Portfolio Sensitivity-2

RiskManager wants to know to which RiskClass his Portfolio is more sensitive to. In order to decide that we can use linear regression. Linear Regression will help us get the weightage we have to give while deciding sensitivity.

Why?
   RiskManager wants to give weightage to RiskClasses based on its impact.

What?
  LinearRegression gives the Weightage(interms of slope coefficient) to very RiskClass.

How?
  By using PROC REG of SAS.

Example:-

DATA PORTFOLIO_EXPOSURE;
INPUT ID  Portfolio $ RiskClass1_Change  RiskClass2_Change  RiskClass3_Change  RiskClass4_Change  Portfolio_Change;
DATALINES;
1  P1  1.0 -2.0 -2.0 0.0   2.41
2  P2 -1.0  4.0  5.0 0.10  1.21
3  P3  4.0 -2.0 -1.0 0.01  3.12
:
;

PROC REG DATA=PORTFOLIO_EXPOSURE;
MODEL Portfolio_Change= RiskClass1_Change  RiskClass2_Change  RiskClass3_Change  RiskClass4_Change;
RUN;

Result:-

Parameter Estimates
Variable	DF	Parameter Estimate	Standard Error	t Value	Pr > |t|
Intercept	1	1.58537	1.06256	1.49	0.1596
RiskClass1_Change	1	0.03163	0.12608	0.25	0.8058
RiskClass2_Change	1	-0.22882	0.31150	-0.73	0.4756
RiskClass3_Change	1	0.16831	0.23670	0.71	0.4896
RiskClass4_Change	1	0.18083	1.66170	0.11	0.9150

Conclustion:-
Portfolio Returns are more sensitive to  RiskClasses RiskClass4 & RiskClass2.