Explaining Global Happiness: Evidence from Decision Trees and Necessary Condition Analysis

In this paper, the factors linked to happiness are analyzed on the basis of the World Happiness Report (WHR) model, introducing a methodological approach that differs from traditional econometric techniques. More specifically, this study examines how the core variables of the WHR model interact in relation to happiness and whether some of them also emerge as necessary conditions within the framework of necessary condition analysis (NCA) for attaining higher happiness levels. Using Gallup World Poll data for the 2022–2024 period, the Cantril Ladder is employed as a measure of subjective well-being, and gross domestic product (GDP) per capita, social support, healthy life expectancy, freedom, generosity, and perceived corruption are considered explanatory variables. This study makes two contributions. First, it applies a decision tree regression model to identify interactions among the correlates of happiness while also facilitating the classification of countries into homogeneous groups according to their well-being configurations. This approach improves interpretability relative to linear models because it does not require prior specification of those interactions. Second, this paper incorporates necessary condition analysis to distinguish between factors that are merely influential and those that emerge as necessary conditions for attaining certain levels of happiness. These assessments make it possible to identify minimum thresholds in key variables, introducing a necessary-condition logic. The results show that social support and GDP per capita emerge as the main structuring variables in the tree and are strongly associated with differences in happiness, whereas freedom emerges as a prominent condition in the NCA results. The findings also show that some factors with low correlation may still play a relevant role in specific contexts because of nonlinear effects and interactions. Overall, the results of this study offer an analytical reinterpretation of the WHR model by combining structural segmentation and threshold identification, advancing the understanding of happiness as a multidimensional, nonlinear phenomenon associated with specific configurations of factors.