Help can be obtained on the following SHAZAM commands and topics:

ARIMA AUTO BAYES BOX CHECKOUT COEF COINT COMPRESS CONFID Continuation Lines COPY DELETE DEMO DERIV DIAGNOS DIM DISPLAY DISTRIB DLAG DO DUMP END ENDIF ENDO EQ Error Codes EXEC FC FILE FLS FORMAT GAMES GENR GEN1 GLS GME HET IF IF1 INDEX INST INTEG LAMBDA LOGIT LP MATRIX MENU MLE NAMES New Features NL NONPAR OLS PAR PAUSE PC PLOT POOL PRINT PROBIT PROC PROCEND READ RENAME RESTORE RESTRICT REWIND ROBUST SAVE SET SIZE SKIPIF SAMPLE SORT STAT STOP SYSTEM Temporary Variables TEST TIME TITLE TOBIT WRITE 2SLS

Additional information on SHAZAM can be found:

Info Books


Information on SHAZAM is available as follows:

      SHAZAM
      Department of Economics
      University of British Columbia
      Vancouver, BC  V6T 1Z1
      CANADA
      URL:  http://shazam.econ.ubc.ca/ or http://econometrics.com
      
      Information and ordering details:
      FAX   : 1-604-608-5511
      E-mail: info@shazam.econ.ubc.ca

      Technical Help:
      FAX   : 1-604-822-5915
      Phone : 1-604-822-9436
      E-mail: help@shazam.econ.ubc.ca

      Professional Edition for Windows:
      E-mail: pro@shazam.econ.ubc.ca


Books

Documentation on all SHAZAM commands is available in:

SHAZAM User's Reference Manual Version Northwest Econometrics, 2004, ISBN 0-9687709-0-8.

Addition instructions on using SHAZAM are found in:

Kenneth J. White and Steven A. Theobald
Basic Econometrics: A Computer Handbook Using SHAZAM, McGraw-Hill, 1994, ISBN 0-07-069864-3. This computer handbook accompanies D.N. Gujarati, 1994.

Kenneth J. White and Linda T.M. Bui
The Practice of Econometrics: A Computer Handbook Using SHAZAM, Addison-Wesley, 1991, ISBN 0-201-50048-5. This computer handbook accompanies E.R. Berndt, 1991.

Kenneth J. White, Shirley A. Haun, and David J. Gow
Introduction To The Theory And Practice Of Econometrics: A Computer Handbook Using SHAZAM And SAS, Second Edition, Wiley, 1988, ISBN 0-471-85946-X. This computer handbook accompanies G.G. Judge, R.C. Hill, W.E, Griffiths, H. Lutkepohl, and T.C. Lee, 1988.

Kenneth J. White, S. Donna Wong, Diana Whistler, R. Quentin Grafton, and Meridith Scantlen
Econometric Models & Economic Forecasts: A Computer Handbook Using SHAZAM, McGraw-Hill, 1991, ISBN 0-07-050101-7. This computer handbook accompanies R.S. Pindyck and D.L. Rubinfeld, 1991.

D. Chotikapanich and W.E. Griffiths
Learning SHAZAM: A Computer Handbook For Econometrics, Wiley, 1993, ISBN 0-471-58592-0. This computer handbook accompanies Griffiths, Hill, and Judge, 1993.

T.J. Coelli and W.E. Griffiths
Computer And Exercise Solutions Manual, Wiley, 1989, ISBN 0-471-63821-8. This computer handbook accompanies G.G. Judge, W.E. Griffiths, R.C. Hill, H. Lutkepohl, and T.C. Lee, 1985.

T.T. Nguyen Statistics With SHAZAM, Narada Press, 1993, ISBN 1-895938-00-7.


New Features

The following new features should be noted:

October 1996 : The LOG option on the PROBIT or LOGIT command computes elasticities at means and weighted aggregate elasticities assuming the explanatory variables are log-transformed variables.

December 1996 : The HETCOV option on the POOL command gives pooled OLS estimation with "Panel Corrected Standard Errors" (PCSE).


COINT Command

The COINT command implements tests for unit roots and cointegration.

In general, the format is:

COINT vars / options

where vars is a list of variable names.

The available options are:

  DN                 Use a divisor of N (rather than N-K) when estimating
                     the variance used in calculating the t-statistics.

  DUMP               Gives output of interest to SHAZAM consultants.

  LOG                Takes logs of the data.

  MAX                Gives more detailed ouptut.

  BEG= END=          Specifies the BEGinning and ENDing observations to be
                     used in the calculations. This option overrides the
                     SAMPLE command and defaults to the sample range in effect.

  NDIFF=             Specifies the order of differencing to transform the
                     the data for unit root tests.

  NLAG=              Specifies the number of lag terms in the Dickey-Fuller
                     regressions or the truncation lag parameter for the
                     Phillips tests. If this is not specified then the order
                     is set as the highest significant lag order from either
                     the ACF or PACF of the first differenced series.

  RESID=             Specifies a matrix to save the residuals from the
                     regression equations used for the unit root tests.
                     The first column is the residuals from the constant,
                     no trend regression and the second column is the
                     residuals from the constant, trend regression for the
                     first series. The third and fourth column are the
                     residuals for the second series (if requested).
                     For example, to inspect the ACF of the residuals the
                     following SHAZAM commands could be used.

        ?demo
        SAMPLE 1 17
        COINT CONSUME PRICE / NLAG=2 RESID=EMAT SIGLEVEL=5
        * Adjust the sample period for the observations lost due to lags
        GEN1 BEG=4
        GEN1 END=17
        GEN1 NOBS=END-BEG+1
        SAMPLE 1 NOBS
        DO #=1,4
        COPY EMAT E / FROW=BEG;END FCOL=#;# TROW=1;NOBS TCOL=1;1
        * Now inspect the residuals with the ARIMA command
        ARIMA E
        ENDO
        STOP

  SIGLEVEL=          Specifies the significance level for the critical
                     values. Use SIGLEVEL=5 for 5% critical values and
                     use SIGLEVEL=10 (the default) for 10% critical values.

  TESTSTAT=          Saves the test statistics in the variable specified.

  TYPE=              Specifies the type of tests to calculate. The available
                     options are:

                        DF   For Dickey-Fuller unit root tests.
                             This is the default.
                        PP   For Phillips-Perron unit root tests.
                        RESD For Dickey-Fuller tests on the residuals of
                             cointegrating regressions.
                        RESP For Phillips tests on the residuals of
                             cointegrating regressions.


SAMPLE command

The SAMPLE command specifies the sample range in effect.

In general, the format is:

SAMPLE beg end

where beg and end are numbers specifying the beginning and ending observation in the sample range.

The expanded form (not available with the READ command) is:

SAMPLE beg1 end1 beg2 end2 ...

The SAMPLE command may be called SMPL for compatibility with old versions and other programs.


READ Command

The READ command reads data.

In general, the format is:

READ vars / options

or

READ (unit) vars / options

or

READ (filename) vars / options

The available options are:

  BINARY             The data is in double precision binary.

  BYVAR              The data is read variable by variable.

  CLOSE              Closes a file that was opened for a READ or WRITE
                     command.

  DIF                Data is read from a DIF file.

  EOF                Prevents SHAZAM from reading any further observations.

  FORMAT             The data is read according to the Fortran type.

  LIST               Lists all the data read in.

  NOREWIND           Prevents the read unit from rewinding.

  REWIND             Rewinds the input data file before reading any data.

  TSP                Data is in TSP format.

  BEG= END=          Resets the sample size for the given READ command.

  ROWS= COLS=        Specifies the dimensions of the matrix to be read in.

  SKIPLINES=         Skips the number of lines specified before reading
                     in any data.


WRITE Command

The WRITE command writes data.

In general, the format is:

WRITE(unit) vars / options

or

WRITE(filename) vars / options

The available options as described for the READ command are:

  BYVAR, CLOSE, DIF, FORMAT, NAMES, WIDE, BEG=, and END=.
    
  BINARY             Writes the specified variables in double precision binary.

  REWIND             Used to rewind the output data file before and after
                     data is written.


STOP Command

The STOP command terminates execution and returns you to the system.


DEMO Command

The DEMO command teaches you how to use SHAZAM, just keep typing DEMO and it will slowly teach you some SHAZAM basics.


MENU Command

The MENU command displays a list of SHAZAM commands and indicates which commands are valid to type at any point in time.


PRINT Command

The PRINT command prints variables.

In general, the format is:

PRINT vars / options

The available options are:

  BYVAR              Prints one variable at a time.

  FORMAT             The currently defined FORMAT is used.

  NEWLINE            These options will start a new line, page or sheet

  NEWPAGE            before printing.  The NEWSHEET option will only

  NEWSHEET           work on page printers which use a : as the new
                     sheet control character.

  NONAMES            The heading of variable names is deleted.

  WIDE               Uses 120 columns, NOWIDE will use 80 columns.

  BEG= END=          Resets the sample size for the given PRINT command.
                     You may print a series of vectors, in which case they
                     will appear as columns on the page.  If you only print
                     one vector it will be presented in horizontal format.


FORMAT Command

The FORMAT command is used to specify a FORMAT to be used in a later READ , WRITE , or PRINT command with the FORMAT option.

In general, the format is:

FORMAT(Fortran type format)

For example,

     FORMAT(3F5.2,4F12.1,2X,F1.0)

If the FORMAT is used for a PRINT statement the user should be careful not to print in column 1 since that column is used for print carriage control.


SET Command

The SET commands allows you to turn on or off and set certain options.

In general, the format to turn the options on is:

SET option

The format to turn off the options is:

SET NOoption

The available options are:

  BATCH              Tells SHAZAM it is a Batch run.

  CC                 Carriage control - some commands will skip to a new
                     page.

  COLOR/NOCOLOR      Used on IBM PC Compatible machines with the Color EGA
                     monitor to obtain a color background.

  CPUTIME            Prints the amount of computer time used in the run.

  DELETE             In Distributed Lag models SHAZAM automatically
                     deletes a number of observations equal to the longest
                     lag before estimation. Use SET NODELETE to disable this.

  DOECHO             Echos commands at each pass thru DO LOOP.

  DUMP               DUMPs a lot of output.

  ECHO               Echos SHAZAM commands.

  FRENCH             All commands may be typed in French.

  LASTCOM            Creates a variable called C$ which will contain
                     the previous command typed.

  LCUC               Converts all lower case characters to upper case.
                     Use SET NOLCUC to distinguish upper and lower case.

  MAX                Turns on maximum output.

  NOCC               Turns off the carriage control.

  OPTIONS            Turns on the display of all options for each command.

  OUTPUT             Turns on the output for all following commands.
                     Use SET NOOUTPUT to suppress all output.

  PAUSE              Causes a pause after each command is executed.

  RANFIX             Prevents random number generator from being set by
                     the system clock.

  SAMPLE/NOSAMPLE    Allows use of the omitted observations in expanded
                     form of the SAMPLE command to be turned on and off.

  SCREEN             Echos the SHAZAM output to the terminal screen when
                     the output has also been redirected to a file
                     assigned to Unit 6. (Works on some computers)

  SKIP               Puts SKIPIFs back.

