Sie sind auf Seite 1von 30

Portfolio

Performance
Evaluation
Part A
FNCE90056 Investment Management
Semester 1 2020

1
What we are going to look at …
Key Learning Outcomes
• Discuss the importance of performance evaluation
• Identify and discuss the properties of a valid benchmark
• Review the analysis of portfolio returns into components due to the market, to style, and to active
management
• Characterize and distinguish among the components of portfolio evaluation: performance
measurement, performance attribution and performance appraisal
• Distinguish time-weighted and money-weighted rates of return
• Compute, compare, and contrast the following risk-adjusted performance appraisal measures in
their ex-post forms: alpha, Treynor measure, Sharpe ratio, information ratio and M2
• Distinguish between macro and micro performance attribution
• Distinguish between the effect of the interest rate environment and the effect of active
management on fixed-income portfolio returns
• Explore persistence in portfolio performance
2
Portfolio Management
What is fund managers objective?

To have more funds (assets) to manage.

• Managers are paid a % of the value of Funds Under Management (FUM)

How is this achieved?


• Attract new funds to manage
• Increase value of funds under management
– Earn positive returns

How do fund managers add value?


• Asset allocation
• Stock selection
• Market timing
• Diversification
3
The Evaluation of Portfolio Performance

Returns:
What factors drive returns?

• market performance
• risk factors
• skill of manager.

Objective is to develop measures of the fund manager’s skill.

4
Winners and Losers

Skill

Lucky Lucky
Luck Skilled Unskilled

Unlucky Unlucky
Skilled Unskilled

5
The Use of Past Performance

Why the focus on past performance ?


• Investors want to allocate their capital resources among the most
effective and efficient mangers of these funds.
– Past performance is the track record of the manager
– Is it a useful guide to the future?
– Is there persistence in performance?

Survey results
• Most important factor of fund investment is performance (54%); Risk
(only 17%).
• ASIC study found 70% of advertising campaigns use past performance to
lure new clients (of these 65% are based on 1 year of data!).
• Important to specify benchmarks. 6
The Scope of Performance Evaluation

7
Evaluating Fund Performance

Two Approaches:

1. Returns-based analysis
• External

2. Decision-based analysis
• Internal

8
Returns-based Analysis

Three general methods to evaluate portfolio performance?

• Excess returns: calculate the difference between portfolio returns and the returns from a return-
generating model like the CAPM.

• Relative return ratios: measure return per unit of risk

• Scaled return methods: adjust the portfolio return for risk so that it can be directly compared to the
benchmark return

9
How do we know if a fund manager has performed
well?

Components of Performance Evaluation

Performance measurement – what was the performance

Performance attribution – why did the performance occur

Performance appraisal – was the performance due to luck or skill

10
Calculating Portfolio Return

Percentage change in market value over a period of time, after accounting for external
cash flows (contributions and withdrawals)

If no external cash flows, return = difference between ending and beginning value,
divided by beginning value

MV1 − ( MV0 + CF )
With cash flow at beginning of measurement period: rt =
MV0 + CF

( MV1 + CF ) − MV0
With cash flow at end of measurement period: rt =
MV0

11
Time Weighted Rate of Return (TWR)

Reflects the compound rate of growth over a stated evaluation period of


one unit of money invested in the account

Interim external cash flows result in sub-periods that begin with each cash
flow, and TWR must link the sub-periods together (chain-linking) such that:

• R(twr) = (1+r(1)) X (1+r(2) X … X (1+r(n)) -1

For example, during a year there are subsequent sub-periods of 1%, 8% and
-3% returns. The time weighted return for the year is 1.01 X 1.08 X 0.097 –
1 = 0.058 = 5.8%

12
Money-Weighted Rate of Return (MWR)

Measures compound growth rate in the value of all funds invested in the account over the evaluation period

Equivalent to internal rate of return (IRR)

MWR is the growth rate that solves:


• MV1 = MV0(1+R)m + CF(1+R)m-L(1) + … + CFn(1+R)m-L(n) where:
– M = number of time units in the sub-period
– L(i) is the number of time units by which the ith cash flow is separated from the beginning of the evaluation period

13
Time Weighted versus Money-Weighted Return
MWR represents average growth rate of all money invested, while TWR represents growth of a single unit
invested

MWR is sensitive to timing of external cash flows, TWR is not affected

Time weighted return is preferable for evaluating a manager who does not have control over the size or
timing of cash flows, and is generally required by GIPS

MWR may be more appropriate in circumstances where manager can influence cash flow size and timing,
such as private equity funds that take down commitments for specific investments

14
Data Quality Issues for Rate of Return Calculations

Underlying valuations of illiquid or infrequently traded securities may be suspect, invalidating


reported rates of return

In some cases, matrix pricing estimates value based on quoted prices for securities (bonds) with
similar attributes

