29-C.ppt

Don’t Kill the Golden Goose

Conference of Consulting Actuaries

2006 Conference Annual Meeting

October 24, 2006

Westin Mission Hills ResortRancho Mirage, California

M. Barton Waring

Chief Investment Officer for Investment Policy & Strategy, Emeritus

[email protected]

2

StratRschInst-060424

The pension funding crisis

Defined benefit pension plans are in danger United Kingdom, Australian experience CIEBA survey of US corporate DB plans: mark-to-market makes them risky!

The perception is that pension funding risks and costs are unmanageable Today’s investment policy practices aren’t effective at controlling

pension funding risk Today’s active management policy practices are inefficient Today’s contribution practices can be improved Today’s pension costs seem too high too high to be sustainable

But the perception is not correct: We do have the tools to fix DB plan problems

If we want to save defined benefit plans, we have to be “on a mission” to adopt and use these new tools

3


DC plans aren’t the answer

Flexible, yes Transportable, yes

But they seldom if ever grow large balances A couple of hundred thousand dollars won’t support much

of lifestyle in retirement ($44,000 median balance!) Only a small augmentation of social security But they are fully funded! (Or are they?)

DC plans aren’t really retirement plans

Your goose may already be cooked!

4


DB plans have many advantages: Golden Geese

A big advantage: Higher (implicit) savings rate

Professional management and state of the art risk control (policy asset allocation)

Lower fees and costs (wholesale versus retail)

More skillful manager selection, in many cases

Higher average returns (2%–4% per year)

Another big advantage--the insurance principle: Spread mortality risk across a large group

Allows all to have lifetime protection at reduced cost

Personal example: For DC, I need to fund for 105 year possible life. In DB, I only need 88 years.

A male age 65 retiree needs only 65% as much savings in a DB plan as in an unannuitized DC plan, for the same monthly draw

DB plans are successful in replacing some realistic part of income on retirement; DC plans generally are not (as used today in the US)

5


What is the investment objective for DB pensions?for developing investment policy

Many investment objectives are stated today Asset-only?

Asset-only, followed by monte carlo simulation of the accounting?

Focus on risk/return relationship of contributions, or of pension expense, or of A/L ratio?

Minimize present value of future contributions, or of future normal cost?

Or . . . . ?

6


“The goal of asset allocation analysis should be stated in terms of surplus. The objective is to maximize the risk-adjusted future value of the surplus.”

- from Asset Allocation by William F. Sharpe, Nobel Laureate in Economic Science

The utility function is specified in financial economics

7


A solution: controlling pension funding risk

Surplus optimization Surplus efficient frontier Definitions, and Two Fund Theorem explanation

Economic views of the liability

The “three decisions” for controlling pension funding risk: A case study A practical application of pension funding risk control

Mark-to-market: Problem, or benefit? Transparency turns out to be the key to progress You can’t hedge a book value liability!

8


The portfolio of interest is the assets less the liabilities – the surplus or deficit

Pension Plan “T-Account”

ContributionsExpense (corporate)A/L ratio

Deficit (Surplus) [=PVFC]

Assets Liability (economic measure)

If we control the economic surplus risk, we also control all the accounting risks

Surplus optimization simultaneously satisfies all conventional objectives

9


The new toolkit for managing pension funding risk

A. Surplus optimization: A “two-fund theorem” problem1) The Liability-Matching Asset, or “Hedging Portfolio:” Duration matching controls the interest rate mismatch between the assets and the

liabilities,1 or “surplus duration” Duration is a measure of how a financial asset or liability changes in value when

interest rates change Pension surpluses have dual durations: Inflation sensitivity and real interest

rate sensitivity Think of these net surplus durations as just “factor betas” for explaining surplus

changes with rate changes

2) The “Risky Asset Portfolio:” Controlling “surplus beta,” the net market risk exposure of the assets relative to the liabilities2

Market risk is rewarded, but it is risky! How much market risk do you want to take?

B. Adding alpha through active management: Manager structure optimization manages your active managers, tactical positioning, hedge funds, etc.3,4

Rewarded if skillful!

See Waring, et al, Journal of Portfolio Management, Summer 20041, Fall 20042, Spring 20003, Spring 20034

10


An economic view of the liability is essentialfor doing surplus optimization

Conventional actuarial view of the liability is expressed as if it were in dollars, but it’s a different kind of dollar!

