Over the last two months in particular, I have become interested in the integration of Low Dimensional Factor Modelling with Structural Modelling (nowadays referred as DSGE or Dynamic Stochastic General Equilibrium (models)). You can see many references to my growing interest in this topic in this blog and in my corporate blogs over on; http://www.asymptotix.eu/blogs
The recent publication of the Spring Forecasts by the Directorate General, Economic and Financial Affairs (DG ECFIN) of the European Commission gives me an idea on how to show how this might be put into practice to support stress testing in a European Financial institution (as it were on a micro-prudential basis), I have blogged this idea as a comment on our corporate news item about the Spring Forecasts here;
A comment which would be particularly relevant here, is that on testing this idea in my REvolution R cockpit last evening (with one ear on the activities at Stamford Bridge!) there develops a certain rule set around the kind of LDFM approach you need to estimate economic risk capital based upon the use of series from a model such as ECFIN's AMECO; your models need to be; non-linear, dynamic & over as long a time period as possible (observation frequency is less important, my current testing is telling me) but the challenge is how do you integrate a monetary aggregate into the essentially real-economy oriented DSGE variables, which monetary aggregate? Well for that you need to go back to the theory and pick the approach which best fits your view from all of the papers I have blogged here. Simple really. We will make this mainstream yet! [HINT: Term Structure & as aggregate an M-number as you can find] Oh yes and a Stock Exchange Index helps, I like the DAX; I like the DAX as a global indicator! Problem then is how low is low?