EasyReg International
Free Econometrics Software for Easy Regression Analysis
Herman J. Bierens
Professor Emeritus of Economics
Pennsylvania State University
University Park, PA 16802, USA
EasyReg (Easy Regression) International (shortly EasyReg)
conducts various econometric estimation and
testing tasks on all 32 bit and 64 bit Windows platforms up to Windows 10, simply by clicking the mouse.
EasyReg also works under Windows 8 and 10, but in that case you need to set the
compatibility mode of EasyReg to Windows XP, and possibly run EasyReg as Administrator.
EasyReg was initially programmed in Visual Basic 3, then migrated to Visual Basic 5 (via Visual Basic 4), and
has been further developed in Visual Basic 5 Enterprise Edition.
EasyReg is designed for use in empirical research (including my own), and for teaching econometrics.
In the latter case the user can choose his or her own econometrics level.
EasyReg is called International because it accepts dots and/or commas as
decimal delimiters, regardless of the local number setting of Windows. Moreover, EasyReg works under
nonEnglish versions of Windows as well without the need to adjust the language
setting of Windows, as long as the language involved uses a roman alphabet, for example German, Dutch, Spanish,
Portuguese, and French Windows versions.
However, if Windows uses a language with a nonRoman alphabet such as Japanese and Chinese, and the EasyReg
menus do not open automatically when you move the mouse pointer over them, the problem is likely due to the
language setting for nonUnicode programs (EasyReg is such a program). If so, click here
for detailed instructions on how to change the language setting for nonUnicode programs.
This will fix the problem without affecting the language of Windows itself.
Under Windows Vista and Windows 7 and 8 it is recommended to always check
the language for nonUnicode programs, because if it is set to a language without a Roman alphabet EasyReg
may seem to work but it will not be completely functional.
It has been verified that EasyReg also works well on Apple & Mac computers via
VirtualBox VM.
The same may apply to Linux computers as well, but that has not yet been verified.
Note that most EasyReg tasks require data. Therefore, when you start up EasyReg for the first
time, the menu items that require data are disabled. To learn how to import data in EasyReg,
consult the guided tours and the Help menu.
Previous versions of EasyReg have been favorably reviewed by
 Sephton, Peter S. (1998), "EasyReg: Version 1.12", Journal of Applied Econometrics 13, 203207.
 Choi, Hwansik and Nicholas M. Kiefer (2005), "Software evaluation: EasyReg International", International
Journal of Forecasting 21, 609616.
EasyReg is designed for my own use in research and teaching, as well as a service to the econom(etr)ics
community. If you have your own web page, and you like EasyReg, please put a link on your page to EasyReg, so that
Google
can list it.
Moreover, if you use EasyReg for research, please refer to it in the reference list of your paper as:
 Bierens, H. J. (2015), "EasyReg International", Pennsylvania State University
(http://www.personal.psu.edu/hxb11/EASYREG.HTM)
Note that this is a reference to the EasyReg program itself, and not to a paper or a manual. There is no
EasyReg manual, and there should be no need for a manual, because EasyReg comes with a variety of
guided tours which explain how to use EasyReg, and a help menu with
answers to common questions.
The present version of EasyReg may not be free of bugs. By using EasyReg you automatically agree to not
holding me or the Pennsylvania State University liable for any damage caused by errors
or viruses. As to the latter, all the EasyReg files have been scanned for possible viruses.
If you encounter problems, please let me know.
Why is EasyReg free?
EasyReg is completely free. It may be downloaded freely, without my written consent,
on home computers, laptops and standalone computers and networks of noncommercial institutions and organizations as well as commercial institutions.
EasyReg is now open source. The Visual Basic 5.0 source code can be downloaded from the EasyReg download web page
(see below).
EasyReg was originally designed to promote my own research. I came to realize that getting my research published
in econometric journals is not enough to get it used. But writing a program that only does the Bierens' stuff would
not reach the new generation of economists and econometricians. Therefore, the program should contain more than only
my econometric techniques.
When I taught econometrics at Southern Methodist University in Dallas in the period 19911996, I needed software
that my graduate students could use for their exercises. The existing commercial software was not advanced enough,
or too expensive, or both. Therefore, I added the econometric techniques that I taught in class first to
SimplReg, and later on to EasyReg after I had bought Visual Basic 3.
Meanwhile, working on EasyReg became a hobby: my favorite pastime during rainy weekends.
When I moved to Penn State University, and made EasyReg downloadable from the web, people from all over the world,
from developing countries in Asia and Africa as well as from western Europe and the USA, wrote me emails with
econometric questions, suggestions for additions, or just saying "thank you". For example, an EasyReg user from
Yemen emailed me:
Click here for more "thank you" notes.
It appears that many students and researchers have no access, or cannot afford access, to commercial econometrics
software. By making EasyReg commercial I would let these people down.
There are also less altruistic reasons for keeping EasyReg free:
 By keeping EasyReg free my own econometric work incorporated in EasyReg will get the widest distribution.
 I will never be able to make enough money with a commercial version of EasyReg to be compensated for the time
I have invested in it.
 Going commercial would leave me no time for my own research.
What does EasyReg do?
EasyReg does most of the econometrics tasks available in competing commercial software, in particular:
and even more. Some of the listed software have additional features, though. EasyReg is mainly a pointandclick program,
whereas some of the listed commercial software packages are programmable.
The EasyReg menus
Menu
EasyReg contains modules (accessible via "Menu" and its submenus) which conduct:
Input
 Transformations of variables: linear and multiplicative transformations, log, exp, differences, percentage
changes, lags, moving averages, dummy variables for subsets of observations, seasonal dummy variables, various
indicator functions, and customized transformations.
 Renaming variables
 Deleting variables
 Converting monthly time series to quarterly time series
 Preparing time series data for outofsample forecasting
 Reordering crosssection observations
 Sorting variables by name

