Detecting Structural Breaks in Latent Feeling and Uncertainty of Expectations of Price Levels Over Time

Qualitative data on individuals’ perceptions and evaluations are collected through official surveys using rating scales, often administered repeatedly over multiple time waves. In many cases, only aggregated response distributions are made available, with no access to individual-level data. In this context, we combine Atheoretical Regression Trees with CUB models to analyze rating data and identify structural changes in the main characteristics of the response distributions. The chosen modeling framework is well-suited to parameterize both the latent feeling and the uncertainty underlying the ordinal evaluation, with a time-varying structure. Then, the use of Atheoretical Regression Trees involves introducing an artificial covariate that preserves temporal ordering, enabling the segmentation of feeling and uncertainty measures into homogeneous time intervals, which can then be interpreted in light of political and socio-economic events. We illustrate the proposed approach using data from the Consumer Confidence Survey issued monthly by the Italian National Institute of Statistics (ISTAT), assessing whether and to what extent individuals changed their expectations regarding price levels from 1994 to 2019. The proposed method demonstrates satisfactory and interpretable performance similar to the Bai and Perron’s method which is considered the gold standard.