Betahat is a Windows based econometrics and economic modeling package that uses state of the art algorithms used by analysts to estimate, simulate, and forecast.  Betahat is used by government agencies, businesses, and universities around the world.  Some of the unique features that differentiate Betahat from other econometric software programs are its advanced simulation and model handling features, a dual mode interface (both menu and command mode), and robust methods for reading data and models.  Betahat is used for single equation estimation and forecasting, and large multiequation macroeconometric models.

Betahat runs under Windows and uses many of the familiar user interface devices such as spreadsheets, dialog boxes, and pull down menus. The length of time the average user spends between the installation of the program and the time when the first model is estimated and simulated is very small. One of the key objectives of the program is for the focus of the users attention to be on the research project at hand, not on the computational tool being used.


What's New

July 23, 2012

Version 8.0 of Betahat is released.  Changes include

  1. Increased speed in simulations.  Benchmarks for stochastic simulations are provided on the Benchmarks page. 
  2. There are now 2 versions of Betahat, a single processor version and a multiprocessor version.  The single processor version is now $249, and the multiprocessor version is $995.


June 20, 2011

Version 7.2 of Betahat is released.  Changes include

  1. Increased speed in simulations.
  2. Added HTML output for simulations.  To generate HTML output use BROWSE option on the SIM command.  To view use the Browse tab.
  3. Different algorithm for the CDF function.  The newer algorithm is a little slower but more accurate.
  4. Automatic creation of OLS results to start estimation of PROBIT models.
  5. Simulations now display a matrix that shows how many equations converged to a certain number of digits.  It also displays the names of the equations that have converged the least.
  6. Added command ANALYZE, which is very similar to the SIM command except that it will simulate each of the equations in the system individually.
  7. Added documentation for VARCOV command, used to create a variance/covariance matrix to be used in stochastic simulations.
  8. A different way the random numbers are converted from a uniform distribution to a normal distribution.  The results from stochastic sims will be different than with previous versions of Betahat.


February 11, 2011

Version 7.1 of Betahat is released.  Changes include

  1. Substantial multi-processor support.  At this time multiple processors can be used estimating least squares and AR1 models, and simulations using the newton and stack-newton algorithms.  Select Options - Program Options - Enable Multiprocessor Use.
  2. QR factorization for least squares in addition to SVD and Cholesky.
  3.  Pardiso solver for Newton & Stacked Time algorithms.  The Pardiso solver is very fast and can use multiple processors.
  4. The IMF Mark III model in Betahat format is available on the models page.


May 15, 2010

Version 7.0 of Betahat is released.  Changes include

  1. Added ability to save the residuals from all equations in a simulation with a single option. These residuals can then be used to compute the shocks in a stochastic simulation.
  2. Added ability for Monte Carlo simulations to save solution path information.
  3. Added more convergence information in the simulation results.
  4. Allow tab delimited files for PRN format.
  5. In addition to placing the results in the Results tab, after running STATS, some of the statistics are available via command mode to be used in equations or transformations. The ones that are available are SUM, MIN, MAX, MEAN, MEANABSVAL (mean of the absolute value), VARIANCE, STDDEV, STDDEVM and MEDIAN.
  6. New simulation switch CrossSection.  With this option Betahat does not try to replace a variable with an endogenous variable if cross section data and the variable has a lag (with cross sectional data the lagged data is not the same concept).
  7. Added conditional case simulations.
  8. On the spreadsheet tab there is a new command "Go To Row", and you can now search data by row or column.
  9. On the Graphics tab there is a new graphics engine.
  10. Added 2 methods for performing stochastic simulations where there are fewer observations in history then there are equations.
  11. The degrees of freedom for unit root testing has changed from N-K to N.

We also provide a history of the changes to Betahat if you would like to see what features have been added to Betahat since you purchased your copy.