Output
 Viewing and/or printing the output text file via Windows Notepad.
 Viewing, editing, and printing bitmap output files, via Windows Paint.

Data analysis:

Tabulating data.

Calculating summary statistics of the data: sample mean and standard error, minimum, maximum, effective
sample size, and the quantiles in steps of 10%.

Plotting time series.

Drawing scatter diagrams.

Kernel estimation of univariate and bivariate density functions (two versions: standard kernel density
estimation, and Bierens' SMINK estimation).

Auto and crosscorrelation functions for time series. In the autocorrelation case also the BoxPierce Q statistics,
the LjungBox Q statistics, and the partial autocorrelations can be computed.

Periodogram of a time series

Correlation matrix and its eigenvalues.

Unit root and stationarity tests:
 Augmented DickeyFuller tests,
 Phillips and PhillipsPerron tests,
 Bierens' unit root tests on the basis of higher order sample autocorrelations,
 Bierens' unit root tests against nonlinear trend stationarity,
 Breitung's nonparametric unit root tests,
 BierensGuo tests of the (linear trend) stationarity hypothesis against the unit root (with drift) hypothesis.
 KPSS tests of the (linear trend) stationarity hypothesis against the unit root (with drift) hypothesis.
 Bierens' test for complex conjugate unit roots.

Single equation models:

Linear regression analysis, which in the time series case can be extended to models with ARMA errors and/or
GARCH errors.

Discrete dependent variables modeling:
 Logit and Probit, Poison regression, Binomial Logit, and Negative binomial Logit,
including the BierensWang Simulated Integrated Conditional Moment (SICM) test.
 Bierens, H. J., and L. Wang (2012): "Integrated Conditional Moment Tests for Parametric
Conditional Distributions", Econometric Theory 28, 328362.
[PDF1 (paper), PDF2 (corrigendum)]
 Bierens, H. J. (2015): "Addendum to Integrated Conditional Moment Tests for Parametric
Conditional Distributions".
[PDF]
 Multinomial Logit, Ordered Logit
 Tobit analysis.
 Heckman's sample selection model (for crosssection data only).
 Rightcensored and intervalcensored proportional hazard models (for crosssection data only).
 Quantile regression.
 Twostage least squares/instrumental variables estimation.
 Kernel estimation of the error density of linear and nonlinear regression models, quantile regression
models, and twostage least squares models.
 Nonparametric kernel regression with one or two explanatory variables.
 Bierens' nonparametric SMINK regression with one or two explanatory variables.
 Tests for ARCH (linear regression models only).
 Outofsample forecasting with linear and nonlinear regression models.
 ARIMA model specification, estimation, and forecasting.
 Wald and Ftests of linear parameter restrictions.
 KieferVogelsangBunzel (KVB) test of linear parameter restrictions in linear time series regression
models with serially correlated errors.
 Integrated Conditional Moment (ICM) test of model correctness of regression models and crosssection
quantile models:

