Prerequisite for Systematic Tactical Process
The key to the success of a tactical asset allocation strategy is forecasting with a good degree of confidence the return for each asset class. At Performance Analytics, we use our proprietary PAR Model™ to forecast equity returns over a six-month period. It provides answers that investment managers need for a systematic process:
- With a forecasting model, skill can be measured using the Information Coefficient (IC). Without it, computing performance alone relies only on one market outcome.
- With a statistical model, factors are weighed consistently according to their coefficients. Without it, “indicator” weights are arbitrary.
- The model provides clarity – the return forecast, and the factors that determine it.
Below you can see an example of the relevant factors, and their contributions to the model’s return forecast.
PAR Model™ is based on dynamic multi-factor regression, using independent variables in three categories: economic, valuation, and market factors. Only variables that have been proven to be statistically significant in explaining stock returns over time are applied in the model. These include factors commonly found in related literature such as the S&P 500 P/E ratio and Industrial Production, but the set of factors is much broader (22 factors for our S&P 500 model, for example).
The model is dynamic – it adjusts to changing market conditions. Factors are retained in the final equation by automatic specification procedure, based on their significance.
The PAR Model’s™ six-month return forecasts for the S&P 500 can be seen in the chart below. The current version of the model has been in place since the beginning of 2012; earlier results are obtained in back-testing.