  SKIPMISS           Skips observations with a value equal to the MISSVALUE=

  STATUS             With some computers displays a status line.

  TALK               Tells SHAZAM you are at a terminal and are typing
                     in SHAZAM commands interactively.

  TERMINAL           Tells SHAZAM you are at a terminal and commands have
                     all ready been put in a file assigned to Unit 5.

  TIMER              Turns on the timer.

  TRACE              Traces subroutine flow.

  WARN/NOWARN        Warning messages are normally printed for illegal
                     operations in GENR, MATRIX, IF, SKIPIF and ENDIF
                     statements.

  WARNSKIP/          Warning messages are printed for every observation
      NOWARNSKIP     skipped by a SKIPIF command.

  WIDE               WIDE assumes that up to 120 columns are available.

  COMLEN=xxx         Sets maximum length of command.

  MAXCOL=xxx         Specifes the maximum number of a column or
                     observation used after a ":".

  MISSVALU=XXX       Specifies Missing value code, Default=-99999

  OUTUNIT=           When the output is going to more than one file.

  RANSEED=xxx        Sets the seed for the random number generator.  To turn
                     on an option you would say something like SET DUMP.  Then
                     to turn it off you would say SET NODUMP.


DISPLAY Command

The DISPLAY command will display the current value of any option that you have specified with the SET command.

In general, the format is:

DISPLAY option

The CPUTIME option shows the CPU time used.


OLS Command

The OLS command will perform OLS regressions.

In general, the format is:

OLS depvar indeps / options

The available options are:

  ANOVA              Prints analysis of variance tables and model selection
                     tests.

  AUXRSQR            Prints the R-squared statistic for the auxiliary
                     regressions.

  DFBETAS            Used with the INFLUENCE option if you want the
                     DFBETAS computed.

  DLAG               Used to compute Durbin's h statistic when there
                     is a LAGged Dependent variable.

  DN                 Uses a divisor of N instead of N-K for SIGMA**2.

  DUMP               DUMPS large amounts of data on the regression.

  DWPVALUE           Computes the exact Durbin Watson probability.

  GF                 Prints Goodness of Fit tests for normality of
                     residuals.

  GNU                Plots residuals and fitted values with GNUPLOT.

  HETCOV             Uses Heteroskedastic-Consistent covariance matrix.

  INFLUENCE          Computes the Belsey-Kuh-Welsch Influence Diagnostics.

  LINLOG             Used when the dependent variable is linear, but the
                     independent variable are in LOG form.

  LIST               Lists the residuals and predicted values of the
                     dependent variable and residual statistics.

  LM                 Gives the Jargue-Bera Lagrange Multiplier test for
                     normality of residuals.

  LOGLIN             Used when the dependent variable is in LOG form, but
                     the independent variables are linear.

  LOGLOG             Used when the dependent variable and all the
                     independent variables are in log form.

  MAX                Prints analysis of variance tables, variance-covariance
                     matrix, correlation matrix, residuals, residual
                     statistics, goodness of fit test for normality,
                     Jarque-Bera test, and plots residuals.

  NOCONSTANT         Removes the intercept from the equation.

  NOMULSIGSQ         When NOMULSIGSQ is specified the estimated covariance
                     matrix will just be X'X inverse.

  NONORM             Weights are NOT to be normalized.

  PCOR               Prints the correlation matrix of coefficients.

  PCOV               Prints the covariance matrix of coefficients.

  PLUSH              Prints the LUSH residuals.

  REPLICATE          Used with WEIGHT= if your weights indicate a sample
                     replication factor.

  RESTRICT           Forces restrictions into the regression.

  RSTAT              Prints and plots residuals and residual statistics.

  UT                 Used with WEIGHT= if you want the residuals and
                     predicted values to be UnTransformed.

  AUTCOV=            Autocorrelation equivalent to HETCOV option.

  BEG= END=          Specifies the sample size for the given command.

  COEF=              Stores the coefficients in another variable.

  COV=               Saves the variance-covariance matrix.

  FE=,FX=            Specify F values in Stepwise regression.

  HATDIAG=           Saves the diagonal of the so-called Hat Matrix.

  IDVAR=             Specifies a variable which contains names to be
                     printed out next to the residuals when the LIST
                     option is used.

  INCOEF=            Used to specify a vector of coefficients to input if
                     you do not want to estimate the mode.

  INCOVAR=           Used with the INCOEF= option to specify covariance matrix.

  INDW=              Specifies a value for the Durbin-Watson test statistic.

  INSIG2=            Used with the INCOEF= option described above to specify
                     a value of sigma - squared.

  METHOD=            Specifies the METHOD; GS (Gram Schmidt) (default),
                     HH (Householder), or NORMAL (Normal Equations).

  ORDER=             Specifies the order to test with the DWPVALUE option.

  PCINFO=            The matrix of Information obtained from a previous
                     PC command for regression on Principal Components.

  PCOMP=             The matrix of Principal Components if this is a
                     regression on Principal Components.

  PE=,PX=            Specify probability levels in Stepwise regression.

  PREDICT=           Saves the predicted values of the dependent variables.

  RESID=             Saves the values of the residuals from the regression.

  RIDGE=             Specifies a value of k to be used in Ridge regression.

  STDERR=            Saves the values of the standard errors of the coefficients.

  TRATIO=            Saves the values of the t-ratios on the coefficients.

  WEIGHT=            Specifies a weight for a Weighted Least Squares regression.

The temporary variables available on the OLS command are:

   $ADR2, $ANF, $DF, $ERR, $CDF, $DW, $K, $LLF, $N, $PI, $R2, $R2OP, $RAW,
   $RHO, $SIG2, $SSE, $SSR, $SST, $ZANF, $ZSSR, and $ZSST.

ANOVA option: $FPE, $LAIC, $LSC, $GCV, $HQ, $RICE, $SHIB, $SC, and $AIC.

DLAG option: $DURH

LM option: $JB


GLS Command

The GLS command does Generalized Least Squares Regressions.

In general, the format is:

GLS depvar indeps / options

The user must specify one of PMATRIX= , OMEGA= , or OMINV= as one of the GLS options to tell SHAZAM which matrix to use for estimation.

The available options are:

  BLUP               Predicted values are adjusted using info obtained
                     from previous period residuals according to the
                     transformation specified by the P matrix.

  DUMP               Prints the Omega, Omega Inverse and P matrices used.

  FULLMAT            If the user is passing the full matrix.

  NOMULSIGSQ         When the matrix is the actual covariance matrix this
                     option is specified so the matrix is not multiplied by
                     sigma squared.

  UT                 Estimated coefficients will be used with the original
                     data to compute predicted values and residuals that
                     are not transformed.

  PMATRIX=,          Specifies the matrix to be used for estimation.
  OMEGA=,
  OMINV=

  RESID=             Saves the residuals according to UT and BLUP.

Other options available are:

   ANOVA, DLAG, DN, GF, LIST, MAX, NOCONSTANT, PCOR, PCOV, RSTAT, BEG=,
   END=, COEF=, COV=, PREDICT=, STDERR=, and TRATIO=
as defined for the OLS command.

For available temporary variables see the OLS command.


GME Command

The GME command implements Generalized Entropy methods.

In general, the format of the GME command is:

GME depvar indeps / options

where options as described for OLS :

   LINLOG, LIST, LOGLIN, LOGLOG, NOCONSTANT, PCOV, RSTAT, BEG=, END=, 
   COEF=, COV=, PREDICT=, RESID=, STDERR=, and TRATIO=.

Options as described for NL are:

   CONV=, ITER=, PITER=, and START=.

Other options are:

  DEVIATION          Variables are converted to deviations from means. This
                     forces the NOCONSTANT option.

  LOGEPS=            Small constant to avoid numerical underflow.

  QPRIOR=            Matrix of priors for supports specified in ZENTROPY=.

  UPRIOR=            Matrix of priors for supports specified in VENTROPY=.

  VENTROPY=          Matrix of support points for the unknown disturbances.

  ZENTROPY=          Matrix of support points for the unknown coefficients.


2SLS Command

In general, the format of the 2SLS command is:

2SLS depvar rhsvars (exogs) / options

where exogs is a list of exogenous variables. SHAZAM automatically includes a constant term in exogs unless the NOCONEXOG option is specified.

The available options as described for the OLS command are:

   DN, DUMP, GF, LIST, MAX, NOCONSTANT, PCOR, PCOV, RESTRICT, RSTAT
   BEG=, END=, COEF=, COV=, PREDICT=, and RESID=.


INST Command

In general, the format of the INST command is:

INST depvar rhsvars (instvars) / options

The INST command is identical to the 2SLS command.


SYSTEM Command

To set up a system for estimation, the following commands are used:

SYSTEM neq exogs / options
OLS depvar indeps
.
OLS depvar indeps

The exogs is only required for Three Stage Least Squares in a Simultaneous equation model and must NOT be specified for systems of Seemingly Unrelated Regressions.

The available options are:

  DN                 Uses N as the divisor for the covariance matrix.

  DUMP               For SHAZAM consultants.

  FULL               Gives full equation output.

  LIST               Used as described for the OLS command.
  RSTAT
  MAX
  GF

  NOCONEXOG          Do not include a constant in the list of exogs variables.

  NOCONSTANT         Suppresses the intercept in all equations.

  PCOR,PCOV          Print the correlation and covariance matrices of all
                     variables in the system after convergence.

  PINVEV             Prints the inverse of the exogenous variables matrix
                     when the 3SLS system is being used.

  PSIGMA             Prints the residual covariance matrix (SIGMA) after
                     each iteration.

  RESTRICT           Used when the SYSTEM command is to be followed by
                     RESTRICT commands and an END command.

  COEF=              Saves the values of the COEFficients (excluding the
                     intercept).

  COEFMAT=           Saves the values of the COEFficients (including the
                     intercept) in a matrix with each column of the
                     matrix representing one equation.

  CONV=              Specifies a convergence criterion to stop the
                     iterative procedure.

  COV=               Saves the COVariance matrix.

  ITER=              Specifies the maximum number of iterations.

  OUT=,IN=           Used to output a file of useful information on
                     unit 11 at each iteration, so that the system can
                     be restarted in another run at the same point by
                     inputting the file from unit 11.

  PITER=             Specifies the frequency with which ITERations are
                     to be Printed.

  PREDICT=           Saves the predicted values of the dependent variable.

  RESID=             Saves the residuals from the regression.

  SIGMA=             Save the residual covariance matrix (SIGMA).


PROBIT Command

The PROBIT command does Probit regressions.

In general, the format is:

PROBIT depvar indeps / options

The available options on the PROBIT command are:

  DUMP               Of interest to SHAZAM consultants.

  LIST               Prints the index, I, the predicted and observed
                     values of the dependent variable, a plot of the
                     predicted values and the residual statistics.

  LOG                Computes elasticities assuming log-transformed variables.

  PCOV               Prints the estimated asymptotic covariance matrix.

  RSTAT              Prints residual statistics.

  CONV=              Sets the convergence criterion for the log-likelihood
                     function.

  IMR=               Saves the computed Inverse Mill's Ratio.

