodPeak-to-Trough Decline U.S. stocks (real total return)1911–192051% U.S. stocks (DJIA, daily)1929–193289 Long U.S. Treasury bonds (real total return)1941–198167 U.S. stocks1973–197449 U.K. stocks (real total return)1972–197474 Gold1980–198562 Oil1980–198671 Japanese stocks1990–200982 U.S. stocks (S&P 500)2000–200249 U.S. stocks (NASDAQ)2000–200278 U.S. stocks (S&P 500)2007–200957 23 Source: Siegel, Laurence B. 2010. “Black Swan or Black Turkey?” Financial Analysts Journal, vol. 66, no. 4 (July/August).
24
Variability of Variability Trailing 252 day standard deviations- DJIA
25
Correlation of Large US Equities with European Equities
26
Correlation of Stocks and Bonds
27
Correlation of Stocks and Real Estate
28
Correlation of Stocks and Commodities
29
Pairwise Correlation of Price Changes 29 Source: R. Sullivan and J. Xiong, Forthcoming 2012, Financial Analysts Journal
30
Asymmetrical Correlations Chua, Kritzman & Page, 1999, JPM and others. Downside correlations are higher than upside correlations. Portfolios should be designed like airplanes – able to withstand turbulence whenever it arises as it is unpredictable. Need to factor the correlation asymmetry into optimization – design portfolios with asset classes with less correlation on the downside and more correlation on the upside. Stress testing portfolios using downside correlations.
31
“Asset Allocation in a Crisis’ Jacobson 2009 SSRN, 2010 JFP and CFAM 2010 –Lessons Learned (mainly aimed at PW) Strategic asset allocation should not be static asset allocation. Awareness of economic and market conditions should inform portfolio allocations. Risk management/allocation goes beyond style and size factors – Other factors such as liquidity matter. “Asset” Allocation not just across asset classes but across risk factors, Beta risk, Alpha risk, liquidity. Still a work in progress to implement. “Insurance like” allocations
32
32 Source: R. Sullivan and J. Xiong, Forthcoming 2012, Financial Analysts Journal Higher Correlations – Crisis Related or systematic? Rise of Index Investing
33
33 Rise of ETFs
34
Observations Markets and models are not perfect. That does not mean we should not diversify our portfolios – portfolios that included high quality government bonds performed well in the 2008 crisis and this year. Research by Ibbotson and others continues on how to best incorporate non-normal distributions into asset allocation. Be prepared for continued high correlations Stress test portfolios
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.