EValue’s Insight asset model provides users with a rich understanding of the financial outcomes that may materialise in the capital markets, the risks involved and the investment opportunities currently available, optimising the outcomes of each user’s goals. This means using our robust model’s output in an accessible format appropriate for the investment time scale and objectives particular to each user’s circumstances. We make sure our model’s output is up-to-date and representative by incorporating current prices directly into our model so that it remains consistent to changes in the market.
Absolutely, change is an inevitable feature of investment markets. Our models are built to expect and reflect certain changes but there is always more. A significant example would be our implementation of a shadow interest rate model in response to the current low interest rate environment.
More broadly, we continuously refine our approach to embrace emerging patterns of market behaviour, where appropriate, to update any longer established patterns, risks and investment outcomes, which underpin our methodology
As we're modelling future behaviour of investment markets to demonstrate the range and likelihood of consumer outcomes, our model needs to reflect what could plausibly happen over various time horizons and not simply play back what has happened in the past.
Therefore, our Insight model produces an array of scenarios showing coherent and consistent forecasts across a set of asset classes for up to 75 years into the future. These scenarios provide projections of investment strategies in nominal, real or income terms in a range of currencies. Our forecasts also incorporate the evolution of asset price, income and total return as well as interest, inflation and exchange rates.
Updated versions of the model are released every quarter and, more frequently, if called for by market events or client requirements. In the event of an unscheduled update e.g. due to market events such as a market crash, an updated version of the model could be released in a matter of days.
The equity risk premium is term dependent, reflecting the actual expected behaviour in the market. The below graph displays the equity risk premium in excess of cash returns over time as of 1 April 2018. Projected returns generated by scenarios will vary around this expectation over the chosen duration for a given market.
The risk premium for fixed income returns is term dependent, reflecting the actual expected behaviour in the market. The below graph displays the fixed income risk premium in excess of cash returns over time as of 1 April 2018. Projected returns generated by scenarios will vary around this expectation over the chosen duration for each type of fixed income investment.
The various equity, bond (fixed interest), and other asset class assumptions are derived from a stochastic process, which means that we use a range of economic scenarios and show the possible outcomes from these. Unlike a deterministic forecast which uses the same, single value each year, we run 1,000 different scenarios to take into account all relevant economic changes during the investment term. Expected future returns for each asset class are then obtained from all of the scenarios and these returns will vary by term. All of our stochastic returns are updated quarterly to take account of changes to factors affecting the forecasts including current economic conditions.
As RPI is the basis of inflation in the market, we model RPI directly. For instance, if the current rate of RPI is around 4%, our model reflects this and projects a path of inflation from this starting figure. Our methodology looks directly at the difference in yields between fixed coupon and index linked government bonds to derive a rate of RPI inflation consistent with market expectations over future durations – e.g. above, below or remaining at 4% in 15 or 25 years' time. Whilst this anchors our inflation scenarios our modelling also allows for future circumstances where inflation is different to the central expectation implied by current market pricing.
We take extra care to interpret the level of RPI inflation expectations. In fact, historical data shows that market estimates of inflation consistently under-estimate actual inflation by approximately 0.5% a year. For this reason, our standard projections correct for this by building in a 0.5% uplift against the market expected RPI inflation.
The return on an asset class is made up of two components: the risk-free return and the risk premium for each asset class. EValue has adopted the widely used Black Litterman optimisation model, which uses risk correlations and market weightings to set consistent asset class risk premia. The process involves establishing a future “steady state” in which there is no incentive, in an efficient market, for switching between asset classes.
In such a “steady state”, the return on emerging market equity would warrant a higher return based on higher measured volatility, but a lower return based on its relatively small market capitalisation and the fact that emerging market equity is not highly correlated with other equity markets.
By comparing the return justified by the “steady state” analysis, with current dividend yields, we can obtain a measure of the relative over or under valuation of an asset class.
When it comes to property growth, our model assumes that property prices capitalise rental income. As a consequence, if interest rates rise in a particular scenario, so will the yield on property with a high correlation. So currently, for example, the combination of weak rental growth, high current prices and the prospect of rising interest rates mean that our model presents a relatively weak outlook for property compared to historic relativities.
As a simple summary, our UK equity risk premium has been calibrated to a little over 3% and, at the time of writing, the current 10-year real interest rate against RPI is negative at -1.53%. This means that, at present, a reasonable estimate for 10-year real returns from a UK equity is about 1.5%. Although equity premium levels are ultimately a matter of judgement, our current estimates are not outliers and we take comfort that our equity model has performed in line with market results over time.