Determined with wrong discount rate, with smoothing. “Sasquatches,” The bottom line: dollars and sasquatches are different units Unlike dollars, sasquatches can’t be plotted on the same graph as things that

happen in actual, real dollars of value

An economic measure of the liability, by construction, is in genuine dollars, not in sasquatches So its returns and risks can be used in surplus optimization, making surplus

optimization “doable” You’re still a non-believer in market discount rates? Set up a laddered portfolio

of coupons and bonds to pay off the liability: It will require an amount = $EL FWIW, there is universal agreement among actual financial economists that the

government bond curve provides the right discount rates for fully funded plans

11


Pension expense volatility can be controlledIf you control investment risk to the surplus, you control expense risk

Using surplus optimization (the “two fund” version) together with economic accounting, we can confirm that if we control investment risk we control expense risk:

normal and supplemental costs

0,1 0,1

0L

PE NC SC

r L

income returns

0L Lr D L

capital gains returns

Liability

0Ar A

income returns

0A Ar D A


Assets #1:

Assets #2: (excess return over hedge

Hedging Portfolio

*0

Risky Asset porPortfol

tfolio)io

Ar A

If the liability is matched and all interest rate risks are hedged with a Hedging Portfolio, there is no investment risk except that taken intentionally in the Risky Asset Portfolio

Pension expense (level and volatility) are then reduced to just that of normal cost, supplemental cost, and the risky excess return of the Risky Asset Portfolio – exposure to the latter being in your complete control.

Pension expense risk can be dramatically reduced!

Supplemental costs do have some risks, but these might be managed to some degree with better actuarial tables and continuous improvement of decrements. They can’t be eliminated.

12


Contribution volatility can be controlledIf you control investment risk to the surplus, you control contribution risk

Economic views of the contribution are similar to those for expense, but we add in any starting deficit (surplus):

0

1 1 1

1 0 0

0

beginning deficit

normal and supplemental cos

,1 0,1

ts

Contrib

Contrib

Lr L

L A

L A

NC SC

income returns

0L Lr D L


Liability

0Ar A

income returns

0A Ar D A


Assets Hedging Portfolio #1:

*

0

Risky Asset PorAssets #2: (excess return o

tfolver

iohedge)

Ar A

Pension expense risk can be dramatically reduced!

Economically, the required contribution is simply the shortfall of the assets against the economic measure of the liability, at period end (time 1). In turn, this is just the shortfall at time 0, adjusted by normal and service costs, interest costs, and asset returns, i.e., by pension expense.This is all identical to the

contribution calculation, other than for the inclusion of the starting economic deficit. The only “risky” terms, again, str supplemental cost and the risky asset portfolio. And so again, investment policy can control nearly all risks!

13


What is surplus optimization?First, look at asset-only optimization

Expected risk

Exp

ecte

d r

etu

rn

Asset-only frontier

A stylized view:

Cash

Bonds

Large Cap Equity

International Equity

Small Cap Equity

14


What is surplus optimization?Add in a liability-matching portfolio (beta factors only), decide risk level

Surplus Frontier

Expected risk

Exp

ecte

d r

etu

rn

Liability

The Hedging Portfolio

Surplus beta decision: The Risky Asset Portfolio

Asset only frontier

How much surplus beta risk?

15


What is surplus optimization? Consider alpha from active management

Expected risk

Exp

ecte

d r

etu

rn

Liability

How much surplus beta risk?

Active riskE

xpec

ted

alp

ha Active frontier

… and how much alpha risk?

Active risk decision

16


Section Summary: Managing DB pension funding risksUsing surplus optimization and the economic liability

Three key investment policy decisions are always presented: 1. Extending portfolio duration (the liability match,

or Hedging Portfolio)

2. Reconsidering the stock-bond mix and its risk/return tradeoff (the Risky Asset Portfolio)

3. Using active management (beating the zero sum game?)

17


Controlling pension costs What is really happening “Under the hood” of conventional accounting

There is an economic “pension budget identity” that reveals that contributions and periodic normal costs have present values equal to that of the liability (ab initio):

A plan’s periodic normal cost is controlled solely by the total present value of the benefit level, not by the accounting!

To control costs, control the benefit level And to control the benefit level, both management and labor

need good valuations, transparency

balance sheet: income statement: cash flow statement: economic liability economic normal cost; expense economic contributions

PVFBP PVFNC PVFC

How could it be any different?

18


A case study Asset-only optimization is the usual tool

Current holdings appear “efficient” in asset-only spaceE

xpec

ted

ass

et r

etu

rn

Expected asset risk

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%

11%

12%

0% 5% 10% 15% 20% 25% 30%

Alternatives

TIPS [D=9]

Cash

Domestic EquityInternational Equity

Nominal Bonds [D=5]

Current Policy

19


Today’s Hedging Portfolio: The assets aren’t well matched to the liabilities!

