FIN 367 Performance Evaluation PDF

Performance Evaluation University of Texas At Austin FIN 367: Investment Management Shikhar Singla 1  Announcements Guest Speaker: Rodolfo Martell, ▪ Monday, 12/2, from 5:00 – 6:30PM ▪ Kepos Capital, Rodolfo Martell is formerly Head of Portfolio Strategy at Pluri...

Performance Evaluation University of Texas At Austin FIN 367: Investment Management Shikhar Singla 1  Announcements Guest Speaker: Rodolfo Martell, ▪ Monday, 12/2, from 5:00 – 6:30PM ▪ Kepos Capital, Rodolfo Martell is formerly Head of Portfolio Strategy at Pluribus Labs, Managing Director in AQR’s global stock selection group and the sole strategist on the firm’s Data Governance Committee. Before AQR, he was a Managing Director at QMA (the quantitative investment management arm of Prudential Financial) where he was a portfolio manager, global lead strategist and co-chair of the ESG group. Before QMA he was a Portfolio Manager and Strategist in BlackRock’s Systematic Active Equities group (which he joined through its acquisition of Barclays Global Investors (BGI) in 2009). Earned a PhD in finance from The Ohio State University’s Fisher College of Business. ▪ This replaced Class Mon/Tues Final Exam ▪ Comprehensive exam – roughly 60% of topics from throughout the semester and 40% since the last midterm ▪ Similar format as previous exams, might have 2 ▪ You may bring three note sheets (front and back) Note these do not need to be those used on the previous exam ▪ Many Practice problems on Canvas University of Texas At Austin FIN 367: Investment Management Shikhar Singla 2  Lecture Agenda Performance Evaluation ▪ Methodology Pre-lecture video Performance Evidence ▪ Mutual Funds Alternative Investments ▪ Issues with Alternatives ▪ Hedge Funds ▪ Private Equity ▪ Venture Capital University of Texas At Austin FIN 367: Investment Management Shikhar Singla 3 Performance Evaluation  Thought Questions Investor perspective: 1. How should I evaluate fund / portfolio performance? 2. Should I invest in actively managed mutual funds? 3. Should I invest in hedge funds or other Alternatives? 4. Which funds should I invest in? Economist perspective: 1. Can fund managers outperform the market on a risk-adjusted basis? 2. Are markets efficient? Caveat: Managers outperforming the market is evidence against market efficiency, but lack of outperformance does not mean markets are efficient University of Texas At Austin FIN 367: Investment Management Shikhar Singla 5  Evaluating Returns Fund or portfolio had high realized returns; we need to ask (at least) three questions 1. Risk adjustment: How risky was the fund / portfolio? Was the high return compensation for risk or did it exceed the risk premium investors should demand? 2. Performance attribution: How did the fund / portfolio do relative to other portfolios with similar investment styles? Could I have achieved the same results myself? 3. Luck vs. skill: Is past performance representative of what we should expect in the future? University of Texas At Austin FIN 367: Investment Management Shikhar Singla 6  Performance Evaluation Methods 1. Universe Comparison 2. Sharpe Ratio 3. Treynor Measure 4. Jensen’s alpha ▪ With CAPM ▪ With 3-factor model, 4-factor model, 5-factor model, … 5. Information Ratio University of Texas At Austin FIN 367: Investment Management Shikhar Singla 7  Universe comparison What is it? ▪ Simplest and most popular way to adjust returns ▪ Compare returns to benchmark and universe of similar funds E.g., “fund beat its benchmark each of the past three years” or “Fund consistently performed in the top quartile” ▪ The comparison universe is a benchmark composed of a group of funds or portfolios with similar risk characteristics, such as growth stock funds or high-yield bond funds. What does it tell us? ▪ Risk adjustment: Reasonable approach if comparable funds and benchmark have similar systematic risk ▪ Performance attribution: I could have invested in the index or alternative funds (opportunity cost), so maybe relative performance is what I should care about ▪ Luck or skill: Subtracts off common market performance, which is likely due to luck University of Texas At Austin FIN 367: Investment Management Shikhar Singla 8  Universe Comparison University of Texas At Austin FIN 367: Investment Management Shikhar