Panel Quantile Regressions and the Subjective Well-Being in Urban China: Evidence from RUMiC Data

Abstract

Using RUMiC data and a simple panel quantile regression method, this paper accounts for the time-invariant individual specific characteristics and investigates the heterogeneous effects of factors on the distribution of subjective well-being (SWB, measured by GHQ-12) in urban China. Comparing results from the pooled regression and fixed effect regression, we find that most results from pooled regressions are likely overestimated. Panel quantile regression results show that income affects the least happy 10 % group twice as much as the happiest 10 % people. Being unemployed brings down the happiness level by 0.97 points on average, but individuals with high SWB seem to be not significantly affected. The complete picture of the relationship between SWB and various factors would have been veiled without using quantile techniques. This paper therefore not only contributes to the happiness literature in China, but also adds evidence to stylized findings going beyond the average.

Does urbanization matter in the expenditure‐happiness nexus?

“…In the wellbeing literature, there are very few studies making use of a quantile regression approach to investigate the impact of socio‐economic variables on the distribution of subjective well‐being. Going beyond the average, the heterogeneous effects of determinants on the happiness distribution have been studied by Binder and Coad (2011), Binder and Freytag (2013), Yuan and Golpelwar (2013), Hand (2018) and D'Ambrosio et al (2020) using quantile regression, Fang and Sakellariou (2016) using unconditional quantile regression, and Binder and Coad (2015) and Fang (2017) using panel quantile regression techniques. Results from these studies highlight clear non‐linearities in the effects of socio‐economic determinants along the happiness distribution.…”