Modeling Life Satisfaction Based on Subgroups in Panel Data


Kızılarslan Ş., Camkıran C.

V. Uluslararası Uygulamalı İstatistik Kongresi, İstanbul, Türkiye, 21 - 23 Mayıs 2024, ss.40, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.40
  • Van Yüzüncü Yıl Üniversitesi Adresli: Hayır

Özet

Life satisfaction is directly influenced by the income levels of individuals or, at the macro level, by the economic growth of nations. It is well-established that income level has an augmentative effect on life satisfaction. However, empirical studies indicate that this augmentative effect is not continuous. The Easterlin Paradox posits that the relationship between life satisfaction and income follows a logarithmic function, suggesting that the augmentative effect of income on life satisfaction is stabled at higher income levels. Particularly in macro-level analyses utilizing panel data, the heterogeneous characteristics of countries complicate the deduction of common interpretations. In this study, a clustering-based modeling approach was utilized to overcome the challenges posed by heterogeneity in panel data. The relationship between life satisfaction and economic growth was examined using panel data from 74 countries for the years 2010 to 2022. Additionally, socioeconomic factors such as inflation, unemployment, and urbanization rate were included as control variables in the model. The countries in dataset were stratified into homogeneous subgroups within which separate models were estimated, thereby revealing heterogeneous parameters between clusters. Upon analysis, it was determined that the countries were divided into two clusters, and the effects of variables significantly differed between them. Accordingly, predicted values of life satisfaction have increasing trend in first cluster while decreasing trend in the other. To explore the source of these differences, the clustering process was reconducted focusing on economic growth and inflation, and consistent results were obtained. This circumstance revealed that the divergent cluster outcomes for life satisfaction essentially stem from the differences in economic growth rates and inflation levels of the countries.