deMarco 1/24/02 1

Strategy or Policy?a brief look at a Risk-Based Approach

Michael deMarcoPutnam Investments

Northfield 14th Annual Research ConferenceDecember 10, 2001

deMarco 1/24/02 2

Asset AllocationDecisions

Proper Measurementof the Objective(pension fund liability)

Proper Measurementof the Asset Classes

(large cap equity)

deMarco 1/24/02 3

Asset Allocation: portfolio construction & rebalancing

Do we have the right target for our equity allocation?

What about the range . . . is it too narrow

or too tight?

What should trigger

rebalancing?

Should rebalancing frequency be

sensitive to market volatility?

Is frequent rebalancing too costly?

Should policy mix include . . . . . . small cap?. . . international?

deMarco 1/24/02 4

Audit the asset allocation decision process

� Who is key decision maker ?� Policy committee / Investment consultant� Investment staff / Investment manager(s)

� When made ?� Decision horizon (parameter estimate or forecast)?� Decision review frequency (trigger?)

� How is it made ?� Resources applied ?

� skill set; quality of info (detail / timeliness); quality of interpretation� Evaluation: “performance analysis” of the Asset Allocation Decision

deMarco 1/24/02 5

� Measure the value-added ( -lost) from:� Choice of Policy

� choice of objective for portfolio construction� “benchmark”

� Choice of Strategy�Active Mix: degree; extent/shape of collar�Fixed Mix: speed (and cost) of rebalancing

� Choice of implementation (cash market/derivatives; cost/delay)� How good is the liability model ?

� Decision (parameter estimate or forecast) horizon� How much realized Surplus Volatility comes from …

� mis-specification ? … errors in assumptions ?

Audit the asset allocation decision process

deMarco 1/24/02 6

Asset AllocationDecisions

Measurement of theObjective

(pension fund liability)

Measurement of theAsset Classes(large cap equity)

Assumptions Results

data

structure & parameters

deMarco 1/24/02 7

Is asset allocation a policy decision or a managed decision?

19991986 19911990 19981997

Rawls and Izakson: "Why is Everyone Talking About

Risk Allocation?"

Jahnke: "The Asset Allocation

Hoax"

Brinson, Hood, and Beebower: "Determinants

of Portfolio Performance"

Ezra, Hensel, and Ilkiw: "The

Importance of the Asset Allocation

Decision"

Ibbotson and Kaplan: "Does Asset Allocation Policy

Explain 40%, 90%, or 100% of Performance?"

Mezrich and Kumble: "Exploiting Changes in

Correlation and Volatility for Stock-Bond

Allocation"

1988

Sharpe: "Asset Allocation"

(Warren, Managing Investment Portfolios)

2000 2001

Jacquier & Marcus: "Asset

Allocation Models &

Market Volatility"

Fleming, Kirby & Ostdiek:

"Economic Value of Volatility

Timing"

deMarco 1/24/02 8

What have we learned from the debate ?

� Pension plans have stuck to fixed-mix allocations since the ERISA.

� “It is … agreed that the asset allocation decision is by far the most important one made by an investor.”

� “timing” & “selection” have as much impact as “allocation” when a naïve policy-benchmark is used.

� Fixed-mix allocations can produce extremes of excessively high or low risk in an investment fund.

� Trailing volatility is a useful predictor of forward volatility & correlations.

deMarco 1/24/02 9

Fixed-mix Allocation Policies* Assuming a Stable Risk Environment

* From “indefinite time horizons”

8

10

12

14

16

18

20

Jan-

45

Jan-

47

Jan-

49

Jan-

51

Jan-

53

Jan-

55

Jan-

57

Jan-

59

Jan-

61

Jan-

63

Jan-

65

Jan-

67

Jan-

69

Jan-

71

Jan-

73

Jan-

75

Jan-

77

Jan-

79

Jan-

81

Jan-

83

Jan-

85

Jan-

87

Jan-

89

Jan-

91

Jan-

93

Jan-

95

Jan-

97

Jan-

99

Jan-

01

Vola

tility

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

Correlation

Annualized Equity Volatility, 5-year MA (ave = 13.9) 5-year Stock-Bond Correlation (ave = +0.21)

