Correlation, co-dependency and risk aggregation [16]

Go to: Summary | Previous | Next   
Bullet points include: Vast reduction in number of parameters that need estimating (if large universe). If m instruments, but just q factors (much smaller than m) then estimate only (q2+q+2*m)/2 not (m2+m)/2. Identifying factor structure may also help us understand better what drives aggregate (market) behaviour. So how might we identify factors? An entire risk model vendor industry involved in this activity

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile