28- DEVELOPING AN INVESTMENT STRATEGY (2)
Active and passive investment management
Active – A method where the investment manager has few restrictions on investment choice.
This method is expected to produce greater returns despite extra dealing costs /
research costs and risks of poor judgment.
Producing greater returns will not be possible if the investment market is efficient.
Thus, active investment management is appropriate only if the investor believes that
the investment market is in fact inefficient.
Passive – This involves holding assets closely reflecting those underlying an index or
specified benchmark. The investment manager has little freedom of choice.
There remains the risk of tracking errors occurring and the possibility of a poorly
performing index or benchmark.
Measuring active risk
Active risk is most commonly measured as a tracking error.
Historic tracking error is the annualized standard deviation of the difference between
actual fund performance and benchmark performance (a.k.a backwards-looking
tracking error).
Forward-looking tracking error (a prospective measure) involves modelling the
future experience of the fund based on its current holdings and likely future volatility
and correlations to other holdings.
The forward predictions are based on volatility & correlation data that is derived from
past performance. Hence there is an element of backward-looking here too.
Active Money Position
Another measure of active risk is the active money position. This is the difference
between the actual position and the benchmark portfolio. For example, if a manager
holds 2% of the portfolio in Welcome shares when the benchmark index holds
1.5%, then the active money position for that share is 0.5%.
Risk budgeting
Risk budgeting is a process that establishes how much risk should be taken and where it is
most efficient to take the risk (in order to maximize return).
With regard to investment risks, the risk budgeting process has two parts:
1. Deciding how to allocate the maximum permitted overall risk between active risk and
strategic risk
2. Allocating the active risk budget across the component portfolios (equities or bonds)
Matching
In its purest form matching of assets and liabilities involves structuring the flow of income
and maturity proceeds from the assets so that they will coincide precisely with the outgo in
respect of the liabilities.
Common problems with the precise matching of assets to liabilities in practice include:
Uncertainty in the timing and/or amounts of either assets or liabilities
Assets of long enough term may not exist
Income from the assets may exceed liability outgo in the early years
Actuarial techniques – asset-liability models
An asset-liability model can be used to help set an investment strategy in line with a stated
objective.
The objectives should include:
A quantifiable and measurable performance target
Defined performance horizons
Quantified confidence levels for achieving the target.
A single deterministic run may be helpful for showing whether the assets are appropriate by
term, but it does not show whether the assets are of the right type, i.e. fixed or real.
To do this, we would need to carry out a large number of re-runs of the model based on many
different sets of assumptions.
A stochastic model allows for the random nature of some of the model parameters. If the
assumptions underlying the model are realistic, then a clearer picture of the appropriateness
of the assets is possible.
Non-actuarial techniques
Other techniques for determining an investment strategy are:
1)
Mean-variance optimization without reference to the liabilities provides a method of
determining the efficient frontier for a particular investor.
This is the set of portfolios that provide the lowest variance of returns for a given level of
expected return, or equivalently the highest level of expected return for a given variance.
Assumptions:
Investors characterize investments purely in terms of their expected returns and
variances of returns over a single-period time horizon.
Ignore all other factors, including liabilities, when selecting investments.
The extension of portfolio theory to take account of an investor’s liabilities is straightforward.
Instead of the return on a portfolio of assets at the end of a single period or, equivalently the
value of the portfolio, we can consider the size of the excess of assets over liabilities i.e. the
surplus. This can be written as:
n
S= A ∑ xi ( 1+ Ri ) −L
i=1
where:
S is the surplus at the end of the period
A is the value of the assets at the start of the period
x iis the proportion invested in security i
Ri is the return on security i
L is the projected value of the liabilities at the end of the period.
Mean-variance portfolio theory can then be applied to minimize the variance of the
surplus for a given expected return, treating the liability as a negative asset.
