Equity & Social Inclusion

The UESI’s approach to equity and social inclusion measures the distribution of environmental harms and benefits within a city, and assesses the relationship between environmental performance and income distribution.

We developed graphic and numeric representations of the distributive equity of selected environmental outcomes (EO) for each city. Using the concept of Concentration and Lorenz Curves, used in public health and economics, we provide a metric - the Environmental Concentration Index (ECI) - that numerically represents the distribution of the environmental outcome in relation to a scenario of perfect equity. In other words, this metric reveals how environmental outcomes are distributed and whether they are disproportionately allocated to the poorest or richest segments of the population. Additionally, we developed a typology that classifies cities based on (1) their overall income inequality and (2) the distribution of environmental outcomes.

Description

The Sustainable Development Goals (SDGs) and social and environmental equity

The United Nations’ Sustainable Development Goal 11 (SDG 11) articulates aspirations to make cities inclusive, resilient and sustainable. Specifically, it aims to ensure that people have access to “adequate, safe and affordable” housing and basic services (Target 11.1) and to “accessible and sustainable transport systems” (Target 11.2), while reducing adverse environmental impacts (Target 11.6). It also seeks to foster inclusion by enhancing citizens’ capacity to participate in urban planning and governance (Target 11.3).1 Other goals, such as SDGs 5 and 10, seek to promote gender equality and reduce inequalities among and within countries.

Cities play a great role in achieving the SDGs, given their prominent political and economic role, as well as the fact that over 50% of the global population now lives in cities. The SDG 17 and the SDG 11 Monitoring Framework emphasizes the need to disaggregate the SDG indicators (e.g., the proportion of people living in slums and the percentage of people with access to public transport) by income, sex, race, ethnicity, disability status and age, to help ensure overall progress does not leave particular groups behind.2

Environmental equity in the UESI

Building on preliminary research and previous versions,3 the UESI framework focuses on distributive equity, a key component of environmental justice. By focusing on a set of core issue areas – exposure to air pollution, urban heat island effect, distance to public transportation, and tree cover deficit – we show how environmental burdens vary across neighborhoods of different income levels within cities allowing stakeholders to identify the distributions of environmental hazards and exposures in relation to demographic considerations. While such demographic considerations may include multiple socioeconomic variables to highlight cumulative social vulnerabilities and intersectionalities from education levels, minority/ethnicity statuses, or age compositions, data on these variables at the desired neighborhood scale is only available and complete for some cities. Balancing data quality, availability and coverage, the UESI focuses primarily on understanding how urban residents with varying income levels may be affected by environmental conditions.

To ensure a consistent analysis across all cities, we utilize income (both household and average individual income), when available at the neighborhood scale, as the primary socioeconomic indicator. Income data for each city comes in different currencies, years and units (e.g., household vs. individual income, monthly vs. yearly income, number of people within an income bracket vs. mean/median income within a neighborhood). We standardize the income data to represent the per capita income for each district maintaining the national currency of each city. In addition, it is worth mentioning that some neighborhoods had no data available due to statistical confidentiality and thus were excluded from the analysis. When income data is unavailable, we rely on GDP per capita from globally available datasets. However, since GDP reflects productivity and tends to show less variation within urban areas, the inequality analysis may underestimate the actual inequality in those cities.

Calculating equity and social inclusion in the UESI

Drawing from the extensive literature and tools for analyzing environmental outcomes, we developed an approach to assess equity by considering both environmental and socioeconomic factors.4 5 6 7 8 9The UESI approach draws heavily on the use of graphical representations, such as Concentration and Lorenz Curves, to capture the distribution of environmental outcomes and income across a city. In addition, the UESI approach includes a numerical representation of the distribution of income and environmental outcomes. Together, these representations shed light on the relationship between the distribution of environmental outcomes and income within cities.

One caveat to our analysis is that primary demographic and economic data are reported at the neighborhood level. We therefore calculate measures of these characteristics’ distribution across neighborhoods, weighted by population, rather than across individuals per se. This approach is equivalent to assuming that all individuals within a neighborhood are identical in terms of both environmental and economic characteristics. We recognize, however, that administrative boundaries often encompass highly heterogeneous populations; some of the wealthiest live juxtaposed to the poorest.10 Given the granularity of the data, our approach should be interpreted as a first attempt to quantify this issue in global cities.

Graphical and numerical representations of income and environmental outcomes

We use Lorenz and Concentration curves to analyze the distribution of income and environmental outcomes such as Air Pollution (Average Exposure to PM_2.5 and NO₂), Urban Heat Island Intensity (CUHI and SUHI); Public Transit (Distance Public Transit Station); and Tree Cover (Tree Cover per capita) for each city. In these plots, the x-axis is the cumulative proportion of a city’s population ranked by income, and the y-axis is the cumulative proportion of income (Lorenz Curve) or environmental outcome (Concentration Curve) available in the entire city. If income and environmental outcomes were distributed equally across a city’s total population, the Lorenz and Concentration curves would look like the 45-degree line.

By definition, the Lorenz income curve will never be above the 45-degree line (i.e., the line of perfect equity). The distance between the income Lorenz curve and the 45-degree line of perfect equity indicates the degree of income distribution inequality: a greater distance between the two lines indicates a more unequal distribution of income.

Environmental Concentration curves can be interpreted differently based on whether they are above or below the line of perfect equity. If the curve is fully located above the line of perfect equity, it indicates that the negative environmental outcome is more heavily allocated to those with less income.11 On the contrary, if the curve is fully located below the line of perfect equity it indicates that the environmental outcome is more heavily allocated to those with more income . However, it's important to note that Environmental Concentration curves can have sections located at both sides of the 45 degree line; in this case the interpretation will not be as straightforward as it was mentioned before, and the interpretation would rely on additional metrics that we will explore in the following sections. Figure 1 provides a graphical representation of the Environmental Concentration Curves and their interpretation.

Figure 1. Examples of Environmental Concentration curves and their interpretation. The left panel shows an Environmental Concentration Curve that falls above the line of perfect equity, indicating that poorer populations face higher UHI intensity. The right panel shows a concentration curve that falls simultaneously on both sides of the line of equity, making it difficult to obtain a conclusion about the tree cover burden allocation based exclusively on the curve. In this situation, we use a numeric approach to determine allocation to different income groups, which we will explain in the next section.

The specific interpretations of the concentration curves for different indicators follow below:

PM_2.5 Equity: Cumulative proportion of total exposure to PM_2.5 concentration for the entire population (Negative Environmental Outcome)
NO₂ Equity: Cumulative proportion of total exposure to NO₂ concentration for the entire population (Negative Environmental Outcome)
UHI Equity: Cumulative proportion of total exposure to UHI Intensity for the entire population (Negative Environmental Outcome)
Tree Cover Equity: Cumulative proportion of the total Tree Cover per capita for the entire population (Negative Environmental Outcome)
Distance to Public Transit Equity: Cumulative proportion of total distance to nearest public transportation station for the entire population (Positive Environmental Outcome)

An example of two sample Environmental Concentration Curves, and the Income Lorenz Curves for Johannesburg can be seen in Figure 2. It is important to note that the Concentration curves do not indicate whether the cumulative exposure to an indicator (UHI intensity, air pollutants, etc.) for a given city is generally large or small. They simply indicate whether the distribution of these environmental burdens is more or less equally distributed to certain income groups.

Figure 2. Results of the distribution of Income Lorenz Curve (Red), UHI Intensity and NO2 Exposure Concentration Curves (Blue) for the city of Johannesburg. The Environmental Concentration Curve results indicate that UHI Intensity is concentrated in the low-income populations in the city. By contrast, the NO2 exposure is only slightly more concentrated in high-income populations, and still within a marginal distance from the line of perfect equity. Finally, the Income Curve indicates that there is an important inequality in the distribution of income, due to the distance of the curve to the 45 degree line of equity.

With the exception of the UESI,12 most of the recent studies analyzing distributional equity in the environmental field use Lorenz Curves and the Gini Index to assess inequality in the distribution of some environmental outcomes like GHG Emissions13 or water related risk.14 Although methodologically robust, the construction and interpretation of these curves can be challenging from a decision-maker’s perspective. Therefore, building on the literature around the use of concentration indices as an accepted measure of inequalities, particularly around health-related outcomes,15 16 17 18 and considering that it considers both population and income distributions, unlike the Gini index which only considers the former, we use this approach to explore the relationship between income and environmental outcome distributions.

The UESI’s approach to numerically quantify inequality uses the concentration curves presented in the previous section, and calculates a summary measure called the Environmental Concentration Index (ECI) using the following formula:

ECI = 1 - 2*AUCenv

Where ECI is the concentration index for the environmental outcome (env) and AUC is the area under its corresponding curve.19

A concentration index value can range from -1 (i.e, the environmental burden is allocated to the poorest individual) to 1 (i.e., the environmental burden is allocated to the wealthiest person).20 The ECI helps us to summarize inequality in a city using a single numeric value, which is particularly useful for when conclusions are difficult to obtain from the graphical representations alone.

Given its definition, the absolute value of the ECI is the net inequality of a city, where the numeric value indicates whether richer or poorer populations are more burdened by that inequality.21 As a result, while useful by themselves, the ECI values should be analyzed in conjunction with the Lorenz Equity curves to have a more accurate interpretation of the results around the presence of different pockets of inequality.

Complementary to the ECI, the UESI also calculates the Gini Coefficient for each city. The Gini Income Coefficient, a commonly-used metric of income inequality, uses the Lorenz Curve to quantify income inequality. This calculation is done using the same formula as the ECI, as both of them are based on the area under the curve (AUC) value. A graphic example of these calculations is shown in Figure 3.

Figure 3. Example of Environmental concentration for Surface Urban Heat Island and the Lorenz Income Curve for the city of Los Angeles.

Typology of relationships between income and environmental outcomes

Even though the ECI is an important metric to describe the relationship between income and environmental outcomes, the picture is still incomplete because it doesn’t consider income inequality. To contextualize these results in a more meaningful way, and to understand the interplay between environmental and socioeconomic inequalities, we developed a typology to categorize where cities fall in relation to each other using both the ECI and the Gini Coefficients.

The typology’s four quadrants are defined by two axes. The x-axis denotes the Environmental Concentration Index of an environmental outcome, and the y-axis denotes income inequality as expressed through the Gini Coefficient. The UESI uses a Gini Coefficient value of 0.1622 and an ECI value of 0 to separate the quadrants.

To better understand the potential interaction between both distributions, it is important to consider that ECI and Gini Coefficients are not interlinked. Pollution and environmental degradation have a direct impact on livelihoods and health, particularly of those who have less capacity to cope with those negative environmental drivers, such as the poor23 This negative feedback loop can create a poverty trap, which hinders the capacity of citizens to escape improve their livelihoods and economic capacity.24 Thus, looking at the space where both environmental and income inequality occurs can help identify spaces of particular concern to address both environmental performance and equity.

For example, a group of people disproportionately burdened by air pollution (a negative environmental outcome) may be further economically disadvantaged through additional healthcare costs resulting from pollution-related respiratory problems. Conversely, a segment of the population that has less access to Tree Cover (a positive environmental outcome) would have to spend time and money to travel to another area with higher tree cover if they expect to enjoy the same benefits tree cover affords.

This impact of the environmental allocation on the income distribution, which we call environmental pressure, is also a potential source of inequality unaccounted for by most analyses. When environmental pressure is placed on poorer populations, it can exacerbate the inequalities between the poorer and richer residents of a city by increasing the resources that the poorest need to invest to compensate for the negative impacts, or for gaining access to the positive ones (See Box 1 - Urban Heat Island Inequity and Energy Burden in US Cities). The degrees to which environmental burdens exacerbate income inequality are described by the scenarios detailed in the quadrants in Figure 4. The interpretation of the ECI values related to the quadrants will depend on the type of environmental outcome (EO) analyzed, positive or negative.

Figure 4: The four-quadrant typology of the UESI and the scenarios it defines.

While most projects that utilize the Lorenz and Concentration curves use data points from individuals, the UESI uses neighborhood-level data to make comparisons. While making our equity calculations on the neighborhood level may disregard heterogeneity within neighborhoods, the UESI provides a more aggregated reflection of inequalities than those that can happen at the micro or individual scale.

Results

The results of this analysis have highlighted important aspects of inequality and its relation to the environmental burdens. The results suggest that while some cities in the Global South, such as São Paulo, are unequal in their income distribution, income inequality occurs in both developing and developed world cities. For instance, some U.S. cities like Boston and Atlanta have relatively high income inequality, while some cities in the Global South, like Jakarta, demonstrate lower income inequality.

In general, most cities included in the analysis are located in the low-income inequality quadrants (top right and left quadrants) of the typology, while the other cities present higher than average income inequality environmental (bottom right and left quadrants). The environmental inequality, on the other hand, varies heavily depending on the environmental outcome in question but for most of them we can find that the poorer citizens are disproportionately affected by a negative environmental outcome. For detailed analyses, results and interpretations of equity calculations, please refer to the corresponding chapters.