  INDEX=             Saves the index.

  ITER=              Sets maximum number of iterations.

  PITER=             Specifies the frequency with which the ITERations will
                     be Printed.

  PREDICT=           Stores the predicted probabilities.

Other options available are:

   MAX, NOCONSTANT, NONORM, PCOR, BEG=, END=, COEF=, COV=, STDERR=, TRATIO=,
   and WEIGHT=
as defined for OLS .


LOGIT Command

The LOGIT command does Logit regressions. The format and options are the same as those for PROBIT . (However, the IMR= option may not be valid).


TOBIT Command

The TOBIT command does TOBIT regressions. In general, the format is:

TOBIT depvar indeps / options

The available options are:

  DUMP               Of interest to SHAZAM consultants.

  LIST               Prints the index, I, the density and cumulative
                     probabilities corresponding to the index I and
                     the observed and expected and conditional values of
                     the dependent variable.

  PCOR               Prints the correlation matrix of the normalized
                     coefficients.

  PCOV               Prints the extimated asymptotic covariance matrix
                     of the normalized coefficients.

  UPPER              Used if it is an upper rather than a lower limit.

  CONV=              Sets the convergence criterion for the log-likelihood
                     function.

  INDEX=             Saves the INDEX.

  ITER=              Sets the maximum number of iterations allowed.

  LIMIT=             Specifies the limiting value of the dependent variable.

  PITER=             Specifies the frequency with which iterations are to be
                     printed.

  PREDICT=           Saves the predicted values of the dependent variable.

Other options available are:

   MAX, NOCONSTANT, NONORM, BEG=, END=, COEF=, COV=, STDERR=, TRATIO=,
   and WEIGHT=
as defined for OLS .


TEST Command

The TEST command does hypothesis testing.

In general, the format is:

TEST equation


END Command

The END command is used in: non-linear estimation to mark the end of the EQ and COEF commands; conjunction with all estimation procedures when joint tests are being done with TEST commands - in this case the END command marks the end of the TEST commands that form the joint test; conjunction with some estimation procedures when RESTRICT commands are being used - in this case the END command marks the end of the RESTRICT commands.


PLOT Command

The PLOT command plots variables. Not all options are available on all machines.

In general, the format is:

PLOT depvars indep / options

The available options are:

  ALTERNATE          Alternates the symbols 'X' and 'O' in  plotting
                     columns of the histogram when the HISTO option is used.

  DASH               Gives dashed lines.

  GNU                Prepares GNUPLOT graphs (SEE BELOW for options).


 ...MSDOS-PC OPTIONS
  GRAPHICS           These options are used only on PC Compatible Personal

  EGA                Computers running MSDOS operating system to obtain
  VGA                graphics on your system.
  HERCULES


  ... Macintosh
  GOAWAY             Used with the GRAPHICS option if you wish the Plot
                     Window to go away after the plot appears on the screen.

  GRAPHICS           Creates a Plot Window and draws a Graphics plot. To remove
                     the Plot, click the GO AWAY Box.  You may also use the
                     LINE and LINEONLY options.

  .....
  HISTO              Plots histograms for the variables specified.

  HOLD                Prevents the plot from being printed.

  LINE               Used with the GRAPHICS, EGA, HERCULES, or TOSHIBA
                     options to draw a line connecting data points and
                     plot a symbol at each point.

  LINEONLY           Used with the GRAPHICS, EGA, HERCULES, or TOSHIBA
                     options to draw a line connecting the data points.

  NOBLANK            Prints the previously held plot as well as the current
                     one.

  NOPRETTY           Tells SHAZAM not to attempt to make axes pretty and
                     just use the range of the data directly.

  NOSAME             Uses a separate page for each plot.

  NOWIDE             80 columns are available for the plot.

  PAUSE              Used with one of the graphics options on IBM-PC Compatible
                     or Macintosh computers to temporarily stop the screen
                     after a graphics plot appears.

  PRINT              Used with one of the graphics options on IBM-PC
                     Compatible computers to PRINT the plot on a printer.

  RANGE              Uses the entire range of the data for plotting
                     the histogram.

  SAME               Plots depvars against the indep on the SAME plot.

  TIME               Plots the variables listed against time.

  WIDE               120 columns are available for the plot.

  BEG= END=          Sets the sample size for the given PLOT command.

  CHARSIZE=          Used with the GRAPHICS option to specify a character size.

  DEVICE=            Available with the GRAPHICS option.

  GROUPS=            When the HISTO option is used SHAZAM normally places
                     data into 6 groups.  This option can be used to specify
                     up to 60 groups.  Suggested values for the GROUPS= are 2,3,
                     4,5,6,10,12,15,30 or 60.

  OUTPUT=            Available with the GRAPHICS option.

  PORT=              Available with the GRAPHICS option.

  SYMBOL=            Uses the user specified symbol.

  XMIN=,XMAX=        Specifies the desired range for the X and Y axis.

  YMIN=,YMAX=        Must be used with the NOPRETTY option.

The GNU option prepares an interface to the GNUPLOT graphics package. When this option is used the following PLOT options are available:

   HISTO, LINE, LINEONLY, TIME, BEG=, END=, and GROUPS=.

Additional options are:

  APPEND

  KEY / NOKEY

  COMMFILE=

  DATAFILE=

  DEVICE=            Use DEVICE=POSTSCRIPT to send output to a PostScript file.

  OUTPUT=

  PORT=


DUMP Command

The DUMP command will dump common blocks and other information useful only to SHAZAM consultants. DUMP DATA will give a listing of all the current variables, their addresses, increment, type, total number of observations, and second if a matrix.


REWIND Command

The REWIND command is used to rewind the READ or WRITE units. REWIND is also available as an option on the READ or WRITE commands.

The format of the REWIND command is:

REWIND unit


SIZE Command

The SIZE command sets the maximum number of variables. This command is not available on all computers.

In general, the format is:

SIZE maximum


INDEX Command

To calculate Indexes it is necessary to have prices and quantities for at least two commodities.

In general, the format is:

INDEX p1 q1 p2 q2 p3 q3 ... pn qn / options

The available options are:

  ALTERN             Normally, the variable list is as described above,
                     where p and q alternate in the list.

  CHAIN              Chains the Laspeyres, Paasche or Fisher indices using
                     the method described above.  The Divisia index is
                     always chained.

  EXPEND             Indicates that the quantity variables measure
                     expenditures rather than quantities.

  NOALTERN           Lists all the prices then all the quantities.

  NOLIST             The price and quantity indexes are not printed.

  BASE=              Specifies the observation request for a base
                     year to use for the index.

  DIVISIA=           Specify the type of index to be calculated and
  PAASCHE=           the variable into which the result should be placed.
  LASPEYRE=          If a quantity index is desired, the option name
  FISHER=            should begin with a Q.  Option names not beginning
  QDIVISIA=          with a Q refer to price indexes.
  QPAASCHE=
  QLASPEYRES=
  QFISHER=

If there is a zero price or quantity in any year the commodity will be ignored for that year. To compute a quantity indexes reverse the p and q variables on the INDEX command.


STAT Command

The STAT command will print means, standard deviations, variances, minimums, and maximums for the variables listed.

In general, the format is:

STAT vars / options

The available options are:

  ALL                The basic statistics are computed for ALL the variables
                     in the data.

  ANOVA              Prints an ANalysis Of VAriance table and an F value
                     that tests the null hypothesis that the means of
                     all the variables listed on the given STAT command
                     are the same.

  BARTLETT           Perform Bartlett's Homogeneity of Variance Test.

  DN                 Uses N rather than N-1 in computing variances.

  MATRIX             Uses any matrices contained in the list vars
                     as single variables.

  MAX                Prints PCOR, PCOV, PCP, PCPDEV and PRANKCOR matrices.

  PCOR               Prints a CORrelation matrix of the variables listed.

  PCOV               Prints a COVariance matrix of the variables listed.

  PCP                Prints a CrossProduct matrix of the variables listed.

  PCPDEV             Prints a CrossProduct matrix of variables in
                     DEViations from the means.

  PFREQ              Prints tables of FREQuencies.

  PMEDIAN            Prints the MEDIAN, mode and quartiles for each variable.

  PRANKCOR           PRints a matrix of Spearman's RANK CORrelation
                     coefficients.

  REPLICATE          Used with WEIGHT= when weights are replication factors.

  WIDE=              Prints the Coefficient of Variation and the Constant-
                     Digits statistic.

  BEG= END=          Specify the beginning and end observations to be used
                     in the given STAT command.

  COR=               Saves the Correlation Matrix.

  COV=               Saves the Covariance Matrix.

  CP=                Saves the CrossProduct Matrix.

  CPDEV=             Saves the CrossProduct Matrix in deviations from
                     the mean.

  MAXIM=             Saves the MAXIMums as a vector.

  MEAN=              Saves the MEANS as a vector.

  MEDIAN=            Saves the MEDIANS as a vector.

  MINIM=             Saves the MINIMums as a vector.

  MODE=              Saves the MODES as a vector.

  RANKCOR=           Saves the RANK CORrelation Matrix.

  STDEV=             Saves the Standard Deviations as a vector.

  SUMS=              Stores the sum of each variable as a vector.

  VAR=               Saves the VARiances as a vector.

  WEIGHT=            Specifies a variable to be used as WEIGHTs if
                     weighted descriptive statistics are desired.


DISTRIB Command

The DISTRIB command will compute probabilities for variables.

In general, the format is:

DISTRIB vars / options

The available options are:

  INVERSE            Returns critical values of the distributions instead of
                     probabilities.  This option may not be used with
                     EDGE  or IMHOF.

  LLF                Computes the Log of the Likelihood function for the
                     data and prints it and stores it in the temporary
                     variable $LLF.

  NOLIST             Suppresses list of probabilities.

  ACCURACY=          Level of accuracy for TYPE=DAVIES.

  BEG= END=          Specifies the sample range to be used.

  BIGN=              Specifies the population size for the Hypergeometric
                     distribution.

  BIGX=              Specifies the population number of successes for the
                     Hypergeometric distribution.

  C=                 Specifies non-centrality parameter for non-central
                     F distribution.

  CDF=               Saves the cumulative distribution probability.

  CRITICAL=          Saves the critical values when the INVERSE option is used.

  DF=                Specifies the degrees of freedom.  Used only with
                     TYPE=T and TYPE=CHI.

  DFVEC=             Specifies vector of Degrees of Freedom for TYPE=DAVIES.

  DF1=,DF2=          Specifies the degrees of freedom for the numerator and
                     the denominator.  Used only with TYPE=F.

  EIGENVAL=          Specifies the eigenvalue to be used with the
                     TYPE=IMHOF option.

  H=                 Specifies the precision parameter for the t-distribution.

  K=                 Parameter K for TYPE=BURRII, BURRIII, BURRXII

  KURTOSIS=          Specifies the population KURTOSIS parameter for an EDGE
                     approximation.

  LAMBDA=            A LAMBDA vector with TYPE=DAVIES.

  LIMIT=             Integration terms with TYPE=DAVIES.

