Data Management
Data can be entered in a variety of ways: directly

from the keyboard, from raw ASCII files, or from files previously prepared and saved from Microfit by the user. Files may then be manipulated as required, saved and retrieved quickly in the binary format unique to this program, or output to ASCII files as raw data.
Data Transformations
Microfit facilitates data transformations with

its functions allowing the computation of new series as algebraic transforms of existing data using standard arithmetic operators. Moreover, it permits the generation of time trends, seasonal dummies, and random numbers from UNIFORM and NORMAL distributions.
General Statistics
Microfit computes the following:

standard deviations
measures of skewness and Kurtosis
the coefficient of variation
the minimum and maximum of a series
autocorrelation coefficients with standard errors
Box-Pierce and Ljung-Box statistics
the matrix of correlation coefficients
Estimation Techniques
Ordinary Least Squares (OLS)
Recursive Least Squares
Generalized Instrumental Variables (IV)
Two-Stage Least Squares (2SLS)
Cochrane-Orcutt iterative technique (CO)
Exact Minimum Likelihood (ML) and
Variable (IV) estimation of models with serially
correlated errors
Non-linear Least Squares (NLS)
Non-linear Two-stage Least Squares and the
generalized methods of moments
Johansen method of maximum likelihood
estimation of co-integrated systems
Rolling (moving window) regression (OLS)
Rolling (moving window) regression (2SLS)
Moving-average maximum likelihood estimation
Moving-average estimation using instrumental
Microfit Navigation Bar


Static and Dynamic Forecasts
Microfit produces static or dynamic forecasts of

single and multivariate regression models, and forecast errors and useful summary statistics. Should the regression include lagged dependent variables, the program automatically computes dynamic forecasts: otherwise static forecasts are generated. The plot of actual and forecast values is also given.
Dynamic Simulation
By using the SIM command the user can

simulate dynamically any non-linear single-equation model. A backward dynamic simulator command is also incorporated as a useful additional feature.

©Oxford University Press Ltd & Camfit Data Limited, 1996.