The MODLER Statistical Information and Modeling System is a time-tested, integrated, and
comprehensive software package that permits a variety of operations to be performed. For
many years, it has been used worldwide by economists in banks, corporations, national and
international governmental organizations, and universities and other research institutions,
among other places. Historically, it is one of the first interactive software packages
ever developed. It is also the first integrated econometric modeling package to be ported to the microcomputer,
in 1984 providing the foundation for the first personal computer based econometric forecasting service,
featuring then the well-known Wharton Quarterly Econometric Model of the United States.
At the most advanced level, MODLER enables the estimation, construction, testing, maintenance,
and solution of econometric models containing from 1 to 1000 or more equations, depending upon
the specific version used. These models can be linear or non-linear, simultaneous or recursive,
and dynamic or static. More generally, the software incorporates comprehensive data base
management facilities, a wide range of general data analysis options, high quality screen
and hard-copy graphics displays, and powerful report generation capabilities, including
direct links to spreadsheet and desktop publishing or advanced word processing packages.
It is thus designed to provide both large-scale number crunching capabilities and at the
same time the means to produce attractive and informative reports and other presentation materials.
The program runs on standard Wintel-compatible microcomputers, including notebooks and sub-notebooks. Compatible with Windows XP, Windows 7 and Windows 8. A version designed specifically for use with Windows 10 will be released shortly. Legacy support for earlier operating systems is also available.
MODLER can be licensed as a single system or in quantity. Prices include one year's updates,
improvements, and unlimited telephone/Email support. On-site training can be provided and an
optional annually renewable AlphaPlus support agreement is offered that includes the same
level of technical support, updates, and improvements in the second and subsequent years.
The construction and solution of econometric models are data intensive operations,
the size of a model increases. MODLER offers scalable time series database maintenance facilities
equal to the demands of any likely task. To provide flexibility and security, it permits the
creation and maintenance of both permanent data banks and temporary workspace (Memory) files.
Each data bank can contain as many as 10,000 time series, each series with a maximum of 1600
observations; as many as 15 such data banks can be accessed simultaneously. Memory Files can store
up to 500 series with the same characteristics. In addition, powerful import and export
facilities are available to permit the transfer of data between MODLER and other environments,
including government sources, the vendors who supply data bases in its data bank format, and
the multiplicity of Internet websites, as well as third party spreadsheet and other software
Internally, MODLER provides an extensive set of data management functions, operators and
features, including Census X-11 quarterly and monthly seasonal adjustment, as well as a wide
range of mathematical, statistical, relational, and logical operators; optionally MODLER also
supports weekly seasonal adjustment. The platform has the further capability to generate easily
such analytic aids as pseudo random variates, correlation matrices and several types of growth
triangles. The standard data observation frequencies include weekly, bi-weekly, ten day, lunar,
monthly, bi-monthly, quarterly, semi-annual, annual, and various fiscal and agricultural year
frequencies. Under user control, automatic frequency conversion is fully supported.
MODLER offers a number of access aids. Data Banks and Memory Files are all component-level
documented. Moreover, this documentation, which includes that for individual series, can be
displayed whole or in part in order to assist in the identification and retrieval of these series.
In addition, both series names and documentation can be keyword searched to find particular items.
Index and Find buttons appear on dialogue boxes whenever such information is likely to be helpful.
As appropriate, a Macro button also appears to allow macro file segments to be included in current
commands, thus obviating needless re-keying, not to mention context specific help.
Presentation quality data display features include line graphs, bar charts, dated points,
mixed line-bar, and stacked bar charts. Each may be specified as plots of one or more series
or expressions against time, or as plots of series or expressions against some other series.
Log and semi-log scale plots are supported as well as plots with dual y-axis scales, grid
lines and other features. Full screen editing of screen graphics displays is also supported,
allowing labels, notes and comments to be placed anywhere. In addition, displays may be saved,
generally to the Clipboard, but also JPG, GIF and other such files in order to integrate both
text and graphics into reports or support the generation of Web pages. It is additionally
possible to autogenerate template macros that permit graphs to be reproduced later using the
same or an updated data set.
Within MODLER, the report generation facilities permit the creation of various types of tables,
including both mixed frequency and those with embedded formulas for row and column oriented
transformations. These features may be combined. Tables may include titles, subtitles (left
justified, right justified, or centered), and blank lines. Tables and portions of tables may
be set off with ruling lines. In general, the tables produced are time-labelled, with dates
usually displayed horizontally, although transposed tables can also be produced. Such tables
may show successive observations or selected observations for even widely separated dates.
A particular MODLER feature is the capability to directly address the cells of a standard
spreadsheet, thus permitting time series observations to be copied to those cells or retrieved.
Such observations may be converted on-the-fly from one observation frequency to another; it is
also possible to embed transformations, so as to copy the results of those transformations to
the spreadsheet cells. This facility permits the creation of tabular reports instantly that
otherwise would require extensive spreadsheet macro programming. It may be operated interactively
or templates (macros) may be created more or less as an automatic by-product and used repeatedly
MODLER also produces specialty tables, such as solution comparison tables, displaying differences
and percentage differences variable-by-variable for different model solutions. Other specialty
tables display standard error statistics.