Bierens, H. J. (1982): "Consistent Model Specification Tests", Journal of Econometrics 20, 105134.
[PDF]
 Bierens, H. J. (2015): "Addendum to Consistent Model Specification Tests".
[PDF]

Bierens, H. J. (1984): "Model Specification Testing of Time Series Regressions",
Journal of Econometrics 26, 323353.
[PDF]
 Bierens, H.J. (2015): "Addendum to Model Specification Testing of Time Series Regressions".
[PDF]

Bierens, H. J., and W. Ploberger (1997): "Asymptotic Theory of
Integrated Conditional Moment Tests", Econometrica 65, 11291151.
[PDF]
 Bierens, H. J. (2014): "Addendum to Asymptotic Theory of
Integrated Conditional Moment Tests". [PDF]

Bierens, H. J., and D. Ginther (2001): "Integrated Conditional
Moment Testing of Quantile Regression Models", Empirical Economics
26, 307324.
[PDF]
 (G)ARCH estimation and testing of financial time series, in particular stock returns (No seasonality. Daily and weekly time series should be declared as annual time series).
This module also contains the option to test the validity of the model via the WSICM test in:
 Bierens, H. J., and L. Wang (2014): "Weighted Simulated Integrated Conditional Moment Tests
for Parametric Conditional Distributions of Stationary Time Series Processes", forthcoming in Econometric Reviews.
[PDF]

Multiple equations models:
 Linear General Method of Moments (GMM) estimation of a system of linear equations with possibly common coefficients,
including seemly unrelated regression (SUR) estimation, and estimation of fixed effect or pooled panel data models.
 VAR innovation response analysis: Sims' nonstructural VAR, and Bernanke's structural VAR analysis, including
asymptotic innovation response confidence bands.
 VAR forecasting.
 Johansen's cointegration analysis, including innovation response analysis with cointegrated vector error correction
models (without confidence bands, though), and the BierensMartins test for timevarying cointegration:
 Bierens, H. J., and L. F. Martins (2010), "Time Varying Cointegration", Econometric Theory
26, 1453–1490. [PDF]
 Breitung's nonparametric cointegrationtest.
 Bierens' nonparametric cointegration test.
 Bierens' nonparametric nonlinear cotrending test.

Userdefined nonlinear models:
All the econometric tasks in the previous menu items can be conducted simply by pointandclick. The following
modules, however, allow you to program the functional form of the models involved. The programming itself is still
done by pointandclick, though.
 Nonlinear regression analysis, which in the time series case can be extended to nonlinear models with ARMA errors.
 Maximum likelihood.
 Nonlinear General Method of Moments estimation.

