Active Management

Anisha Ghosh∗
∗ McGill University

Asset Management – September 10th 2018

1/6 Ghosh (2018) Asset Management

Alpha
Definition
Alpha is the average return in excess of a benchmark:
T
1 X ex
α ≡ rt
T
t=1
T
1X
rt − rtbmk


=
T
t=1

If the benchmark is passive (i.e., can be produced without any particular

investment knowledge), rtex is referred to as active returns:
rt = rt − rtbmk + rtbmk


| {z } |{z}
Active Return Benchmark Return

The benchmark returns, rtbmk are typically the result of a strategic asset
allocation decision.
The active returns arise when asset classes in the benchmark are timed
(called tactical asset allocation) or when securities are picked in each
asset class (called security selection).

2/6 Ghosh (2018) Asset Management

Alpha
Definition
Alpha is the average return in excess of a benchmark:
T
1 X ex
α ≡ rt
T
t=1
T
1X
rt − rtbmk


=
T
t=1

If the benchmark is passive (i.e., can be produced without any particular

investment knowledge), rtex is referred to as active returns:
rt = rt − rtbmk + rtbmk


| {z } |{z}
Active Return Benchmark Return

The benchmark returns, rtbmk are typically the result of a strategic asset
allocation decision.
The active returns arise when asset classes in the benchmark are timed
(called tactical asset allocation) or when securities are picked in each
asset class (called security selection).

2/6 Ghosh (2018) Asset Management

Alpha
Definition
Alpha is the average return in excess of a benchmark:
T
1 X ex
α ≡ rt
T
t=1
T
1X
rt − rtbmk


=
T
t=1

If the benchmark is passive (i.e., can be produced without any particular

investment knowledge), rtex is referred to as active returns:
rt = rt − rtbmk + rtbmk


| {z } |{z}
Active Return Benchmark Return

The benchmark returns, rtbmk are typically the result of a strategic asset
allocation decision.
The active returns arise when asset classes in the benchmark are timed
(called tactical asset allocation) or when securities are picked in each
asset class (called security selection).

2/6 Ghosh (2018) Asset Management

Alpha
Definition
Alpha is the average return in excess of a benchmark:
T
1 X ex
α ≡ rt
T
t=1
T
1X
rt − rtbmk


=
T
t=1

If the benchmark is passive (i.e., can be produced without any particular

investment knowledge), rtex is referred to as active returns:
rt = rt − rtbmk + rtbmk


| {z } |{z}
Active Return Benchmark Return

The benchmark returns, rtbmk are typically the result of a strategic asset
allocation decision.
The active returns arise when asset classes in the benchmark are timed
(called tactical asset allocation) or when securities are picked in each
asset class (called security selection).

2/6 Ghosh (2018) Asset Management

Tracking Error

Definition
Tracking error is the standard deviation of the excess return rtex ; it
measures how disperse the manager’s returns are relative to the
benchmark.
Tracking Error = σ (rtex )

Tracking error constraints are often imposed to ensure a manager does

not stray too far from the benchmark.
If the benchmark is risk-adjusted (e.g., using an asset pricing model like
the CAPM), the tracking error is referred to as idiosyncratic volatility.

3/6 Ghosh (2018) Asset Management

Tracking Error

Definition
Tracking error is the standard deviation of the excess return rtex ; it
measures how disperse the manager’s returns are relative to the
benchmark.
Tracking Error = σ (rtex )

Tracking error constraints are often imposed to ensure a manager does

not stray too far from the benchmark.
If the benchmark is risk-adjusted (e.g., using an asset pricing model like
the CAPM), the tracking error is referred to as idiosyncratic volatility.

3/6 Ghosh (2018) Asset Management

Tracking Error

Definition
Tracking error is the standard deviation of the excess return rtex ; it
measures how disperse the manager’s returns are relative to the
benchmark.
Tracking Error = σ (rtex )

Tracking error constraints are often imposed to ensure a manager does

not stray too far from the benchmark.
If the benchmark is risk-adjusted (e.g., using an asset pricing model like
the CAPM), the tracking error is referred to as idiosyncratic volatility.

3/6 Ghosh (2018) Asset Management

Information Ratio

Definition
The information ratio is the ratio of alpha to tracking error:
α
Information Ratio = IR =
σ (rtex )

Alpha can, for example, be produced by a manager by taking large

amounts of risk.
The information ratio divides the alpha by the risk taken so it is the
average excess return per unit of risk.
When the benchmark is the risk free rate rtf , the alpha is the average
return in excess of the risk free rate and the information ratio coincides
with the Sharpe ratio:

α ≡ rt − rtf
rt − rtf
Sharpe Ratio = IR =
σ (rt )

4/6 Ghosh (2018) Asset Management

Information Ratio

Definition
The information ratio is the ratio of alpha to tracking error:
α
Information Ratio = IR =
σ (rtex )

Alpha can, for example, be produced by a manager by taking large

amounts of risk.
The information ratio divides the alpha by the risk taken so it is the
average excess return per unit of risk.
When the benchmark is the risk free rate rtf , the alpha is the average
return in excess of the risk free rate and the information ratio coincides
with the Sharpe ratio:

α ≡ rt − rtf
rt − rtf
Sharpe Ratio = IR =
σ (rt )

4/6 Ghosh (2018) Asset Management

Information Ratio

Definition
The information ratio is the ratio of alpha to tracking error:
α
Information Ratio = IR =
σ (rtex )

