TABLE OF CONTENTS In this chapter, we consider additional topics related to applications of VAR and CVAR for corporate-level risk management. We first discuss aggregation issues. We then discuss implied correlations between business unit P&Ls. We end with a consideration of aged inventory.
Large banks and broker-dealers have a complex internal structure involving hundreds of products dealt with on many desks. The desks are arranged in a hierarchy into business units and/or divisions [1]. For corporate risk management over this entire structure, it is necessary to aggregate the risks of the individual components. A plethora of problems or difficulties can arise, both technical and non-technical.
Some technical difficulties, not necessarily in order of importance and certainly not complete, include:
Data for time series: Availability, consistency, completeness etc.
Systems: Hundreds of feeds, thousands of variables, legacy issues etc.
Risk measures: Availability, timeliness, consistency, completeness etc.
Calculation: Level of sophistication, huge correlation matrices etc.
We have spent a fair amount of time in this book discussing these technical issues in some detail. Other formidable difficulties are non-technical, including budgets, priorities, time limitations, personnel, communication, sociology, etc. Moreover in this age of acquisitions and mergers, the corporate structure can change [2], requiring flexibility. Regulator requirements also exist that exert pressure. The bottom line is that these real-life issues can make corporate-level risk aggregation a gigantic, long, painful effort.
Conceptually there is no difficulty in writing down VAR aggregation. In Ch. 27,...
