💰 quantitative finance

GAGI: A Gini-Adjusted GDP-per-Capita Index for Distribution-Aware Macroeconomic Welfare Monitoring

This paper introduces the Gini-Adjusted GDP per Capita Index (GAGI), a transparent and reproducible welfare metric that adjusts GDP for inequality and inflation, demonstrating through G7 data that welfare-adjusted prosperity has increasingly diverged from headline GDP growth, particularly after 2022.

Original authors: Sivasathivel Kandasamy

Published 2026-06-10

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Sivasathivel Kandasamy

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Core Problem: The "Average" Lie

Imagine you and your friend go out for dinner. You spend $100, and your friend spends $1,000. If someone asks, "How much did you spend on average?" the answer is $550. That number sounds like a moderate expense. But it hides the truth: one person is struggling to pay their share, while the other is barely noticing the cost.

This is how GDP per capita (the standard measure of a country’s wealth) works. It takes the total money a country makes and divides it by the number of people. It gives you an "average." But just like the dinner bill, that average is blind to two huge problems:

Inequality: Who actually got the money? Did the rich get richer while the middle class stayed the same?
Inflation: Did the prices of groceries and rent go up so much that the money people did earn doesn’t buy as much as it used to?

The author argues that because GDP ignores these two things, it’s a bad tool for checking if people are actually living better lives, especially now that Artificial Intelligence (AI) is changing how we work.

The Solution: GAGI (The "Real Welfare" Scorecard)

The author created a new metric called GAGI (Gini-Adjusted GDP per Capita Index). Think of GAGI as a "Reality Check" filter for GDP.

Instead of just looking at the total pot of money, GAGI does two adjustments:

The Fairness Filter: It shrinks the score if the country is very unequal. If the Gini coefficient (a measure of inequality) is high, GAGI goes down.
The Purchasing Power Filter: It shrinks the score if inflation is high. If prices are rising faster than wages, GAGI goes down.

The Analogy: Imagine GDP is the size of a pizza. GAGI asks: "How big is the slice that the average person actually gets to eat, and how hungry are they after paying for the delivery fee (inflation)?" If the pizza is huge but one person eats 90% of it, and the delivery fee is expensive, the "GAGI" score for the rest of the table is terrible, even though the pizza itself was large.

What the Data Shows (2010–2026)

The author applied this "Reality Check" to the G7 countries (the US, UK, Canada, France, Germany, Japan, and Italy) from 2010 to 2026. Here is what they found:

1. The "Gap" is Widening
In every G7 country, the standard GDP score is higher than the GAGI score. This means that while the countries are technically making more money, the actual well-being of the average person isn’t keeping up. Since 2022, this gap has gotten much bigger.

2. The US: "Automation Without a Safety Net"
The United States shows the biggest gap.

What’s happening: AI is replacing jobs fast. The paper cites data showing that layoffs blamed on AI in the US jumped 13 times between 2023 and 2025.
The Result: Because the US has high inequality and weak support systems for displaced workers (like low spending on retraining or unemployment benefits), the "GAGI" score drops significantly. The wealth from AI is going to the top, not the middle.

3. Japan and Italy: "Stagnation Without Automation"
These countries have low AI investment, but their GAGI scores are also flat or falling.

The Lesson: Just not using AI doesn’t save you. If your economy isn’t growing and productivity is stuck, people’s lives still get harder. This is a different kind of failure than the US, but the result (low welfare) is similar.

4. The Nordic Model: "Automation With a Safety Net"
The author looked at Denmark and Sweden (not in the G7, but used for comparison).

What’s happening: These countries use more robots and automation than the US.
The Result: Their GAGI scores remain stable or improve. Why? Because they have strong "absorption" mechanisms: high taxes on the rich, strong unions, and generous government spending to retrain workers and support the unemployed. They automate and share the benefits.

The Warning for the Future

The paper looks ahead to 2035 with three scenarios:

Business as Usual (The US Path): If we keep automating without fixing inequality, the "GAGI" score will drop dangerously low by the early 2030s. The economy might look rich on paper, but society will be unstable.
Governance-Constrained (The Nordic Path): If governments actively manage inequality and support workers, the GAGI score stays healthy, even with high automation.

The Bottom Line

The author’s main point is that GDP is a broken compass for the AI age. It tells us the ship is moving forward, but it doesn’t tell us if the passengers are sinking.

GAGI is proposed as a new "dashboard light" for governments. If the GAGI light turns red, it means that even if the economy is growing, the average citizen is falling behind due to inequality and inflation. The paper argues that to survive the AI revolution, countries need to stop watching just the GDP number and start watching the GAGI number, ensuring that the benefits of automation are shared, not just concentrated at the top.

Technical Summary: GAGI – A Gini-Adjusted GDP-per-Capita Index for Distribution-Aware Macroeconomic Welfare Monitoring

Problem Statement
The paper argues that Gross Domestic Product (GDP) per capita, the standard metric for tracking economic prosperity, is insufficient for monitoring welfare in the context of AI-driven automation. GDP is blind to two critical determinants of lived prosperity: income distribution and purchasing power (inflation). Consequently, a country can report record GDP growth while median household living standards stagnate or decline. The authors identify a gap in the macroeconomic monitoring toolkit: the lack of an operational, reproducible, and transparent statistic that serves as an early-warning trigger for distributional harm. Existing inequality-adjusted measures (e.g., Atkinson index, inequality-adjusted HDI) are often retrospective research artifacts rather than real-time governance tools. The specific problem addressed is the need for a minimal, publicly computable index that aggregates growth, distribution, and price levels to track welfare-adjusted prosperity, particularly in the context of accelerating AI deployment.

