A COMPARATIVE-STUDY OF ALTERNATIVE METHODS OF QUANTIFYING QUALITATIVE SURVEY RESPONSES USING NAPM DATA

This article examines the effectiveness of a subjective probability approach to quantify three-category qualitative responses from expectations surveys. Since one cannot make a judgment on the validity of estimated expectations by comparing them with ex post outcomes, we use the National Association of Purchasing Managers survey data on materials buying prices and compare them with the Producer Price Index for intermediate materials and components for manufacturing. We experimented with normal, logistic, and scaled-t distributions and found that the approach could explain approximately 60% of the variation in the monthly series. Marginal improvement in the explanatory power of the procedure was noted when the imperceptible parameter was allowed to be asymmetric and variable over time.