  MEAN=              Specifies the population mean value.  This option is
                     used only with TYPE=NORMAL and TYPE=EDGE.

  N=                 Saves the sample size for the Hypergeometric distribution.

  NEIGEN=            Specifies the number of Eigenvalues to be used with
                     the TYPE=IMHOF option if the entire vector is not required.

  NONCEN=            Specifies the non-centrality parameter value of
                     the Davies distribution.

  PDF=               Saves the densities in the variable specified.

  P=,N=              Specifies the parameters of the BINOMIAL distribution.

  P=,Q=              Specifies the parameters of BETA or GAMMA distribution.

  SKEWNESS=          Specifies the population skewness coefficient for an
                     EDGE approximation.

  S=,V=              Specifies the parameters of Inverted Gamma distribution.

  TYPE=              Specifies the type of distributions.  The choices
                     are:  BETA, BINOMIAL, BURRII, BURRIII, BURRXII, CHI,
                     DAVIES, EDGE, F, GAMMA, HYPERGEO, IG2, IMHOF, NORMAL and T.

  VAR=               Specifies the variance.  Used only with TYPE=BETA,
                     TYPE=EDGE and TYPE=NORMAL.

The types of distribution available on the TYPE= option are:

  BETA               Beta Distribution
                     P= and Q=, or MEAN= and VAR= must be used.  The beta
                     variable must be in the 0-1 interval.

  BINOMIAL           Binomial Distribution
                     When the BINOMIAL distribution is specified the
                     N= and P= options must be used.

  BURRII, BURRIII,
      BURRXII        The Burr Distribution.

  CHI                Chi Squared Distribution
                     DF= must be used.

  DAVIES             Davies Distribution.

  EDGE               Edgeworth Approximation
                     MEAN=, VAR=, SKEWNESS=, and KURTOSIS= should be used.

  F                  F Distribution
                     DF1= and DF2= must be used.
                     C= must be used for a Non-Central F Distribution.

  GAMMA              Gamma Distribution, parameters are P= and Q=.

  HYPERGEO           Hypergeometric Distribution, parameters are
                     BIGN=, N=, and BIGX=.

  IG2                Inverted Gamma, parameters are S= and V=.

  IMHOF              Imhof Distribution
                     EIGENVAL= must be used.

  NORMAL             Normal Distribution
                     MEAN= and VAR= should be used with it.

  T                  Student's t Distribution
                     DF= must be used.


GENR Command

The GENR command will create new variables from old ones and do a variety of data transformations. The SAMPLE command in effect defines the observation range used in GENR commands.

In general, the format of the GENR command is:

GENR newvar = f(oldvars)

The following mathematical operators may be used:

   addition(+), subtraction(-), multiplication(*), division(/), exponentiation(**),
   .EQ.,.NE.,.GE.,.GT.,.LE.,.LT.(relations), .NOT.,.AND.,.OR.(logical operators).

Operators are listed according to priority-level. Unary functions are first priority.

The available unary functions are:

   ABS(x), DUM(x), EXP(x), INT(x), LAG(x), LAG(x,n), LGAM(x), LOG(x), MAX(x,y),
   MIN(x,y), MOD(x,y), NCDF(x), NOR(x), SAMP(x), SEAS(x), SIN(x), SQRT(x),
   SUM(x), SUM(x,n), TIME(x), UNI(x).


GEN1 Command

The GEN1 command will generate a constant.

In general, the format is:

GEN1 equation

For example:

     GEN1 NOB=$N

Saves the value of the number of observations from a previous regression in the variable NOB. (Also type HELP TEMP ).


DERIV Command

The DERIV command will differentiate an equation with respect to any variable and return the derivates as another variable.

In general, the format is:

DERIV variable resultvar = equation

where variable is the variable in the equation for which the derivative should be taken with respect to and the result of the derivative will be placed in the variable resultvar.


IF Command

The IF command is a conditional form of the GENR command. It will perform the implied GENR option if the condition is true.

In general, the format is:

IF(expression)

where expression is an expression in parentheses to be evaluated.


IF1 Command

Equivalent to a SAMPLE 1 1 and an IF command.


INTEG Command

The INTEG command will perform numeric univariate integration on an equation.

In general, the format is:

INTEG variable lower upper resultvar = equation

where integral of the equation with respect to the variable will be taken using the range specified in lower and upper. The result of the integration will be placed in the variable resultvar.


SKIPIF Command

The SKIPIF command is used to specify conditions under which observations are to be skipped. (The observations will still be in the data and will be used in all GENR statements, but will not be used in estimation.)

In general, the format is:

SKIPIF(expression)

The observation will be skipped if the value of the expression is positive; otherwise, the observations is retained. To completely eliminate all SKIPIF commands in effect:

DELETE SKIP$


MATRIX Command

The MATRIX command will do matrix operations.

In general, the format is:

MATRIX newmat = f(newmat)

The following logical operators are valid on the MATRIX command: negation(-), multiplication(*), addition(+), subtraction(-), transpose('), Kronecker multiplication(@), Hadamard division(/), and Concatenation (|).

Regular matrix rules apply on the MATRIX command. The available options are:

  CHOL(matrix)       Cholesky's decomposition of symmetric positive.

  DET(matrix)        Determinant of matrix.

  DIAG(matrix)       Diagonal.

  EIGVAL(matrix)     The eigenvalues of matrix are taken.

  EIGVEC(matrix)     The eigenvector of matrix is taken.

  EXP(matrix)        Exponential operator is applied to each element
                     of matrix.

  FACT(matrix)       Factors the inverse of the matrix.

  IDEN(ndim)         Identity matrix is created.

  IDEN(ndim,ndiag)   A matrix with ndim rows and columns is created
                     with a diagonal of ones on the ndiag lower
                     diagonal. (ndiag=1 gives an identity matrix)

  INT(matrix)        Integer truncation of each element of matrix.

  INV(matrix)        Inverse of matrix.

  LAG(matrix,n)      Each column of the matrix is lagged n times.

  LOG(matrix)        Natural log of each element is taken.

  NCDF(matrix)       Normal cumulative density function.

  NOR(nrow,ncol)     Normally distributed numbers.

  RANK(matrix)       Rank of matrix is calculated.

  SAMP(matrix,nrows) A new matrix with n rows is created by random
                     sampling with replacement from the old matrix.

  SEAS(nob,nseas)    A series of seasonal dummy variables is created.

  SIN(matrix)        Sine of each element of matrix is taken.

  SQRT(matrix)       Square root of each element of matrix is taken.

  SVD(matrix)        Singular Value Decomposition of matrix is
                     performed.

  SYM(matrix)        Creates a symmetric matrix from a square full
                     matrix.

  TIME(nob,x)        A vector with nob observations is created with
                     values equal to a time index plus x.

  TRACE(matrix)      Trace of matrix is calculated.

  TRI(matrix)        Creates a lower triangular matrix from a
                     square full matrix.

  UNI(nrow,ncol)     Random numbers between 0 and 1 is created.

  VEC(matrix)        Stacks columns of a matrix into a longer vector.

  VEC(matrix,nrows)  Unstacks a vector into a matrix with nrows.


COPY Command

The COPY command has the following format:

COPY fromvar(s) tovar / options

The available options are:

  FROW=first;last    Tells SHAZAM the first and last rows to copy.

  TROW=first;last    Tells SHAZAM the first and last rows of the
                     matrix you are copying into to be filled.

  FCOL=first;last    Tells SHAZAM the first and last columns to copy.

  TCOL=first;last    Tells SHAZAM the first and last columns of the
                     matrix you are copying into to be filled.

If no options are specified, all of fromvar(s) will be copied into tovar .


SORT Command

The SORT command will sort the variables specified in ascending order according to the sort variable specified.

The format of the SORT command is:

SORT sortvar vars / options

The available options are:

  DESC               Sorts the specified variables in descending order.

  LIST               Lists all the sorted data, ie. vars and sortvar.

  BEG= END=          Specifies the sample range for the given SORT
                     command.


TIME Command

The TIME command specifies the beginning year and frequency for a time series.

The format of the TIME command is:

TIME beg freq var

For example:

     SAMPLE 1 48
     TIME 1981 12 YEARMO

The data 1981.01, 1981.02, etc will be put in YEARMO.


TITLE Command

The TITLE command is used to specify a title which will be printed at the top of each page of output.

The format of the TITLE command is:

TITLE title


DIM Command

The DIM command dimensions a vector, or a matrix so that data may be placed in it piece by piece.

The format of the DIM command is:

DIM var size var size...


EQ Command

The EQ command is used in conjunction with the NL command.


COEF Command

The COEF command is used in conjunction with the NL command.


GAMES

Global Thermonuclear War


COMPRESS Command

The COMPRESS command is used when you have deleted some variables and have run out of space and need to squeeze out all the unused space.


SAVE Command

The SAVE Command writes all variables in a file for use in a later SHAZAM run. Later you can use the RESTORE command to read back all variables that have been SAVED. This command is experimental.


PAUSE Command

The PAUSE command causes the machine to PAUSE.


DELETE Command

The DELETE command is used to delete variables.

In general, the format is:

DELETE vars

The format to DELETE all variables is:

DELETE / ALL

At the end of a SHAZAM procedure use:

DELETE / ALL_


NAMES Command

The NAMES command will print out the NAMES of all variables that have been read or generated in the particular SHAZAM run.

The format of the NAMES command is:

NAMES

or

NAMES LIST

which will print a table of names, the typed of variable and the size of the variable.


RENAME Command

The RENAME command is used to rename variables that already exist in the current run.

The format of the RENAME command is:

RENAME oldname newname


RESTORE Command

The RESTORE command reads all variables that have been written with the SAVE command. Any variables currently defined with the same name as one being RESTORED will be deleted. Variables that do not have the same name as a RESTORED variable will still exist. This command will also read data from the CITIBASE tape.


DO Command

The DO command provides some DO-loop capabilities that are useful for repeated calculations.

For example:

     DO #=1,20,2
     GENR LX#=LOG(X#)
     ENDO

The value of the DO character is substituted in all statements in the loop. Any special character (eg. #, ?, %, $, !) other than letters or numbers may be used for the DO character. Be sure to end the DO loop with the ENDO statement. DO-loops can be nested up to 8 levels. Values are available in the temporary variables $DO, $DO2 - $DO8. Type HELP ENDO and HELP ENDIF for more information.


ENDO Command

The ENDO command is used to end DO loops.


ENDIF Command

The ENDIF command will terminate execution of a DO loop if the condition is true for any observation. If the ENDIF is used outside of a DO loop and the condition is true for any observation then the run will stop.

In general, the format is:

ENDIF(expression)

where expression is the expression to be evaluated.


AUTO Command

The AUTO command estimates models with autoregressive errors or moving average errors.

In general, the format is:

AUTO depvar indeps / options

The available options are:

  DLAG               Used with first-order models to indicate that the
                     first independent variable is a LAGged Dependent
                     variable.

  DN                 Divide by N instead of N-K to get Variances.

  DROP               Drops first observation in estimation.

  GS                 Uses a grid search to estimate rho.

  MISS               Maximum likelihood estimates are adjusted for any
                     missing observations that were deleted with SKIPIF
                     commands.