deMarco 1/24/02 10

Fixed Allocations Track a W

ide Range of R

isk

20 25 30 35 40 45Jan-45

Jan-47

Jan-49

Jan-51

Jan-53

Jan-55

Jan-57

Jan-59

Jan-61

Jan-63

Jan-65

Jan-67

Jan-69

Jan-71

Jan-73

Jan-75

Jan-77

Jan-79

Jan-81

Jan-83

Jan-85

Jan-87

Jan-89

Jan-91

Jan-93

Jan-95

Jan-97

Jan-99

Jan-01

Total Risk

-0.25

0.00

0.25

0.50

0.75

1.00

Sharpe Ratio

Annualized Total Risk, 5 yr MA

Sharpe ratio, 5 yr MA

deMarco 1/24/02 11

Targeted Policies Specify Level & R

ange of Risk

20 25 30 35 40 45Jan-45

Jan-47

Jan-49

Jan-51

Jan-53

Jan-55

Jan-57

Jan-59

Jan-61

Jan-63

Jan-65

Jan-67

Jan-69

Jan-71

Jan-73

Jan-75

Jan-77

Jan-79

Jan-81

Jan-83

Jan-85

Jan-87

Jan-89

Jan-91

Jan-93

Jan-95

Jan-97

Jan-99

Jan-01

Total Risk

-0.25

0.00

0.25

0.50

0.75

1.00

Sharpe Ratio

Annualized Total R

isk, 5yr MA

Sharpe ratio, 5yr M

A

deMarco 1/24/02 12

Unleveraged Portfolios can Still Drop to Low Risk Levels

0

20

40

60

80

100

120

8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62

60:40-Mix Annualized Total Risk 10% Risk Tgt Annualized Total Risk

Freq

uenc

y (#

of m

onth

s)

Annualized Volatility Level

deMarco 1/24/02 13

Targeted Risk Policies have Better Results

20

25

30

35

40

45

Jan-

45

Jan-

47

Jan-

49

Jan-

51

Jan-

53

Jan-

55

Jan-

57

Jan-

59

Jan-

61

Jan-

63

Jan-

65

Jan-

67

Jan-

69

Jan-

71

Jan-

73

Jan-

75

Jan-

77

Jan-

79

Jan-

81

Jan-

83

Jan-

85

Jan-

87

Jan-

89

Jan-

91

Jan-

93

Jan-

95

Jan-

97

Jan-

99

Jan-

01

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

10% risk target vs. 60:40-bounded Sharpe ratios

10% risk target: Annualized Total Risk, 5yr MA

60:40-bounded: Annualized Total Risk, 5 yr MA

deMarco 1/24/02 14

Risk “Transitions” Tend to be Abrupt

1

2

3

4

5

6

7

8

9Ja

n-45

Jan-

48

Jan-

51

Jan-

54

Jan-

57

Jan-

60

Jan-

63

Jan-

66

Jan-

69

Jan-

72

Jan-

75

Jan-

78

Jan-

81

Jan-

84

Jan-

87

Jan-

90

Jan-

93

Jan-

96

Jan-

99

1-Ye

ar E

quity

Vol

atili

ty

-1.0

-0.8

-0.6

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1-Year Stock-Bond Correlation

1-year Equity Volatility 1-year Stock-Bond Correlation

deMarco 1/24/02 15

Short-horizon Rebalancing Rules are Required to Catch Volatility Storms

-0.10

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

Jan-

45

Jan-

47

Jan-

49

Jan-

51

Jan-

53

Jan-

55

Jan-

57

Jan-

59

Jan-

61

Jan-

63

Jan-

65

Jan-

67

Jan-

69

Jan-

71

Jan-

73

Jan-

75

Jan-

77

Jan-

79

Jan-

81

Jan-

83

Jan-

85

Jan-

87

Jan-

89

Jan-

91

Jan-

93

Jan-

95

Jan-

97

Jan-

99

Jan-

01

-0.25

0.00

0.25

10% risk target vs. 60:40-bounded Sharpe ratios 10% risk target Sharpe ratio, 5yr MA 60:40-bounded Sharpe ratio, 5 yr MA