In practice it will be necessary to decide how to place values on the liabilities and to
determine, not only the expected value of the liability at the end of the period, but also its
variance and co-variances with the assets. One way of doing this is to use a stochastic asset-
liability model.
2)
Basing asset allocations on market capitalizations, i.e. Index-Tracking
This is what index tracking does within each particular asset market.
Here the investment performance should broadly track that of the underlying investment
market(s).
3)
Shadowing the strategies of other comparable institutional investors:
Instead of tracking the index, the aim is to match the performance of your competitors. Such
commercial matching will reduce the potential for both significant under performance and
over performance compared to your competitors, i.e. reduce the relative performance risk.
Liability hedging
Liability hedging is where the assets are chosen in such a way as to perform in the same way
as the liabilities.
In other words, hedging against (or matching) all of the unpredictable changes in the
liabilities that arise from unpredictable changes in the factors that influence liability values,
e.g. interest rates, inflation levels.
Approximate liability hedging
In most situations, hedging liabilities with respect to all factors that affect liability values will
not be possible. In such circumstances, the investor might try to hedge its liabilities with
respect to specific factors that affect liability values.
Full liability hedging
In full liability hedging, assets are chosen in such a way that they can hedge unit-linked
liabilities.
In most cases the problem is “solved” by establishing a portfolio of assets, determining a unit
price by reference to the value of the asset portfolio, and then using this price to value units
held, allocated or realized.
The value of the liabilities is then said to be implied by the values of the assets.
A problem arises if the assets held are not (or cannot be) the same as those underlying the
value of the liabilities.
Immunization
Immunization is the investment of the assets in such a way that the PV of the assets minus the
PV of the liabilities is immune to a general small change in the rate of interest.
The purpose of immunization is the same as the purpose of matching
Immunization might be used in circumstances when pure matching is not possible.
Immunization relates to ensuring that the present value of assets is no less than that of the
liabilities, rather than matching the dates and amounts of the individual cash flows.
There are three conditions in classical immunization theory (according to Redington)
1. The [Link] of the liability-outgo and asset-proceeds are equal.
2. The (discounted) mean term of the value of the asset-proceeds must equal the mean
term of the value of the liability-outgo.
3. The spread (or convexity) about the mean term of the value of the asset-proceeds
should be greater than the spread of the value of the liability-outgo.
There are a number of theoretical and practical problems with immunization:
1. Immunization is generally aimed at meeting fixed monetary liabilities. Even if index
linked bonds are used to match index linked liabilities, still there would be time lag
associated with indexation.
2. Immunization removes mismatching profits apart from a second-order effect. It also
rules out investment in assets with uncertain returns such as equities and property.
3. The theory relies upon small changes in interest rates. The fund may not be protected
against large changes.
4. The theory assumes a flat yield. In practice the yield curve changes shape from time
to time.
5. In practice, the portfolio must be constantly rebalanced.
Need to achieve correct balance of:
Equal discounted mean term
Greater spread of asset proceeds
This is because the formulae for duration and convexity both depend upon the times
to each payment, which are continuously changing.
6. The theory ignores dealing costs.
7. Assets of a suitably long discounted mean term may not exist.
8. The timing of asset proceeds and liability outgo may not be known.
Portfolio construction
Portfolios are typically constructed to meet two objectives of:
1. Reducing risk
2. Achieving high long-term returns
The process of quantifying risk often involves dividing risk into:
Strategic risk – risk that the strategic benchmark performing badly relative to value of
the fund’s liabilities
Active risk – risk that the assets chosen by investment manager perform poorly
relative to the benchmarks given to the managers.
Structural risk – where the aggregate of the individual investment manager
benchmarks does not equal the total benchmark for the fund.
Pure matching is not always possible:
Assets / Liabilities not known with certainty (e.g. mortality / expenses unknown)
Investment and Reinvestment income unknown
Required assets not available (type or term)
Regulatory rules may inhibit pure matching
New business strain increasing expenses earlier on.