  ML                 Maximum Likelihood Estimation

  NOPITER            Suppresses the printing of the iterations.

  PAGAN              Used with ORDER=1 or ORDER=2 to estimate the model
                     using Pagan's ]1974( procedure.

  CONV=              Sets a convergence criterion when the Cochrane-Orcutt
                     by procedure is being used.

  GAP=               Used when the missing observations were included in
                     the data.

  ITER=              Used when doing an iterative Cochrane-Orcutt
                     procedure to control the number of iterations.

  NMISS=             Used with the GAP= option to indicate how many
                     observations are missing from the data.

  NUMARMA=           Used with PAGAN option to specify the number of AR
                     or MA coefficients to be estimated.

  ORDER=n            Used to estimate models with second or higher order
                     autocorrelation.

  ORDER=-n           Model will be estimated using a moving-average error
                     model instead of an autoregressive error model.
                     DLAG, DROP, GAP, GS, MISS, ML, RESTRICT, RHO and SRHO
                     may not be used.

  RHO=               Specifies rho desired by user.

  SRHO=              Specifies rho for second order.

Other available options are:

   ANOVA, DUMP, GF, LINLOG, LIST, LOGLIN, LOGLOG, MAX, NOCONSTANT, PCOR,
   PCOV, RESTRICT, RSTAT, BEG=, END=, COEF=, COV=, PREDICT=, RESID=,
   STDERR=, and TRATIO=
as defined for OLS .

The available temporary variables on the AUTO command are:

   $DF, $DW, $ERR, $K, $LLF, $N, $R2, $R2OP, $RAW, $RHO, $SIG2, $SSE,
   $SSR, $SST, $ZDF, ZSSR, and $ZSST.


RESTRICT Command

The RESTRICT command is used to specify restrictions that are to be imposed in estimation. When this command is used the RESTRICT option must be specified on the estimation command. Restrictions must be linear. Restrictions cannot be used with TOBIT , LOGIT , PROBIT , or NL commands.

An example of the RESTRICT command on an OLS command is:

OLS CONSUME INCOME PRICE / RESTRICT
RESTRICT INCOME+PRICE=0
END


DLAG Command

Polynomial Distributed Lags (or Almon Lags) can be specified on AUTO , BOX , GLS , OLS , and POOL commands. SHAZAM allows the user to specify a lag length, order of polynomial and endpoint constraints on any independent variable in the model. For illustration, the OLS command is used:

OLS depvar indep(first.last,order,endcon) / options


NL Command

The NL command estimates non-linear models.

In general, the format is:

NL neq exogs / NCOEF= options
EQ equation
COEF values
END

where exogs are the list of instrumental variables for N2SLS or N3SLS .

The available options are:

  ACROSS             Corrects for autocorrelation across equations
                     as well as within a single equation.

  AUTO               Corrects the equations for first-order
                     autocorrelation or for higher order if specified on
                     the ORDER= option.

  DRHO               Gives a different value of rho for each equation.

  DUMP               Dumps the internal code that SHAZAM has generated
                     for the EQ commands.

  EVAL               Evaluates the likelihood function for the starting
                     value prints out the answer without doing any
                     estimation if ITER=0 is specified.

  GENRVAR            Takes a vector of coefficients and generates a set
                     of scalar variables.

  LOGDEN             Tells SHAZAM that the equation given on the EQ
                     command is the LOG-DENSITY for a single observation.

  MAXFUNC            Maximizes the function given on the EQ command.

  MINFUNC            Minimizes the equation given on the EQ command.

  NOCONEXOG          Do not include a constant in the list of exogs variables.

  NOPSIGMA           Suppress printing of the sigma matrix.

  NUMCOV             Uses numeric difference method to compute derivatives
                     in the algorithm.

  NUMERIC            Uses the numeric difference method to compute
                     derivative used in the algorithm.

  OPGCOV             Uses the outer-product of the Gradient method to
                     compute the covariance matrix.

  PCOV               Prints an estimate of the covariance matrix of
                     coefficients after convergence.

  SAME               Will run the previous NL regression without
                     repeating the EQ commands.

  SOLVE              Solves the equations given on the EQ command as a set
                     of nonlinear simultaneous equations.

  AUTCOV=            Lag length used with the GMM= option.

  CONV=              Specifies the convergence criterion for the coefficients.

  IN=                Reads back the coefficients that were saved with the
                     OUT option.

  ITER=              Specifies the maximum number of iterations.

  GMM=               Specifies the weighting matrix for Generalized Method
                     of Moments estimation.

  METHOD=            Default is Davidon-Fletcher-Powell algorithm.
                     An alternative is the Broyden-Fletcher-Goldfarb-Shanno (BFGS).

  NCOEF=             Specifies the number of coefficients to be estimated.

  ORDER=             Specifies the order of autocorrelation to be
                     corrected for when the AUTO option is used.

  OUT=               Writes out on unit 3 the values of the coefficients
                     after each iteration.  Useful for restarting the model
                     in another run with the IN option described above.

  PITER=             Specifies the frequency with which iterations
                     will be printed.

  SIGMA=             Saves the sigma matrix in the variable specified.

  START=             Uses the values in the specified variable as starting
                     values for the estimation.

  STEPSIZE=          Specifies the stepsize to use with the NUMCOV option
                     to control the differential in numeric derivatives.

  ZMATRIX=           Specifies a matrix to create and use to store the
                     derivatives of the nonlinear function.

Also available are:

   LIST, RSTAT, BEG=, END=, COEF=, COV=, PREDICT=, RESID=, STDERR=, and
   TRATIO=
as described for OLS .


NONPAR Command

The NONPAR command implements nonparametric density estimation and regression smoothing.

In general, the format for density estimation is:

NONPAR vars / DENSITY options

In general, the format for regression smoothing is:

NONPAR depvar indeps / options

The available options are:

  DENSITY            Specifies kernel density estimation is required.

  GNU                Plots results with GNUPLOT.

  PCOEF              For regression smoothing prints the coefficients.

  BRHO=              Specifies the parameter for the boundary modified kernel.

  COEF=              For regression smoothing saves the coefficients in a matrix.

  DELTA=             Used with METHOD=LOWESS to group observations.

  FCSE=              Saves prediction standard errors in the variable specified.

  HATDIAG=           Saves the diagonal elements of the smoother matrix.

  INCOVAR=           Specifies an input covariance matrix in lower triangular
                     form.

  ITER=              Used with METHOD=LOWESS to specify the number of iterations.

  METHOD=            The options are EPAN, MULTI, NORMAL, QUARTIC, TRIANG,
                     UNIFORM, and LOWESS.

  RWEIGHTS=          For METHOD=LOWESS saves the robustness weights.

  SIGMA=             Saves conditional standard errors in the variable specified.

  SMATRIX=           Saves the N x N smoother matrix.

  SMOOTH=            Specifies the value of the smoothing parameter.

Also available are:

     LIST, BEG=, END=, PREDICT=, and RESID=.


FC Command

The FC command uses a set of regression coefficients and a set of data to compute predicted values of the dependent variable and compare the predictions to the actual values.

The format of the FC command when you are using the estimated coefficients from the previous regression, is:

FC / options

The format when reading in all the coefficients is:

FC depvar indeps / options COEF=

The available options are:

  AFCSE              Used with BLUP or IBLUP option.

  BLUP,IBLUP         Used in autoregressive models when the user wishes the
                     predicted values to be adjusted with the lagged residual,
                     thereby giving the best linear unbiased predictions.

  DYNAMIC            Performs DYNAMIC forecasts for models with a lagged
                     dependent variable.

  MAX                Equivalent to the LIST, GF, and RSTAT options.

  PERCENT            Displays the percentage change in the actual and
                     predicted values.

  CSNUM=             Specifies the cross-section to use on a Pooled
                     Cross-Section Time-Series model.

  ESTEND=            Specifies the last observation of the estimation
                     for the AUTO and POOL models.

  FCSE=              Saves the Forecast Standard Errors.

  NC=                Specifies the number of cross-sections in a Pooled
                     Cross-Section Time-Series model.

When the COEF= option is given the following options may be used:

  NOCONSTANT         Forced through the origin.  No intercept should be
                     included.

  UPPER              Used when a TOBIT model is being run with the FC command.

  LIMIT=             Used when MODEL=TOBIT is specified.

  MODEL=             Specifies the model as : AUTO, BOX, LOGIT, OLS, POOL,
                     PROBIT or TOBIT.

  ORDER=             Specifies the order for the model for MODEL=AUTO.

  POOLSE=            Specifies the standard error of the cross-section in a
                     Pooled Cross-Section Time-Series model when POOL=TOBIT.

  RHO=               Specifies a value of the first-order autoregressive
                     parameter for the regression.

  SRHO=              Used with the RHO= option to specify a value of the
                     second order rho in a second order autoregressive model.

Also available on the FC command are the:

   GF, LIST, BEG=, END=, PREDICT=, and RESID=

options as defined for OLS .


BOX Command

The BOX command does Box-Cox regressions.

In general, the format is:

BOX depvar indeps / options

The available options are:

  ACCUR              Iterates to an accuracy of .001 instead of .01 for lambda.

  ALL                Extends the Box-Cox model so that all variables receive
                     the same power transformation.

  AUTO               Will simultaneously estimate lambda and the first-order
                     autocorrelation parameter rho.  With this option the DROP,
                     NMISS= and GAP= options as described for the AUTO command
                     can be used.

  DN                 Uses N instead of N-K as the divisor for SIGMA**2.

  DUMP               Gives some intermediate output on the iterations.

  FULL               Combined Box-Cox and Box-Tidwell estimation is
                     done and all variables have different lambdas.

  NOCONSTANT         Forces the regression through the origin.

  TIDWELL            Box-Tidwell regression is done.

  UT                 Untransforms the observed and predicted dependent variables.

  COEF=              Saves the coefficients, the lambda for each independent
                     variable and the rho if there is autocorrelation in the
                     matrix specified.

  LAMBDA=            Specifies a fixed value for lambda.

  LAMS=              Used to do a manual grid search of lambda.  The starting
  LAME=              value (LAMS=), the ending value (LAME=) and an increment
  LAMI=              (LAMI=) of lambda should be specified.

  RHO=               Specifies the rho to be used when the AUTO option is
                     specified.

Other available options are:

  ANOVA, GF, LIST, MAX, PCOR, PCOV, RESTRICT, RSTAT, BEG=, END=, COV=,
  PREDICT=, and RESID=
as defined for OLS .

The available temporary variables on the BOX command are:

  $ANF, $DF, $DW, $ERR, $K, $LLF, $N, $R2, $R2OP, $RAW, $RHO, $SIG2, $SSE,
  $SSR, $SST, $ZANF, $ZDF, $ZSSR, and $ZSST.


MLE Command

The MLE command does Maximum Likelihood Estimation of some regression models with non-normal errors.

