Indian Association of Alternative Investment Funds (IAAIF)
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Portfolio Construction with Alternative Investments Rohan Misra, CFA, FRM Partner & CEO Transparency. Safety. Performance.
AGENDA PART 1: PORTFOLIO CREATION FINDING THE EQUILIBRIUM 1. ASSET ALLOCATION 2. ACTIVE VS PASSIVE BALANCE 3. MANAGER/FUND SELECTION PART 2: PORTFOLIO PERFORMANCE EVALUATION AND REBALANCING 3
Finding the equilibrium… …By mixing a number of poorly correlated assets, and - • Maximize target return for a given level of risk • Minimize risk for a targeted level of return 5
2 ways to construct portfolios Bottom Up Top down • Used by private Asset THE EQUILIBRIUM FINDING • Favoured by investors Allocation professional investors • Adhoc – objectives and • Begins by exploring Active Vs risk not factored investment risk Passive Mix • Susceptible to • Creates a “buy high sell framework to low” behaviour Manager/Fund decide investments Selection based on investor’s objectives 6
Setting objectives Return ILLUSTRATIVE EXAMPLE Return Target: Typical Objectives: • Maximize return for5%aplus inflation given level of;risk after fees • Minimize risk for aRisk Budget targeted andofRisk level Definition: return Risk Time Max loss 20% Budget Time Horizon : 5 years Examples of other considerations 1. Interim/Terminal Goals: financing a second home purchase 2. Constraints: dedicated assets (residential home), asset class restrictions , short selling restrictions 7
PART 1: PORTFOLIO FINDING CREATION THE EQUILIBRIUM 1. ASSET ALLOCATION 2. ACTIVE VS PASSIVE BALANCE 3. MANAGER/FUND SELECTION 8
What is an asset allocation (AA) Portfolio strategy that involves setting target allocations for various asset classes and attempts to balance risk versus reward, according to the investor’s risk budget, goals and time horizon Strategic Asset Allocation (SAA) Driven by the long-term investment objectives of the investor, with a typical time frame of > 1Y Tactical Asset Allocation (TAA) Represents short-term tilts away from the SAA that are driven by visible opportunities and risks 9
Why begin with AA? Decomposition of Time-Series Total Return Variations 150 100 R square % 50 0 -50 BHB Equity Funds BHB Balanced HEI & IK Equity HEI & IK Balanced Funds Funds Funds Active Management Asset Allocation Policy Market Movement Interaction Effect BHB: Brinson, Hood, Beebower, Determinants of Portfolio performance, 1986 IK: Ibbotson & Kaplan, Does AA explain 40,90 or 100% of performance?, 2000 HEI: Henzel, Ezra, Ilkiw, The importance of the AA decision, 1991 10
Choice of asset classes and their mix is key • For portfolios with market exposures, e.g. long only portfolios, market movement and asset allocation policy mainly drive return variability • Market movement is a function of the chosen asset classes • Asset allocation policy defines how we have mixed the chosen asset classes • Can we improve an asset allocation by including alternatives like Hedge Funds, PE, Real Estate and Commodities?? 11
Asset class choices Asset Class Proxy Index Currency Freq. MSCI All Country World Net Total Equities USD Monthly Return Index Bloomberg Barclays US Govt Total Bonds USD Monthly Return Index Unhedged USD Commodities Bloomberg CMCI Total Return Index USD Monthly HFRI Fund Of Hedge Funds Hedge Funds USD Monthly Composite Index FTSE EPRA NARIET Developed Total Real Estate USD Monthly Return Index Cambridge Associates US Private Private Equity USD Quarterly Equity Index All indices are assumed for illustration purposes only, All data starting Jan-2000 12
Summary statistics Hedge Real Private Equity Bonds Com Funds Estate Equity Ann. Return 3.5% 4.8% 6.4% 3.2% 9.5% 9.7% Ann. Volatility 16.0% 4.2% 16.2% 5.1% 19.1% 10.4% Max DrawDown 54.9% 4.6% 57.1% 22.2% 67.2% 25.2% Return/ Volatility 0.22 1.16 0.40 0.64 0.50 1.01 Equity returns biased downwards as sample begins in Tech bubble, 13 Source: B&B Analytics
