Scribd
Upload
54 views10 pages

6242 Lecture

The document discusses creating customized benchmarks for evaluating money managers. It provides 3 main ways to create customized benchmarks: 1) Matching a manager's specific stock selections based on their past picks and investment criteria. 2) Matching a manager's portfolio based on underlying risk factors and exposures over time. 3) Matching a manager's returns pattern by regressing it against various style index returns. The document also discusses problems with over-customization of benchmarks and the benefits of broader benchmarks for comparison across managers. It explores using factor models to group common risk factors and attributes that impact stock returns.

Uploaded by

api-3699305
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
54 views10 pages

6242 Lecture

The document discusses creating customized benchmarks for evaluating money managers. It provides 3 main ways to create customized benchmarks: 1) Matching a manager's specific stock selections based on their past picks and investment criteria. 2) Matching a manager's portfolio based on underlying risk factors and exposures over time. 3) Matching a manager's returns pattern by regressing it against various style index returns. The document also discusses problems with over-customization of benchmarks and the benefits of broader benchmarks for comparison across managers. It explores using factor models to group common risk factors and attributes that impact stock returns.

Uploaded by

api-3699305
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 10

Investment Analysis

and
Portfolio Management

Active Portfolio Management


Customized Benchmark
• Broad-based index may not be appropriate for all
money managers as they may consider only a
subset of the universe of stocks for investment
choice
• Also the benchmark needs to be weighted in the
same fashion as the proportion in which manager
normally invests in different components of his/her
portfolio
• Required for judging the stock selection skill of the
money manager
• Aggregate of all customized benchmarks may be
used to choose mix of managers and manage risk
exposure compared to a broad-based index 2
Ways to create
Customized Benchmark
• Matching Manager’s stock selection
– Based on analysis of manager’s past pickings
– A series of screens (sequencing is important) may be
used to isolate the manager’s domain from the
universe of stocks
– Helped by communication with manager or
assessment of his/her selection criteria
– Weighting the stocks – while normal indices are
capitalization weighted, managers usually do equal
weighting
– Resulting subset may be evaluated by comparing with
the manager’s portfolio as to portfolio characteristics 3
Ways to create
Customized Benchmark contd. …
• Matching manager’s portfolio based on
underlying factors
– Matching of portfolio characteristics based on factor
exposures in a factor model
– Subset is usually bigger than stock-matching
– Screening criteria is obtained by risk factor
exposures over time
– Existing factors proxy for missing factor
– A persistent unintentional bet may be construed as a
valid factor
4
Ways to create
Customized Benchmark contd. ….
• Matching manager’s portfolio by style
– Returns pattern of the portfolio is regressed against
the returns of a few style indices so as to arrive at the
mix of individual indices which match the portfolio
returns
– Alphas are completely captured by indices and error
terms and benchmarking results in identifying broad
asset categories
– Alternatively, portfolio characteristics which tend to
generate return, are attempted to be matched with the
mix of individual index characteristics
– Both results highly dependent upon choice of indices 5
Problems with
Customized Benchmark
• Over-customization will lead to
– Elimination of Alpha
– Non-identification of returns due to sectoral
under/over weightage
– Non-identification of returns due to higher/lower bets
in certain stocks
• Benchmarks with equally weighted stocks creates
difficulty for fund management
• Broad-based benchmarks enable comparison
across managers 6
Factor Models Revisited
• Types of Factor Models
– Statistical Models
– Macroeconomic Factor Models
– Fundamental Factor Models
• Individual factors are grouped under categories
(conversely, multiple proxies are identified for
each attribute)
• Factor betas (regression coefficients) are obtained
through multiple regression between stock returns
for each period and unexpected changes in
macrofactor variables or fundamental factor levels 7
Factor Models visit continues
• Betas are usually subjected to cross-sectional
standardization to make comparison feasible
• Factor returns are calculated by regressing the
periodic returns on universe of stocks against that
particular factor values
• Estimated returns are simply the aggregate of all
(factor betas x factor returns)
• For predicting returns, factor returns and factor
betas need to be predicted using methods ranging
from simple average to statistical time series
analysis
8
Factor Models visit continues
• Portfolio’s sensitivity to a given factor risk is a
weighted average of the factor sensitivity of the
stocks in the portfolio
• Based on prediction as to future factor values,
portfolios may be formed to match a benchmark
in all risk dimensions, tilt in a particular direction
and earn a specified excess return

9
Commonality in the determinants of
expected stock returns - Haugen
• Attribute Risk – payoff to risk variable positive
• Attribute Liquidity – higher liquidity leads to
lower bid-ask spread and cost of trading – payoff
negative
• Attribute Cheapness – relative magnitude of
market price to current cash flows - payoff
negative
• Attribute Growth Potential – probability for faster
than average growth in future – payoff positive
• Technical Factors – history of stock returns -
payoff negative (1 month), positive (6-12 months),
negative (3-5 years) 10

You might also like