A Note on the Bias in the Herfindahl Based on Count Data