New Composite Indicator for the Business Tendency Survey

• Activity Number: 521
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Michel Grun-Rehomme*+ and Olga Vasyechko
• Companies: and University Paris 1
• Keywords: Business Tendency Survey ; Time Series ; Balance of Opinion ; Robust Statistics ; Composite Indicators

Abstract

Short-term surveys provide precious information for economic fluctuations analysis. In short-term surveys like the Business Tendency Survey, most of the questions are qualitative and concern the evolution of different economic factors of the business activity. They provide economic information on the present situation and short-term perspectives. Usually, the respondents have to choose between three possible evolutions: increase (improvement, favourable, level higher than the normal), stability (normal) or decrease (unfavourable, level lower than the normal). The balance of opinion is defined as the difference between the proportion of respondents expressing a positive opinion and the proportion expressing a negative opinion. To analyze these types of surveys, the methods are well standardized and use both the multidimensional approach and time series (scoring, dynamic factor analysis, etc.)

In this paper, we propose a new method of calculating a robust composite indicator based on range median statistics, and on a lexicographical order relation of the individual data. A confidence interval is constructed around these statistics. The indicator's advantage is a simplicity of ca