Miscellaneous modules:
Miscellaneous modules are modules based on my work in progress, in particular seminonparametric survival analysis.
File
The "File" menu has the following submenus:
Current data
 Viewing the number setting of Windows.
 Viewing the current startup folder.
 Viewing the current input file.
 Editing the input file comments.
 View or edit the EasyReg database:
EasyReg comes with its own database containing various data to play with. You may add variables from
the current data file to this database, delete variables from the database, or restore the database from backup.
The database files involved are EASYREG.DB1 and EASYREG.DB2, respectively, in the EasyReg folder, with
corresponding backup files ERDBBAK.DB1 and ERDBBAK.DB2. Only the latter two files are included on the setup disk.
If you run EasyReg for the first time, the backup files ERDBBAK.DB1 and ERDBBAK.DB2 are automatically copied
to EASYREG.DB1 and EASYREG.DB2, respectively. If you upgrade an older version of EasyReg, the latter is not done,
in order to prevent the existing database files from being overwritten. (The files EASYREG.DB1 and EASYREG.DB2 are
random access files. If you try to import these files in a text editor or wordprocessor, they will be damaged.)
If the current folder does not contain a subfolder EASYREG.DAT, you can only view, restore, or backup
the database, but not retrieve variables from the database. For retrieving variables from the database,
see below.
Get data
EasyReg creates a subfolder EASYREG.DAT, containing the random access data files INPUT1.RAN and INPUT2.RAN,
the output text file OUTPUT.TXT, picture files (*.BMP), and various auxiliary files (*.TMP), in the folder where
EasyReg has been started up. The auxiliary files (with extension TMP) are created during the EasyReg session and
deleted automatically when you quit. If you start up EasyReg in a folder already containing the subfolder EASYREG.DAT,
you can continue the previous session. All output is stored and can be printed via Windows.
When you activate EasyReg, it will automatically move to the last folder in which you have activated EasyReg,
so that you may continue the previous EasyReg session.
This submenu has the following items:
 Searching the hard disk for Excel files in CSV format.
 Searching the hard disk for input files in EasyReg space delimited text format.
 Retrieving data from the EasyReg database.
 Creating artificial data.
 Continue with a previous EasyReg session.
How to get data
This item opens a guided tour for first time EasyReg users, explaining stepbystep
how to import data in EasyReg.
This menu answers most of the questions which may arise when you startup EasyReg
for the first time. This help menu item is also available online:
Online
help menu.
Back to top.
Level
The Level menu enables you to set the econometrics level. EasyReg is designed to
be used in undergraduate, intermediate, and advanced econometrics courses, as well as for empirical economic research.
Therefore, the user can startup the program at various levels, so that the topics and features that have not yet been
taught in class so far will be suppressed. If you use EasyReg for empirical economic research, choose the "Research level"
option, otherwise choose the appropriate "Student level" option. However, the "Student level" only applies during the
EasyReg session. The "Research level" is the default.

Novice student level
 Bivariate least squares regression only.
 + tvalues, Rsquare, TSS, RSS, standard error.
 + Outofsample forecasting
 + Multivariate least squares regression. Variance matrix.
 + Lagged dependent and/or independent variables
 + Dummy variables (as regressors)
 + BreushPagan heteroskedasticity test.
 + Heteroskedasticity consistent tvalues.
 + JarqueBera/SalmonKiefer normality test.
 + DurbinWatson autocorrelation test.

Intermediate student level
 + F and Wald tests of joint significance and linear parameter restrictions.
 + Probit/Logit analysis.
 + Poisson regression.
 + Binomial Logit regression.
 + Negative Binomial Logit regression.
 + Twostage least squares.
 + Linear time series regression models with ARMA errors.
 + Tests for ARCH.
 + Augmented DickeyFuller unit root tests.
You cannot choose sublevels.If you choose this option, then you choose all of the above.

Advanced student level:
 + Everything else, except the "Bierens stuff" listed below.

Research level
The research level includes the advanced level, plus the approached in the following papers (See my
downloadable papers web page):

Bierens, H.J. (1982): "Consistent Model Specification Tests", Journal of Econometrics 20, 105134.
 Bierens, H.J. (2015): "Addendum to Consistent Model Specification Tests".
[PDF]

Bierens, H.J. (1983): "Sample Moments Integrating Normal Kernel Estimators of Multivariate Density and
Regression Functions", Sankhya, Series B, 45, 160192

Bierens, H.J. (1984): "Model Specification Testing of Time Series Regressions",
Journal of Econometrics 26, 323353.
 Bierens, H.J. (2015): "Addendum to Model Specification Testing of Time Series Regressions".
[PDF]

Bierens, H.J. and S.Guo (1993): "Testing Stationarity and Trend Stationarity Against the Unit Root Hypothesis",
Econometric Reviews 12,132.

Bierens, H.J. (1993): "Higher Order Autocorrelations and the Unit Root Hypothesis", Journal of Econometrics
57, 137160.

Bierens, H.J. (1997): "Testing the Unit Root Hypothesis Against Nonlinear Trend Stationarity, With an
Application to the Price Level and Interest Rate in the U.S.", Journal of Econometrics 81, 2964.

Bierens, H.J. (1997): "Nonparametric Cointegration Analysis", Journal of Econometrics 77, 379404.

Bierens, H.J. and W.Ploberger (1997): "Asymptotic Theory of Integrated Conditional Moment Tests, Econometrica
65, 11291151.

Bierens, H.J. (2000): "Nonparametric Nonlinear CoTrending Analysis, with an Application to Interest and Inflation
in the United States", Journal of Business & Economic Statistics 18, 323337.