deMarco 1/24/02 16

A Slow Rebalancing Rule May Weaken Performance

-12

-9

-6

-3

0

3

6

9

12Ja

n-45

Jan-

47

Jan-

49

Jan-

51

Jan-

53

Jan-

55

Jan-

57

Jan-

59

Jan-

61

Jan-

63

Jan-

65

Jan-

67

Jan-

69

Jan-

71

Jan-

73

Jan-

75

Jan-

77

Jan-

79

Jan-

81

Jan-

83

Jan-

85

Jan-

87

Jan-

89

Jan-

91

Jan-

93

Jan-

95

Jan-

97

Jan-

99

Jan-

01

-50

-40

-30

-20

-10

0

10

20

30

40

50

10% risk target vs. 60:40-bounded Equity Allocation

10% risk target vs. 60:40-bounded Excess Returns, 5yr MA

deMarco 1/24/02 17

Risk Target vs. Asset-mix Target50 year comparison

Risk Target Fixed-mix TargetReturn 13.03% 11.17%

Total Risk 32.56% 30.49%

Sharpe ratio 0.41 0.37

Best One-year Return 61.89 (to Dec '54) 48.71 (to June '83)

Worst One-year Return - 29.18 (to Sept ''74) - 26.97 (to Sept ''74)

# of times Rebalanced 29 28

Max. Stock Allocation 100.0 65.0

Min. Stock Allocation 38.3 55.2

deMarco 1/24/02 18

Low Cost Implementation

Data source: Goldman Sachs

Futures Contracts

Average Daily

Volume(US$ billions)

Impact Ceiling @

5%(US$ millions)

Implied Size of Investment

Fund (2)

(US$ millions)

S&P 500 23.5 1,175 54,210Nasdaq 100 4.5 225 51,903Russell 2000 0.6 180 (1) 47,059US Composite 53,167EuroSTOXX 50 3.8 190 42,222FTSE 3.1 155 34,444CAC 3.4 170 37,778DAX 6.6 330 73,333MIB 2.9 145 32,222IBEX 1.4 70 15,556Nikkie 225 4.5 225 50,000Hang Seng 1.5 75 16,667ASX SPI 3.1 155 34,444non-US Composite 43,344(1) : adjusted for realized liquidity(2) : to conduct a 5% reallocation

deMarco 1/24/02 19

MODELS

Measurement of theAsset Classes

(large cap equity)

Assumptions Results

data

structure & parameters

deMarco 1/24/02 20

Measuring the Asset Class

� Asset allocation is by proxies for the asset classes; i.e., country / market / sector indexes: S&P, Russell, MSCI, FT, Lehman, Salomon Brothers.

� The basis for market-cap weighted indexes is Efficient Market Theory which assumes economic equilibrium; i.e., asset prices = fair value.

� Actual asset prices can be distorted for extended periods by sentiment (or momentum) and by Central Bank actions.

deMarco 1/24/02 21

Asset Price D

istortion from a "H

ot Stock"

0 10 20 30 40 50 60 70 80 90

May-96

Jul-96

Sep-96

Nov-96

Jan-97

Mar-97

May-97

Jul-97

Sep-97

Nov-97

Jan-98

Mar-98

May-98

Jul-98

Sep-98

Nov-98

Jan-99Mar-99

May-99

Jul-99

Sep-99

Nov-99

Jan-00

Mar-00

May-00

Jul-00

Sep-00

Nov-00

Jan-01

Mar-01

May-01

Jul-01

Sep-01

Index & Equity Price

0% 5% 10%

15%

20%

25%

30%

35%

40%

45%

Nortel Weight in TS100

Nortel

TS100 Index / 10

Nortel W

eight in TS100 Index

deMarco 1/24/02 22

Asset Price Distortion from a "Hot Sector"

