On the slope information of fertility measures
P. C. Roger Cheng, National Central University, Taiwan
When cohort fertility continues to decrease (increase), most of its predictors tend to yield positively (negatively) biased estimates, indicating the importance of predicting turning points so that the bias can be corrected in an accurate direction. However, a direct appeal to the cohort fertility curve is shown possible but of little worth. This paper, therefore, turns to a simple but definite relationship between cohort fertility and Bongaarts and Feeney (1998)'s period adjusted measure, and then proposes a procedure of utilizing slope information in producing helpful signals suggesting whether and how to correct the bias. Empirical evidence from Canada, the U.S., and 23 European countries suggests the validity of this procedure.