The X12 Procedure

PROC X12 Statement

PROC X12 options ;

The PROC X12 statement provides information about the time series to be processed by PROC X12. Either the DATE= or the START= option must be specified. If both options are specified, then a syntax error results and the X12 procedure is not executed.

The original series is displayed in Table A1. If there are missing values in the original series and a regARIMA model is specified or automatically selected, then Table MV1 is displayed. Table MV1 contains the original series with missing values replaced by the predicted values from the fitted model. If outliers are identified and Table A19 is added in the TABLES statement, then the outlier adjusted series is displayed in Table A19. Table B1 is displayed when the original data are altered (for example, through an ARIMA model estimation, prior adjustment factor, or regression) or the series is extended with forecasts.

Although the X-12-ARIMA method handles missing values, there are some restrictions. In order for PROC X12 to process the series, no month or quarter can contain missing values for all years. For instance, if the third quarter contained only missing values for all years, then processing is skipped for that series. In addition, if more than half the values for a month or a quarter are missing, then a warning message is displayed in the log file, and other errors might occur later in processing. If a series contains many missing values, other methods of missing value replacement should be considered prior to seasonally adjusting the series.

You can specify the following options in the PROC X12 statement:

AUXDATA=SAS-data-set

specifies an auxiliary input data set that contains user-defined variables, which are specified in the INPUT statement, the USERVAR= option in the REGRESSION statement, or the USERDEFINED statement. The AUXDATA= data set can also contain the date variable, which is specified in the DATE= option in the PROC X12 statement. If the date variable is present, then the date variable is used to align the observations in the auxiliary data set to the observations in the series that is being processed. The date values must be sorted in ascending order with no gaps or duplications, and the interval must match the interval of the series. If the date variable is not present or valid, then observations in the auxiliary data set are matched by observation number to the series that is being processed. The auxiliary data set does not support BY-group processing. The variables in the auxiliary data set are applied to all BY groups, where the dates of the BY group correspond to the dates of the auxiliary data set. Example 37.11 shows the use of the AUXDATA= data set.

DATA=SAS-data-set

specifies the input SAS data set to use. If this option is omitted, the most recently created SAS data set is used.

DATE=variable DATEVAR=variable

specifies a variable that gives the date for each observation. Unless specified in the SPAN= option, the starting and ending dates are obtained from the first and last values of the BY group for the DATE= variable, which must contain SAS date or datetime values. The procedure checks values of the DATE= variable to ensure that the input observations are sequenced correctly in ascending order. If the INTERVAL= option or the SEASONS= option is specified, the values of the date variable must be consistent with the specified seasonality or interval. If neither the INTERVAL= option nor the SEASONS= option is specified, then the procedure tries to determine the type of data from the values of the date variable. This variable is automatically added to the OUT= data set if a data set is requested in an OUTPUT statement, and the date values for the variable are extrapolated if necessary. If the DATE= option is not specified, the START= option must be specified.

INEVENT=SAS-data-set

specifies the input data set that defines any user-defined event variables. This option can be omitted if events are not specified or if only SAS predefined events are specified in an EVENT statement. For more information about the format of this data set, see the section INEVENT= Data Set.

INTERVAL=interval

specifies the frequency of the input time series. If the input data consist of quarterly observations, then INTERVAL=QTR should be used. If the input data consist of monthly observations, then INTERVAL=MONTH should be used. If the INTERVAL= option is not specified and SEASONS=4, then INTERVAL=QTR is assumed; likewise, SEASONS=12 implies INTERVAL=MONTH. If both the INTERVAL= option and the SEASONS= option are specified, the values should not be conflicting. If neither the INTERVAL= option nor the SEASONS= option is specified and the START= option is specified, then the data are assumed to be monthly. If a date variable is specified using the DATE= option, it is not necessary to specify the INTERVAL= option or the SEASONS= option; however, if specified, the values of the INTERVAL= option or the SEASONS= option should not be in conflict with the values of the date variable. See Chapter 4: Date Intervals, Formats, and Functions, for more details about intervals.

MDLINFOIN=SAS-data-set

specifies an optional input data set that contains model information that can replace the information contained in the TRANSFORM, REGRESSION, ARIMA, and AUTOMDL statements. The MDLINFOIN= data set can contain BY-group and series names. It is useful for providing specific information about each series to be seasonally adjusted. See the section MDLINFOIN= and MDLINFOOUT= Data Sets for details.

MDLINFOOUT=SAS-data-set

specifies the optional output data set that contains the transformation, regression, and ARIMA information related to each seasonally adjusted series. The data set is sorted by the BY-group variables, if any, and by series names. The MDLINFOOUT= data set can be used as input for the MDLINFOIN= option. See the section MDLINFOIN= and MDLINFOOUT= Data Sets for details.

NOPRINT

suppresses any printed output.