Alpha can, for example, be produced by a manager by taking large

amounts of risk.
The information ratio divides the alpha by the risk taken so it is the
average excess return per unit of risk.
When the benchmark is the risk free rate rtf , the alpha is the average
return in excess of the risk free rate and the information ratio coincides
with the Sharpe ratio:

α ≡ rt − rtf
rt − rtf
Sharpe Ratio = IR =
σ (rt )

4/6 Ghosh (2018) Asset Management

Information Ratio

Definition
The information ratio is the ratio of alpha to tracking error:
α
Information Ratio = IR =
σ (rtex )

Alpha can, for example, be produced by a manager by taking large

amounts of risk.
The information ratio divides the alpha by the risk taken so it is the
average excess return per unit of risk.
When the benchmark is the risk free rate rtf , the alpha is the average
return in excess of the risk free rate and the information ratio coincides
with the Sharpe ratio:

α ≡ rt − rtf
rt − rtf
Sharpe Ratio = IR =
σ (rt )

4/6 Ghosh (2018) Asset Management

Benchmarks
The concept of alpha requires first defining a benchmark against
which alpha can be measured.

Example: Martingale Asset Management’s Low Volatility Strategy based

on Russell 1000 universe of large stocks.
CAPM regression implies

rt − rtf = 0.0344 + 0.7272 rtR1000 − rtf + εt

 

⇒ α = 3.44% per year is the average excess return of the strategy

relative to a market-adjusted portfolio.
If we assume a naive benchmark of just the Russell 1000, the alpha
is only 1.50% per year:

rt = 0.0150 + rtR1000 + εt

The naive benchmark falsely assumes that the beta of the strategy
is 1 when, in fact, it is 0.73.

5/6 Ghosh (2018) Asset Management

Benchmarks
The concept of alpha requires first defining a benchmark against
which alpha can be measured.

Example: Martingale Asset Management’s Low Volatility Strategy based

on Russell 1000 universe of large stocks.
CAPM regression implies

rt − rtf = 0.0344 + 0.7272 rtR1000 − rtf + εt

 

⇒ α = 3.44% per year is the average excess return of the strategy

relative to a market-adjusted portfolio.
If we assume a naive benchmark of just the Russell 1000, the alpha
is only 1.50% per year:

rt = 0.0150 + rtR1000 + εt

The naive benchmark falsely assumes that the beta of the strategy
is 1 when, in fact, it is 0.73.

5/6 Ghosh (2018) Asset Management

Benchmarks
The concept of alpha requires first defining a benchmark against
which alpha can be measured.

Example: Martingale Asset Management’s Low Volatility Strategy based

on Russell 1000 universe of large stocks.
CAPM regression implies

rt − rtf = 0.0344 + 0.7272 rtR1000 − rtf + εt

 

⇒ α = 3.44% per year is the average excess return of the strategy

relative to a market-adjusted portfolio.
If we assume a naive benchmark of just the Russell 1000, the alpha
is only 1.50% per year:

rt = 0.0150 + rtR1000 + εt

The naive benchmark falsely assumes that the beta of the strategy
is 1 when, in fact, it is 0.73.

5/6 Ghosh (2018) Asset Management

Benchmarks
The concept of alpha requires first defining a benchmark against
which alpha can be measured.

Example: Martingale Asset Management’s Low Volatility Strategy based

on Russell 1000 universe of large stocks.
CAPM regression implies

rt − rtf = 0.0344 + 0.7272 rtR1000 − rtf + εt

 

⇒ α = 3.44% per year is the average excess return of the strategy

relative to a market-adjusted portfolio.
If we assume a naive benchmark of just the Russell 1000, the alpha
is only 1.50% per year:

rt = 0.0150 + rtR1000 + εt

The naive benchmark falsely assumes that the beta of the strategy
is 1 when, in fact, it is 0.73.

5/6 Ghosh (2018) Asset Management

Ideal Benchmarks

1 Well defined: verifiable and free of ambiguity about its

contents.
2 Tradeable: Otherwise the computed alphas do not represent
implementable returns on investment strategies.
3 Replicable: by both the asset owner and the funds manager.
4 Adjusted for risk: The particular risk adjustment used can
make a big difference in the alpha and information ratio of a
strategy.

6/6 Ghosh (2018) Asset Management

Ideal Benchmarks

1 Well defined: verifiable and free of ambiguity about its

contents.
2 Tradeable: Otherwise the computed alphas do not represent
implementable returns on investment strategies.
3 Replicable: by both the asset owner and the funds manager.
4 Adjusted for risk: The particular risk adjustment used can
make a big difference in the alpha and information ratio of a
strategy.

6/6 Ghosh (2018) Asset Management

Ideal Benchmarks

1 Well defined: verifiable and free of ambiguity about its

contents.
2 Tradeable: Otherwise the computed alphas do not represent
implementable returns on investment strategies.
3 Replicable: by both the asset owner and the funds manager.
4 Adjusted for risk: The particular risk adjustment used can
make a big difference in the alpha and information ratio of a
strategy.

6/6 Ghosh (2018) Asset Management

Ideal Benchmarks

1 Well defined: verifiable and free of ambiguity about its

contents.
2 Tradeable: Otherwise the computed alphas do not represent
implementable returns on investment strategies.
3 Replicable: by both the asset owner and the funds manager.
4 Adjusted for risk: The particular risk adjustment used can
make a big difference in the alpha and information ratio of a
strategy.

6/6 Ghosh (2018) Asset Management