Singla 9  Sharpe Ratio Purpose: ranks performance based on the portfolio’s risk premium per unit of risk Formula 𝑅ത𝑝 − 𝑅𝑓 𝑆ℎ𝑎𝑟𝑝𝑒 𝑅𝑎𝑡𝑖𝑜 = 𝜎𝑝 ▪ Where 𝑅ത𝑝 is the average portfolio return, Rf is the risk free return, and 𝜎𝑝 is the volatility of the portfolio Interpretation ▪ Higher Sharpe Ratio = better risk adjusted performance ▪ Used to rank portfolio performance from highest to lowest What does it tell us? ▪ Risk adjustment: This is the correct risk vs. return metric if: 1. The fund represents investor’s entire investment portfolio 2. Investor is one-period mean-variance optimizer (CAPM assumptions) ▪ Performance attribution: Not much value ▪ Luck or skill: Not much value University of Texas At Austin FIN 367: Investment Management Shikhar Singla 10  Treynor Ratio Purpose: ranks performance based on the portfolio’s risk premium per unit of systematic risk (beta) Formula 𝑅ത𝑝 − 𝑅𝑓 𝑇𝑟𝑒𝑦𝑛𝑜𝑟 𝑅𝑎𝑡𝑖𝑜 = 𝛽𝑝 ▪ Where ഥ𝑅𝑝 is the average portfolio return, Rf is the risk free return, and 𝛽𝑝 is the beta of the portfolio Interpretation ▪ Higher Treynor Ratio = better risk-adjusted performance relative to market risk ▪ Used to rank portfolio performance from highest to lowest What does it tell us? ▪ Risk adjustment: This is the correct risk vs. return metric if: 1. Fund is part of investor’s overall investment portfolio 2. Risk is measured by CAPM beta ▪ Performance attribution and luck vs. skill: Improvement over Sharpe ratio because we adjust for beta but still does not adjust for market returns or exposure to other factors University of Texas At Austin FIN 367: Investment Management Shikhar Singla 11  Information Ratio Purpose: measures a portfolio’s excess return relative to a benchmark, adjusted for the volatility of those excess returns Formula 𝑅ത𝑝 − 𝑅ത𝐵 𝐴𝑙𝑝ℎ𝑎 𝐼𝑛𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑅𝑎𝑡𝑖𝑜 = = 𝜎 𝑅𝑝 −𝑅𝐵 𝑇𝑟𝑎𝑐𝑘𝑖𝑛𝑔 𝐸𝑟𝑟𝑜𝑟 ▪ Where 𝑅ത𝑝 is the average portfolio return, 𝑅ത𝑝 is the average benchmark return, and 𝜎 𝑅𝑝 −𝑅𝐵 is the standard deviation of excess returns, which is also known as the tracking error Interpretation ▪ Higher information ratio = better risk-adjusted performance relative to the benchmark ▪ Used to assess the consistency of the portfolio’s excess returns over the benchmark What does it tell us? ▪ Risk adjustment: Tells us how much we can increase Sharpe ratio of market portfolio by adding portfolio P Conceptually, P’s alpha is good for us, but investing in P also adds idiosyncratic risk to our portfolio ▪ Performance attribution: Similar to alpha alone ▪ Luck or skill: High information ratio implies that alpha is high and tracking error is low; this gives us more confidence that the alpha is real as opposed to luck / noise University of Texas At Austin FIN 367: Investment Management Shikhar Singla 12  Which Approach is Preferred or Best? Can depends on what you are trying to measure, but… Alpha is frequently the most appealing because it adjusts excess returns for systematic risk and known factors (performance attribution) and is flexible to adding factors University of Texas At Austin FIN 367: Investment Management Shikhar Singla 13  Jensen’s Alpha Purpose: Measures a portfolio’s excess return relative to the expected return predicted by the risk model, such as the Capital Asset Pricing Model, Fama French 3, etc Formula using CAPM 𝐽𝑒𝑛𝑠𝑒𝑛′ 𝑠 𝐴𝑙𝑝ℎ𝑎 = 𝑅ത𝑝 − 𝑅𝑓 + 𝛽𝑝 𝑅ത𝑀 − 𝑅𝑓 ▪ Where 𝑅ത𝑝 is the average portfolio return, 𝑅𝑓 is the risk-free return, 𝛽𝑝 is the portfolio’s beta to the market, and 𝑅ത𝑀 is the average market return Formula using the Fama French 3 Model 𝐽𝑒𝑛𝑠𝑒𝑛′ 𝑠 𝐴𝑙𝑝ℎ𝑎 = 𝑅ത𝑝 − 𝑅𝑓 + 𝛽𝑝 𝑅ത𝑀 − 𝑅𝑓 + 𝛽𝑆𝑀𝐵 𝑓𝑆𝑀𝐵 + 𝛽𝐻𝑀𝐿 𝑓𝐻𝑀𝐿 ▪ Where 𝑅ത𝑝 is the average portfolio return, 𝑅𝑓 is the risk-free return, 𝛽𝑝 is the portfolio’s beta to the market, 𝑅ത𝑀 is the average market return, 𝛽𝑆𝑀𝐵 is the small-minus-big beta, 𝑓𝑆𝑀𝐵 is the SMB risk premium, 𝛽𝐻𝑀𝐿 is the high-minus-low beta, and 𝑓𝐻𝑀𝐿 is the HML risk premium. University of Texas At Austin FIN 367: Investment Management Shikhar Singla 14  Jensen’s Alpha (continued) Interpretation ▪ Positive alpha = portfolio outperformed the market after adjusting for risk ▪ Negative alpha = portfolio underperformed the market after adjusting for risk Advantages ▪ Adjusts for risk and market performance ▪ Indicates the value added by the manager Disadvantages ▪ Alpha estimates are frequently noisy We need a