Methodology
The authors introduce the Gini-Adjusted GDP per Capita Index (GAGI). The metric is defined for country $k$ and year $t$ as:

$GAGI_{k,t} = \frac{(1 - G_{k,t}) \times GDP\text{-}pc_{k,t} / \pi_{k,t}}{(1 - G_{k,2010}) \times GDP\text{-}pc_{k,2010} / \pi_{k,2010}}$

Where:

$G_{k,t}$ is the Gini coefficient (inequality measure).
$GDP\text{-}pc_{k,t}$ is GDP per capita in constant 2015 USD.
$\pi_{k,t}$ is the Consumer Price Index (CPI, 2010 = 100).
The denominator normalizes the index to a 2010 baseline (value of 1.0).

Key Design Choices:

Inequality Adjustment: The term $(1 - G)$ rescales GDP by the inverse of inequality.
Inflation Adjustment: The authors divide constant-price GDP by the CPI. They justify this by noting that the GDP deflator (production-side) and CPI (consumption-side) price different baskets. The wedge between them represents a cost-of-living effect not captured by production-side real GDP.
Simplicity and Auditability: GAGI uses only three publicly available annual inputs (World Bank, IMF, national statistics), requiring no proprietary data or complex modeling assumptions. This allows it to function as a governance trigger.
Data Tiers: The paper distinguishes between finalized data (through 2023–2024), preliminary estimates (2024–2025), extrapolated values (2026), and scenario projections (2025–2035).

Key Contributions

Operational Metric: GAGI is presented not as a new welfare concept, but as an operational instrument designed for annual computation and cross-country comparison. It is intended as a plug-in empirical proxy for the "Economic Stability" component of the Human Utility Factor (HUF) framework.
Empirical Application to G7: The paper applies GAGI to G7 economies (2010–2026) to demonstrate its utility in revealing welfare trends invisible to GDP.
Regime Identification: The analysis identifies distinct policy regimes based on GAGI trajectories: "automation without absorption" (USA), "structural stagnation without automation" (Italy, Japan), and "automation with absorption" (Nordic comparators: Denmark, Sweden).
Scenario Projections: The paper presents three scenario-based projections (Business-as-Usual, Governance-Constrained, Full Complementarity) for G7-average GAGI through 2035 to bound the welfare stakes of unconstrained automation.

Results

GAGI–GDP Wedge: In every G7 economy, the GDP per capita index exceeds the GAGI index throughout 2010–2026. This divergence widens significantly after 2022, coinciding with accelerated AI investment and deployment.
US-Specific Displacement: The United States exhibits the largest and most sustained GAGI–GDP divergence. Employer-attributed AI layoffs in the US rose 13-fold between 2023 and 2025 (from 4,247 to 54,836). Workers aged 22–25 in high-AI-exposure occupations saw a 13% employment decline.
Productivity–Wage Divergence: Global labor productivity in high-income countries outpaced real wage growth by 14 percentage points cumulatively (1999–2024). In the US, the labor share of income fell from 63.3% (2000) to 59.1% (2022).
Regime Separation:
- USA: High AI investment, rising Gini, low active labor-market policy (ALMP) spending (0.10% of GDP), resulting in welfare deterioration despite GDP growth.
- Italy/Japan: Low AI investment but near-flat GAGI, indicating structural stagnation unrelated to automation intensity.
- Nordic Comparators (Denmark/Sweden): High robot density but stable GAGI due to high ALMP spending (2.05% and 1.27% of GDP, respectively) and high collective-bargaining coverage, demonstrating "automation with absorption."
Middle-Skill Hollowing: The share of middle-wage employment fell significantly in the US (38.4% to 26.4%, 2000–2024) and EU-27 (47.3% to 41.1%, 2010–2023).
Projections: Under a "Business-as-Usual" scenario (Scenario A), G7-average GAGI is projected to decline, potentially crossing an indicative stability threshold (GAGI ≈ 0.85) in the early 2030s. A "Governance-Constrained" scenario (Scenario B), which pins the Gini coefficient via ALMP activation, avoids this decline.

Significance and Claims
The paper claims that GAGI is a necessary complement to GDP-based monitoring. It argues that any instrument tracking only aggregate output will systematically miss the distributional harm caused by automation. The significance of GAGI lies in its ability to make year-on-year and cross-country changes in welfare-adjusted prosperity legible to regulators. The authors posit that the choice of governance regime (specifically, the presence of absorption mechanisms like ALMP and redistribution) determines whether automation leads to welfare divergence or stability. GAGI is framed as a minimal, auditable tool to detect these divergences early, serving as an empirical proxy for economic stability in broader welfare-optimization frameworks. The paper does not claim causality between AI and welfare decline but highlights the temporal coincidence and the distinct trajectories revealed by the index.

The Core Problem: The "Average" Lie

The Solution: GAGI (The "Real Welfare" Scorecard)

What the Data Shows (2010–2026)

The Warning for the Future

The Bottom Line

More like this