Asset Class Durations Real duration Inflation duration

Composition Duration $ Dur.(m) Duration $ Dur.(m)

Equity and Equity-like 75.0% 8.00 $820 0.00 $0

Nominal bonds(5.25 year nominal duration) 22.5% 5.25 $161 5.25 $161

TIPS (9 year real duration) 2.5% 9.00 $31 0.00 $0

Aggregate Plan Assets 100% 7.41 $1,012 1.18 $161

Liability Durations Valuation Real duration Inflation duration

Weight Duration $ Dur.(m) Duration $ Dur.(m)

1) Retired/Inactive 55.2% 12.87 $926 12.67 $912

2) Current employees 44.8% 22.89 $1,335 8.41 $490

Aggregate Plan 100% 17.36 $2,261 10.76 $1,402

20


How much net (surplus) interest rate risk is left over?

What this means for this plan sponsor: These surplus durations represent uncompensated risk For a 1% decrease in the real rate, surplus will go down by roughly 9.6%, or $1,249m For a 1% decrease in the inflation rate, surplus will go down by roughly 9.5%, or

$1,241m

Summary: Today’s plans have large bets on both inflation rate increases and real rate increases

Surplus calculations based on using the market-valued benefit security A/L of 105%.

Real duration Inflation durationDuration $ Dur. (m) Duration $ Dur. (m)

Liability 17.36 $2,261 10.76 $1,402Assets 7.41 $1,012 1.18 $161

Surplus duration (9.58) ($1,249) (9.52) ($1,241)

21


Asset model

We use BGI assumptions for this analysis. Assumptions can be modified to incorporate the companies expectations, however for modest changes the results will be materially similar.* Represents the company’s existing nominal bond portfolio with a nominal duration extension to 45 years.** Represents the company’s existing TIPS portfolio with a real duration extension to 45 years.

Arithmetic

Asset class Expected return (%) Expected risk (%)

Domestic Equity 8.75 15.50

International Equity 8.75 16.25

Bonds (5 dur) 4.75 6.00

Bonds (15 dur) 5.35 14.00

Bonds (45 dur)* 7.15 38.00

TIPS (9 dur) 4.50 6.00

TIPS (15 dur) 4.86 9.30

TIPS (45 dur)** 6.66 25.80

Alternatives 11.20 30.00

Cash 3.25 1.50

Dom Eq Int’l Eq Bonds TIPS Alts CashDomestic Equity 1.00International Equity 0.65 1.00Bonds 0.20 0.15 1.00TIPS 0.28 0.20 0.70 1.00Alternatives 0.65 0.40 0.20 0.28 1.00Cash 0.00 0.00 0.00 0.00 0.00 1.00

22


Detail: The surplus efficient frontier

In “surplus graphical space”

Exp

ecte

d s

urp

lus

retu

rn

Expected surplus risk

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

0% 1% 3% 5% 7% 9% 11% 13% 15%

Case 1 (Minimum surplus risk, 4% Equity)

Case 5 (75% Equity)

Case 4 (60% Equity)

Case 3 (45% Equity)

Case 2 (30% Equity)

Current Policy

More complete hedging of liability

Less complete hedging of liability

23


The ratio of domestic equity to international equity to alternatives is constrained to be in a 50 : 15 : 10 ratio, consistent with current policy.

Optimal mixes from the surplus efficient frontier

Surplus optimal mixes

Asset class: Current Policy Case1 Case 2 Case 3 Case 4Case 5

Equity-Like 75.0% 4.3% 30.0% 45.0% 60.0%75.0

Domestic Equity 50.0% 3.3% 20.0% 30.0% 40.0%50.0%

International Equity 15.0% 1.0% 6.0% 9.0% 12.0%15.0%

Alternatives 10.0% 0.0% 4.0% 6.0% 8.0%10.0%

Bonds 25.0% 95.7% 70.0% 55.0% 40.0%25.0%

Nominal Bonds 22.5% 60.6% 50.8% 43.6% 34.9%24.3%

Duration 5.25 16.9 20.2 23.5 29.342.1

TIPS 2.5% 35.2% 19.3% 11.4% 5.1%0.7%

Duration 9.00 16.9 20.2 23.5 29.342.1

Expected surplus return 2.76% 0.20% 1.36% 2.02% 2.68%3.32%

Expected surplus risk 13.24% 0.08% 3.86% 6.09% 8.32%10.55%

Duration mismatch

Real -9.58 0 (Matched) 0 (Matched) 0 (Matched) 0 (Matched)0 (Matched)

Inflation -9.52 0 (Matched) 0 (Matched) 0 (Matched) 0 (Matched)0 (Matched)