Risk/Return Profiles 12.0% PE 10.0% RE 8.0% 6.0% COM Bonds 4.0% EQ 2.0% HF 0.0% 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% Note! : Profiles may be potentially biased due to the chosen sample since 2000 Source: B&B Analytics 14
In-sample correlations Hedge Real Private Equity Bonds Com Funds Estate Equity Equity 1.00 Bonds -0.28 1.00 Com 0.56 -0.17 1.00 Hedge Funds 0.67 -0.13 0.58 1.00 Real Estate 0.80 -0.03 0.49 0.58 1.00 Private Equity 0.52 -0.29 0.35 0.41 0.40 1.00 Average 0.45 -0.18 0.36 0.42 0.45 0.28 Correlations Source: B&B Analytics 15
Impact of introducing Alts in a 50 Eq/50 bonds portfolio 45 Eq; 45 Eq; 45 Eq; 45 Eq; 50 Eq; 45 Bonds; 45 Bonds; 45 Bonds; 45 Bonds; 50 Bonds 10 Com 10 HF 10 RE 10 PE Ann Return 3.9% 4.1% 3.8% 4.3% 4.4% Volatility 7.7% 7.9% 7.3% 8.6% 7.2% Ann. Return/Vol 0.50 0.51 0.52 0.50 0.61 Max 30% 31% 29% 35% 29% DrawDown Returns adjusted to the vol level of 50 Eq / 50 Bonds Portoflio Ann. Return 3.9% 4.0% 4.0% 3.9% 4.7% Source: B&B Analytics, portfolios rebalanced monthly 16
Data biases and other vagaries common to alternatives 17
1. Survivorship bias • Generally accepted notion: Indices, in particular hedge funds ones, are distorted because ‘closed’ funds no longer contribute – and remaining funds overstate the average • Some nuanced studies look at reason for closure and find the index understates actual performance 1. ‘Closures’ due to negative performance 2. ‘Closures’ to new investors due to strong performance • Jury is still out there! • Partial remedy – Use a Hedge Fund FoF Index Returns 1Y 2Y 3Y HFR Global Asset Wtd Composite 5.38% 8.31% 23.68% HFR Fund of Funds Composite 4.16% 4.94% 17.25% Source: B&B Analytics, HFRI 18
2. Smoothed & stale prices • Esp. Relevant in the context of private equity and real estate • Appraiser lacks confidence in the new evidence regarding valuation - instead attaches too much weight on the most recent empirical evidence • Reported valuation lags true market valuation • Positive serial correlation is introduced into returns REDUCED VARIABILITY IN RETURNS ACROSS TIME à UNDERSTATES RISK NEEDS TO BE CORRECTED! 19
Correcting for smoothed prices • Vast body of academic work PE Beta P-Value Significant exists Intercept 1.66 0.01 Yes Lag1 0.32 0.00 Yes • FGW 1993 is a one widely Lag2 0.17 0.09 No adopted approach Lag3 -0.02 0.87 No Lag4 0.02 0.80 No • Removes serial autocorrelation Beta P-Value Significant to unsmooth the time series Intercept 2.04 0.00 Yes • Time series is regressed against Lag1 0.38 0.00 Yes lagged values to identify r(t) = (r*(t) – 0.38r*(t-1))/w statistically significant variables. US Private Equity* Volatility • Unsmoothed series is obtained Smoothed Series 9.4% Unsmoothed Series 15.7% by removing these variables. Return 9.7% Return/New Vol 0.62 *Source: Cambridge Associate, FGW : Fisher Geltner Webb 20
3. Non-normality and tail risk • Most models assume that returns follow a “normal distribution” – Chance of move > 3 s.d. < 1/300 – Skewness = 0 ; return symmetry – Kurtosis = 3 • In reality most asset classes exhibit negative skew and excess kurtosis … i.e left tail risk 21
This can be seen in our data Hedge Real Private Equity Bonds Com Funds Estate Equity Skewness -0.89 -0.20 -1.08 -1.15 -1.54 -0.57 Excess Kurtosis 2.47 1.36 4.07 4.26 7.15 1.94 60 Real Estate Monthly Return Distribution 40 20 0 -30% -28% -25% -23% -20% -18% -15% -13% -10% -8% -5% -3% 0% 3% 5% 8% 10% 13% 15% 18% 20% 23% Directly incorporating volatility into models will underestimate risk and lead to incorrect allocations Source: B&B Analytics 22
Adjusting for non-normality Two approaches: • Adjust the risk (std. deviation) of each asset class or investment to capture the higher moments of skewness and kurtosis before determining the optimal portfolio weights • Directly adjust the portfolio risk measure in the asset weighting process (i.e the optimization process) to incorporate the skewness and kurtosis of the portfolio for a given combination of weights • Second approach is convenient • Both approaches are appropriate only if the historical distribution appropriately captures higher moments Source: B&B Analytics 23