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5D

ec-9

8

Feb-

99

Apr

-99

Jun-

99

Aug

-99

Oct

-99

Dec

-99

Feb-

00

Apr

-00

Jun-

00

Aug

-00

Oct

-00

Dec

-00

Feb-

01

Apr

-01

Jun-

01

Aug

-01

Oct

-01

NDX Electronic TechnologyTechnology Services Consumer Non-DurablesRetail Trade

deMarco 1/24/02 23

Asset Price Distortion from Central Bank Policy

1,000

1,500

2,000

2,500

3,000

3,500

4,000

4,500

5,000

5,500

Dec

-85

Jun-

86

Dec

-86

Jun-

87

Dec

-87

Jun-

88

Dec

-88

Jun-

89

Dec

-89

Jun-

90

Dec

-90

Jun-

91

Dec

-91

Jun-

92

Dec

-92

Jun-

93

Dec

-93

Jun-

94

Dec

-94

NK

Y In

dex/

10

75

100

125

150

175

200

JPY Currency

NKY Index / 10JNMBRESB IndexJPY Curncy

Bank of Japan -- Monetary Base:

Reserve Balances

Yen

Nikkei 225

deMarco 1/24/02 24

Measurement of theInvestment Objective

(pension fund liability)

MODELSAssumptions Results

data

structure & parameters

deMarco 1/24/02 25

Approaches to measuring the pension fund liability:� Financial (FAS 87 “pension expense”)

� VBO: ABO less unvested staff� ABO: accrual from actual service & salary� PBO: ABO plus salaries projected to retirement� PV of Benefits: PBO plus years of service projected to retirement

� Regulatory (ERISA & IRS “funding status”)� Actuarial Accrued Liability� Actuarial value of plan assets (multi-year smoothing & phase in)

� Economic� IBO: PBO plus post-retirement indexing to inflation� EBO: IBO plus demographic recomposition of workforce

Specification of the Investment Objective

deMarco 1/24/02 26

Measuring the Objective� Magnitude & riskiness of investment objective influences

“aggressiveness” of portfolio … and risk tolerance.� Need for equity in the asset mix (& its risk composition)

influenced by current status & goals for surplus or wealth.�

Source: D. Ezra, FAJ, Jan-Feb 1991.

Comparative Risk Levels ofAlternative Measures of Pension Fund Liabilities

AnnualizedStandard Deviation

ABO 14.76 %PBO 17.39 %IBO 30.28 %T-bills 0.97 %T-bonds 13.93 %Equity 19.12 %

deMarco 1/24/02 27

How you measure the objective sets the required risk?

Required features

Desired enhancements

Inflation adjustmentEarly retirement window

et ceteraBlend

Low risk

High risk

ActualPortfolio

“Firewall”

VBO

ABO

PBO

EBO

“Plan Liabilities” Investment Fund Structure

model “plan liabilities”

IBO

AAL

deMarco 1/24/02 28

Cap

ital M

arke

ts E

nviro

nmen

tU

nfav

orab

leFa

vora

ble

Episodic ReliablePerceived Skill Level

Tolerance for risk is conditional

degree ofConviction in taking decisions

Cautio

us

Highly

Confid

ent

deMarco 1/24/02 29

6/15/94 6/15/94 10/28/95 3/11/97 7/24/98 12/6/99

Asian currency

crisis

Russia / LTCM

0

20

40

60

80

100

120

140%RISK AVERSION LEVEL IMPLIED BY S&P 500 OPTIONS

RIS

K A

VER

SIO

N C

OEF

FIC

IEN

TRisk tolerance changes with the capital market environment

Source: Putnam.

deMarco 1/24/02 30

How much risk should you take ?

0

20

40

60

80

100

120

140

160

0 10 20 30 40 50 60Risk

S&P 500, 3rd Quarter 1998

S&P 500, 4th Quarter 1998

Is the market environment rewarding risk-taking?

deMarco 1/24/02 31

Fina

ncia

l fle

xibi

lity

(liqu

idity

)

Financial endurance (earning power & stability)

Conviction:the financial capacity to take risk

Financia

l stren

gth(w

ealth)

Ability to make up for mistakes

deMarco 1/24/02 32

0

5

10

15

20

25

0510152025

Skill can add value over the naïve “model” even for 5-year horizons.

