Mohenjo-daro, the planned Bronze Age city of the Indus Civilisation in what is now Sindh, Pakistan, scored a Gini coefficient of 0.44 for the sizes of its houses, and within one neighbourhood that figure fell from about 0.39 around 2500 BC to about 0.23 by around 2100 BC. The measurements cover 309 reconstructed residences. Adam S. Green of the University of York and his co-authors Iqtedar Alam and Cameron Petrie read the drop as a city becoming more equal over time.
AI, explained every day: The Deep View is the go-to daily newsletter for 600,000+ founders, knowledge workers and professionals who want to stay up to date with artificial intelligence.
We may earn a commission if you sign up.
That is a narrower claim than it first sounds, and the way a Gini coefficient is built explains why.
The 309 houses behind the number
The team used records from early twentieth-century excavations and a geographic information system to reconstruct the layouts of 309 residences at Mohenjo-daro, then calculated Gini coefficients from differences in residential floor area. The study, titled Inequality declined in the Bronze Age city of Mohenjo-daro, appeared in Antiquity, volume 100, issue 413, with the DOI 10.15184/aqy.2026.10359. It was first published online on 18 May 2026, and the city it studies dates to the Indus Civilisation, roughly 2600 to 1900 BC.
Green has said the team measured all of the excavated houses and entered them in the GINI database, a collaborative project on long-term inequality that holds data from more than 50,000 residences. In a first-person summary for Popular Archaeology, he described house area as a reasonable but imperfect stand-in for wealth or income.
What a Gini coefficient of 0.44 means
A Gini coefficient runs from 0, where every household has the same, to 1, where a tiny share holds nearly everything. Applied to house sizes, it asks how unevenly floor area is spread across the sampled homes. Mohenjo-daro’s 0.44 is described as significantly lower than several other ancient cities: Knossos on Crete scored 0.86 and Palenque in the Maya lowlands 0.75, while Ur and Ugarit both came in above 0.6. A figure of 0.44 is still well above zero, so the city was not perfectly equal, a point Green makes himself.
The trend sits in one place. In the DK-G South neighbourhood, which has the clearest chronology of the areas studied, the coefficient was about 0.39 around 2500 BC and about 0.23 around 2100 BC. Median residential floor area in that neighbourhood first fell and then rose in later phases, when inequality was lowest, which the authors take as a sign that prosperity grew alongside equality, although the link is uncertain because chronological data are limited.
What a house-size Gini can and cannot tell
What it can tell is how dispersed dwelling sizes were. Floor area is one of the few things archaeologists can measure the same way across thousands of buildings and across cities separated by oceans, which is why the comparison with Knossos or Palenque is possible at all. A coefficient that falls over time means the houses in the sample became more similar in size, and that is a real, checkable pattern in the plan of a city.
What it cannot tell is the rest. Residential floor area is an indirect measure that does not capture all differences in wealth, status or access to resources, as the HeritageDaily report of the study states. A merchant with modest rooms and a store of goods, a household that rented, or people who lived outside the excavated houses would not register. The number also describes the surviving, excavated sample, not every person who lived in the city, and a Gini carries no information about why the sizes converged.
There is also a technical limit that applies to any Gini built on a sample. The coefficient compares each house with the others in the set, so it depends on which houses made it into the set, and excavations of a Bronze Age city never uncover everything. Two cities with the same coefficient could differ enormously in how large their biggest houses were, how many people lived in small ones, or how much wealth sat in forms that leave no floor plan, such as livestock, goods in storage or jewellery. For that reason the useful comparison is the direction of change within one comparable neighbourhood, which is the form of evidence the study leans on, rather than a ranking of whole cities from one number.
On the why, the authors offer an interpretation rather than a measurement. They point to shared infrastructure, including public buildings, drainage systems and standardised weights and measures, to seals found in homes instead of palaces or temples, and to residential buildings that increasingly followed the planned street network. From this they suggest collective governance that maintained public goods may have limited economic inequality, challenging the assumption that urbanisation inevitably led to greater social inequality. Green adds that the exact mechanisms are unknown and could include equal access to land and fuel for bricks, or community limits on house size.
The paper is also archived by the White Rose university repository, where the published version can be read, and the repository record gives the submission and acceptance dates.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
AI, explained every day: Get The Deep View →
We may earn a commission if you sign up.
More from Morning Overview
