EV Asset Model

Group 11001

Background

The EV Asset Model is a state-space asset model of markets and economies that describes in probabilistic terms how possibly hidden, or indirectly observed states of the economy evolve. Underlying all asset class and economic forecasts is an interest rate model capable of handling the whole range of historical rates and the current low rate environment with high accuracy.

The model is an assumption setting tool so that capital market assumptions (CMAs) emerge from the model rather than being set as an input. This makes it extremely powerful for comparing projections and assessing their relevance to current circumstances and is one of the purposes of the model. The model generates and identifies risk premia which are essential for forecasts and especially for asset allocations.

Methodology

At a high level, the EV Asset Model describes the distribution of returns consumed by an economic scenario generator that samples from it. The modelling employs a Monte Carlo sampling process producing thousands of realistically correlated scenarios, both in aggregate and at an individual level. The model represents actual investment journeys making it robust for customisation and application to investments, e.g., projections and portfolio optimisation.

The EV Asset Model typically produces 10,000 scenarios that are updated monthly. As an integrated and consistent model, within each individual investment scenario, there are consistent relationships between the movements of different asset classes. It is a global multi-asset model and, as such, permits investment views across economies, in various currencies, over multiple investment horizons and on different bases (e.g. in nominal or real terms).

The role intended for ​the model​ is that of a complete and objective view of prospects for investments and other financial considerations. The model is economical, with a clear link between changes in market conditions and changes in outlook. This is facilitated by the data-driven approach taken, which minimises the need for human intervention.

Interest rate modelling

When it comes to investment forecasting (and by proxy, asset allocations), everything depends on interest rates, especially fixed income returns. Modelling interest rates is of fundamental importance when modelling return and risk over the long term. When rates increase, future return expectations for cash and bond returns rise. Moreover, to estimate the risk of bonds over different investment horizons, it is essential to consider how the adverse price impact of interest rate rises is offset by improved reinvestment returns over the time horizon of interest.

For these reasons, the EV Asset Model starts by modelling interest rates. Returns for cash and fixed income assets are directly determined by the evolution of interest rates. All other asset classes have some dependency on interest rates. We expect growth rates to be higher when interest rates are high, so equity dividend growth rates and property rent growth rates are linked to interest rates. Property rental yields are expected to depend on finance cost, so they are linked to interest rates. Interest rates also determine market expectations of future exchange rates, so currency strengths are related to interest rates.

Capturing term structure of risk

The link between short term and long term risk in practice is weak; this is particularly true of the dynamics of fixed-income investments. We capture their term structure of risk by correctly accounting for the term structure of interest rates. The interest rate model covers the whole range from the historical highs seen in the ‘70s and ‘80s to the current low-rate environment that has persisted for over a decade. It captures the term structure of cash and bond returns exceptionally accurately. In particular, the notion of cash risk growing faster with investment term than the risk of bonds and eventually surpassing it in a long time has clear implications for multi-asset allocations.

Stochastic volatility

Equity markets exhibit volatility clustering - long periods of low volatility interspersed with shorter periods of high volatility - and this is taken into account using stochastic volatility, which also captures the skew and fat tails characteristic of equity returns distributions.

Experienced team

We have a long track record of more than 10 years with almost £10 billion invested to asset allocations from the EV Asset Model and have forecast investment returns with enviable accuracy over this period. The model has consistently made correct calls in the lead up to recent financial crises and navigated them with great success.