From 1.85 Billion to 4.7 Billion: How Homi Kharas Predicted the Rise of the Global Middle Class

In 2009 — the height of the global financial crisis — most economists had their eyes fixed on a collapsing Western middle class. However, one researcher was looking east. Homi Kharas, co-founder of World Data Lab and globally recognized authority on the middle class, published a landmark paper with the OECD that shifted how the world thinks about consumer growth. Seventeen years later, the forecasts have held up remarkably well.

When Kharas began his research in 2009, the conventional narrative was one of Western decline. The American and European middle class, long the engine of global consumer spending, was under severe strain. But Kharas had spent years working in East Asia, where rapid growth was building a new consumer base from a low starting point.

"I wanted to do a simple experiment to see how much of East Asia's growth could offset and mitigate the reduction in growth happening in the United States and Europe," Kharas explains. His conclusion pointed to a long-term shift in the global economy's center of gravity away from the West and toward Asia, accelerated by China's accession to the World Trade Organization in 2001.

At the time of publication, Kharas estimated 1.85 billion people belonged to the global middle class. He projected this figure would reach roughly 3.25 billion by 2020. The actual count, as later calculated by World Data Lab: 3.4 billion — slightly above the original forecast.

The 2030 projection originally stood at 4.9 billion. World Data Lab now forecasts 4.7 billion: a gap of approximately 200 million people, attributable largely to the economic disruption caused by the COVID-19 pandemic. For a forecast made over a decade in advance, the margin is remarkably narrow.

The middle class is where economic activity concentrates. According to Kharas, the global middle class accounts for around 60% of all consumer spending, with the affluent contributing roughly a third. Together, they dominate global consumption. Understanding this group — where they live, how old they are, what they spend on — is foundational to understanding the direction of the global economy.

This is precisely the territory that World Data Lab maps. The platform goes beyond headcounts to track spending levels, age-group breakdowns, generational differences, urban versus secondary-city patterns, and category-level expenditure, from housing to health to digital services.

Kharas is careful to distinguish between conceptual clarity and execution difficulty. "It's easy to do, but very hard to implement," he says. The underlying drivers — demographics, long-term economic growth rates, urbanization — are stable enough in aggregate to support multi-decade projections. But translating those aggregate trends into granular, actionable data across hundreds of countries and thousands of cities is an entirely different challenge.

World Data Lab's methodology draws on household survey data, geospatial satellite imagery, and large-scale open datasets like Google Maps and OpenStreetMaps. Survey data tends to under-represent the tails of the income distribution — both the very poor and the very wealthy — so multiple data streams are used to validate and calibrate the models. The result is a continuously updated, globally consistent dataset that improves with every iteration.

One area where this empirical approach is especially valuable is climate emissions forecasting. Many models rely on government commitments and policy pledges. World Data Lab takes a different route: measuring what is actually being implemented, not what has been promised. "We use that to make our forecasts rather than just the promises of what people say they're going to do," Kharas notes — a distinction with significant implications for investment strategy, policy planning, and business risk assessment.

A key insight from the conversation is the tension between scale and actionability. Aggregate global trends are useful as orientation, but they rarely drive decisions. The real value lies in disaggregation: which cities are growing fastest, which age cohorts are entering peak spending years, which product categories are set to expand in which markets. The more granular the forecast, the harder it is to maintain precision — but also the more useful it becomes for the companies, governments, and investors who rely on it.

After seventeen years of tracked forecasts and continuous model refinement, World Data Lab's methodology has demonstrated a rare quality in economics: it works. Not because the future is predetermined, but because the right signals — demographics, structural growth, urbanization — move slowly enough to be measured, modeled, and trusted.