long time series to measure the alpha of any given fund Extra challenging because ▪ Funds have limited lifespans ▪ Fund may change over time ▪ Beta may be misestimated Particularly problematic for non-linear returns, low-probability events, and illiquid assets ▪ Selection and survivorship bias If you are evaluating a fund because it has had high returns, you are data snooping – out of hundreds of mutual funds it is highly likely that a few will have high alphas by sheer chance University of Texas At Austin FIN 367: Investment Management Shikhar Singla 15  Section Summary Alpha and performance relative to a benchmark are the most common (and probably best) ways to evaluate fund performance Noise and selection bias makes evaluating individual funds difficult if not impossible ▪ The best we can do is probably evaluating portfolios of funds that meet certain characteristics (e.g., funds that have performed well over some horizon) University of Texas At Austin FIN 367: Investment Management Shikhar Singla 16 Performance Evidence  Mutual Fund Performance On average, mutual funds underperform the market by about 1% per year (Jensen, 1968) Mounting Evidence since mostly confirms that active Mutual Funds Underperform by about 1% a year ▪ Daniel, Grinblatt, Titman, and Wermers (1997) ▪ Underperform roughly by the amount of their fees! ▪ This result has been shown to be incredibly robust Are some active mutual funds better than others? ▪ Underperformance is persistent Past losers tend to keep losing (in part because high fees and transaction costs are persistent) ▪ No good evidence of persistent overperformance After controlling for investment style, past winners have the same future returns as other funds (Carhart, 1997), DGTW (1997) Bessembinder, Cooper, and Zhang (2024) ▪ “We tabulate an aggregate wealth loss of $1.02 trillion to mutual fund investors over our 30-year sample” ▪ No Consistent Winners University of Texas At Austin FIN 367: Investment Management Shikhar Singla 18  Active Public Manager Snapshot University of Texas At Austin FIN 367: Investment Management Shikhar Singla 19  Investor Behavior University of Texas At Austin FIN 367: Investment Management Shikhar Singla 20  What about during the bad times? Should active managers outperform when times are bad? Pástor and Vorsatz use a similar approach to evaluate how mutual funds performed during the COVID-19 crisis They find funds tended to underperform the S&P 500 74% of active funds underperform the benchmark Average -5.6% return during 10- week period (-29% annualized) compared to S&P 500 University of Texas At Austin FIN 367: Investment Management Shikhar Singla 21 Alternative Investments  Private vs Public Performance According to JPM, Private investments have outperformed public investments This is the common belief among industry. But, this calculation does not seem to take into account, leverage and risk adjustment. ▪ i.e. You could also earn a higher return by investing in the 2X levered S&P 500 ETF. But, this is a passive strategy that takes no skill just risk University of Texas At Austin FIN 367: Investment Management Shikhar Singla 23  Private vs Public Performance Private investments tend to have much wider performance dispersion than public investments. Could indicate that manager selection in privates is much more important than in publics But, could also be driven by being so much harder to measure risk-adjust returns in private assets University of Texas At Austin FIN 367: Investment Management Shikhar Singla 24  Private vs Public Reward vs Risk Private investments on the face of it have a higher return per unit risk than public investments (higher Sharpe Ratio)! This has driven the huge movement of Pensions into Alternatives But… Why do they have lower Standard deviation? Actual less volatility? Asset returns are not frequently traded Volatility could be low simply because they are illiquid. ▪ → Question: Do HF, VC, and PE on average outperform? University of Texas At Austin FIN 367: Investment Management Shikhar Singla 25  Challenges with Alternative Indices Valuation Differences ▪ Infrequent pricing → few transactions ▪ Vintage year → how do you allocate across vintage year? Equal weight, size weight, other? ▪ Subjective valuation → valuations tend to rely on models Liquidity ▪ Illiquid → most alternative investments are not easily traded ▪ Return impact → valuation mark may not represent a value that could have been traded Data Availability and Timing ▪ Limited Data → not all managers report their performance. ▪ Lack of consistent reporting standards Diverse Investment Structures ▪ Varied structures ▪ Complexity Performance Smoothing ▪ Appraisal Smoothing → valuations based on appraisals tend to smooth returns ▪ Lag Effects → valuations tend to lag behind market conditions University of Texas At Austin FIN 367: Investment Management Shikhar Singla 26  Challenges with Alternative Indices (continued) Benchmarking Challenges ▪ Lack of comparable benchmarks ▪ Customized benchmarks can be very subjective Regulatory and Compliance Differences ▪ Varied regulations ▪ Varied disclosure requirements Risk Measurement ▪ Previously mentioned impacts on valuations can adversely impact return calculations, which impact risk calculations ▪ Non-normal return distribution Example: many people believe private equity portfolio returns follow a power law distribution, not a normal distribution. Implies that 20% of the holdings generate 80% of the profits. Fee Structures ▪ Complex fee structures make it difficult to calculate net performance after fees University of Texas At Austin FIN 367: Investment Management Shikhar Singla 27  Estimating betas are harder with Alternative Investing We estimate betas with linear regressions of excess portfolio returns on contemporaneous excess market (and other factor) returns What can go wrong? ▪ Linear model may not work well for non-linear returns (e.g., options) ▪ Can miss risk associated with low probability events E.g., issuing flood insurance looks like a very safe investment until a 50-year flood hits ▪ Illiquid assets might not have good return (price) information in all periods E.g., if asset didn’t trade today, price will be a day old and return will be calculated as zero ▪ Implication: Will look like return has zero correlation with market Problem is even worse if fund purposefully smoothe’s reported returns ▪ Some hedge funds appear to do this (Daniel, Agarwal, and Naik (2011) Overall, this is a big issue for Hedge Funds, PE and VC! ▪ It is difficult to control for but simple metrics that do not control for this can find positive alphas ▪ Newer studies try to control for these issues and find it makes a huge difference University of Texas At Austin FIN 367: Investment Management Shikhar Singla 28  Simple Smoothing Example Imagine PE firm with 2 assets, one liquid one illiquid: Asset Price Liquid $5.00 Illiquid $5.00 Imagine a market downturn and all assets lose 20% of their value, but since the illiquid asset is not traded the price does not get updated: True Value of Portfolio Reported Portfolio Asset Price Asset Price Liquid $4.00 Liquid $4.00 Illiquid $4.00 Illiquid $5.00 $9.00−$10.00 The reported return is $10.00 = −10%, but if it’s assets were continuously traded it’s real return $8.00−$10.00 would be $10.00 = −20% The PE firm seems to be outperforming the market, but this is just an artifact of stale pricing! University of Texas At Austin FIN 367: Investment Management Shikhar Singla 29  Estimation errors in beta change alpha Rp - Rf Rm - Rf True line University of Texas At Austin FIN 367: Investment Management Shikhar Singla 30  Estimation errors in beta change alpha R p - Rf p R m - Rf True line regression line University of Texas At Austin FIN 367: Investment Management Shikhar Singla 31  Do Hedge funds generate alpha? Industry: Yes, of course. Otherwise they would not be so popular. Academic literature: ▪ Older view: hedge funds outperform ▪ More recent view: Probably not Fund of funds probably don’t have positive alpha (because of additional fees) Growing evidence is skeptical of hedge fund performance ▪ Asness, Krail, and Liew (2001), Amin and Kat (2003), Griffin and Xu (2009) HF Performance decreased over time. Bollen, Joenväärä, Kauppila and (2021) ▪ Hedge funds are risky (i.e., they are not as “hedged” as you might think) University of Texas At Austin FIN 367: Investment Management Shikhar Singla 32  Bollen, Joenväärä & Kauppila (2021): Hedge Fund Performance: End of an Era? Gather large dataset combining six commercial databases of returns Unsmooth returns and correct for survivorship bias