PM 5: Portfolio performance: empirical evidence

?
  • Created by: charlie
  • Created on: 20-05-18 18:07
X-Sectional evidence: Underperformance (not normal distribution)
High % of actively managed funds are outperformed by the benchmark
1 of 21
X-Sectional evidence: Survivorship bias (upwards bias)
Many samples include only surviving funds (those that die are removed as time goes on)
2 of 21
X-Sectional evidence: Fama 1) Aggregated returns on US Funds (steps)
1) Cap weighted portfolio of all active funds 2) Use FF-3 Factor model for risk-adjustment
3 of 21
X-Sectional evidence: Fama 1) Aggregated returns on US Funds (results)
Net (after fees + expenses) alpha = -0.8% / Gross (before fees + expenses) alpha = 0% / Sensitivity of portfolio to market returns = 1
4 of 21
X-Sectional evidence: Fama 2) Are winners skilled or lucky? (steps)
1) create a benchmark distribution (no skill/ all randomness) 2) compare wit actual distribution
5 of 21
X-Sectional evidence: Fama 2) Are winners skilled or lucky? (step 1)
1) Create clone population of funds (mean alpha/ t(alpha) = 0) 2) Use bootstrap methodology (random sample of months) 3) Construct alpha + t(alpha) for samples 4) repeat 10,000 times
6 of 21
X-Sectional evidence: Fama 2) Are winners skilled or lucky? (INVESTOR POV: net results
fund managers don't have enough skill to cover costs/ percentile for t(alpha) of actual < percentile for t(alpha) of randomness
7 of 21
X-Sectional evidence: Fama 2) Are winners skilled or lucky? (ECONOMIC POV: gross)
small % of 'true losers' in extreme left tail (do worse than if due just to chance) small % of 'true winners' in extreme right tail (95th + percentile do better than if just due to chance)
8 of 21
Performance persistence evidence: investment flows + compensation
Both +vely correlated + convex in returns (significant upside potential in gains + limited downside potential in losses)
9 of 21
Performance persistence evidence: persistence is short lived
Best performers: disappear fastest (
10 of 21
Performance persistence evidence: measured over shorter horizons
1) ST persistence at higher frequencies 2) Hold ST info 3) High turnover (changing holdings) (4) Investors cant predict LT
11 of 21
Performance persistence evidence: reasons
1) Lack of real skills 2) Winner funds are victims of own success (limited superior investment opportunities = as AUM increase = harder to generate alpha)
12 of 21
Performance persistence evidence: AUM trade-off
Economies of scale (lower fees due to fixed costs) Vs Diseconomies of scale (lower alpha due to investment opportunities)
13 of 21
Other evidence of skill: Looking past alpha
Performance positively correlated with: 1) Education 2) Fund concentration ('best ideas' generate significant alpha but diluted by less successful positions)
14 of 21
Other evidence of skill: less successful positions added due to...
1) Regulation (concentration risk) 2) Volatility decrease 3) Sharpe increase 4) AUM increase (get more as fixed fee%)
15 of 21
Other evidence of skill: agency issue (conflicts of interest)
MANAGERS: like to increase fees (increase AUM by investing in more than just 'best ideas') INVESTORS: like increased returns/ alpha (will be optimal fund size)
16 of 21
Empirical issues (statistical issues)
What do we use as rf rate?/ What do we use as (M)?/ Survivorship bias/ Inaccurate risk measurement/ fees arent fully transparent/ Stat. significance (outperformance or luck) takes a long time
17 of 21
Conceptual issues (model issues)
CAPM model for expected returns isnt well-specified = erodes confidence on performance measures (Treynor & Jenson-alpha)/ Recent improvements: extension to include alpha-type measures (Carhart 4th momentum factor)
18 of 21
Hedge Funds: Similarities
1) Distribution of returns is non-normal (use of derivatives/ shorting) 2) Mean-variance analysis doesnt work well 3) Constantly changing portfolio composition (can't fit linear characteristic lines)
19 of 21
Hedge Funds: Increases performance measures
1) Increases Sharpe/ M: reducing total risk (illiquid assets/ stale prices) 2) Increases Treynore/ Jenson: reducing systematic risk (Reducing covariance/ correlation between asset + market)
20 of 21
Hedge Funds: data-induced bias is worse due to lack of transparency
1) Survivorship bias (Hedge funds have higher returns) 2) Backfill bias (Hedge funds only report when returns are good)
21 of 21

Other cards in this set

Card 2

Front

X-Sectional evidence: Survivorship bias (upwards bias)

Back

Many samples include only surviving funds (those that die are removed as time goes on)

Card 3

Front

X-Sectional evidence: Fama 1) Aggregated returns on US Funds (steps)

Back

Preview of the front of card 3

Card 4

Front

X-Sectional evidence: Fama 1) Aggregated returns on US Funds (results)

Back

Preview of the front of card 4

Card 5

Front

X-Sectional evidence: Fama 2) Are winners skilled or lucky? (steps)

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Economics resources:

See all Economics resources »See all Portfolio Management resources »