In general, the format of the MLE command is:

MLE depvar indeps / options

   The ANOVA, DUMP, GF, LINLOG, LIST, LOGLIN, LOGLOG, MAX, NOCONSTANT, 
   NONORM, PCOR, PCOV, RSTAT, BEG=, END=, COEF=, COV=, PREDICT=, RESID=,
   STDERR=, TRATIO=, and WEIGHT=

options as defined for OLS are available on the MLE command. Also,

   CONV=, IN=, ITER=, OUT=, and PITER=

as defined for NL may be used on the MLE command.

The following additional options are available.

  LM                 Performs a Lagrange Multiplier test of the estimated
                     models against a less restricted model.

  METHOD=            Specifies the nonlinear algorithm to use.

  TYPE=              Specifies the type of distribution to be assumed for
                     the errors as EXP (the default), WEIBULL, EWEIBULL,
                     BETA, GAMMA, EGAMM, GG, EGG, LOGNORM, or EXTREMEV.

The available temporary variables on the MLE command are:

   $DF, $DW, $ERR, $K, $LLF, $N, $R2, $R2OP, $RAW, $RHO, $SIG2, $SSE, $SSR,
   $SST, $ZDF, $ZSSR, and $ZSST.


ARIMA Command

The ARIMA command estimates Univariate ARIMA models. There are 3 forms of the command: IDENTIFICATION, ESTIMATION, and FORECASTING. The specified options determine which form of the ARIMA command is in effect.

IDENTIFICATION PHASE

In general, the format is:

ARIMA vars / options

where vars is a list of variables. The available options are:

  ALL                Compute all orders up to and including NDIFF and NSDIFF.

  IAC                Compute inverse autocorrelations.
                     (reference: Cleveland, Technometrics, 1972, pp.277-293)

  GNU                Prepare GNUPLOT plots for the PLOTAC, PLOTDATA, and
                     PLOTPAC options.

  LOG                Take logs of the data.

  PLOTAC             Plot autocorrelation function.

  PLOTDATA           Plot data.

  PLOTPAC            Plot partial autocorrelation function.

  WIDE/NOWIDE        Controls size of terminal screen output.

  ACF=               Saves the AutoCorrelation Function in the variable
                     specified.

  BEG= END=          First and last observation to be used.

  NDIFF=             Order of differencing to transform the data.

  NLAG=              The number of lags for autocorrelations.  The default is 24.

  NLAGP=             The number of lags for partial autocorrelations.  The
                     default is 12.  (The value for NLAGP= must not exceed
                     that for NLAG=).

  NSDIFF=            Order of Seasonal Differencing.
                     If this is specified then NSPAN= must be set.

  NSPAN=             The number of periods for the seasonal cycle.
                     For example, set NSPAN=4 for quarterly data and
                     NSPAN=12 for monthly data.

  PACF=              Saves the Partial AutoCorrelation Function in the
                     variable specified.

  TESTSTAT=          Saves the Ljung-Box-Pierce statistics (computed at every
                     lag) in the variable specified.

The available temporary variables are:

     $N - the number of observations used in the identification phase.

ESTIMATION PHASE

In general, the format is:

ARIMA var / NAR= NMA= options

where var is a variable. The available options as used for the IDENTIFICATION phase are:

   LOG, WIDE, BEG=, END=, NDIFF=, NSDIFF=, and NSPAN=.

The available options as used for the OLS command are:

   ANOVA, PCOR, PCOV, COV=, STDERR=, and TRATIO=.

Other options are:

  DN                 Computes the estimated variance of the regression by
                     dividing SIGMA**2 by N instead of (N-K).

  GNU                Prepare GNUPLOT plots for the PLOTRES option.

  NOCONSTANT         No intercept in model.

  PITER              Print every iteration.

  PLOTRES            Plot the residuals.

  RESTRICT           Use zero starting values as zero restrictions.

  START              Starting values for coefficients follow the ARIMA command.

  ACF=               Saves the AutoCorrelation Function of the estimated
                     residuals in the variable specified.

  COEF=              Save Coefficients in variable specified.

  ITER=              Maximum number of iterations. The default is 50.

  NAR=               Order of the AR process. (REQUIRED)

  NMA=               Order of the MA process (REQUIRED).

  NSAR=              Order of the Seasonal AR process.
                     If this is specified then NSPAN= must be set.

  NSMA=              Order of the Seasonal MA process.
                     If this is specified then NSPAN= must be set.

  PREDICT=           Save Predicted values in variable specified.

  RESID=             Save Residuals in variable specified.

  START=             A vector of starting values.

  TESTSTAT=          Saves the Ljung-Box-Pierce statistics (computed at every
                     lag - up to 60 lags) in the variable specified.

Following model estimation the available temporary variables as for the OLS command are:

   $ERR, $K, $N, $R2, $SIG2, $SSE, $SSR, and $SST.

FORECASTING PHASE

In general, the format is:

ARIMA var / COEF= NAR= NMA= FBEG= FEND= options

where var is a variable. The available options as used for the IDENTIFICATION and ESTIMATION phase are:

   LOG, NOCONSTANT, BEG=, END=, NAR=, NDIFF=, NMA=, NSAR=, NSDIFF=, NSMA=,
   NSPAN=, PREDICT=, and RESID=.

Other options are:

  GNU                Prepare GNUPLOT plots for the PLOTFORC option.

  PLOTFORC           Plot the forecast with error bounds.

  STOCHAST           For first differenced models compute the
                     Beveridge/Nelson stochastic trend.
                     (reference: Journal of Monetary Economics, 1981).

  COEF=              Input coefficient variable. If this is not specified
                     then coefficients must be entered on the line following
                     the ARIMA command.

  FBEG=              The origin date of the forecast (required).
  FEND=              Last observation to forecast (required).
                     (The maximum number of forecasts is 200.)

  FCSE=              Saves the forecast standard errors in the variable.

  SIGMA=             Used in calculating the forecast standard errors.
                     By default the data for the current sample period will
                     be used to estimate SIGMA. A recommended approach is
                     following estimation enter GEN1 S=SQRT($SIG2).
                     Then use the option SIGMA=S for the ARIMA forecasting.


BAYES Command

The BAYES command is used to specify the number of replications desired, and is also followed by the RESTRICT commands that specify the inequality restrictions. Following the RESTRICT commands, an END command must be placed. Finally, the BAYES command must itself follow an estimation command such as OLS .

In general, the format is:

BAYES / options NSAMP=

The available options on the BAYES command are:

  NOANTITHET         Omits the ANTITHETIC replications.

  NORMAL             Assumes the coefficients are normally distributed
                     rather than t distributed.

  PSIGMA             Prints the original covariance matrix.

  DF=                Specifies the degrees of freedom for the t-distribution.

  NSAMP=             Used to specify the number of desired replications.

  OUTUNIT=           Specifies a Unit to use to write all the replicated
                     coefficients.


CONFID Command

SHAZAM automatically computes the Confidence Intervals for any coefficients from a previous estimation with the CONFID command.

In general, the format is:

CONFID co1 co2 ... / options

The available PLOT options are:

   EGA, GNU, GOAWAY, GRAPHICS, HERCULES, HOLD, NOBLANK, PAUSE, PRINT, SYMBOL,
   VGA, WIDE/NOWIDE, MODE=, XMAX=, XMIN=, YMAX=, and YMIN=.

Other available options are:

  NOFPLOT            Deletes the joint confidence region plot if two
                     coefficients were specified.

  NOMID              Used when you are computing the FPLOT.

  NORMAL             Normal distribution used in computing confidence regions.

  NOTPLOT            Deletes the computation of the confidence intervals for
                     individual coefficients based on the t or normal
                     distribution and suppresses the drawing of the (+) symbol.

  DF=                Specifies the Degrees of Freedom required.

  FCRIT=             Specifies a critical value from an F distribution to
                     use for the joint confidence region plot.

  POINTS=            Specifies the number of points desired in the
                     construction of the confidence ellipse for the FPLOT.

  TCRIT=             Allows the calculation of intervals other than the
                     95% interval.


DIAGNOS Command

The DIAGNOS command performs diagnostic tests after OLS estimation.

In general, the format is:

DIAGNOS / options

The available options are:

  ACF                Prints the Autocorrelation Function of residuals
                     and associated test statistics for various orders of
                     autocorrelation.

  BACKWARD           Used with the RECUR option if you want to compute
                     BACKWARDS recursive residuals.

  BOOTLIST           Used with the BOOTSAMP= option to print the entire
                     list of Bootstrapped coefficients for every
                     generated sample.

  CHOWTEST           Used to obtain a set of sequential Chow test
                     statistics and sequential Goldfeld-Quandt test
                     statistics which split the sample in 2 pieces at
                     every possible point.

  CTEST              Computes Pinkse's C-test for serial independence of
                     the residuals and saves the results in $CTES.

  GNU                Prepares gnuplot plots of the recursive residuals when
                     the MAX or RECUR options are specified.

  HET                Runs a series of tests for heteroskedasticity.

  JACKKNIFE          Runs a series of regressions, successively omitting
                     a different observation to get the "Jackknife
                     Coefficient Estimates".

  LIST               Prints a table of observed and predicted values of
                     the dependent variable, and regression residuals.

  MAX                This option is equivalent to specifying the LIST,
                     RECUR, ACF, RECEST, RECRESID, BACKWARD, CHOWTEST,
                     RESET and HET options.

  RECEST/            Used with the RECUR option if you want to print the
   NORECEST          recursive estimated residuals.

  RECRESID/          Used with the RECUR option if you want to print the
   NORECRESID        recursive residuals.

  RECUR              Performs Recursive Estimation by running a series
                     of regressions adding one observation per regression.
                     Recursive residuals and CUSUM tests are printed.

  RESET              Used to compute the Ramsey RESET specification tests.

  WIDE/NOWIDE        NOWIDE reduces the width of output to 80 columns from
                     120 columns.

  BOOTSAMP=          Bootstraps the previous OLS regression.

  BOOTUNIT=          Writes out the generated coefficients for each sample
                     in a Bootstrap experiment on the unit specified.

  CHOWONE=           Specifies the breakpoint for the Chow test.

  GQOBS=             Used with the CHOWTEST option to specify the number
                     of observations to be omitted for the Goldfeld-Quandt
                     test.

  MHET=              Used with CHOWTEST or RECUR options to specify M, the
                     number of residuals to use in Harvey's Recursive
                     Residual Exact Heteroskedasticity test.

  RECUNIT=           Used with the RECUR option if you want to write the
                     recursive estimates and residuals on the Fortran Unit
                     number specified.

  SIGLEVEL=          Used with RECUR option to specify the significance
                     level desired for the CUSUM and CUMSUMSQ tests.


PC Command

The PC command does Principal Components and Factor Analysis.

In general, the format of the PC command is:

PC vars / options

The available options are:

  COR                Analysis is done on the correlation matrix.

  LIST               Lists the matrix of Principal Components.

  MAX                Equivalent to specifying the PEVEC,PRM,PFM,LIST
                     and NCODE=1 options.

  PCOLLIN            Prints the proportion of the variance of the
                     estimate accounted for by each principal component.

  PEVEC              Print the matrix of EigenVECtors.

  PFM                Prints the Factor Loading Matrix for the retained
                     factors.

  PRM                Computes and prints the Rotated Factor Matrix by a
                     Varimax rotation.

