Low well-being among middle-aged people: inherent or external factors

Well-being is increasingly recognized as a fundamental goal at both individual and societal levels. Significant differences in well-being across age groups have long been detected and noted. However, the primary factors contributing to this disparity remain unknown. Here, leveraging an extensive global survey dataset and advanced machine-learning techniques, this study investigates the reasons underlying notably low levels of well-being among middle-aged individuals. Utilizing an exogenous switching treatment effect model (ESTEM) enhanced by machine learning, we analyze over 1.9 million individual observations from 168 countries collected between 2009 and 2022. Our results empirically confirm a U-shaped relationship between age and subjective well-being, indicating that middle-aged individuals consistently experience the lowest well-being. Further analysis reveals that middle-aged people receive significantly harsher external treatments compared to younger and older age groups, highlighting external societal conditions as critical contributors to the midlife crisis phenomenon. Conversely, elderly populations inherently experience higher subjective well-being. Temporal analyses indicate that external treatments for younger and middle-aged groups are becoming increasingly stringent relative to those for the elderly. By systematically mapping these treatment effects and intrinsic differences among age groups, this study provides critical insights to inform targeted policies and social programs designed to improve quality of life, thereby supporting equitable improvements in human well-being across the lifespan.