Portfolio Math 24
Recap: essence of portfolio construction • Maximize target return for a given level of risk • Minimize risk for a targeted level of return How to measure portfolio return and what is portfolio risk? 25
Portfolio Return Weighted average of the expected returns of portfolio components Example: 2 asset class portfolio 26
Variance as portfolio risk • Not a simple weighted average of individual component risks. • Requires an estimate of expected covariances between assets – which first requires an estimate of the volatilities of all assets and the correlations between them or Example: 2 asset class portfolio (wi)’COV(wi) Portfolio Volatility: √0.0504 = 22.4% 27
VaR as a portfolio risk measure • Investor’s don’t think in volatility terms! • Focus is on downside risk • Value at risk (VaR) focuses on the left tail of the return distribution • Interpretation: 95% chance that portfolio loss will not exceed X% over a given time horizon; 95% represents confidence 28
Parametric VaR • Assumes normal distribution – requires only mean and standard deviation of returns to calculate VaR • VaR (confidence) = Mean - Std. Dev * z • ‘z’ is the normal z-score corresponding to the confidence level (e.g. 1.65 for 95%, 1.96 for 99%) • Simple but practically limited – does not include negative skew and fat tails! 29
Modified Parametric VaR • Formulaic adjustment to parametric VaR for the empirically observed skew and kurtosis of the portfolio return distribution • Somewhat better at capturing non-normality by transforming the normal ‘z’ to a modified ‘Z’ incorporating higher moments 30
Drawdown • Max DD measures peak to trough loss – very conservative • Difficult to implement in practice and needs to be simplified to an either inception to date drawdowns or rolling drawdowns 31
Optimal Portfolio Mix Traditional Markowitz Approaches based on Modern Portfolio Theory 32
Optimization Setup • Maximize Expected Portfolio Return subject to an absolute constraint on risk 0% and < 35% (the latter to ensure diversification) 33
Expected Return adjustments 11.0% 10.0% PE 9.0% 8.0% RE 7.0% 6.0% 5.0% EQ COM 4.0% Bonds HF 3.0% 2.0% 1.0% 0.0% 0.0% 5.0% 10.0% 15.0% 20.0% 25.0% Adjustments done largely to reflect current thinking around capital market assumptions Source: B&B Analytics, for illustrative purposes only 34
Capital constraints (0,35%) • Allocation to bonds and hedge funds increases as more conservative approaches (VaR and mVaR) are used 100% 80% 35.0% 35.0% 35.0% 60% 19.5% 30.2% 22.6% 40% 35.0% 35.0% 20% 34.8% 10.5% 7.4% 0% Mean Variance Mean VaR Mean mVaR (Parametric) (Parametric) (Parametric) Equity Bonds Commodities Hedge Funds Real Estate Private Equity 35
Portfolio Metrics Mean Variance Mean VaR Mean mVaR (Parametric) (Parametric) (Parametric) Annualized Return 6.8% 6.1% 6.0% Annualized Risk 10.0% 10.0% 10.0% Return/Risk 0.68 0.61 0.60 Drawdown 30.3% 20.3% 19.2% 36
Unconstrained efficient frontiers 12.0% 10.0% 8.0% 6.0% 4.0% Return Vs Volatility (Parametric) Return Vs VaR (Parametric) Return Vs mVaR (Parametric) 37
Optimal Portfolio Mix Other Approaches 38
Practical issues with traditional approaches Estimating many inputs for N asset classes • N expected returns • N expected volatilities • N(N-1)/2 expected correlations Capital Allocation Risk Allocation Vs Bonds Equities Bonds Equities 39
Practical issues with traditional approaches Sensitivity of weights to changing equity return assumptions 100% 18.3% Unconstrained Weights 28.3% 80% 39.4% 51.1% 67.2% 67.2% 67.2% 67.2% 63.0% 31.5% 60% 30.5% 30.1% 40% 30.6% 32.0% 41.2% 50.3% 20% 32.8% 32.8% 32.8% 32.8% 30.5% 18.3% 0% 5.0% 7.00% 7.50% 8.00% 8.50% 9.00% 9.50% 10.00% 10.50% 11.00% Equity Return Assumption Equity Bonds Commodities Hedge Funds Real Estate Private Equity 40
