Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
In business-cycle research, smoothing data is an essential first step to evaluate the extent to which model-generated moments stand up to their empirical counterparts. We put to test McDermott's (1997) modified version of Hodrick and Prescott's (1997) smoothing filter. On the one hand, our simulations suggest that relative to other filters, the modified HP-filter replicates better artificially generated series with known properties. On the other hand, using true data we find that autoregressive properties of smoothed series are not affected by the choice of smoothing HP filters, but the same does not hold when it comes to multivariate analysis. The later result is especially strong for annual data. We report results for a large set of countries.