Bierens, H.J. (2001): "Complex Unit Roots and Business Cycles: Are They Real?", Econometric Theory17,
962983.

Bierens, H.J., and D.K. Ginther (2001): "Integrated Conditional Moment Testing of Quantile Regression Models",
Empirical Economics 26, 307324.

Bierens, H.J., and L. Wang (2012): "Integrated Conditional Moment Tests for Parametric Conditional Distributions",
Econometric Theory 28, 328362.
 Bierens, H.J., and L.F. Martins (2010): "Time Varying Cointegration", Econometric Theory
26, 1453–1490.
 Bierens, H.J., and L. Wang (2014): "Weighted Simulated Integrated Conditional Moment Tests
for Parametric Conditional Distributions of Stationary Time Series Processes", forthcoming in Econometric Reviews.
[PDF]
The research level option is the default option when EasyReg is started up. The chosen student level is
applicable only during the current session, and can be changed during the session.
Tools
EasyReg contains various Tools modules:

Teaching tools
Demonstration of
 the difference between the concepts of convergence in probability and convergence in distribution,
 the size and power of the ttest,
 plot of the density of bivariate normal distribution for different marginal variances,
 various Wiener processes, and
 spurious time series regression results.

Distribution tools
 Computation of binomial numbers and the probabilities of the binomial, hypergeometric, negative binomial,
and the Poisson distribution.
 Computation of either the cumulative distribution function F(x) and corresponding pvalue in a given x,
or the critical value at a given significance level, for the following standard distributions:
 Standard normal
 standard Cauchy
 Student t
 Chisquare
 F
and the null distributions of the (Augmented) DickeyFuller and PhillipsPerron unit root tests. All
test statistics which have one of these (asymptotic) null distributions are automatically endowed with
(asymptotic) pvalues.
Matrix tools
For a given matrix A, his module computes A', AA', A'A, A^2, trace(A),
A^1, det(A) (the latter four if A is square), the eigenvalues and eigenvectors
of A if A is symmetric, diag(A) if A is a row or column vector, and for
conformable matrices A and B: A+B, AB, and AB.
Plot a function
This module allows you to plot a function of one or two variables from data on file.
Microsoft Excel tools
 Opening Excel to test whether Excel is installed.
 Taking a subsample from an Excel CSV file if this file is too large to fit in EasyReg.
Make an EasyReg standalone CD or USB drive
 How to run EasyReg from CD or USB drive without installing it on the host computer.
 Prepare this CD or USB drive.
WWW
The first item of the WWW menu states the conditions under
which EasyReg users may email me with questions and comments.
The next seven items give access (via the default web browser) to
The last two items are the upgrade modules:
 Upgrading EasyReg to the current version.
 Upgrading and downloading the EasyReg guided tours.
Tours
The Tours
menu gives access (via the default web browser) to the available guided tours. These tours are in the form of
HTM files with associated JPG picture files, located in the EasyReg folder, provided that you have downloaded them.
See the EasyReg download web page for the available guided tours. They are also available
online.
Online guided tours.
Some of these guided tours have been translated in French by Dr. Nasser Ary Tanimoune, University of Ottawa, Canada.
Back to top.
Data size limitations
All arrays in EasyReg are declared with variable dimensions. However, the number of observations is stored
in an integer variable, which in Visual Basic 5 is limited to a maximum value of 32767. Consequently, if you
try to import a data set with more than 32767 observations you will get an error message, and EasyReg will
crash. On the other hand, such large data sets are rare in empirical econometrics. Therefore, in practice the
amount of data that EasyReg can handle depends on the memory capacity of your computer.
Structure and appearance of EasyReg
In order to allocate the maximum amount of memory to the data, EasyReg is chopped up in separate EXE files
conducting specific tasks. The main program, EASYREG.EXE, is only a shell around these other EXE files, containing
the menus. The specific EXE files will not run standalone. The EasyReg EXE files are Multiple Document Interface
(MDI) programs with nonmovable child forms. The control buttons of the child windows have been removed in order
to prevent the user from accidentally closing or reactivating a child window. The reasons for the latter are twofold:

