The Cumulative Residual Entropy in Assessing the Duration of Unemployment

Aim: The aim of the article was to use cumulative residual entropy (CRE) to assess the informational value of data on deregistration from the Poviat Labour Office in Szczecin (Poland) from 2007 to 2024. The events of declining cooperation with the labour office and starting work were analysed. Methodology: The study used survival analysis methods with the assumption of an exponential distribution of duration. The CRE was calculated for the specified distribution parameters. Hierarchical clustering was used to identify clusters of years with similar CRE values. Using the dynamic time warping method, the CRE time series and the unemployment rate were compared. Results: The study revealed a similarity between the registered unemployment rate and the development of entropy of unemployment duration. High unemployment rates corresponded to high entropy values, and vice versa. During periods of shock (caused by crises) in the labour market, the CRE for both reasons for deregistration took on extreme values. Implications and recommendations: Analyses related to the duration in unemployment are more informative when the unemployment rate is lower. The labour market, as a system, is then characterised by less uncertainty. Originality/value: There is a lack of research on the application of entropy by using survival analysis in studies regarding the labour market. The study took into account the relationship between the hazard value and the CRE for the exponential duration distribution.