Asset-only return 7.99% 5.43% 6.59% 7.25% 7.91%8.55%

Asset-only risk 12.07% 12.40% 12.95% 13.73% 14.78%16.06%

24


The impact of funding ratio on the surplus frontierConstrained and unconstrained

2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0%

Surplus Standard Deviation

Su

rplu

s R

etu

rn

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

-2.5%

-2.0%

-1.5%

-1.0%

-0.5%

A/L= 100%

A/L = 75%

A/L = 50%

Unconstrained

Constrained

25


Asset return distributions: a reminderLeft: Asset “tulip,” minimum surplus variance portfolio (mostly fixed income, asset beta .405)Right: Asset “tulip,” 100% equity portfolio (asset beta 1.44)

$100

$1,000

$10,000

0 5 10 15 20

Years

0 5 10 15 20Years

$100

$1,000

$10,000

Seeking higher returns (steeper slope) means accepting a wider distribution of ending wealth

95%

75%

Mean

Median

25%

5%

26


Current policy: 75% Equity-like asset classesFunded ratio (A/L) distributions, over time

Current holdings modestly improve the expected funded ratio over time but imply an unattractive downside scenario

This and all other forecasts are focused exclusively on financial risk/return tradeoffs and exclude the impact of future cash flows and all other unhedgeable risks such as mortality risk and other experience risks.

Exp

ecte

d f

un

din

g r

atio

(A

/L)

260%

176%

134%

102%

69%

95%

75%

50%

25%

5%

Time (years)

50%

100%

150%

200%

250%

0 2 4 6 8 10

105%

27


Current policy: 75% Equity-like asset classesAn equivalent picture, but showing impact in dollars of surplus space

Exp

ecte

d d

oll

ars

of

su

rplu

s ($

b)

Time (years)

0 2 4 6 8 10

$25.6b

$13.4b

$6.8b

$0.6b

-$9.6b

Investments paying

for the planInvestments causing

higher contributions

30

20

10

0

-10

$0.3b

95%

75%

50%

25%

5%

28


Improving investment policyImpact of extending duration to hold a proper Hedging Porfolio

Exp

ecte

d s

urp

lus

retu

rn


0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%


Case 5 (75% Equity)

Case 4 (60% Equity)

Case 3 (45% Equity)

Case 2 (30% Equity)

Current Policy

0% 1% 3% 5% 7% 9% 11% 13% 15%

Extending the dual durations of the bond portfolio improves the expected surplus returns and reduces the expected volatility of the surplus Expected surplus return = 3.32% Expected surplus risk = 10.55%

29


95%

75%

50%

25%

5%

Impact of extending durationFunded ratio (A/L) distributions over time

Time (years)

Exp

ecte

d f

un

din

g r

atio

(A

/L)

224%

166%

135%

110%

82%

260%

176%

134%

102%

69%

(Current policy in background)

50%

100%

150%

200%

250%

105%

0 2 4 6 8 10

30


Improving investment policyImpact of reducing stock-bond mix to 60/40, in addition to extending duration

Exp

ecte

d s

urp

lus

retu

rn


0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%


Case 5 (75% Equity)

Case 4 (60% Equity)

Case 3 (45% Equity)

Case 2 (30% Equity)

Current Policy

0% 1% 3% 5% 7% 9% 11% 13% 15%

Reducing the stock-bond mix to 60/40 (from 75/25) reduces the volatility of surplus but has minimal impact on the surplus return Expected surplus return = 2.68% Expected surplus risk = 8.32%

31


95%

75%

50%

25%

5%

Impact of extending duration, changing stock-bond mixImproved funded ratio (A/L) distributions over time

Time (years)

Exp

ecte

d f

un

din

g r

atio

(A

/L)

193%

152%

129%

110%

87%

260%

176%

134%

102%

69%

250

200

150

100

50

105%

(Current policy in background)

0 2 4 6 8 10

32


Improving investment policyImpact of all changes (extending duration, reducing stock-bond mix, and incorporating active management)

Exp

ecte

d s

urp

lus

retu

rn


Case 1(Minimum surplus risk, 4% Equity)

Case 4 (60% Equity)

Case 3 (45% Equity)

Case 2 (30% Equity)

Current Policy

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0%

Case 4A (60% Equity, including alpha)

0% 1% 3% 5% 7% 9% 11% 13% 15%

A 1% expected alpha overlay (with 1% expected active risk) on the entire plan improves expected surplus return with only a small change in surplus risk Expected surplus return = 3.68% Expected surplus risk = 8.38%

Case 4 (60% Equity)

Case 5 (75% Equity)

33


50%

75%

100%

125%

150%

175%

200%

225%

Impact of adding active managementFunded ratio (A/L) distributions over time

This and all other forecasts are focused exclusively on financial risk/return tradeoffs and exclude the impact of of future cash flows and all other unhedgeable risks such as mortality risk and other experience risks.