The structure of EasyReg is linear: the various tasks have to be conducted in a particular order.
 MDI programs which keep all the child windows visible are often confusing: You need a manual to determine which
window to activate.
The default size of EasyReg is 640x480 pixels. You can run EasyReg
in full screen mode by clicking the "maximize" control in the upperright corner of the EasyReg main window. EasyReg
will then be maximized, and all the windows, controls, fonts and menus will be proportionally enlarged. This is quite
different from other Windows applications, which require reading glasses if you run them in a higher screen resolution
than 800 x 600 pixels. EasyReg will remember the last setting. If you close EasyReg while in full screen mode, it will start
in full screen mode the next time.
Under full screen mode some of the control buttons of EasyReg may be covered by the taskbar. If so, open Control
Panel > Taskbar and Start Menu > Taskbar, and check Autohide.
Running EasyReg from a network server
Single user mode
EasyReg runs default in single user mode: In this mode it is not possible to activate more than
one instance of EasyReg. The reason is the following:
When you startup EasyReg (in the default single user's mode), it creates a subfolder TEMP
and a further subfolder TEMP\REGISTR in the EasyReg system folder. The folder TEMP is used for storage
of the setting of EasyReg, such as the last position and size of the EasyReg main window, and the path of the
last folder where you have started up EasyReg. The registration information (in particular your name) is stored
in subfolder REGISTR. Therefore, if it would be possible to run multiple instances of EasyReg, the files in
subfolder TEMP, in particular the file STARTDIR which contains the path of the folder were the current
EASYREG.DAT subfolder is located, will be messed up, so that neither instance of EasyReg will run correctly.
Moreover, if EasyReg is installed on a server, the EasyReg system folder is usually write protected, so that
EasyReg cannot create and/or write to the subfolders TEMP and TEMP\REGISTR in the system folder.
The network solution
If EasyReg is installed on a network server, it should be startedup by executing EASYREGN.EXE instead of EASYREG.EXE.
Program EASYREGN.EXE opens with the following window:
In the Department of Economics at Penn State, EasyReg has been installed on server "godzilla" in folder EasyReg.
Each graduate economics student and economics faculty has his or her own folder on network drive H (which is actually
a virtual drive of another server). My folder is h:[HBierens] and I have stored
the settings of EasyReg in folder "h:[HBierens]\EasyReg Setting". This folder is used by EasyReg in the same way as the folder
TEMP in the single user mode. Therefore, in order to run the network version of EasyReg, I have to click drive H, double
click folder "EasyReg Setting", and then click the "Start EasyReg" button. If the folder involved does not yet exist,
you may type the path of the user's folder in the text box. EasyReg will then create this folder. It is not allowed to
store the EasyReg setting in the root folder (h:[HBierens] in my case);
you have to use an existing subfolder or create one.
When you click "Start EasyReg", the path to this folder is passed on to EASYREG.EXE, and then EasyReg will run
in almost the same way as in single user mode. The only exceptions are:
 The users can no longer change the database; they can only read it, and retrieve data from it.
 The menu item "Upgrade EasyReg to the latest version" in the WWW menu is not available.
 The menu item "Upgrade EasyReg guided tours" in the WWW menu is not available.
Network administrators, please take notice of the following:
Before EasyReg can run from the network server, you have to initialize it.
Download

Download EasyReg International and its Visual Basic 5.0 source code from the Penn State website.

Download EasyReg International from the EasyReg mirror
site in Brazil, maintained by Professor Claudio Shikida, Federal University of Pelotas, Brazil.
Remarks
 The EasyReg mirror site in Brazil is less frequently updated than the Penn State site.
Therefore, after downloading EasyReg from this mirror site, open the menu items
 Upgrade EasyReg to the latest version
 Upgrade EasyReg guided tours
in the WWW menu to check for upgrades. However, before you use these menu items, consult the section
"Upgrading an older version of EasyReg International" of the EasyReg download web page.
 If you are an econometrics instructor using EasyReg for teaching, and downloading EasyReg in your
country from the Penn State web site is too slow, you may consider creating an EasyReg mirror site at
your institution, similar to the mirror sites in Brazil and Colombia. If so, I will send you a
CD ROM with the EasyReg setup files.
Links

Number and origins of registered EasyReg users. Effective March 28, 2011, the online
registration of EasyReg has been discontinued!

After you have downloaded EasyReg it is recommended that you check out regularly which bugs
have shown up so far and which new features have been added.

Back to my home (page)