Document decline in Hedge fund alpha and Sharpe ratio’s post 2008 No evidence one can pick winning funds Evidence that increased regulation and changes in market characteristics explain decline University Finance of Texas 367 - Lecture 23At Austin PortfolioFIN 367: Investment Performance Management Evaluation Shikhar Singla 33 33  Does Private Equity Outperform Private equity returns are not observable: ▪ What’s the return? ▪ What’s the variance? ▪ What’s the covariance with factor X? This makes it difficult to risk-adjust returns (i.e., compare returns relative to systematic market risk) Industry commonly reports IRR ▪ What’s the right benchmark? S&P500, small-cap value, etc. ▪ Reported volatility is downward biased University of Texas At Austin FIN 367: Investment Management Shikhar Singla 34  Kortweg (2019): Risk Adjustment in Private Equity Return Large review of literature on PE performance Risk-adjusted return estimates vary substantially by method, time period, and data source Average leveraged buyout investments have generally earned positive risk-adjusted returns both before and after fees, compared with a levered stock portfolio, but alpha has been decreasing with time and with better measures of risk adjusted returns Close to zero University of Texas At Austin FIN 367: Investment Management Shikhar Singla 35  Kortweg (2019): Risk Adjustment in Private Equity Return VC funds earned positive risk- adjusted returns before the turn of the millennium, but net-of-fee returns have been zero or negative since University of Texas At Austin FIN 367: Investment Management Shikhar Singla 36  Does private equity outperform? PE funds have about the same returns as public equity indices since at least 2006, and yet have collected $230B in performance fees High fees persist due to “multiple layers of agency conflicts and the complexity of measuring risk and returns” Phalippou (2020) University of Texas At Austin FIN 367: Investment Management Shikhar Singla 37  Stafford (2022): PE does not outperform Obtains data on fund-level “Burgiss” is the name of the data provider cashflows Constructs a portfolio of public stocks to match the cashflows Adds a similar amount of leverage as typical LBOs Find IRR of new portfolio is 3% higher than PE’s University of Texas At Austin FIN 367: Investment Management Shikhar Singla 38  Does VC outperform? Korteweg and Nagel (2016) develop new tools to risk-adjust fund returns They note that current measures do not adjust for the high beta of venture capital and thus mechanically overestimate the risk adjusted return Once VC’s beta is properly measured and accounted for, on risk-adjusted basis, they find slightly negative excess returns of VC funds University of Texas At Austin FIN 367: Investment Management Shikhar Singla 39  Boyer et al. (2019): VC does not outperform Boyer et al (2019) use a novel data source to overcome previous issues with measuring VC performance A secondary market for stakes in VC funds developed in early 2000’s Thus, Boyer et al can use the same methods for VC funds as researchers would use for standard public equities They find that VC has 𝛽 > 1 and an 𝛼 ≈ 0 or even negative University of Texas At Austin FIN 367: Investment Management Shikhar Singla 40  Real estate is highly correlated with S&P 500 University of Texas At Austin FIN 367: Investment Management Shikhar Singla 41  Summary Mutual funds generally underperform the market ▪ Average mutual fund has alpha of around -1% per year The underperformance mainly comes from their high fees. This has been known for a long time and still true Leading the push to Passive management ▪ Past good performance is mainly due to luck and tends not to persist (past winners ≠ future winners) Does Alternative Management deliver alpha? ▪ It’s hard to measure but.. ▪ Hedge Funds– no ▪ PE – No ▪ VC – No ▪ REITs – Not made to capture alpha. Does get Real estate exposure. But correlated w market just like the above. ▪ So why do people pay so much in the fees for these funds? Maybe because traditional performance evaluation doesn’t work as well with these funds because they don’t actively report and returns are smoothed. Evidence seems to consistently point to the benefits of Passive Management University of Texas At Austin FIN 367: Investment Management Shikhar Singla 42