The SSM Procedure

Functional Summary

Table 34.1 summarizes the statements and options that control the SSM procedure. Most commonly needed scenarios are listed; see the individual statements for additional details.

Table 34.1: Functional Summary

Description

Statement

Option

Data Set Options

   

Specifies the input data set

PROC SSM

DATA=

Writes series and component forecasts to an output data set

OUTPUT

OUT=

Model Specification Options

   

Specifies the index variable

ID

 

Defines variables as model parameters

PARMS

 

Specifies a response variable and the associated observation equation

MODEL

 

Specifies a state subsection

STATE

 

Specifies the transition matrix of a state subsection

STATE

T

Specifies the disturbance covariance matrix of a state subsection

STATE

COV

Specifies the size of the diffuse initial condition of a state subsection

STATE

A1

Specifies the initial covariance matrix of a state subsection

STATE

COV1

Specifies a state subsection for a predefined structural model

STATE

TYPE=

Specifies the regressors in a state equation

STATE

W

Specifies the input vector in a state equation

STATE

SINPUT=

Specifies a component

COMPONENT

 

Specifies a predefined trend component

TREND

 

Likelihood Optimization Process Control Options

 

Specifies the optimization technique

PROC SSM

OPTIMIZER(TECH=)

Limits the number of iterations

PROC SSM

OPTIMIZER(MAXITER=)

Outlier and Structural Break Detection Options

 

Turns on the search for additive outliers (AO)

 

Default

Turns on the search for structural breaks in a state subsection

STATE

CHECKBREAK

Turns on the search for structural breaks in a state subsection associated with a trend

TREND

CHECKBREAK

Specifies the significance level for additive outlier tests

OUTPUT

AO(ALPHA= )

Limits the reported number of additive outliers

OUTPUT

AO(MAXNUM= )

Limits the reported number of additive outliers to a percentage of the series length

OUTPUT

AO(MAXPCT= )

Specifies the significance level for structural break tests

OUTPUT

BREAK(ALPHA= )

Limits the reported number of structural breaks

OUTPUT

BREAK(MAXNUM= )

Limits the reported number of structural breaks to a percentage of the series length

OUTPUT

BREAK(MAXPCT= )

Turns on the search for maximal state shock

OUTPUT

MAXSHOCK

Graphical Residual and Outlier Analysis Options

 

Creates a panel of plots that consists of residual normality plots

PROC SSM

PLOTS=RESIDUAL(NORMAL)

Creates the standardized residual plot against time

PROC SSM

PLOTS=RESIDUAL(STD)

Creates a panel of plots that consists of prediction error normality plots

PROC SSM

PLOTS=AO(NORMAL)

Creates the standardized prediction error plot against time

PROC SSM

PLOTS=AO(STD)

Creates the plot of maximal state shock chi-square statistics against time

PROC SSM

PLOTS=MAXSHOCK

Output Control Options

 

Specifies the significance level of the forecast confidence limits

OUTPUT

ALPHA=

Prints the prediction error sum of squares table

OUTPUT

PRESS

Specifies a linear combination of components to be output

EVAL

 

Global Printing and Plotting Options

 

Turns off all printing for the procedure

PROC SSM

NOPRINT

Turns on all printing options for the procedure

PROC SSM

PRINTALL

Turns off all plotting for the procedure

PROC SSM

PLOTS=NONE

Turns on all plotting options for the procedure

PROC SSM

PLOTS=ALL

Printing State Equation System Matrix Options

 

Prints the transition matrix that is associated with a state subsection

STATE

PRINT=T

Prints the disturbance covariance matrix that is associated with a state subsection

STATE

PRINT=COV

Prints the initial covariance matrix that is associated with a state subsection

STATE

PRINT=COV1

Prints the autoregressive coefficient matrix that is associated with a state subsection

STATE

PRINT=AR

Prints the moving average coefficient matrix that is associated with a state subsection

STATE

PRINT=MA

Printing Component, Series Forecast, and Smoothed Estimate Options

 

Prints the series forecasts

MODEL

PRINT=FILTER

Prints the full-sample estimates of missing series values

MODEL

PRINT=SMOOTH

Prints the smoothed trend estimate

TREND

PRINT=SMOOTH

Prints the filtered trend estimate

TREND

PRINT=FILTER

Prints the smoothed component estimate

COMPONENT

PRINT=SMOOTH

Prints the filtered component estimate

COMPONENT

PRINT=FILTER

Prints the smoothed component estimate

EVAL

PRINT=SMOOTH

Prints the filtered component estimate

EVAL

PRINT=FILTER

BY Groups

   

Specifies BY-group processing

BY