General Conclusions

1
General Conclusions

• The standard pension funds approach to asset
  allocation is incomplete
• Ignores interest rate risk - a large
  uncompensated source of volatility
• In the search for higher returns, over-allocates
  to equities, resulting in portfolios with lower
  reward/risk expectations than are readily
  obtainable
• Limits the use of leverage, resulting in
  suboptimal capital allocations, both to beta and
  alpha sources
• Tend to concentrate alpha sources in the same
  locations as beta sources
• Makes only modest attempts to budget risk
• Pension plans would be better served by
• Hedging interest rate risk to a desired level
  using capital efficient futures/swaps
• Independently determining
• The optimal exposure to beta sources the
  maximum Sharpe Ratio portfolio
• A desired level of overall total fund volatility
• Replicating the optimal beta portfolio
  synthetically using the minimal amount of capital
  necessary
• Allocating scarce capital to alpha-generating
  active managers (if identified), offsetting any
  unintended beta exposures in the process

2
Modern Investment Management

• Pension plans face three primary types of risk in
  their investment programs
• Interest rate
• Beta (systematic)
• Active
• Modern investment management plays two roles in
  quantifying and controlling exposures to these
  risks
• Creates a risk budget so as to more efficiently
  allocate capital to attractive sources of returns
  and eliminate unintended (and uncompensated)
  risks
• Monitors and manages the investment program at
  the total plan and manager levels so as to
  efficiently implement the risk budget
• Risk should be considered to be a scarce resource
  and carefully allocated

Unique problem for each sponsor
Same problem for all sponsors
3
Typical Sponsors Investment Program

• Focus is almost entirely on asset-only
  optimization
• Liabilities will take care of themselves in the
  long-run
• Determine the long-term policy asset allocations
  to equities and other asset categories
• Policy allocation reflects the risk-reward
  preferences of the sponsor
• Select asset category targets specify
  risk/return characteristics
• Conduct some sort of heavily constrained implicit
  optimization
• Rebalance back to policy targets periodically
• Determine the level of active risk appropriate
  within each asset class
• Select investment managers within each asset
  class based on their perceived ability to add
  value relative to an appropriate benchmark
• Allocate to managers within each asset class
• Achieve active management risk targets within the
  asset class
• Be aware of misfit risk and reduce it to
  acceptable levels (perhaps through use of a
  completeness fund)

4
An Alternative Approach

• Determine a surplus risk target
• Develop a framework for separating interest rate,
  beta, and active risk
• Minimize unproductive interest rate risk
• Determine an optimal combination of beta risk
  exposures
• Lever this optimal beta portfolio to reach a
  desired total beta risk level
• Identify and retain skillful active managers,
  regardless of asset category exposure
• Delink the source of manager alpha returns from
  manager beta returns and manage this beta
  exposure as necessary
• Adjust the optimal portfolios risk level as
  appropriate to accommodate the active risk level

5
Interest Rate Risk

• Interest rate risk is the exposure of the plans
  surplus to changes in interest rates
• Pension plan assets and liabilities are typically
  mismatched in terms of interest rate sensitivity
• The large allocation to equities and intermediate
  bonds causes plan assets to have relatively short
  durations on the order of 1-2 years
• Meanwhile, pension plan liabilities often have
  relatively long durations on the order of 10-15
  years
• The result is most pension plans are short
  duration
• Pension surpluses are therefore inversely related
  to changes in interest rates
• Implicitly, pension plans are betting on a rise
  in interest rates
• With their orientation toward asset-only
  optimization, most sponsors pay little attention
  to interest rate risk
• Focus is on long-run cost minimization, with
  large allocations to high return equity assets
• As interest rate risk is uncompensated, and with
  the advent of inexpensive risk management tools,
  sponsors should be positioned so as to be neutral
  to interest rate movements

6
Managing Interest Rate RiskAlternative Approaches
Likely Effect Likely Effect Approach On
Returns On Risk Comments
Do Nothing None None Leaves uncompensated risk
exposure in place Extend Duration Small Small
Most common approach of Bonds increase decrease
Slightly improved liability match Uncompensated
risk exposure remains Reduces opportunities
for active mgmt Increase Bond Decrease Decrease Re
duced equity allocation lowers Allocation
expected return Very large allocation
needed Increased credit risk Use
Derivatives Increase Decrease Leaves underlying
asset mix unchanged Overlay New asset with
positive expected return Improves risk-reward
profile of fund Introduces (risk-reducing)
leverage Issues regarding how gains/losses
perceived Cash flow timing issues
7
Thinking about Asset Mix in a World of Interest
Rate Risk Hedging and Optimal Beta Portfolios