NOTRIMMISS

suppresses the default, by which leading and trailing missing values are trimmed from each variable listed (or implied) in the VAR statement. If you specify the NOTRIMMISS option, PROC X12 treats leading and trailing missing values in the same manner as it treats embedded missing values. For information about the treatment of embedded missing values, see the section Missing Values. Missing values are not supported in the regression variables that you specify in the REGRESSION, INPUT, or USERDEFINED statement; therefore, leading and trailing missing values are always trimmed from user-defined regressors even if you specify NOTRIMMISS.

OUTSTAT=SAS-data-set

specifies an optional output data set which contains the summary statistics that related to each seasonally adjusted series. The data set is sorted by the BY-group variables, if any, and by series names. See the section OUTSTAT= Data Set for details.

PERIODOGRAM

specifies that the PERIODOGRAM rather than the spectrum of the series be plotted in the G tables and plots. If PERIODOGRAM is not specified, then the spectrum is plotted in the G tables.

PLOTS<(global-plot-options)> <= plot-request <(options)>> PLOTS<(global-plot-options)> <= (plot-request <(options)> <... plot-request <(options)>>)>

controls the plots that are produced through ODS Graphics. When you specify only one plot request, you can omit the parentheses around the plot request.

Following are some examples of the PLOTS= option:

plots=none
plots=all
plots=residual(none)
plots(only)=(series(acf pacf) residual(hist))

ODS Graphics must be enabled before you request plots. For example:

ods graphics on;

proc x12 data=sales date=date;
   var sales;
   identify diff=(0,1) sdiff=(0,1);
run;

Since no specific plot is requested in this program, the default plots associated with the PROC X12 and IDENTIFY statements are produced.

For general information about ODS Graphics, see Chapter 21: Statistical Graphics Using ODS in SAS/STAT User's Guide. If you have enabled ODS Graphics but do not specify any specific plot request, then the default plots that are associated with each of the PROC X12 statements used in the program are produced. Line printer plots are suppressed when ODS Graphics is enabled.

If NONE is specified in an option, then no plots are produced for that option. If ALL is specified without NONE in an option, then all plots are produced for that option.

Global Plot Options:The global-plot-options apply to all relevant plots that are generated by the X12 procedure. The following global-plot-option is supported:

ONLY: suppresses the default plots. Only the plots specifically requested are produced.

Specific Plot Options:The following list describes the specific plots and their options:

ALL

produces all plots that are appropriate for the particular analysis.

NONE

suppresses all plots.

ADJUSTED(<sa-plot-options>) SA(<sa-plot-options>)

produces plots of the seasonally adjusted series that results from the decomposition specified in the X11 statement. The SPECTRUM plot is produced by default.

The following sa-plot-options are available:

ALL: produces all seasonally adjusted plots.
NONE: suppresses all seasonally adjusted plots.
SPECTRUM: produces the spectral plot of Table G1. Table G1 is calculated based on the modified seasonally adjusted series (Table E2). The data is first-differenced and transformed as specified in the TRANSFORM statement. By default, the type of spectral estimate used to calculate the spectral plot is the spectrum. If the PERIODOGRAM option is specified in the PROC X12 statement, then the periodogram of the series is used to calculate the spectral plot.

FORECAST(<forecast-plot-options>)

produces the regARIMA model forecast plots if the FORECAST statement is specified. The FORECAST plot is produced by default. The following forecast-plot-options are available:

ALL: produces all the forecast plots that are appropriate for the particular analysis.
FORECAST: plots the actual time series and its one-step-ahead forecast over the historical period, and plots the forecast and its confidence bands over the forecast horizon. The OUT1STEP option must be specified in the FORECAST statement in order for the X12 procedure to calculate the one-step-ahead forecasts.
FORECASTONLY: plots the forecast and its confidence bands over the forecast horizon only.
MODELFORECASTS: plots the one-step-ahead model forecast and its confidence bands in the historical period; plots the forecast and its confidence bands over the forecast horizon. The OUT1STEP option must be specified in the FORECAST statement in order for the X12 procedure to calculate the one-step-ahead forecasts.
MODELS: plots the one-step-ahead model forecast and its confidence bands in the historical period. The OUT1STEP option must be specified in the FORECAST statement in order for the X12 procedure to calculate the one-step-ahead forecasts.
NONE: suppresses all the forecast plots.
TRANSFORECAST: plots the transformed time series and its one-step-ahead forecast over the historical period; plots the forecast and its confidence bands over the forecast horizon. The OUT1STEP option must be specified in the FORECAST statement in order for the X12 procedure to calculate the one-step-ahead forecasts. The TRANSFORECAST plot is available only if the data have been transformed using the TRANSFORM statement.
TRANSFORECASTONLY: plots the forecast of the transformed series and its confidence bands over the forecast horizon only. The TRANSFORECASTONLY plot is available only if the data have been transformed using the TRANSFORM statement.
TRANSMODELFORECASTS: plots the one-step-ahead model forecast of the transformed series and its confidence bands in the historical period; plots the forecast and its confidence bands over the forecast horizon. The OUT1STEP option must be specified in the FORECAST statement in order for the X12 procedure to calculate the one-step-ahead forecasts. The TRANSMODELFORECASTS plot is available only if the data have been transformed using the TRANSFORM statement.
TRANSMODELS: plots the one-step-ahead model forecast of the transformed series and its confidence bands in the historical period. The OUT1STEP option must be specified in the FORECAST statement in order for the X12 procedure to calculate the one-step-ahead forecasts. The TRANSMODELS plot is available only if the data have been transformed using the TRANSFORM statement.

IRREGULAR(<ic-plot-options>) IC(<ic-plot-options>)

produces plots of the irregular series that results from the decomposition specified in the X11 statement. The SPECTRUM plot is produced by default.

The following ic-plot-options are available:

ALL: produces all irregular plots.
NONE: suppresses all irregular plots.
SPECTRUM: produces the spectral plot of Table G2. Table G2 is calculated based on the modified irregular series (Table E3). The data is first-differenced and transformed as specified in the TRANSFORM statement. By default, the type of spectral estimate used to calculate the spectral plot is the spectrum. If the PERIODOGRAM option is specified in the PROC X12 statement, then the periodogram of the series is used to calculate the spectral plot.

RESIDUAL(<residual-plot-options>)

produces the regARIMA model residual series plots if the CHECK statement is specified. The ACF, PACF, HIST, SQACF, and SPECTRUM plots are produced by default. The following residual-plot-options are available:

ACF: produces the plot of residual autocorrelations.
ALL: produces all the residual diagnostics plots that are appropriate for the particular analysis.
HIST: produces the histogram of the residuals and also the residual outliers and residual statistics tables that describe the residual histogram.
NONE: suppresses all the residual diagnostics plots.
PACF: produces the plot of residual partial-autocorrelations if PRINT=PACF is specified in the CHECK statement.
SPECTRUM: produces the spectral plot of Table GRs. Table GRs is calculated based on the regARIMA model residual series. By default, the type of spectral estimate used to calculate the spectral plot is the spectrum. If the PERIODOGRAM option is specified in the PROC X12 statement, then the periodogram of the series is used to calculate the spectral plot.
SQACF: produces the plot of squared residual autocorrelations.

SERIES(<series-plot-options>)

produces plots that are associated with the identification stage of the modeling. The ACF, PACF, and SPECTRUM plots are produced by default. The following series-plot-options are available:

ACF: produces the plot of autocorrelations.
ALL: produces all the plots that are associated with the identification stage.
NONE: suppresses all plots that are associated with the identification stage.
PACF: produces the plot of partial-autocorrelations.
SPECTRUM: produces the spectral plot of Table G0. Table G0 is calculated based on either Table A1, A19, B1 or E1, as specified by the SPECTRUMSERIES= option. The original data is first-differenced and transformed as specified in the TRANSFORM statement. By default, the type of spectral estimate that is used to calculate the spectral plot is the spectrum. If the PERIODOGRAM option is specified in the PROC X12 statement, then the periodogram of the series is used to calculate the spectral plot.

SEASONS=number

specifies the number of observations in a seasonal cycle. If the SEASONS= option is not specified and INTERVAL=QTR, then SEASONS=4 is assumed. If the SEASONS= option is not specified and INTERVAL=MONTH, then SEASONS=12 is assumed. If the SEASONS= option is specified, its value should not conflict with the values of the INTERVAL= option or the values of the date variable. See the preceding descriptions for the START=, DATE=, and INTERVAL= options for more details.

SPAN=(mmmyy ,mmmyy ) SPAN=(’yyQq’ ,’yyQq’ )

specifies the dates of the first and last observations to define a subset for processing. A single date in parentheses is interpreted to be the starting date of the subset. To specify only the ending date, use SPAN=(,mmmyy). If the starting or ending date is omitted, then the first or last date, respectively, of the input data set or BY group is assumed. Because the dates are input as strings and the quarterly dates begin with a numeric character, the specification for a quarterly date must be enclosed in quotation marks. A four-digit year can be specified; if a two-digit year is specified, the value specified in the YEARCUTOFF= SAS system option applies.

SPECTRUMSERIES=table-name

specifies the table name of the series that is used in the spectrum of the original series (Table G0). The table names that can be specified are A1, A19, B1, or E1. The default is B1.

START=mmmyy START=’yyQq’ STARTDATE=mmmyy STARTDATE=’yyQq’

specifies the date of the first observation. Unless the SPAN= option is used, the starting and ending dates are the dates of the first and last observations, respectively. Either this option or the DATE= option is required. When using this option, use either the INTERVAL= option or the SEASONS= option to specify monthly or quarterly data. If neither the INTERVAL= option nor the SEASONS= option is present, monthly data are assumed. Because the dates are input as strings and the quarterly dates begin with a numeric character, the specification for a quarterly date must be enclosed in quotation marks. A four-digit year can be specified; if a two-digit year is specified, the value specified in the YEARCUTOFF= SAS system option applies. When using the START= option with BY processing, the start date is applied to the first observation in each BY group.