The regression methods currently supported include:
- Ordinary Least Squares
- Restricted Least Squares
- Ordinary Ridge Regression
- Weighted Least Squares
- Generalized Least Squares
- Two Stage Least Squares
- Polynomial Distributed Lags
- Autoregressive corrections
These options may be combined whenever the algorithmic operations involved are inherently
compatible. Furthermore, commands may include embedded expressions that can even combine
mathematical, logical, relational, and statistical operators and functions, both for dependent
and regressor variables.
The usual range of supporting regression statistics and tests are provided, including the
basic t-statistics, R-Square, and Durbin-Watson, plus a very wide range of other useful statistics
such as the Log Likelihood function, Durbin h, Schwarz Criterion, Akaike Information Criterion,
and tests such as unit root, structural stability and specification tests. MODLER also
incorporates various estimation support facilities, including regressor term DROP and ADD
commands, the capability to save residual and predicted sample period values, and provisions
for making disk or hard copies of most displays; it is also possible to archive both commands and
MODLER is evolutionary in its offerings but is not intended to be a specialty regression
package; by design, its focus is on the standard methods commonly used for simultaneous
equation models. However, its data link facilities encourage, permit and support the
joint use of specialist packages such as MicroFit and TSP, thus offering its
users the best of both worlds.
The model construction facilities include the ability to autocode estimated equations, to replace
or edit existing model equations, and automatically to normalize and renormalize equations
analytically with respect to any current period included variables provided of course that
an appropriate inverse exists. Equations can also be normalized or renormalized manually.
Once estimated, models can be changed in various ways. Equations can be copied from one
model to another or from one part of a model to another. Alternatively, equations may be
directly inserted. With some editing, models estimated in other packages can be used; in
particular, it is quite easy to convert models created earlier in such packages as AREMOS,
EPS, EViews, TROLL, and XSIM.
Within MODLER, Models and model equations can be edited using a built-in onscreen command
text editor. Models can also be constructed by merging smaller MODLER models, which can be
simply groups of equations as any group of equations can be treated as a model. These conventions
allow models to be built and tested piece-by-piece, which, among other things, greatly aids
in testing and debugging them.
MODLER incorporates comprehensive model solution facilities. Among these are the capability to
treat a model as a satellite of one or more other models and the ability to specify it as
containing functionally distinct pre-simultaneous, simultaneous, and post-simultaneous equations.
The ability to make these distinctions allows even huge models to be solved almost instantly,
although with today's CPU's increased speed is not so much the issue as the constructive
Specific types of models that can be solved optionally include both rational expectations
models and those incorporating input-output structures. Models can be solved as sets of
independent structural equations or as simultaneous systems; both single and multiperiod
dynamic solutions are allowed. MODLER and its data files can also be used in conjunction
with third-party specialized packages such as WinSolve.
To support the solution of models, a wide range of solution modifications and adjustments
are available, including both additive and multiplicative adjustments, the ability to
exogenize variables for all or part of a solution period, and the ability to specify target
values for some or all the endogenous variable solution values. MODLER will in addition
produce ragged edge solutions, in the case that the most recent historical observations
are available only for certain endogenous variables.
Solution methods include both Gauss-Seidel and Jacobi, damped or undamped. MODLER includes
facilities for monitoring solutions in progress, as well as to display, contrast, and compare
solutions. Model performance may also be exhibited in terms of RMSE, absolute errors, Theil
prediction statistics, and other measures.
Solutions can be displayed graphically, as well as in tables. Solution tables can be
created so as to present the results exactly as the user wishes; however, during the
solution process, results can be displayed immediately using default tables that present
the variables in solution order.
As indicated earlier, MODLER is designed to be used with other software, both other members of
its own software family and third party packages: data can be imported and exported using
standard data interchange file formats, both fixed field and comma separated, including ASCII,
Lotus 1-2-3, and TSD. It can also directly read from and write to spreadsheet cells.
MODLER's DataClp technology additionally permits the capture of text based data from web pages,
textfiles and other sources. This facility is extremely flexible in operation: the files can
take the form of documents or those that are formatted as data tables. As indicated, they can
be located on your machine or they may originate as Internet Web pages. The observations can
be captured singly or as a chronological group.
As needed, MODLER can also be network integrated into corporate data based systems that are
based on standard DBMS technologies, although this usually requires specific customization.
Virtually all forms of MODLER output can be incorporated into textual reports, spreadsheet
tables, and other displays. Data can be directly copied cell-by-cell to spreadsheet files,
and to the files of other relevant number-oriented packages. Provisions also exist to export
directly to desktop publishing and advanced word processing packages; the packages specifically
supported include Pagemaker, Ventura Publisher, Word, Word Pro, and Word Perfect. Custom
versions of MODLER are also available that support professional publishing systems.
Depending upon the specific application, MODLER can be operated either using commands or
the Windows interface by means of menus, icons, and dialog boxes. Using these Windows
facilities, it is also possible to generate automatically macro commands that can be captured
in order to perform repetitive periodic operations. In general, commands can be entered
interactively or formed into macros that can be preserved from one session to the next.
Users who have programming skills, or access to these, can also create programs using
Visual Basic, Delphi, or even programming languages such as C/C++ or Fortran, that incorporate
MODLER offers several types of online help facilities, both general and context specific.
The standard help file offers multiple help levels, so as to provide the appropriate amount
of help needed in each individual case; application specific help files can be organized in
the same way. Users can produce hard copy using the incorporated Print command and can also
annotate and otherwise personalize the help pages.