                                                    9 February 1998

Example 13: The two spec files listed below show options for 
            adjusting for trading day and festival effects as well 
            as seasonal effects an Industrial Production Index for 
            Israel.  The first spec file imports trading day and
            festival adjustment factors (obtained via an external
            regression program).  The second spec file
            illustrates the use of the x11regression spec to
            calculate such calendar effect adjustments adjustments
            completely within X-12-ARIMA, from a regression model
            of the irregular component whose regressors are read in
            from a file. 

            Note:  We lacked full information about the model
            estimated by the external program and about the methods
            used for dealing with outliers.   For this reason, calendar
            effect factors obtained via x11regression, which are 
            in Tables C 16C and 16C.A of the output of the 
            second spec file, are occasionally somewhat
            different from the factors obtained by the external
            program, printed in Table A.2 of the output of the first
            spec file.


     
# Example 13: tpr.spc  

# Spec file for obtaining a trading day and festival effect adjustment
# of the industrial index series TPR50, (along with a seasonal adjustment) 
# by division of the series by prior adjustment factors stored in the
# file israel.dat under the name PPR50.    

series {file="israel.dat"
        format="2r"
        name="TPR50"
        title="TPR50"  
        precision=3
        decimals=1
}
transform {file="israel.dat"
           format="2r"
           name="PPR50"
           type=permanent
           title="PERMANENT TRADING DAY AND FESTIVAL ADJUSTMENT FACTORS"
           function=log
}
automdl {identify=all}
forecast{ }
outlier { }
x11 {seasonalma=x11default
     trendma=13
     mode=mult
}



# Example 13: tprxreg.spc 

# The x11regression spec is used to fit a user-defined model to the 
# irregular component to obtain an adjustment for trading day and
# festival day effects of an industrial production index of Israel 
# in addition to a seasonal adjustment.  With I(t) denoting the
# irregular value at time t, R(t) the regression expression and 
# e(t) noise, the model has the form
#
#         I(t)-1.0 = R(t) + e(t). 
#
# The series of values to be subtracted from the I(t)'s, the constant 
# series with value 1.0 in this case, is input via the file
# argument of the x11regression spec. The holiday variables in this 
# example represent the day of occurance of the festival in the
# month (fd) and a constant term for the four festival months (mu), 
# with each month's value divided by the length of the month.

series {file="israel.dat"
        format="2r"
        name="TPR50"
        title="TPR50"  
        precision=3
        decimals=1
}
transform {
           function=log
}
automdl {identify=all
}
forecast{ }
outlier { critical=3.5 }
x11 {seasonalma=x11default
     trendma=13
     mode=mult
}
x11regression{
          user=(td1 td2 td3 td4 td5 td6
		fd3 fd4 fd9 fd10
		mu3 mu4 mu9 mu10
                )
          usertype=(td td td td td td 
                    holiday holiday holiday holiday 
                    holiday holiday holiday holiday
                    )
          start=1975.jan
          file="isrfinal.xrg"
          umstart=1975.jan
          umfile="isrusrmu.xrg"
     	  critical=3.5
}

