The ECM Procedure

specifies the output data table that contains the summary statistics of the total loss sample. CAS-libref.data-table is a two-level name, where CAS-libref refers to the caslib and session identifier, and data-table specifies the name of the input data table. For more information about this two-level name, see the DATA= option and the section Using CAS Sessions and CAS Engine Librefs. The CAS-libref must be identical to the CAS-libref that you specify in the DATA= option.

You can control the statistics that appear in this data table by specifying the PCTLPTS= option, the TVARPTS= option, and different statistic-keyword values.

specifies the maximum number of decimal places to use when creating the names of the variables for the percentile values in the PCTLPTS= or TVARPTS= options. For example, for a percentile value of 99.9995, if you specify PCTLNDEC=4, then the suffix of the variable is 99_9995. Specifying a small number might result in duplicate variable names. If that happens, PROC ECM writes an error message and exits.

By default, PROC ECM computes the required number of decimal places by using the precision of the percentiles that you specify in the PCTLPTS= option or the TVARPTS= option.

specifies the prefix to use for creating the names of the variables that contain the percentile estimates. The percentile estimate for a value p is written to a variable that has the name of the form prefixt, where t is the text representation of p that is formed by replacing the decimal point with an underscore (_). For example, if PCTLPRE=P, then the estimate of the 97.5th percentile is written to a variable whose name is P97_5. By default, PCTLPRE=P.

specifies the percentiles of the total loss that you want PROC ECM to compute. The percentile-option can have one of the following two forms:

specifies the prefix to use for creating the names of the variables that contain the tail value-at-risk (TVar) estimates. The TVaR estimate for a percentile value p is written to a variable that has the name of the form prefixt, where t is the text representation of p that is formed by replacing the decimal point with an underscore (_). For example, if TVARPRE=TVaR, then the tail value-at-risk estimate for the 97.5th percentile is written to a variable whose name is TVaR97_5. By default, TVARPRE=TVaR.

specifies the percentile values for which you want PROC ECM to compute the tail value-at-risk (TVaR) of the total loss, where tvar-percentile-list is a comma-separated list of percentile values, each of which must belong to the (0,100) open interval. You can also use a list notation of the form "<number1> to <number2> by <increment>". For example, the following two options are equivalent:

tvarpts=90, 95, 96.25, 97.5, 98.75, 99.5
tvarpts=90, 95 to 99 by 1.25, 99.5

For each value in the tvar-percentile-list, PROC ECM also reports the percentile estimates, which are the estimates of the value-at-risk (VaR) for the corresponding TVaR estimate.

If you specify a data table in the OUTSUM= option, PROC ECM adds the TVaR estimates to that table. You can control the name of each TVaR estimate’s variable by specifying the TVARPRE= option.

writes the statistic that corresponds to the statistic-keyword to a variable named variable-name. For the predefined statistics that correspond to percentiles, the variable-name takes precedence over the default name that PROC ECM forms by using the PCTLPRE= option. All variable names in the OUTSUM= data table have a limit of 32 characters.

If you do not specify the variable-name, then PROC ECM writes the statistic to a variable named statistic-keyword.

You can specify the following statistic-keywords, which result in the specified statistic being written to the OUTSUM= table: