Private clients buy performance, it is said. This is not what they buy from Fowler Drew. Once they see what we’re doing it becomes obvious they are buying something to do with the satisfaction their money can provide, not performance. Test drive The Planner yourself and you will see why. Like our clients, you might never view performance in the same way.
The liability-driven difference
Organising private wealth on the basis of individual goals with defined outcomes is the equivalent of liability-driven investing in the institutional world. Though it’s happening decades later, and only in a few firms so far, it marks a sea change in how portfolios are designed and managed and who has credit for what decisions. When the portfolio is defined by the nature, horizon and quantum of the client’s liabilities, together with the client’s required certainty of outcome, performance can no longer be a race or a competition.
With liability-driven investing, as applied in pension funds and insurance companies and as we apply it at Fowler Drew, the assets fall into one of two categories:
- Hedging assets that near-perfectly match the amount and time horizon of the liabilities
- Return-seeking or risky assets that may or may not meet the liabilities in full or exceed them.
Obviously, clients are indifferent to the performance of the hedging assets along the way: all that matters is the value at maturity, when the money is needed to meet the liability. That is the definition of a hedge. Performance matters for risky assets because the outcome does depends, to some extent, on the path of returns along the way. However, it is still outcomes that matter most, as this is where the problems arise if there is a shortfall and additional benefits materialise if a surplus. So what a liability-driven, or outcomes-driven, approach does is shift the client’s focus onto the outcomes of a plan as a whole, and on the consequences, which only they can visualise, of better or worse outcomes.
The path of returns matters only for the risky assets and only to the extent it can effect the outcomes. The main information in actual performance of the risky assets is therefore how it affects projections of outcomes for the combination of risky and risk free assets. This is usually most meaningful if expressed as a projected surplus or shortfall over liabilities: do you have more or less money than required to satisfy the objective?
If the mix of hedges and return-seeking assets is dynamic, changing with age and therefore the duration of the plan, the effects of past returns on projected shortfall or surplus may well be less important that the assumptions about future path effects when ‘derisking’, or moving money from risky to hedged. This dynamic approach to the split between hedging and return seeking is a feature of Fowler Drew portfolios where the plan duration will shorten with time, such as a ‘drawdown’ portfolio funding a spending goal. The outputs of probabilistic modelling are used both to manage the assets and to reproject the shortfall or surplus. This is what The Planner demonstrates.
Performance in a goal-based portfolio
A Fowler Drew goal-based portfolio is virtual: a combination of real holdings in different accounts with (for a couple) different legal owners that are assigned at outset to a single, highly-specified goal. They are then managed as a distinct portfolio to deliver the planned outcomes – equivalent to the liabilities. For instance, two spouse’s SIPPs, their ISAs and part of their general investment accounts are assigned to a spending goal, as required, while the balance, as surplus, is assigned to an intergenerational or ‘bequest’ goal. The spending goal is defined by cash flows: savings in and draw out. The surplus goal is defined by certainty of real value at a rolling horizon.
Performance reporting is only relevant at the level of the virtual goal-based portfolio, not the individual accounts (though we report both).
- It is mainly explained by the high-level separation of each portfolio into a risky or return-seeking allocation and a risk free allocation
- It is explained to a lesser extent by the mix of individual equity markets in the risky component
- It is barely explained at all by selection as we only use index funds.
Collectively, asset allocation across all Fowler Drew clients (and therefore ‘firm’ performance) mathematically results from:
- The random mix within the firm of spending plans (defined by yearly planned outcomes) and intergenerational capital
- Within spending plans: the random joint effect of their weighted-average duration (capital assigned to funding planned outcomes in discreet years) and weighted-average risk approach
- Within intergenerational goals: the client’s horizon and risk approach together typically lead to 100% in risky assets.
Across the firm, portfolio duration for different goals mainly varies between 7 and 50 years. This explains more of the allocation differences than the risk approach. But within any 10-year duration band, other than the nearest and furthest, risk differences explain more of the allocation differences. Obviously, some of these entirely customised portfolios could end up being very similar in asset allocation and return earned, but only randomly.
Implications of firms marketing ‘their’ performance
A wealth manager promoting its firm-wide investment performance, or ‘model portfolio’ performance, is effectively treating clients’ performance as the performance of the firm. It cannot then be fully customising portfolios and certainly not based on specific liability duration or unique multi-faceted risk preferences.
If firm performance matters, it must introduce an agenda difference with clients. The firm sees itself as being in a performance contest with other firms and so is no longer ‘disinterested’ in its client’s preferences and trade-offs because these affect the contest.
For most of my career managing Defined Benefit pension funds, managers competed on the basis of a performance contest. The innovation of liability-driven investing effectively put paid to it.
Equities as the common denominator
In a Fowler Drew portfolio, risky assets are formed of a geographically-diversified mix of equity markets. This 100% equity portfolio forms all of most of our intergenerational goals. It is also a component of all spending plans (later years being 100% equity and middle years a mix of equities and risk free assets: cash and index linked gilts). This equity component is therefore representative of firm-wide selection. And we do measure it.
We maintain a composite performance time series, starting in 2007, for all 100% equity portfolios that hold entirely ‘model’ assets (as opposed to a combination of model and pre-held legacy assets). There is little return dispersion within the composite because the management approach is systematic and the holdings are index funds. (The quarterly range from best to worst of 0.17% to 3.55% arises mainly because of differences in the valuation point for different index funds and ETFs and disappears over long periods.) Though adjusted for cash flows, there are not many for this composite.
You can see the latest equity composite performance here.
In a separate article we consider how actual performance should be benchmarked in goal-based portfolios with defined outcomes, so it clearly differentiates between the contribution of the client, based on how they defined the goal, and the contribution of the manager. Discretionary wealth managers are required by regulation to measure and report quarterly returns and compare them with an ‘appropriate’ benchmark, usually an index that represents the underlying investments. Never obvious, it is particularly challenging with a liability-driven investment approach.