Exp

ecte

d f

un

din

g r

atio

(A

/L) 193%

152%142%

120%

95%

212%

167%

129%

110%

87%

Time (years)

0 2 4 6 8 10

105%

95%

75%

50%

25%

5%

(Case 4 in background)

34


95%

75%

50%

25%

5%

Impact of all changes (extending duration, reducing stock-bond mix, and incorporating active management)Funded ratio (A/L) distributions over time

(Current policy in background)Time (years)

0 2 4 6 8 10

Exp

ecte

d f

un

din

g r

atio

(A

/L)

50%

100%

150%

200%

250%

212%

167%

142%

120%

95%

260%

176%

134%

102%

69%

105%

This and all other forecasts are focused exclusively on financial risk/return tradeoffs and exclude the impact of of future cash flows and all other unhedgeable risks such as mortality risk and other experience risks.

35


How much tolerance for investment risk?The integrated corporate balance sheet

Corporate “T-Account”

S/H Equity

Operating assets Debt

.

Pension assets Pension Liability

If the pension assets are a large part of the total assets, beta risk from the risky asset exposure will have a large effect on S/H equity beta risk

So, bad investment experience in the plan may compound an otherwise bad period for the company

The weighted avg. beta of the assets = weighted avg. beta of the liabilities

36


Risk Tolerance: How much equity?What does bad investment experience mean to you?

To hold more equity increases the expected or average return, but it also increases the cumulative probability of very bad returns over long periods of time Equivalently, contributions and expense can be expected on

average to be smaller, but the probability that they will be larger does go up

Can you afford greater contributions, expense when markets are generally depressed?

Enterprise view: Today’s 75% equity allocations are probably going to come down

37


Exceeding the limits of the possible: Do you feel lucky with your aggressive investment policy?

• Too much pressure is placed on the possibility of getting extraordinary, high returns, in order to solve funding problems

• But “feeling lucky” is a poor substitute for fair expectations, risk control

38


Conclusions: Don’t kill the golden goose!

Prescription: Use core teachings from modern portfolio theory to control true pension funding risks and costs

Pension funding risks are manageable, with 3 tools: The Risky Asset Portfolio: Surplus efficient frontiers help manage equity,

or market, risks The Hedging Portfolio: Dual duration management techniques manage

both types of interest rate risks (real interest rate, and inflation) Active management, skillfully employed, can significantly improve

surplus performance

Costs can be managed if benefit levels are negotiated using economic measures of the liability If costs are controlled, then contributions and expense are also controlled

DC plans are not a good substitute!

39


Biography

M. BARTON WARINGManaging DirectorChief Investment Officer for Investment Policy & Strategy, Emeritus

Barton Waring ran BGI’s Client Advisory Group from 1996 until his recent decision to retire. His research and published articles on investment policy and strategy issues have significantly contributed to the ability of today’s investors to control their risks and enhance their returns, in both beta and alpha dimensions. While most of his client work has been for BGI’s “strategic” clients, the largest of the world’s institutional investors (defined benefit retirement plans, foundations, endowments, social security systems, and central banks), it was also often directed at the needs of individuals in their personal and defined contribution retirement plan accounts. He has published over two dozen articles on surplus asset allocation, manager structure optimization and risk budgeting, as well as many on defined contribution/individual investor investment strategy. Four of these articles have won “outstanding article” awards from their respective journals, and these and many others are widely cited as setting the bar for today’s standards of practice. He serves on the Editorial Advisory Boards for the Journal of Portfolio Management, the Financial Analysts Journal, and the Journal of Investing.

His background prior to BGI also dealt intensively with classical investment strategy and policy issues. He was the manager of the specialist investment strategy consulting firm Ibbotson Associates, co-leader of Towers Perrin’s asset-liability practice and the head of its Central and Western regional asset consulting practices. He started and led the original defined contribution business for Morgan Stanley Asset Management in 1992, implementing the lifestyle fund concepts that he pioneered in 1989 and which he has written about frequently. Barton received his BS degree in economics from the University of Oregon, his JD degree from Lewis and Clark, with honors, and his masters degree in finance from Yale University.

29-C.ppt

Documents

benefit pension plans

pension expense

surplus beta

terms of surplus

pension funding risks

pension funding crisis

surplus duration duration

liabilitymatching asset