Media Briefings

NOT KEEPING UP WITH THE JONESES: Some US city neighbourhoods have income inequality higher than Mexico or Chile

  • Published Date: April 2017

Even close neighbours in American cities have income differences much higher than the national average for inequality – sometime as high as in the most unequal OECD countries. These are among the findings of a new study by Francesco Andreoli and Eugenio Peluso, to be presented at the Royal Economic Society's annual conference at the University of Bristol in April 2017.

The authors use a rich income database constructed from the US population census and the American Community Survey to test inequality in the 50 largest American metro areas. While US cities are traditionally more unequal than the country as a whole, and the difference between rich and poor is higher in cities like New York than in countries such as Chile or Mexico, it has traditionally been assumed that residents of neighbourhoods within those cities have much more uniform incomes. For many of the cities measured, this turns out not to be the case.

The authors also detect an intergenerational trade-off: greater income inequality within individual neighbourhoods – the ‘mixed city’ model – is associated with life expectancy gains for poor adults; but relatively low levels of neighbourhood income inequality – the ‘even city’ ideal – promotes young people’s opportunities for upward mobility.


Income inequality among neighbours of the average American household is high and close to citywide inequality in major US metro areas. The desirability of more urban income mix is questionable, as it is shown to have opposite implications for the long-term economic and health prospects of young and adult residents. These are among the findings of the new study.

Income inequality in major US metropolitan areas is, and has historically been, high. Big urban agglomerations convey in fact sizable wage dispersion that contributes to inequality.

In 2014, for example, the Gini index for New York City was above 0.5, higher than in the most unequal OECD countries (Chile and Mexico) and 25% more than inequality in the country as a whole. It is a commonsense also supported by empirical evidence that poor and rich people segregate across the urban space, making inequality observed in certain places of the city look different from inequality observed in the city as a whole.

New methodology recovers strong evidence that inequality among the neighbours of the average American household is generally high and close to citywide inequality, even for neighbourhoods of small size.

There is nevertheless heterogeneity in neighbourhood inequality across these cities. For example, if neighbourhoods are conceptualised as the portion of land spanning for at most two miles radius around each individual, one can show that within-neighbourhood income inequality in New York City, Chicago or Los Angeles is at least 20% larger than inequality measured in Washington DC.

To get to this result, the authors have introduced a new measurement framework. They propose to shift the focus from the traditional perspective, which takes the administrative partitions of the city as given and measures income inequality in the neighbourhoods thereby defined, to move towards an individualised concept of neighbourhood.

The most intuitive way to represent such a neighbourhood is to draw circles of a given radius around each individual (who can be precisely located on the city map) and to mark as neighbours all people falling within a range from her. In this setting, the distance between residents identifies close or far neighbours.

Income inequality between neighbours and across neighbourhoods is measured by two new indices: the Gini Individual Neighbourhood Inequality (GINI). The GINI-within index is the average degree of income inequality within individual neighbourhoods; the GINI-between index measures the inequality of the average incomes among individual neighbourhoods.

The GINI indices capture different features of the spatial distribution of incomes. They are used by the authors to describe categories of cities identified in urban planning and sociology literature.

Some cities display everywhere relatively low income inequality (the ‘even city’ model) while neighbourhood inequality in other cities is high but evenly distributed throughout the metro area (the ‘mixed city’ model).

Other cities display relatively high inequality between neighbourhoods (the ‘divided city’ model), with either little variability in incomes among close neighbours (the ‘polarised city’ case) or high local income heterogeneity (the ‘unstable city’ case).

The authors use a rich income database constructed from the US population census (1980, 1990 and 2000) and the American Community Survey (ACS, 2010-14) to put patterns of inequality in the 50 largest American metro areas to the test.

Income information is reported on an extremely fine spatial scale (block group level), which has been georeferenced. This makes it possible to keep track of inequality even for a small individual neighbourhood size.

Spatial income inequality within and between neighbourhoods is shown to be sizable in largest American cities. The GINI-within index has been constantly on the rise over the last 35 years, implying that the average resident is facing substantial and steeply increasing income inequality within the neighbourhood.

This pattern might be explained by increasing dispersion in the citywide income distribution, as well as by sorting behaviour of the resident (it is, for example, coherent with implications of recent waves of gentrification observed in major US cities).

Increasing inequality within the neighbourhood has implications for the resident population, but the direction of these implications is not clear-cut. As shown in Figure 1, less inequality within the neighbourhood is associated with positive and large upward mobility gains for children who grew up in poor families. Living in an homogenous local community (as proxied by parental income inequality within the neighbourhood) facilitates social interaction among young peers, as motivated by positive contagion or collective socialisation mechanisms.

Conversely, the authors find that larger income inequality within individual neighbourhoods is associated with life expectancy gains for adult poor residents. The income mix, in fact, might approximate for (otherwise unobservable) positive role models and raising aspirations among adult residents.

An intergenerational trade-off arises: while the ‘mixed city’ model has the potential to improve adult health prospects, the ‘even city’ ideal promotes children’s opportunities for upward mobility. These dimensions should be taken into consideration by urban planners.


So close yet so unequal: Reconsidering spatial inequality in US cities
Francesco Andreoli (LISER) and Eugenio Peluso (University of Verona)

Francesco Andreoli
Luxembourg Institute of Socio-Economic Research (LISER)
11 Porte des Sciences, L-4366 Esch-sur-Alzette, Luxembourg
Telephone: +352 5858551 (office)

Eugenio Peluso
University of Verona, via Cantarane 24, 37129 Verona, Italy

Figure 1 : Inequality within individual neighbourhoods in major US metro areas, neighbourhood effects (upward mobility gains/losses for children living in poor families in 2000, by Commuting Zone) and life expectancy (all American males aged 40 and above in 2000, by Commuting Zone).