Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
We evaluate the usefulness of bias-correction methods for autoregressive (AR) models in enhancing the out-of-sample forecast accuracy. We employ two popular methods, proposed by Hansen (1999) and So and Shin (1999). Our Monte Carlo simulations show that these methods do not necessarily achieve better forecasting performances than the bias-uncorrected least squares (LS) method, because bias correction increases the variance of the estimator. Both the bias and the relative variance tend to decrease as the sample size (T) increases, meaning that larger numbers of observations do not always imply gains from bias-correction. As the degree of persistence increases, the bias becomes greater while the relative variance becomes smaller, which implies a greater gain from correcting for bias for highly persistent data. We also provide real data applications that confirm our major findings overall.