The mining mita

How a 450-year-old colonial institution still shapes Peru today

Based on Dell (2010), Econometrica [PDF]

Built by Max Ghenis

↓ Scroll to explore

The mita boundary

Mita region
Non-mita region

The colonial labor draft

Under the mita, Spanish colonial administrators forced indigenous communities within a designated catchment area to send one-seventh of their adult male population to work in the silver mines of Potosí and the mercury mines of Huancavelica each year. The brutal conditions killed many workers.

A 450-year-old question

Spain abolished the mita in 1812. But does this colonial institution still affect Peru today? Are communities that suffered under the mita still worse off than their neighbors who escaped it?

Why we can't just compare

We could simply compare outcomes in mita vs. non-mita regions. But this would be misleading: maybe the regions differed before 1573. The mita area centered around Cusco—the Inca capital with unique history and geography. Any differences today might reflect pre-existing gaps, not the mita itself.

Focus on the boundary

The solution: compare districts right at the boundary. Communities just inside vs. just outside were nearly identical before 1573—same geography, same climate, same Inca heritage. The only difference was which side of an administrative line they fell on.

From map to data

Watch as each district transforms into a data point. Geographic location gives way to a new axis: distance from the mita boundary. Districts that were neighbors on the map now separate based on their proximity to the colonial divide.

Plotting the data

Each dot represents a district. The x-axis shows distance from the mita boundary, with mita districts on the right (dark) and non-mita districts on the left (gray). The y-axis shows child stunting rates.

Finding the trend

We fit separate regression lines on each side of the boundary. These illustrative lines show the average relationship between distance and stunting within each region. (The paper uses richer polynomial specifications with geographic controls.)

A misleading result

Our simple regression suggests mita districts have less stunting—the opposite of what we'd expect from forced labor! This counterintuitive result shows why we can't trust naive comparisons. Geography, elevation, and other factors confound the relationship.

The controlled estimate

Dell's paper uses polynomial RD with controls for elevation, slope, and other geographic factors. The refined estimate: 6 percentage points higher stunting in mita districts—about 15% above the baseline rate. Colonial exploitation continues to harm children generations later.

Finding #2: Lower consumption

The same pattern appears for household consumption. Mita districts have about 22% lower consumption today—a coefficient of -0.25 log points. The persistent poverty reflects centuries of institutional disadvantage.

Finding #3: Less infrastructure

Mita districts have 36 fewer meters of road per km²—about a third less than the non-mita average of 108 m/km². Dell traces this to land tenure: without the labor draft, large haciendas formed and invested in roads to attract workers and transport goods.

Why do effects persist?

Dell explores several mechanisms. One key channel: the mita blocked hacienda formation outside the catchment area, where large landowners invested in roads and public goods to attract workers. These institutional differences compounded over centuries, producing the disparities we observe today.

History casts a long shadow

Dell's study shows that institutions can shape economic outcomes for centuries. The mita's arbitrary boundary provides a rare opportunity to measure these persistent effects precisely.

Key findings

  • +6pp higher child stunting
  • ~22% lower household consumption
  • ~33% lower road density

These effects persist 200+ years after Spain abolished the mita.