  RAW                Indicates that analysis is to be done on the original
                     RAW cross-product matrix instead of the deviations
                     from the means cross-product matrix.

  SCALE              Indicates that analysis is to be done on a scaled
                     cross-product matrix.

  EVAL=              Specifies the vector in which to save the EigenVALues.

  EVEC=              Specifies the matrix in which to save the EigenVECtors.

  MAXFACT=           Specifies the Maximum number of FACTors to be retained.

  MINEIG=            Specifies the minimum eigenvalue allowed to retain a
                     component.

  NC=                The normalization code which must be specified if a
                     listing of the actual retained components is desired.

                     The codes are:
                       1.  No normalization, variance is lambda/(n-1)
                       2.  Variance is 1/(n-1)
                       3.  Variance is lambda
                       4.  Variance is 1
                     where lambda is the eigenvalue corresponding to the
                     component and n is the number of observations.

  PCINFO=            Saves PC information needed for regressions on
                     Principal Components in the matrix specified.

  PCOMP=             Specifies a matrix into which the principal
                     components will be put.


POOL Command

The POOL command does Pooled Cross Section - Time Series estimation under certain model specifications and conditions.

In general, the format is:

POOL depvar indeps / options NC=

The available options are:

  CORCOEF            Estimates rho using the correlation coefficient form
                     to confine the estimate of rho to [-1 , +1].

  DN                 Uses a divisor of T instead of T-K for the PHI matrix.

  FULL               Estimates the full cross-sectionally correlated and
                     and time wise autoregressive model.

  HETCOV             Pooled OLS estimation with "Panel Corrected Standard 
                     Errors" (PCSE). 

  MULSIGSQ           Multiplies Covariance Matrix by Sigma**2.
                     Use NOMULSIGSQ to turn this off.

  SAME               Forces all values of rho to be equal for each of
                     the cross-sections.

  CONV=              Convergence criteria when ITER= is used.

  ITER=              Maximum number of iterations.

  NCROSS=            Specifies the number of cross-sections in the data.

  NTIME=             Specifies the number of time periods in the data.

  RHO=               Specifies a value of rho to use instead of the
                     estimated values.

Other available options are:

    ANOVA, DLAG, DUMP, GF, LINLOG, LIST, LM, LOGLIN, LOGLOG, MAX, 
    NOCONSTANT, PCOR, PCOV, RESTRICT, RSTAT,  BEG=, END=, COEF=,
    COV=, PREDICT=, RESID=, STDERR=, and TRATIO=

as defined for OLS and BLUP and UT as for GLS .

The available temporary variables on the POOL command are:

    $ANF, $DF, $ERR, $DW, $K, $LLF, $N, $R2, $R2OP, $RAW, $RHO, $SIG2,
    $SSE, $SSR, $SST, $ZANF, $ZDF, $ZSSR, and $ZSST.


ROBUST Command

The ROBUST command performs the Robust Estimation Method.

In general, the format is:

ROBUST depvar indeps / options

The available options are:

  FIVEQUAN           Uses the five quantile values in computing regression
                     quantiles.

  GASTWIRT           Uses the Gastwirth weighting scheme in computing
                     regression quantiles.

  LAE                Default method that specifies the Least Absolute
                     Error model.

  TUKEY              Uses the Tukey Trimean weighting scheme in computing
                     regression quantiles.

  UNCOR              Used with the MULTIT= option and it specifies that
                     the uncorrelated error model is requested.

  CONV=              Used with the MULTIT= option and it specifies a
                     convergence criterion to stop the iterative procedure.

  DIFF=              Used with the MULTIT= option and it specifies the
                     maximum number of iterations allowed in the
                     iterative procedure.

  ITER=              Used with MULTIT to specify the maximum number of
                     iterations allowed in the iterative procedure used
                     to obtain the estimates.

  MULTIT=            Specifies that the Multivariate-t method should be used.

  THETA=             Specifies a value of THETA to use for the Regression
                     Quantile method.

  THETAB=            These options specify the beginning, ending and
  THETAE=            increments to be used for a grid search over THETA
  THETAI=            values for the Regression Quantile method.

  TRIM=              Specifies that Trimmed Least Squares is desired.

Other options available are:

  GNU, LINLOG, LIST, LOGLIN, LOGLOG, MAX, PCOR, PCOV, RSTAT, BEG=, END=,
  COEF=, COV=, PREDICT=, RESID=, STDERR=, and TRATIO=

as defined for OLS .


HET Command

The HET command implements the estimation of models which require correction for heteroskedastic errors.

In general, the format is:

HET depvar indeps (exogs) / options

The available options are:

  PRESAMP            Sets the initial value for the presample data
                     in all iterations.

  ARCH=              Specifies the order q of the ARCH process.

  ARCHM=             Specifies the functional form for the ARCH-in-Mean
                     term. Set ARCHM=0 for the log of the variance function
                     and set ARCHM=0.5 for the square root of the variance
                     function.

  GARCH=             Specifies the order p of the lagged conditional
                     variances in the GARCH process.

  GMATRIX=           Specifies an NxK matrix to use to store the derivatives
                     of the log-likelihood function at each observation.
                     This option is not available with the NUMERIC option.

  MACH=              Specifies the order of the moving average process in
                     models with ARCH errors.

  MODEL=             Specifies the form of heteroskedasticity.
                     Available options are:

                         ARCH   - for ARCH models
                         DEPVAR - Dependent Variable Heteroskedasticity
                         MULT   - Multiplicative Heteroskedasticity
                         STDLIN - Standard Deviation is a linear function
                                  of exogenous variables
                         VARLIN - Variance is a linear function of exogenous
                                  variables

  START=             Specifies a vector of starting values for the estimation.

  STDRESID=          Saves the standardized residuals in the variable specified.

Other available options are:

    DUMP, LINLOG, LIST, LOGLIN, LOGLOG, MAX, NOCONSTANT, PCOR, PCOV, RSTAT,
    BEG=, END=, COEF=, COV=, PREDICT=, RESID=, STDERR=, and TRATIO=

as defined for the OLS command.

    NUMCOV, NUMERIC, OPGCOV, CONV=, ITER=, METHOD=, PITER=, and STEPSIZE=

as defined for the NL command.

Temporary variables that are available are:

    $ERR, $K, $LLF, $N, and $R2OP.


LP Command

The LP command gives the solution to a linear programming problem.

In general, the format is:

LP c A b / options

The available options are:

  UMP                Prints information for SHAZAM consultants.

  MIN                Indicates that the problem is a MINimization problem.

  ITER=              Maximum number of iterations.

  DUAL=              Saves the dual solution in the vector specified.

  DSLACK=            Saves the dual slack variables in the vector specified.

  PRIMAL=            Saves the primal solution in the vector specified.

  PSLACK=            Saves the primal slack variables in the vector specified.


EXEC Command

The EXEC command runs a SHAZAM procedure. Also see PROC and PROCEND .

The format is:

EXEC procname


PROC Command

The PROC command is the first command in a SHAZAM procedure. Also see PROCEND and EXEC .

The format is:

PROC procname .... procedure commands .... DELETE / ALL_ (optional) PROCEND


PROCEND Command

The PROCEND command is the last command in a SHAZAM procedure. Also see PROC and EXEC .


CHECKOUT Command

The CHECKOUT command displays characteristics of the machine and operating system for SHAZAM running at this installation. This is usually only useful for debugging purposes.


PAR Command

The PAR command is used to change the amount of memory used as a workspace in SHAZAM. Par can be set on a system command or it can be changed in a SHAZAM session.

The format is:

PAR number

For example:

     PAR 50

The PAR command is not available for SHAZAMC or SHAZAMN on the PC or Macintosh versions or some mainframe versions. On SHAZAME, SHAZAMO, SHAZAMP see instructions in the manual.


FILE Command

The FILE command is used to assign units to files.

The format is:

FILE unit filename

For example:

     FILE 11 MYDATAFILE

In addition, the FILE command can be used to assign special units, these are:

  FILE CLOSE filename      - Closes the specified filename.

  FILE DELETE filename     - Delete the specified filename.

  FILE HELPDEMO pathname   - Specifies a directory path for the
                             SHAZAM HELP and DEMO files.

  FILE INPUT filename      - If all the commands are in a file,
                             this command is used.

  FILE KEYBOARD filename   - Places a copy of all commands typed
                             on a keyboard into a file.  Not
                             available on all computers.

  FILE LIST filename       - Lists the specified file on the screen.

  FILE OUTPUT filename     - Puts the output in the assigned filename.
                             This should be the first FILE command.

  FILE PATH pathname       - Specifies a directory path to use to
                             search for files.

  FILE PLOTPATH pathname   - Specifies a directory path for the
                             GNUPLOT program.

  FILE PRINT filename      - Prints the file specified on a printer.
                             (Available for Macintosh only)

  FILE PROC filename       - Load the PROC from filename.

  FILE PROCPATH pathname   - Searches the directory path for PROCS.

  FILE SCREEN filename     - Displays the output on the screen and puts
                             the output in a file simultaneously.

  FILE TEMP pathname       - Specifies a directory path for temporary
                             scratch files.


FLS Command

The FLS command implements time-varying linear regression via flexible least squares.

In general, the format is:

FLS depvar indeps / options

where depvar is the dependent variable and indeps is a list of independent variables with coefficients which evolve over time.

The available options are:

  GNU                Plot the time paths of the coefficients with GNUPLOT.

  PCOEF              Print the sequence of coefficient estimates.

  COEF=              Saves the FLS estimates in a N x K matrix.

  DELTA=             Specifies the smoothing weight in the range (0,1).
                     The default value is DELTA=.5.

Other available options:

   MAX, NOCONSTANT, BEG=, END=, PREDICT=, and RESID=.


LAMBDA Command

The LAMBDA restriction command is used with the BOX command to impose restrictions on LAMBDA on individual variables. The RESTRICT option on the BOX command must be specified if the LAMBDA command is used. The END command must follow the restrictions.

For example:

     BOX MONEY INCOME INTRST / RESTRICT ALL
     LAMBDA INCOME=0
     END


Continuation Lines

Commands may be up to 4096 characters long including any continuation lines. To continue a command to the next line you type a & character at the end of the line, for example:

     OLS CONSUME INCOME PRICE / RSTAT PCOV &
     PCOR LIST

This is exactly equivalent to:

     OLS CONSUME INCOME PRICE / RSTAT PCOV PCOR LIST

There is no limit to the number of continuation lines, but the total number of characters in any command must not exceed 4096. Many computers will not accept a single line longer than 80 columns however, continuation lines can be used.


Temporary Variables

Temporary Variables are statistics available for use in GENR or GEN1 statements that have been computed from previous commands.

The names are:

    $ADR2, $ANF, $CDF, $CHI, $CRIT, $CT11, $CT12, $CT22, $DF, $DF1, $DF2,
    $DO, $DO2-$DO8, $DW, $ERR, $F, $K, $LLF, $N, $PDF, $PI, $R2, $R2OP,
    $RAW, $RHO, $SIG2, $SSE, $SSR, $SST, $STES, $T, $VAL, $VAL1, $VAL2,
    $ZANF, $ZDF, and $ZSSR.