Valuations should be reported on a trade-date, fully accrued basis reflecting unsettled trades and
income owed but not yet paid

15
Components of Portfolio Return

Market return – the return on the market index

Style – the difference between the benchmark portfolio and the market index

Active return – the difference between the portfolio and its benchmark

16
Callan Table

17
Factors that lead to abnormal performance

Successful across asset allocations

Superior allocation within each asset class


• Sectors or industries
– Over-weight better performing sectors, under-weight poorer performers

Individual security selection


– Pick the right stocks, those with performance better than expected
Properties of a Valid Benchmark

Unambiguous – clearly defined identities and weights of securities or factor exposures


Investable – can be held without active decisions
Measurable – readily calculable return
Appropriate – consistent with manager’s investment style or expertise
Reflective of current investment opinions – manager has knowledge of the components and opinions of
some sort
Specified in advance – specified prior to the evaluation period and known to all interested parties
Owned – manager accepts accountability for performance relative to the benchmark

19
Types of Benchmarks
Absolute return benchmark
• Can be an objective, but is not investable and therefore does not satisfy benchmark criteria
– For example, CPI + 3%

Broad market indexes


• Meet many criteria for benchmark and are suitable for core approaches
• May not be appropriate if manager does not pursue a broad style

Style Indexes
• Typically represent portions of an asset category

Factor Model-based benchmarks


• Relate one or more systematic sources of return to returns on the fund

Returns-based Benchmarks
• Combine style indexes that track managers historical returns

Manager Universes
• Performance of average manager fails all benchmark criteria other than measurability

20
Types of Benchmarks
Style indexes

• Well known, easy to understand and widely available

• Some weight securities or sectors more highly than managers would consider prudent

• Style definition may be ambiguous or inconsistent with the manager’s investment process

21
Types of Benchmarks
Factor model-based

• Relate one or more systematic sources of return to the returns on a fund

• Single factor model based on market returns would be expected to earn the risk free rate plus the
manager’s beta times the market risk premium

• Multi-factor models can also be created

• Can be useful in performance evaluation because they identify systematic sources of return and provide
insight into manager style

22
Types of Benchmarks
Factor model-based

• Problems applying model as a benchmark:

– Most managers do not think of their process in terms of factor exposures

– Can be difficult to obtain and expensive to use

– Ambiguous

– No specified securities and weights (may not be investable)

23
Types of Benchmarks
Returns based

• constructed using:

– Manager’s series of returns (monthly since inception)


– Series of returns on several style indices over the same period

• Allocation algorithm solves for combination of style indices that best tracks manager’s returns, which
becomes the benchmark

• Meet most benchmark criteria, and are especially useful when returns are the only information available

24
Types of Benchmarks

Returns based

Problems applying model as a benchmark:

• Underlying styles may not reflect manager’s process

• May not discern a pattern for managers who rotate among styles

• Requires many months of data for a statistically valid pattern

25
Types of Benchmarks

Custom security based

• Manager’s research universe weighted in a particular fashion that reflects


manager’s weighting style

• Meets all benchmark criteria

• Is useful for performance evaluation

• Can be expensive to construct and maintain

• Since not composed of published indexes, lack of transparency can be a


concern
26
Constructing a Custom Security Based Benchmark

Identify prominent aspects of manager’s investment process

Select securities consistent with that investment process

Devise a weighting scheme for the benchmark securities, including a cash position

Review the preliminary benchmark and make modifications

Rebalance the benchmark portfolio on a predetermined schedule

27
Types of Benchmarks

Manager Universes
Investors want to know how the managers they hired performed relative to the managers they could have hired instead

Cannot be specified in advance

Median manager is not investable because can only be determined ex-post

Ambiguous

Style of median manager may not be appropriate

Subject to survivor bias, which overstates peer group return

Does use of manager universes motivate Tournament behavior?

28
Tests of Benchmark Quality
Systematic biases – historical beta of account relative to benchmark should be close to 1.0, and correlation between
active and style return should be statistically insignificant

Tracking error – volatility of active returns should be closer to the benchmark than to a market index or alternative
benchmarks

Risk characteristics – exposure to systematic sources of risk should be similar to those of the benchmark over time

Coverage – a high percentage of the securities in the portfolio should be in the benchmark

Turnover – percentage of benchmark’s market value allocated to purchases during a rebalancing has bearing on
whether the benchmark is investable

Positive active positions – too many securities the manager does not want to own indicates a poorly constructed
benchmark
29
Thank you
End of Part A

FNCE90056 Investment Management


Semester 1 2020

Das könnte Ihnen auch gefallen