Risk Parity • Lets throw some darts! • Risk driven approaches – expected returns not required to be estimated • Based on premise that a range of outcomes (risk) is easier to estimate than the outcome (return) • Optimize capital weights so that “risk contributions” of all asset classes are equal • Risk Contribution (RC) of asset i = W(i) X Std. Dev(p) X Beta (i,p) • If an asset’s weight = 20%, its beta with portfolio = 2 , then assuming portolio vol = 10%, the RC = 4% => or 40% of risk comes from the asset 41
Risk Parity Asset Allocation Risk Contribution 7.2% 5.8% 4.7% 20.0% 20.0% Equity Bonds 18.0% Commodities 20.0% 20.0% Hedge Funds 57.5% Real Estate 6.8% 20.0% 20.0% Private Equity Portfolio Evolution 300 Risk Parity Annualized Return 4.1% 200 Annualized Risk 4.1% 100 Return/Risk 0.99 2000 2003 2006 2009 2012 2015 Risk Parity Drawdown 12.8% Criticism: Too much weight to fixed income, requires leverage to scale to traditional 42 portfolio risk budgets
Risk Budgeting • Optimize capital weights so that ”risk contribution” of each asset class falls within the max risk contribution allocated to it • This involves setting risk contribution budgets e.g. RC(1) = X , RC(2) = Y … • Sum [ RC(i) ] = Portfolio Standard Deviation • Robust alternative to using expected returns -> Increase risk budgets if view on asset class is positive , decrease risk budget if the view is negative 43
PART 1: PORTFOLIO FINDING CREATION THE EQUILIBRIUM 1. ASSET ALLOCATION 2. ACTIVE VS PASSIVE BALANCE 3. MANAGER/FUND SELECTION 44
Are alternative betas investable? Asset Class Developed Markets India Thousands of ETFs available by Few ETFs, but many MFs to Equities (EQ) region, market, sector and style choose from Fixed Income Large number of ETFs available by ETFs practically non-existent; (FI) FX, Duration, Rating, Risk country but several MFs to choose from Commodities Many ETF and ETN options Few, largely limited to Gold (COM) on most commodities 1st REIT expected in 2017; RE Real Estate (RE) REIT & CEF/FoF options available PE Funds existing Hedge Funds Several HF Index replication & FoF PMS, AIF funds, Largely single (HF) available managers Private Equity Diversified FoF options available Largely single managers (PE) 45
Unfortunately not • Traditional betas (EQ,FI,COM) cheaply available • Alt managers (HF,PE,RE) mainly seek to deliver alpha • Alt betas not easily available (at least for HF, PE, RE) • High dispersion of Alt returns makes it difficult to replicate benchmarks, FoF investment route solves this problem only partially Solution: Treat Alts (esp HF,PE,RE) as a part of an active management mandate 46
Core & Satellite Approach Satellite • Core: long-term, low-cost investments including ETFs, MFs etc seeking market returns Core (EQ and FI and possibly COM betas) • Satellite: actively managed alpha producing investments seeking to deliver absolute return (HF,PE,RE) Asset Passive Active • A best of both (active & passive) worlds Class (Core) (Satellite) approach EQ ✔ ✔ • Optimize costs (inexpensive core) FI ✔ ✔ • Potential to outperform the asset COM ✔ ✔ allocation RE ✔ • Diversify risk through greater number of HF ✔ holdings PE ✔ 47
But this should be consistent with Asset Allocation 1. E[RC(core + satellite)] = E[R(asset class in SAA)] Where E[R] is expected return and E[RC] is expected risk contribution. Nice in theory, difficult to achieve in practice! 48
PART 1: PORTFOLIO FINDING CREATION THE EQUILIBRIUM 1. ASSET ALLOCATION 2. ACTIVE VS PASSIVE BALANCE 3. INVESTMENT/MANAGER SELECTION 49
Diversification within asset class How many investments should we make within HF, within PE..? Alts typically require high investment sizes Traditional diversification Risk methods are not practical for all investors 1 2 3 6 10 20 25 … Manager selection # investments becomes KEY • Diversify via a fund of fund structure – low investment size, but higher fee trade-off – an outsourcing decision • Cultivate superior manager selection capabilities, i.e. identifying truly uncorrelated alpha generating streams 50
