Exploring the persistence and transience of energy poverty: evidence from a Greek household survey

Abstract The present study uses four rounds of household panel data to investigate consensual-based energy poverty in Greece. Employing dynamic Probit random effects and Wooldridge conditional maximum likelihood (WCML) estimators, we find evidence of genuine state dependence effects in consensual-based energy poverty among Greek households. Poverty persistence (10-12%) effects are also evident in our data. Socioeconomic, demographic, market, household, and climatic characteristics are essential predictors of energy poverty. Around 9–10% of the households seem chronically energy poor, while education, income level, dwelling characteristics, migration background, and employment status affect the chances of suffering and exiting from energy poverty. Empirical results have significant policy implications that could mitigate residential energy poverty.