• Determine the desired level of interest rate risk
  hedging done through swaps versus Hedge Assets
• Segregate Hedge Assets from Risk Assets
• Set a Risk Assets volatility target
• Establish an appropriate policy asset mix within
  the Risk Assets portfolio

8
Roles within Surplus Management
Pay benefits Manage interest rate risk
Liabilities
Hedge Assets
Pay benefits Reduce LT funding costs
Risk Assets Domestic equities Intl
equities Private equity Real assets
Manage interest rate risk
Equitize excess cash Maintain policy
weights among liquid risk assets
Swaps
Strategic Asset Allocation Overlay
9
Hedging Through Swaps vs. Bonds

• Swaps and Hedge Assets protect pension surplus
  against declines in interest rates
• In general, Hedge Assets will be high quality
  fixed income securities
• Swaps are the most cash efficient means to hedge
  interest rate risk
• Swaps provide the greatest flexibility to
  simultaneously invest in Risk Assets and mange
  interest rate risk
• Each dollar invested in Hedge Assets necessarily
  removes a dollar that could be held in Risk
  Assets
• This Hedge Assets-vs.-Swap decision is made in
  conjunction with the Risk Assets allocation
  decision
• If no Risk Assets are desired, virtually all
  hedging can be done with Hedge Assets
• If some Risk Assets are desired, then swaps may
  be required to move allocations from Hedging
  Assets to Risk Assets
• Reasons to hedge interest rate risk through Hedge
  Assets rather than swaps include
• Reduce counterparty risk
• Fewer custodial administrative issues
• Less exposure to discount rate/swap rate basis
  risk

10
Determine the Risk Assets Volatility Target

• All non-Hedge Assets are by definition Risk
  Assets
• The purpose of Risk Assets is to earn returns in
  excess of Libor so as to reduce the cost of
  funding future benefits
• A well-funded pension plan might choose to set a
  Risk Assets allocation of zero
• Essentially the plan invests all of its available
  assets in Hedge Assets, entirely defeasing its
  liabilities and seeking no incremental returns
• For example, if a pension plan were hard frozen,
  it makes little sense to have a Risk Assets
  allocation
• Risk Assets generate surplus risk
• The benefits of holding Risk Assets are long-term
  in nature, but in the short-run they can detract
  from surplus
• The sponsor must determine how much surplus
  volatility is acceptable over the near term
• In particular, the sponsor must have a sense of
  the maximum tolerable surplus erosion
• The desired surplus volatility could fluctuate
  with the funding level
• This surplus volatility target can readily be
  translated into a Risk Assets volatility target
  which will guide the selection of a policy asset
  mix within the Risk Assets

11
Establish a Risk Assets Policy Asset Mix

• Given the Risk Assets volatility target,
  allocations must be made to various types of Risk
  Assets
• All asset classes are potential holdings
• In aggregate, the allocation must be consistent
  with the Risk Assets volatility target
• The allocation should be expressed in terms of
  the contribution of each asset class to Risk
  Assets volatility
• The betas of the asset classes will dominate the
  contribution to the volatility of the Risk Assets
• Active management, if utilized, can also offer a
  contribution
• Generally, however, this contribution will be a
  second-order effect
• Traditionally, sponsors have held Risk Assets
  almost entirely in equity-related asset classes,
  largely US stocks
• As a result, surplus risk is dominated by
  equities
• Alternative approaches to setting the policy
  asset mix for Risk Assets involve a more
  diversified set of beta sources
• Appropriate levering/delevering is needed to meet
  the Risk Assets volatility target

12
Beta Risk

• Beta risk is the exposure of the plans surplus
  to systematic sources of return
• Unlike interest rate risk, a plan sponsor expects
  to be compensated for assuming beta risk
• Characteristics of beta sources
• Few in number
• Relatively correlated with one another (compared
  with alpha sources)
• Relatively low expected reward/risk ratios
  (compared with alpha sources)
• Excess returns have proven reliable over the
  long-run
• The cost of gaining naïve exposure to traditional
  sources of beta risk is generally low
• The attractive long-run equity risk premium has
  traditionally led sponsors to seek most of their
  beta risk from equities