EValue sets cash returns to be consistent with current bond markets at all terms, so both assets will be very similar to each other. If we didn't do this, our model would suggest that there are large, sure gains to be had by trading between the two, which would never hold for any reasonable period in reality. When expected returns are very low, as they have been over short durations for some time now, our model generates a greater number of low rate scenarios most showing only a small fall in rates, balancing rising interest rate scenarios..
Yes. EValue is committed to providing transparent communications of both our assumptions and workings of the Insight asset model.
A recent report detailing the assumptions and expected returns is available here and the most up-to-date versions are automatically provided to our clients.
Our asset allocations are derived directly from the returns output by the Insight asset model. Based on client requirements, we allow for a range of asset allocation strategies, use projected outcomes from each strategy from the asset model and choose the strategy which achieves the highest expected final fund value for a given risk target over the particular investment period. This process is repeated for a number of investment periods to establish optimised asset allocations for each time horizon. Each time the Insight asset model is updated, the best strategy will be rebased to the latest market conditions and communicated to clients in our allocation updates.
Short term goals
Long term goals
We work with our clients to agree what is most appropriate for their needs. Benchmark portfolios are typically identified consisting of asset allocations corresponding to the highest and lowest risk levels for the shortest and longest investment terms under consideration.
Once these are established, the level of risk for each intermediate risk category are also determined. Importantly, the gaps between the lower risk levels are typically smaller than those between the higher risk levels, because the efficient frontier is a curve rather than a straight line, with less difference between higher risks than lower ones. Also, the allocations between shorter and longer term investment horizons are not linear, reflecting the different behaviour of each asset class over time.
The age of an investor will drive the investment term, which applies to their circumstance. The asset allocations are then chosen subject to the appropriate investment term, which will also vary depending on an investor's goals.
We, therefore, stress that both the investor's age and their financial objectives will need to be taken into account before a final recommendation or decision is made on their most suitable investment choices.
There are many ways to derive an asset allocation given a number of asset classes and an investment term. The results produced depend on several factors, including the asset model used to generate the stochastic forecast, any limiting criteria (such as maximum levels of certain assets), the number and volatility of target risk levels, the assets chosen to be included and the term for the target investment. As all of these elements differ between suppliers, the asset allocations which are provided will also vary. We hope that this website and overview will help explain our approach but are always open to discuss with our clients how our core methodology can be adapted for specific purposes.
Not in our standard asset allocations since our model uses the most common asset classes to suit the majority of users. Commodities can have a wide range of characteristics and, therefore, a corresponding wide range of risk levels. Without knowing the actual commodity and market they are traded in, they cannot be accurately included in the forecasts. This does not mean that it is an unsuitable investment class, simply that our standard asset allocations do not explicitly include them.
We are, however, happy to include customised asset classes for specific client applications where we jointly agree the specific characteristics of the asset class being modelled. This can also be extended to other asset classes.
We are very sensitive to this concern. Heavy cash allocations could be a result of either taking a very strong short-term view or the difficulty of fitting a long-term view into market approaches such as a Mean Variance Covariance (MVC) model. Most calibration techniques for an MVC model tend to understate the uncertainty of cash returns in the medium to long term and must choose a compromise level of return which will likely overstate the level of return available. For example, a 15-year bond is quite risky in the short term but very low risk over 15 years while cash is the opposite. Failing to take this into account can produce inappropriate results. The Insight approach does not suffer from this drawback given our modelling approach allows for time horizons, from the very short term through to the long term, when considering asset returns.
Charges clearly directly reduce returns and can even affect some measures of volatility by dampening the range of outcomes from a fund. We, therefore, take into account an average level of charges for asset management costs (but not product charges or adviser fees) for each asset class to make the allocations as realistic as possible.
The most common asset categories are included (e.g. Money markets, Fixed Interest, Developed market equities, Emerging market equities, and Property)
The allocations do not include specialist investment products such as structured products and guarantees. Although it is possible to achieve the same risk target with different proportions of the assets, one recommended solution is given to optimise diversification effects in current market conditions.
The allocations assume that all funds in a certain asset class are modelled in line with the index in that particular class. Our clients, therefore, need to consider the extent to which specific funds might differ from the core assumption, even if they map to the high-level asset, e.g. small cap equities may have different characteristics to the main equity index. We are happy to discuss such differences and how they might be addressed for specific client circumstances, where appropriate.
Asset allocations are created without reference to product selection. Asset allocations are term dependent and are calculated for all future durations. For simplicity, these are grouped into wider term categories in our standard client output.
Risks for specialist fund mandates such as counterparty, liquidity and specific product or high-alpha seeking risks are not considered as standard. Diversification is only considered at the asset category level. For example, allocations do not take account of diversification within asset classes such as different holdings within an equity or bond fund.
The latest quarterly asset allocation reports for each risk set is available here.