The theory of econometric modelling rests on the assumptions of the principles of rational or dynamic expectations building. Models basically bring statistics to the process of belief generation of individuals. A good model should capture past, present and future expectations and it is a continuous process meaning the agent has the ability to learn. In the field of finance agents form expectations of the returns they receive based on factors that they perceive to pose a risk on the assets capacity to generate positive net cash flow. Through learning, agents build beliefs on future price volatility based on past experience on the impact of shocks to these variables and their ultimate impact on the sustainability of future cash flow. Asset prices react to unanticipated news to the prespecified factors and this news is interpreted based on the agents based on past, present and future expectations. However this cannot be done in a vacuum as these prespecified macroeconomic variables behave differently depending on the exchange regime in which the shocks occur.The primary aim of the study is to investigate if the goodness of fit or the fraction of time series variance explained of prespecified macroeconomic factor APT models improves when the exchange regime is factored into the model, therefore re-defining the expectation and information transmission process of asset expected returns within an exchange regime framework.