The Three Ps People Process Performance Can a manager be What sets the Is there actual skill? Is the manager a trusted? manager apart? natural fit? • Track record • Investment • Sustained • Correlation • Background strategy outperformance • Quantitative • Education • Risks • Benign & analysis and • Restrictions adverse misfit risk • Philosophy • Rigor & environments • Will the strategy • Attitude Repeatability • Adaptability continue to • Peer group work at scale? analysis 51
Managing misfit risk • Start by identifying investments HF1 HF2… PE1 PE2… RE1 RE2… EQ FI COM within each Alt asset class that are HF1 expected to beat AA hurdle rate HF2… (HF1,HF2,PE1,PE2 etc..) – keep an eye PE1 on correlations, esp. tail ones PE2… • Pool selected Alt investments with RE1 other investments and estimate an RE2… extended covariance matrix EQ FI • Optimize weights to minimize tracking COM error (wp – wsaa)’COV(wp – wsaa) subject to constraints Simple and straightforward approach, • Sum of weights to investments within especially when number of investments asset class ≅ SAA wt. to asset class are not too large • Portfolio risk = SAA risk Note: This covar matrix assumes that there is only one passive investment within 52 EQ,FI,COM that perfectly replicates the index used in the SAA
Identifying sources of return • A simple approach to analyze the sources of excess return for a fund relative to a comparable style benchmark • Define a peer group of hedge funds with similar style and size • Calculate average peer return over time – benchmark • Calculate fund β w.r.t to benchmark over time via rolling regressions. Then – Style Returns = β* Rb – Timing alpha = Rb *(β - 1) – Selection alpha = Rp - β* Rb – Timing alpha + Selection alpha = Excess Return = Rp - Rb • Analyze stability and superiority of timing and selection returns Implicit assumption is that an appropriate peer group exists! 53
Peer group style attributions 54
PART 2: PERFORMANCE FINDING THE EQUILIBRIUM EVALUATION AND REBALANCING 55
What makes a valid benchmark? • Investable: ability to buy and hold the benchmark • Unambiguous: names and weights of holdings are clearly stated • Measurable: transparent w.r.t calculation • Independent: not be designed by manager – removes conflict • Relevant: should reflect the investment strategy 56
How does this look in the case of Alt indices • Investable: NOT REALLY, EXCEPT COMMODITIES • Unambiguous: YES • Measurable: YES • Independent: YES • Relevant: MAYBE 57
Peer group analysis 58
Peer groups a good benchmark? Convenient - shows the edge, or lack thereof! • Investable: NO • Unambiguous: NO • Measurable: YES • Independent: NO • Relevant: MAYBE Issues: classification bias, survivorship bias, prone to snapshot assessments – end point bias, can be gamed… 59
End point bias The same fund ranked in the top quartile when looking at trailing periods 60
Only one benchmark please! Peer Asset Class Cash/Hurdle Index Group Equities (EQ) ✔ Fixed Income (FI) ✔ Use as a secondary tool to assess Commodities (COM) ✔ performance versus competition, ✔ ✔ and attribute Real Estate (RE) (RE PE/ REFs) (REITS) returns Hedge Funds (HF) ✔ Private Equity (PE) ✔ 61
Basic Performance Comparison Measures 62
Time Weighted Return • Cumulates returns over time • Gives an equal weight to each result, regardless of the dollar amount invested • Returns are calculated daily and geometrically linked over time • Time-weighted methods do not consider the effect of contributions or withdrawals on the portfolio 63
Time Weighted Return Investor 1 invests $1M on Dec 31. On Aug 15 of the following year, his portfolio is valued at $1,162,484. At that point, he adds $100,000, bringing total value to $1,262,484. By the end of the year, portfolio has decreased in value to $1,192,328. 1st period return = ($1,162,484 - $1,000,000) / $1,000,000 = 16.25% 2nd period return = ($1,192,328 - ($1,162,484 + $100,000)) / ($1,162,484 + $100,000) = -5.56% Time-weighted over the two time periods = (1 + 16.25%) x (1 + (- 5.56%)) - 1 = 9.79% 64
