Introduction

The report examines whether consensus on measuring societal wellbeing beyond Gross Domestic Product (GDP) is achievable. It argues that GDP-focused growth is associated with inequality, environmental degradation and social harms. Although hundreds of alternative indicators exist, fragmentation has limited uptake. The study seeks to identify shared conceptual foundations to support agreement on post-GDP metrics centred on sustainable and inclusive wellbeing.

Methods

A database of 213 English-language wellbeing indicators was compiled from academic, governmental, business and NGO sources. Over 400 were initially identified; those with narrow or population-specific scope were excluded. Indicators average 33 components (range 3–295).

Semantic similarity was analysed using S-BERT embeddings and hill-climbing clustering to generate 10, 20, 40 and 80 clusters. Eight synthesising indicators were constructed: four derived (algorithm-selected components) and four created (manually defined summarising components).

Synthesising potential (SP) measures how well an indicator captures the semantic content of others. A piecewise linear model identified diminishing returns from adding components. Popularity was approximated using Google and Google Scholar hits, with weighted analyses testing robustness.

Results

Of the 213 indicators, 70.9% operate at national scale. Most are composite indices (68.5%), with 15.5% dashboards. Proposals originate from academia (85), government (74), NGOs (40) and business (14).

SP increases with the number of components, but gains slow markedly after about 20 components. A breakpoint is identified at 22 components. SP continues rising gradually up to 60–80 components, indicating diminishing returns beyond the initial threshold.

The 20-component synthesising indicator captures substantial thematic breadth while remaining manageable. Its components span human, social, built/financial and natural capital, including life satisfaction, health, education, employment, governance, inequality, infrastructure, per capita GDP, greenhouse gas emissions, air and water quality, and natural capital.

Created synthesising indicators show residual SP values of 0.105 (10 components), 0.120 (19), 0.124 (36) and 0.118 (73). Derived indicators average 0.130 and generally outperform existing indicators with comparable complexity.

Five high-performing existing indicators include Canada’s Quality of Life Framework (19 components; SP 0.442), Measuring What Matters Framework (50; 0.499), European Quality of Life Survey (12; 0.400), Iceland’s Indicators for Measuring Wellbeing (39; 0.474), and Community Indicators Victoria (74; 0.509). Performance does not depend solely on component count.

Popularity varies widely. The Sustainable Development Goals receive 275 million Google hits, yet popularity shows negligible correlation with SP (r² = 0.06). Weighting by popularity does not materially alter thematic clustering or rankings.

Discussion

Despite apparent diversity, substantial conceptual convergence exists across indicators. Increasing components improves representativeness but adds complexity, cost and interpretability challenges. The findings indicate a practical balance at around 20 components.

The semantic method identifies themes but not measurement approaches. Selecting metrics will require expert input, public dialogue and consideration of scale and data availability.

The report also considers dashboards versus single indices. Around 30% of indicators are dashboards and 70% composite indices. Both approaches have roles, but aggregation methods require further refinement, as simple averaging may overlook limiting factors within wellbeing systems.

Conclusions

Strong thematic overlap across 213 indicators suggests international consensus beyond GDP is feasible. A synthesised 20-component framework offers a balanced and operational basis for measuring sustainable and inclusive wellbeing, but effective implementation requires coordinated development of metrics, models and policy alignment beyond GDP growth.