VC - A Program for Estimating Time-Varying Coefficients

VC implements Schlicht's method for estimating a linear regression with time-varying coefficients. The variances are estimated by a moments estimator that performs better than the corresponding likelihood estimator in small samples and coincides with the likelihood estimator in large samples. Instead of the usual parametrization by initial values, an orthogonal parametrization is used, and instead of the one-sided Kalman filter, a statistically superior two-sided filter is implemented. The program runs under Windows.