SDG Indicators Need a Wellbeing Reality Check

• In Event: Advanced Methods for Causal Inference and Power Analysis

Abstract

The 2030 Agenda rests on an implicit assumption: that progress on the Sustainable Development Goals improves human wellbeing. This premise underpins 169 targets, 232 indicators, and the policy interventions designed to advance them. As UN Member States negotiate post-2030 frameworks, the assumption has not been systematically tested against dynamic data. Do countries achieving SDG progress actually see improvements in how their citizens evaluate their lives?We compile panel data for 42 countries from 2014 to 2024, matching annual Cantril Ladder scores from the World Happiness Report with more than 150 SDG indicators per country drawn from the UN Statistics Division SDG Global Database. For each country-indicator pair, we estimate bivariate OLS regressions in two specifications: a levels model testing cross-sectional associations and a first-differences model testing dynamic relationships between year-to-year changes. Benjamini-Hochberg False Discovery Rate correction at q<0.05 addresses multiple hypothesis testing across the full indicator space.Results diverge sharply between specifications. In levels, the analysis reproduces established cross-sectional findings: many SDG indicators correlate strongly with national happiness, with substantial numbers showing R² above 0.60 in individual countries, consistent with prior work by De Neve and Sachs (2020) and Kubiszewski et al. (2022). In first-differences, this explanatory power collapses. The vast majority of indicators fall into the lowest R² bins (0 to 0.1). Across 40 of 42 countries, not a single indicator-happiness association survives FDR correction. Year-to-year SDG progress, across every domain tested, explains almost none of the variation in contemporaneous happiness changes.This extends the Easterlin Paradox from income to comprehensive development measurement. The established pattern, in which cross-sectional correlations between development and wellbeing fail to hold when examining within-country changes over time, generalizes across health, education, infrastructure, governance, and environmental quality. Low-income countries do not show stronger dynamic relationships, ruling out ceiling-effect explanations. The consistency across countries and indicator types suggests a systematic measurement challenge rather than country-specific idiosyncrasies.The paper contributes to post-2030 indicator design debates currently underway ahead of the 2027 SDG Summit. It provides the first systematic test of the SDG-wellbeing assumption across comprehensive indicator sets. It distinguishes cross-sectional structural differences from within-country responsiveness, a distinction routinely blurred in SDG monitoring practice. The findings do not imply that current indicators should be discarded. They suggest that post-2030 frameworks may require distributional disaggregation across income strata and demographic groups, process dimensions alongside outcome targets, and differentiated monitoring tools for short-term change versus long-term development assessment. Indicator systems shape what governments optimize for; extending current architecture through 2040 or 2050 without confronting this evidence risks perpetuating a measurement system whose connection to lived wellbeing improvement remains undemonstrated.

Author

• Mel George, University of Maryland
  Presenting Author