The US Trading Model forecasts equity portfolio risk over daily horizons. First released in 1997

Guy Miller compares Fundamental, Statistical, and 'Hybrid' Equity Factor Risk models. He discusses when the different types work best and when they are likely to fail in risk management and portfolio construction. When statistical factors are used to extend a fundamental factor model, we see modest improvements in risk forecasting. The improvement in portfolio optimization seems even slighter and should be applied only with caution

IFRS Accounting and its Effect on AUE3

Strongly variable risk levels are common in emerging equity markets, and complicate modeling their risks.  Applying daily index returns to a model through the DEWIV methodology often enhances the quality of market risk forecasts — DEWIV has long been a feature of models for developed markets such as Japan and the UK.  Our research indicates that in about half of the 20 emerging markets for which we could obtain daily index returns, implementing DEWIV significantly improved...

Since specific risk is the only risk that can be reduced through diversification, it is crucial to obtain an accurate specific risk forecast.  Stock pickers target specific returns, and the active return on which they base their businesses primarily bears specific risk. Specific risk models are especially vulnerable to misforecasting during periods of rapidly changing risk. Are there specific risk models that can follow such large variations and produce forecasts that do not mislead?...

Since the collapse of the Internet bubble, many Japanese portfolio managers have observed a surprising contrast between trends in tracking error and market volatility: tracking errors have fallen dramatically for many portfolios, while the volatility of the TSE1 index has declined much more gradually.  The decrease in tracking error is related to a phenomenon occurring in markets around the globe.  The cross-sectional dispersion of asset returns within these markets is much smaller...

CHE2 forecasts risk in portfolios of mainland Chinese equities - i.e., in portfolios composed of A- and B-shares. Its predictions utilize daily returns data, so that the model responds quickly to changes in the dynamic Chinese risk environment.

This report introduces the Japan Short-Term Equity Model, JPE3S, a model for near-term (several month) Japanese equity risk. JPE3S responds more quickly to changes in risk levels than the Japan Equity Model, JPE3. In order to make more responsive risk forecasts, JPE3S employs daily returns data and accounts for their serial correlations. Daily data provide denser and more detailed intra-horizon volatility information than would be available from monthly returns, and allow the model to base...

A global model that forecasts risk for portfolios with holdings across several markets will typically disagree with the predictions of a model specifically adapted to a single market. Given a global model and a collection of single market models, we describe an optimal, consistent way to embed the single market forecasts into the global model. The method involves framing the problem as an optimization over the ortogonal group O(n).