Source: Putnam Investments.

Return R

ank

Volatility Rank

Netherlands

Switzerland

Ireland

Malaysia

SingaporeJapan

U.S.A.

deMarco 1/24/02 33

Policy + Active Management

Source: Ontario Teachers' Pension Plan Board.

Trade off the risk you needfor the risk you can tolerate

Policy Mix

Risk of surplus loss as % of assets

2.5 %

1.3 %

0 %-0.5 %

-2.5 %

“Worst 1 year in 100”: 0% 10% 20% 30% 55%

Nominal Bonds

Cash

Equities

Real Return Bonds

Surplusgrowing

Surplusfalling

Net

ass

et g

row

thN

et li

abili

ty g

row

th

One-year event risk . . . falling rising

deMarco 1/24/02 34

Asset AllocationDecisions

Measurement of theObjective

(pension fund liability)

Measurement of theAsset Classes(large cap equity)

Assumptions Results

data

structure & parameters

deMarco 1/24/02 35

Institutions have followed fixed-mix asset allocation policies since the beginning of time. What's the problem?

� Findings� Investment funds go beyond probable limits of risk tolerance

during episodes of extremes in markets risk.� Level risk-taken probably too low . . .

Result: relatively less efficient (Sharpe ratio) and less effective (total return) investment programs.

� Asset-price distortions can persist for extended periods . . . Result: capital market expectations that throw off asset allocation policy-making & strategies.

� Investment fund liabilities are more equity-like and larger than conventional measures represent them to be . . . Result: lower than optimal equity-targeting.

� Implications� Actions

deMarco 1/24/02 36

� Findings� Implications

� Higher long-range targets for equity and higher risk-composition targets (with more small cap & international exposure) for equity.

� Broader policy ranges for equity allocations.� Faster tactical rebalancing strategies to achieve targeted risk levels.� Use of derivatives for low cost asset allocation adjustments and to

preserve health from cash-market investment program.� Actions

� Extend research to examine shorter-horizon rebalancing rules & new ranges for rebalancing.

� Apply Information Ratio concept for proxying assets segments to construct asset composites; i.e., to build a fair-value index.

Institutions have followed fixed-mix asset allocation policies since the beginning of time. What's the problem?

deMarco 1/24/02 37

References� Brinson, Gary P., Randolph L. Hood, and Gilbert L. Beebower. “Determinants of Portfolio

Performance”; Financial Analysts Journal July-August 1986.� deMarco, Michael and Todd Petzel. “Risk Budgeting with Conditional Risk Tolerance”; Risk

Budgeting: A New Approach to Investing, Risk Books, 2000.� Ezra, Don. “Asset Allocation by Surplus Optimization”; Financial Analysts Journal Jan-Feb 1991.� Fleming, Jeff, Chris Kirby, and Barbara Ostdiek. “The Economic Value of Volatility Timing”; The

Journal of Finance vol no 1 February 2001.� Hensel, Chris R., Don Ezra, and John H. Ilkiw. “The Importance of the Asset Allocation Decision”;

Financial Analysts Journal July-August 1991.� Ibbotson, Roger G. and Paul D Kaplan. “Does Asset Allocation Policy Explain 40,90,100 Percent

of Performance?”; Association for Investment Management and Research January/February 2000.� Jacquier, Eric and Alan J. Marcus. “Asset Allocation Models and Market Volatility”; Financial

Analysts Journal March/April 2001.� Mezrich, Joseph J. and Di Kumble. “Exploiting Changes in Correlation and Volatility for Stock-

Bond Allocation”; Quantitative Strategies July 1998.� Rawls, S. Waite III. “Why is Everyone Talking About Risk Allocation”; Risk Budgeting,

Institutional Investor Press 2000. � W. F. Sharpe. “Asset Allocation,” in Maginn and Tuttle, editors, Managing Investment Portfolios,

Boston: Warren, Gorham & Lamont 1990.

deMarco 1/24/02 38

Risk Standards for Institutional Investors and Investment Managers

video clips

� Research Papers

putnaminstitutional

Putnam Institute