We Pretended the Poor Would Catch Up. They Never Did.

There is a way in which the graph tells you everything you need to know about what happened here. The line for GDP per capita climbs steadily upward from 1980 to 2026 — the reassuring arc of progress, the vindication of markets, the proof that growth works. Then you notice the second line, the one measuring the income share of the bottom fifty percent. It does not climb. In some countries it falls. In others it moves sideways for four decades, a flat electrocardiogram of failed promise. The space between the two lines is where the theory lived, and where it died.

I have been thinking about this space. About what we were told would happen there, and what actually did.

The Story We Told Ourselves

The story went like this: Growth comes first. Redistribution can wait. Build the economy, attract investment, liberalize markets, and prosperity will trickle down to those at the bottom. A rising tide lifts all boats. We heard this in Washington and Geneva, in the structural adjustment programs of the 1980s and the Washington Consensus of the 1990s and the Millennium Development Goals of the 2000s. We heard it so many times that it stopped sounding like a theory and started sounding like physics.

The theory had a certain elegance. It absolved us of difficult questions. It meant we did not have to talk about land reform or progressive taxation or labor rights or any of the messy redistributive politics that might upset investors. It meant development could proceed without conflict, without confronting who owns what and who decides and who benefits. Growth was technical. Inequality was political. We chose the technical.

I know what I am talking about here. I have read the World Bank reports and the IMF working papers and the OECD policy briefs. I have watched the PowerPoint presentations in hotel conference rooms in capital cities, the ones with the upward-sloping arrows and the confident projections. The question is not whether growth happened. It did. The question is where it went.

What the Evidence Shows

The randomized controlled trials started arriving in the early 2000s. Esther Duflo and Abhijit Banerjee at MIT, Michael Kremer at Harvard, a generation of development economists who decided to test what actually worked instead of what theory said should work. They measured microcredit schemes and deworming programs and conditional cash transfers and teacher incentives. They found that context mattered. That poor people respond to incentives. That small interventions targeted at specific constraints — bed nets, iodized salt, deworming medication — could produce measurable improvements in health and education outcomes.

They also found what did not work. Microcredit, the darling of the 1990s, produced no significant increase in household income or consumption in six separate trials across six countries. Building schools did not improve learning outcomes if teachers did not show up. Providing free fertilizer increased yields only if farmers had secure land tenure. The interventions that worked were often the ones that directly addressed market failures and power imbalances, not the ones that assumed markets would self-correct.

This was evidence-based development. It was humble about what it knew. It tested interventions rigorously. It acknowledged failure. And yet. And yet it still operated within a framework that treated poverty as a problem of individual households rather than structural inequality. It optimized programs for the poor without asking why there were so many poor people to begin with. It was, in a sense, trickle-down research: help enough individuals, and eventually the structure will change. The evidence suggests otherwise.

The Countries That Took a Different Path

There is another set of data points worth examining. South Korea in the 1960s implemented aggressive land reform, breaking up estates and redistributing to smallholders. Taiwan did the same. Both countries invested heavily in universal primary education before industrialization, not after. Both imposed strict capital controls. Both protected infant industries. Both had authoritarian governments that suppressed labor movements, which complicates the story, but both also ensured that growth was broadly shared through deliberate policy choices about wages, education, and asset ownership.

Brazil in the 2000s tried something different. Bolsa Família, the conditional cash transfer program, reached 14 million families by 2012. It was not microcredit. It was not a loan. It was direct redistribution, funded by progressive taxation, conditional on children attending school and receiving health checkups. Between 2001 and 2011, Brazil's Gini coefficient fell from 0.594 to 0.527 — a significant decline in one of the world's most unequal societies. The poverty rate dropped by half.

Then the commodity boom ended. Fiscal space tightened. Political will evaporated. By 2016, Bolsa Família's real value had eroded. The progress stalled. The lesson here is not that redistribution failed. The lesson is that redistribution requires sustained political commitment and cannot rely on commodity windfalls. It requires a decision that reducing inequality is a goal in itself, not a happy byproduct of growth.

Who Benefits From the Current Arrangement

There is a reason the growth-first story persisted for sixty years despite mounting contrary evidence. It is the same reason trickle-down economics persisted in the United States and austerity persisted in Europe. The theory was convenient for those who already had capital. It justified low taxes on wealth, weak labor protections, minimal regulation, and unlimited capital mobility. It meant development agencies could maintain good relations with finance ministries and foreign investors. It meant avoiding confrontation with elites in recipient countries over land ownership, inheritance laws, capital gains taxes, or any other policy that might actually redistribute power.

I am not suggesting conspiracy. I am suggesting that theories survive because they serve interests. The Washington Consensus was not wrong because economists were stupid. It was wrong because it prioritized the concerns of creditors over debtors, investors over workers, growth over distribution. It assumed the political status quo and optimized policy within those constraints. What it could not do was question the constraints themselves.

The Reckoning

Let us return to the graph. The space between the two lines — between total growth and the income share of the bottom half — represents trillions of dollars of created wealth that did not reach those who needed it most. It represents schools not built, clinics not staffed, farms not irrigated, infrastructure not maintained. It represents political choices dressed up as economic necessity.

The evidence-based development movement got many things right. It brought rigor to evaluation. It tested what worked. It challenged assumptions. But it operated at the level of interventions when the problem was structural. You cannot randomize land reform. You cannot run a controlled trial of progressive taxation. You cannot A/B test a minimum wage law. The questions that matter most in development are not amenable to the experimental method.

Which means we are back to politics. To decisions about who owns the land and who sets the wages and who pays the taxes and who controls the resources. These are not technical questions. They never were. We pretended they were because it was easier than admitting that development requires confronting power.

I think about the graph often now. About the distance between the two lines. About what we told ourselves would happen in that space, and what happened instead. About the children who did not get educated and the sick who did not get treated and the farmers who lost their land because we chose a theory that made inequality acceptable. We told ourselves stories in order to continue. The question is whether we will finally tell a different one.

The answer is not more growth. It is not better interventions. It is not smarter technocrats. The answer is redistribution — of land, of wealth, of power. It always was. We just chose not to see it.