The EMA “Estema” estimation program makes it simple to estimate the
factor model underlying a set of securities returns using EMA’s proprietary
implementation of EM (“Expectation Maximisation”) factor analysis. Users can set
a wide range of variables and can build pure statistical or hybrid models.
Estema can be used for building factor models along the lines of the
Arbitrage Pricing Theory. These factor models can be used for modelling and
controlling portfolio risk. They can also be used to search for mispricing
opportunities by highlighting high residual returns. Arbitrage Pricing Theory
indicates that these residual returns may indicate mispricings that can
potentially be exploited with zero risk. The long-term average return can be
used to examine whether the Arbitrage Pricing Theory holds, and for identifying
buy or sell opportunities.
Key features
Limited data requirements – just (log) total returns
User defined universe, lookback period and frequency of observation
Variable time weights
Variable stock weights
EM algorithm in-fills missing returns data
Easy to use: interfaces with MS Access
OLE compliant – can be set to run from other applications
Please feel free to Contact EMA or telephone +44 (20) 7397 8395 for further details.