How is my portfolio doing on an absolute return basis? • An absolute return measure allows direct alignment with investment objective • No comparison to a benchmark or peer • Relevant for goal based investing agnostic of market or benchmark performance 300 Portfolio Absolute Return Benchmark 250 Portfolio Ann. TWR= 200 6.45% 150 100 50 Jan-00 Jan-02 Jan-04 Jan-06 Jan-08 Jan-10 Jan-12 Jan-14 Jan-16 65
How is my portfolio doing on a relative return basis? • Shows the portfolio is doing relative to SAA benchmark after incorporating for drift and actively set tactical weights 300 Portfolio SAA 250 Portfolio Ann. TWR= 6.45% SAA Ann. TWR = 5.70% 200 150 100 50 Jan-00 Jan-02 Jan-04 Jan-06 Jan-08 Jan-10 Jan-12 Jan-14 Jan-16 66
Sharpe Ratio • Measures portfolio excess return generated over the risk free rate per unit of risk taken • Implies that one is left with the premium that is independent of total risk • Provides easy comparison of portfolios and best used as a ranking metric Sharpe Ratio = (Ra - Rf) / σa Ra is the portfolio return Rf is the risk free rate σa is the standard deviation of the portfolio return 67
Sharpe Ratio • Portfolio Sharpe = (6.45% - 0.60%) / 8.16% = 0.72 • SAA Sharpe = (5.70% - 0.60%) / 7.46% = 0.68 • Pitfalls – It is a ranking criterion only – Negative Sharpe is meaningless – Does not incorporate higher moments – Upward movement is penalized via higher volatility – Doesn’t distinguish between active and passive return – Not so useful when comparing strategies with vastly different trading frequencies (e.g. HFT versus low frequency) 68
Treynor Measure • Measures outperformance over market or benchmark (beta) • Independent of portfolio risk meaning one can compare two portfolios even though they have different betas Treynor Measure= (Ra - Rb) / βa Ra is the portfolio return Rb is the risk free rate βa is the beta of the portfolio 69
Treynor Measure • Portfolio Treynor Measure = (6.45% - 0.60%) / 1.22 = 0.048 • Pitfalls – Doesn’t quantify value added by active portfolio management – It is a ranking criterion only – Unlike Sharpe which applicable to all portfolios, Treynor uses relative market risk or beta and hence is applicable only to well diversified portfolios 70
Jensen’s Alpha • Measure of a security’s excess return with respect to the expected return given by Capital Asset Pricing Model Jensen’s Alpha = Ra - [Rb + βa*(Rm - Rb)] = 6.45% - [0.60% + 1.22*(5.70% - 0.60%)]= -0.39% Ra is the portfolio return Rb is the risk free rate Rm is the market return βa is the beta of the portfolio • Pitfalls: It only allows an absolute measurement of active return 71
Sharpe vs Treynor vs Jensen Sharpe Treynor Jensen’s Return Beta Std. Dev Ratio Measure Alpha Manager 10% 0.90 11% 0.91 0.11 0.03 A Manager 14% 1.03 20% 0.70 0.14 0.06 B Manager 15% 1.02 27% 0.56 0.15 0.07 C Assuming risk free rate of 0% and benchmark return of 8% Don’t forget: this is a snap shot, analyzing across time is crucial to assess 72 stability of these rankings
The rebalancing decision 73
Portfolio Setting SAA Min Max Current Allocation Allocation Allocation Allocation Equity 25% 20% 30% 20.0% Bonds 20% 15% 25% 12.6% Commodities 5% 0% 10% 4.0% Hedge Funds 15% 10% 20% 21.4% Real Estate 10% 10% 20% 11.0% Private Equity 25% 20% 30% 31.0% Should we rebalance? - YES 74
Can we rebalance effectively? MOST PROBABLY NOT • Low trading liquidity (potentially due to illiquid investments) • Subscription/Redemption windows : time taken to subscribe/redeem post request • Lock-ins : investment can’t be redeemed at all • Investment size: accurate rebalancing simply not possible unless portfolio is of significant size • High transaction costs and taxation Factor in these considerations – set wider SAA bands 75
Thank you. 76
You can also read