Definitions of these Temporary Variables can be found in the manual.


Error Codes

   * Error Codes for all commands.

     1010   ...INSUFFICIENT MEMORY-AT LEAST ??? MORE WORDS NEEDED
               THIS REQUIRES AT LEAST PAR=???  CURRENT PAR=???

     1011   ...ERROR..DATA ON VARIABLE ??? DOES NOT EXIST

     1012   ...INVALID OPTION

     1013   ...ERROR.. ??? EXCEEDS MAXCOL=??? USE SET MAXCOL=

     1014   ...ERROR..VARIABLE ??? DOES NOT EXIST

     1015   ...ERROR.. ??? EXCEEDS MAXCOL= ??? USE SET MAXCOL=
     1016   ...SYNTAX ERROR IN LINE ABOVE NEAR $

     1017   ...ERROR ... COMMAND TOO LONG 

   * Error Codes for the OLS command.

     6010   ...ERROR..VARIABLE ??? IS ON BOTH SIDES OF EQUATION

     6011   ...ERROR..IN RIGHT HAND SIDE OF OPTION ???

     6012   ...ERROR..INCOMPATIBLE OPTIONS

     6013   ...INSUFFICIENT NUMBER OF VARIABLES SPECIFIED

     6014   ...ERROR ??? OBSERVATIONS IS NOT ENOUGH FOR ESTIMATION

     6015   ...ERROR..RESTRICT AND RIDGE OPTIONS NOT ALLOWED

     6016   ...ERROR..VARIABLE ??? IS NOT VALID

     6017   ...INSUFFICIENT MEMORY TO CORRECT COVARIANCE MATRIX
                

    * Error Codes for the TEST, RESTRICT, and LAMBDA commands.

    15008   ...INSUFFICIENT MEMORY FOR THIS CALCULATION

    15009   ...SYNTAX ERROR IN LINE ABOVE

    15010

    15011   ...ERROR..??? DOES NOT EXIST

    15012   ...ERROR.. ??? IS NOT IN EQUATION

    15013   ...ERROR..MORE EQUATIONS THAN PARAMETERS

    15014   ...ERROR..THIS COMMAND MUST FOLLOW AN ESTIMATION COMMAND

    15015   ...TEST COULD NOT BE DONE

    15016   ...RESTRICTIONS COULD NOT BE DONE

    15017   ...ABOVE COMMAND IS INVALID FOR THIS PROBLEM

    15018   ...ERROR.COMMAND IS NOT A RESTRICT, TEST, OR END COMMAND
      

    * Error Codes for the PLOT and CONFID commands.

    28010   ...ERROR IN RIGHT HAND SIDE OF OPTION ???

    28011   ...ERROR..INVALID COEFFICIENTS OR OPTIONS

    28012   ...AXIS ERROR: TRY NOPRETTY OPTION ??? ??? ??? ??? 


    * Error Codes for the POOL command.

    29010   ...ERROR..MISSING RIGHT HAND SIDE OF OPTION ???

    29011   ...ERROR..INCOMPATIBLE OPTIONS

    29012   ...ERROR..NCROSS= OR NTIME= NOT SPECIFIED

    29013   ...ERROR..INVALID LIST OF EXOGENOUS VARIABLES

    29014   ...ERROR..TIME-PERIODS NOT OF EQUAL LENGTH

    29015   ...ESTIMATION FAILED DUE TO UNSTABLE RHO = ???

    29016   ...ERROR..FOUND IN ROW ???

    29017   ...MATRIX IS NOT POSITIVE DEFINITE

    29018   ...ITERATION LIMIT EXCEEDED 


    * Error Codes for the SYSTEM and 2SLS command.

    50010   ...ERROR..MISSING RIGHT HAND SIDE OF OPTION ???

    50011   ...ERROR..NO. OF EQUATIONS NOT GIVEN ON SYSTEM COMMAND

    50012   ...ABOVE COMMAND IGNORED, IS NOT ALLOWED AT THIS TIME

    50013   ...ERROR..UNIT ??? IS RESERVED FOR INTERNAL USE

    50014   ...PROBLEM TOO LARGE.TOO MANY VARIABLES OR RESTRICTIONS

    50015   ...ERROR...EQUATION ??? IS UNDERIDENTIFIED

    50016   ...SIGMA.MATRIX IS NOT POSITIVE DEFINITE

    50017   ...MATRIX IS NOT POSITIVE DEFINITE

    50018   ...ERROR..EXOG VAR ?? IS A CONSTANT..A CONSTANT IS
               AUTOMATICALLY INCLUDED

    50019   ...MATRIX OF EXOGENOUS VARIABLES IS NOT POSITIVE DEFINITE
               

    * Error Codes for the NL, MLE, and HET commands.

    52010   ...ERROR..SAME OPTION NEEDS PREVIOUS NL OR MLE COMMAND

    52011   ...ERROR.. START= VARIABLE DOES NOT EXIST

    52012   ...ERROR..LOGLOG, LOGLIN, LINLOG OPTIONS NOT AVAILABLE

    52013   ...ERROR..CONVERGENCE SET TO ZERO

    52014   ...ERROR..IN RIGHT HAND SIDE OF OPTION ??

    52015   ...ERROR.. YOU FORGOT TO SPECIFY NCOEF=

    52016   ...ERROR.. EXOGENOUS VARIABLES NOT ALLOWED

    52017   ...ERROR..ARCH ERRORS NOT ALLOWED

    52018   ...ERROR.. INVALID MODEL= OPTION

    52019   ...ERROR.. TYPE= OPTION NOT ALLOWED

    52020   ...SYNTAX ERROR IN LINE ABOVE

    52021   ...ERROR..ATTEMPT TO USE ??? PARAMETERS FROM UNIT ??? FAILED

    52022   ...ERROR..SET SAMPLE COMMAND TO USE ONLY ONE OBSERVATION

    52023   *** FAILURE TO COMPLETE A LINE SEARCH IN 20 FUNCTION
                EVALUATIONS. THIS IS PROBABLY BECAUSE THE LIKELIHOOD
                FUNCTION HAS NO GLOBAL MAXIMUM

    52024    No Message: Too Many Parameters required, XXIT(10)

    52025   ...COVARIANCE MATRIX AND STD ERRORS ARE INCORRECT

    52026   ...ERROR..DEP VAR IN OB. NO. ??? IS NOT POSITIVE

    52027   ...MAXIMUM NUMBER OF ITERATIONS

    52028   ...ERROR..GMM=??? IS INVALID OR WRONG DIMENSION

    52029   ...HESSIAN IS NOT POSITIVE DEFINITE

    52030   ...GRAD TRANSPOSE TIMES X GREATER THAN OR EQUAL ZERO
               CONV SET TOO SMALL?  


    * Error Codes for the PROBIT, LOGIT and TOBIT commands.

    55010   ...ERROR IN RIGHT HAND SIDE OF OPTION ???

    55011   ...ERROR ??? OBSERVATIONS IS NOT ENOUGH FOR ESTIMATION

    55012   ...MAXIMUM ITERATIONS REACHED

    55013   ...MODEL HAS BLOWN UP WITH -XB= ???

    55014   ...ERROR..DEPENDENT VARIABLE IS A CONSTANT

    55015   ...OBSERVATION ??? HAS ILLEGAL DEP VAR ?? 

    
    * Error Codes for the TOBIT command.

    57012   ...THERE ARE NO OBSERVATIONS AT THE LIMIT

    57013   ...ALL OBSERVATIONS ARE AT THE LIMIT 


    * Error Codes for the PC command.

    65010   ...MISSING RIGHT HAND SIDE OF OPTION ???

    65011   ...DISASTER..IN EIGENVALUE OF ROW ???  


    * Error Codes for the ARIMA command.

    72010   ...ERROR..MISSING RIGHT HAND SIDE OF OPTION ???

    72011   ...ERROR..INVALID TYPE= OPTION

    72012   ...ERROR..INVALID NLAG OR NLAGP

    72013   ...ERROR..SPAN OF SEASONAL CYCLE NOT SET

    72014   ...ERROR..THE FORECAST LIMIT IS 200

    72015   ...ERROR..REQUIRE FEND>FBEG

    72016   ...ERROR..FBEG IS OUTSIDE SAMPLE RANGE

    72017   ...ERROR..START= VARIABLE DOES NOT EXIST

    72018   ...ERROR..NO VARIABLE SPECIFIED FOR ANALYSIS

    72019   ...ERROR..PRE-SAMPLE VALUES EXCEEDS 104

    72020   ...ERROR..TOO FEW OBSERVATIONS

    72021   ...ERROR..TOO FEW COEFFICIENTS OR INVALID COEFFICIENTS

    72022   ...ERROR..ORIGIN DATE IS LESS THAN MAXIMUM AR LAG

    72023   ...ERROR..INVALID PARAMETER LIST

    72024   *** SPECIFICATION ERROR, PLEASE CHANGE ***

    72028   ...NO CONVERGENCE AFTER ??? ITERATIONS
                  INITIAL SUM OF SQS= ???   FINAL SUM OF SQS =???

    72029   ...FAILED TO REDUCE SSE FURTHER AFTER ??? ITERATIONS
                  INITIAL SSE = ???  FINAL SSE = ???
               

    * Error Codes for the COINT command.

    72025   ...ERROR..COINTEGRATION TESTS NEED MORE VARIABLES

    72026   ...ERROR..NDIFF= OPTION NOT AVAILABLE

    72027   ...ERROR..TOO FEW OBSERVATIONS 


    * Error Codes for the ROBUST command.

    77010   ...ERROR..MISSING RIGHT HAND SIDE OF OPTION ???

    77011   ...ILLEGAL THETA VALUES ??? ??? ???

    77012   ...SOLUTION NONUNIQUE

    77013   ...PREMATURE END

    7014    ...EXIT CODE=??? DEGREES OF FREEDOM PROBLEM

    77015   ...TRIM PROPORTION OF ??? TOO LOW OR TOO HIGH WITH ONLY ???
                        OBSERVATIONS

    77016   ...ESTIMATION FAILED WITH DIFF=???  USE DIFF= OPTION

    77017   ...MATRIX IS NOT POSITIVE DEFINITE 


    * Error Codes for the HET command.

    79001   ...THE STARTING VALUES GIVE NEGATIVE VARIANCE OR OVERFLOWS
               

    * Error Codes for the NONPAR command.

    96010   ...ERROR... CONSTANT OPTION NOT ALLOWED
    
    96011   ...ERROR... SMOOTH= OPTION INVALID

    96012   ...ERROR... A REGRESSOR IS REQUIRED

    96013   ...ERROR... ONLY ONE REGRESSOR ALLOWED

    96014   ...ERROR.. xxxx VARIABLE NOT VALID

    96015   ...ERROR... SKIPPED OBSERVATIONS NOT ALLOWED

    96016   ...ERROR.. xxxx OPTION NOT VALID

    96017   ...ERROR.. INCOVAR= ARRAY HAS TOO FEW ELEMENTS

    96020   ...INVALID METHOD= OPTION

    96021   ...ERROR .. DIVIDE BY ZERO