ECM Procedure

The copula simulation data table is expected to contain the output of the copula simulation at a uniform scale. It must contain all the marginal variables that you want to analyze. Each observation contains the probability of a loss value for each of the marginal variables; this probability is expected to be generated by accounting for the dependence structure between the marginal variables.

This is a required data table; you must specify it in the DATA= option in the PROC ECM statement.

The copula table is usually generated by specifying the OUTUNIFORM= option in the SIMULATE statement in the CCOPULA procedure for a dependency structure that is either defined by you (in the DEFINE statement in PROC CCOPULA) or estimated by PROC CCOPULA (according to the FIT statement in PROC CCOPULA).

The marginal data tables that you specify in the MARGINAL statements are expected to contain an empirical sample of the corresponding marginal variable. This is a required data table; you must specify it for each marginal that you want to analyze.

The empirical sample must appear in a variable that has one of the following names:

If a marginal table contains a variable named _DRAWID_, then the table is assumed to contain multiple empirical samples of the marginal variable. The value of the DRAWID= option in the MARGINAL statement or the value of the DRAWID= option in the PROC ECM statement determines which sample to use for the analysis.