13
Beta Risk Allocations
Asset Risk Asset Category Allocation Allocation
US Equity 36 52 Intl Equity 22 25 Private
Equity 5 14 US Bonds 23 4 Intl Bonds 2 0 Real
Estate 5 2 Hedge Funds 2 0 Natural
Resources 5 3 Total 100 100
91
14
Equities vs. Levered Diversified Portfolio1980 -
2005
Annual Annual Portfolio Compound Return Std
Deviation
SP 500 13.2 15.3 Lehman Aggregate 9.3 6.0 60/40
Stock/Bond Mix 11.9 9.8 Levered 60/40 14.7 15.3
Note The levered 60/40 mix also has preferable
skewness and kurtosis characteristics relative to
100 equities
155 stock/bond mix, -55 T-bills
15
Levering the Optimal Portfolio
Assume that a 60/40 stock/bond portfolio has the
maximum SR
Capital Market Line
Levered 60/40 Mix
Equities
Expected Return
Maximum Sharpe Ratio (60/40 Mix)
Sponsors seeking high returns tend to hold
portfolios with large equity allocations this
approach is suboptimal. A more diversified
portfolio, levered to a desired risk level, can
produce better expected reward/risk profiles.
Efficient Frontier
Risk Free Return
Standard Deviation
16
What is the Optimal Beta Portfolio?

• The optimal beta portfolio will likely be a
  well-diversified combination of beta risk sources
• A 60/40 US stock/US bond mix is likely not the
  optimal beta portfolio, as it contains no
  exposure to alternative beta sources
• Sponsor must select an appropriate level of total
  fund volatility and lever the optimal beta
  portfolio to achieve this risk target
• Note that leveraging a balanced collection of
  beta sources to a desired risk level produces a
  different form of risk than traditional
  equity-dominated portfolios
• Traditional portfolio risk is driven by the
  inherent volatility of equities
• Levered balanced portfolio risk is driven by the
  possibility that the beta sources in aggregate
  underperform the risk free return
• Once agreed upon, this optimal beta portfolio
  then becomes the base case for centralized beta
  management
• Capital allocations that generate beta will be
  neutralized to the extent possible
• All deviations should justify themselves on a
  risk-adjusted basis in terms of
• Expected illiquidity premiums (e.g., private
  equity)
• Expected alpha

17
Possible Components of an Optimal Beta Portfolio

• Equities
• Global developed, global emerging, US small and
  mid cap
• Nominal bonds
• Global developed, global emerging
• Inflation hedges
• Commodities, global TIPS
• Diversified credit/default risk premia
• US high yield spread, US corporate credit spread,
  global swap spreads, emerging currencies

18
Levering the Optimal Beta Portfolio
Assume now that a diversified combination of beta
sources, not the 60/40 portfolio, is the max SR
Capital Market Line
Levered Max SR
Equities
Expected Return
Unlevered Max SR
Note the unlevered Max SR portfolio can be
delevered to produce the same risk or expected
return as the 60/40 portfolio or an all-equity
portfolio. In both cases, the levered Max SR
portfolio has a higher expected return for the
same level of risk or less risk for the same
expected return.
Max SR Delevered
60/40 Mix
Efficient Frontier
Risk Free Return
Standard Deviation
19
Total Fund Completeness Strategy

• Begin with optimal beta portfolio
• Establishes optimal market risk position
• This portfolio is held in synthetic form to the
  greatest degree possible to facilitate
  centralized beta management
• Monitor the risk exposures of the sponsors
  active managers
• Partition the active managers risks into beta
  and active risks
• Hedge out the active managers beta risks either
  by
• Adjusting the composition of the optimal
  portfolio or
• Directly shorting the beta embedded in the active
  managers portfolio, using ETFs, futures, or
  swaps
• Combination of the two in order to
  cost-effectively take into account the
  correlations among the beta sources
• Presumably, some sort of centralized beta manager
  would oversee this process

20
Why Aint This Happening?

• Several important constraints
• Its different herd instincts
• Lack of robust risk models
• Legacy asset mix structure
• Focus on hedge funds and active management
• The current market blow-up will drive this
  structure forward
• Active management is at best a second-order
  effect
• The early 2000s taught us interest rate risks
  matter
• Diversified betas matter
• The number of managers offering interest rate
  risk management and optimal beta portfolio
  management is increasing