Box 1. Urban Heat Island Inequity and Energy Burden in US Cities

Urban environmental hazards, represented by the urban heat island effect, are often considered to be directly related to higher energy use. Existing research shows that households tend to use air conditioning more to maintain comfort needs during hot weather and turn on heating devices in cold winters.25 A possible consequence of this hypothesis lies in the fact that urban residents may bear a greater energy burden, which could further trap low-income communities in energy poverty, as a large portion of their income is tightly linked to energy costs.26

To verify whether heat exposure and its consequences are disproportionately distributed, we explored the relationship between heat exposure equity using the UESI’s CUHI Concentration index and their energy burden - summarized as the z-score for each city - using a four-quadrant plot. The CUHI concentration index is calculated using canopy urban heat island and income data, where negative values indicate that it is disproportionately burdening the lowest income earners. Energy burden is defined as the proportion of household income spent on energy costs, with higher energy burdens representing greater difficulties for households in meeting their energy needs.

Using the United States as the study area, the results in Figure 5 show that low-income groups in less affluent U.S. cities are exposed to higher CUHI intensity, indicated by negative CUHI Concentration indices (35 out of 41 cities) on the plot. Further exploration reveals that in 20 of these 35 cities (~60%), the average citizen also experiences higher energy burdens than in other cities. In other words, in cities where low-income groups already face disproportionate environmental hazards, their ability to cope with or adapt to these hazards is likewise relatively lower, as they also face higher energy burdens. Conversely, in cities in the lower left quadrant, although low-income groups face disproportionate heat exposure, their energy burdens are relatively low.

Figure 5 Relationship between U.S. city-wide CUHI concentration index and energy burden.

To further investigate the hypothesis about adaptive capacity of less affluent cities, Figure 6 explores the relationship between the UESI’s Concentration Index for Tree Cover per capita and the energy burden. The results highlight that in most of the cities analyzed, the lowest income earners have less access to green spaces than their wealthier counterparts, indicated by the negative Concentration Index. In particular, the residents in the cities of the top left quadrant also have a higher than average energy burden, which indicates that in addition to the economic burden associated with energy access, they also have less access to natural cooling spaces.

Figure 6. Relationship between U.S. city-wide tree cover per capita concentration index and energy burden.

Looking at these two issues in conjunction, we can identify certain cities such as Evansville, New York, Phoenix or Miami, where lower income citizens face a disproportionate exposure to Urban Heat Island, lower access to Tree Cover and experience higher than average energy burden, which indicates the presence of a particularly stark situation, where environmental and economic conditions can exacerbate negative effects on health and wellbeing to the residents of those cities. Overall, the results highlight the importance of analyzing environmental and economic inequities in conjunction to gain a better understanding on the effect or concurrent inequities that disproportionately impact the citizens with lowest resources.

Accompanying indicators for goal 11 include the “proportion of urban population living in slums, informal settlements, or inadequate housing” (indicator 11.1.1) and the “percentage of cities with a direct participation structure of civil society in urban planning and management which operate regularly and democratically” (indicator 11.3.2). Global indicator framework for the Sustainable Development Goals and targets of the 2030 Agenda for Sustainable Development. Retrieved from: https://unstats.un.org/sdgs/indicators/Global%20Indicator%20Framework_A.RES.71.313%20Annex.pdf.
ibid.
Hsu, A., N. Alexandre, J. Brandt, T. Chakraborty, S. Comess, A. Feierman, T. Huang, S. Janaskie, D. Manya, M. Moroney, N. Moyo, R. Rauber, G. Sherriff, R. Thomas, J. Tong, Y. Xie, A. Weinfurter, Z. Yeo (in alpha order), (2018). 2018 Urban Environment and Social Inclusion Index. New Haven, CT: Yale University. Available: datadrivenlab.org/urban.
Maguire, K., & Sheriff, G. (2011). Comparing distributions of environmental outcomes for regulatory environmental justice analysis. International journal of environmental research and public health, 8(5), 1707-1726.
Sheriff, G., & Maguire, K. (2013). Ranking Distributions of Environmental Outcomes Across Population Groups.
Groot, L. (2010). Carbon Lorenz curves. Resource and Energy Economics, 32(1), 45-64.
Padilla, E., & Serrano, A. (2006). Inequality in CO₂ emissions across countries and its relationship with income inequality: a distributive approach. Energy Policy, 34(14), 1762-1772.
Cantore, N., & Padilla, E. (2007). Equity and CO₂ emissions distribution in climate change integrated assessment modelling (No. 7001). Alma Mater Studiorum University of Bologna, Department of Agricultural Economics and Engineering.
Maguire, K., & Sheriff, G. (2011). Quantifying the Distribution of Environmental Outcomes for Regulatory Environmental Justice Analysis (No. 201102). National Center for Environmental Economics, US Environmental Protection Agency.
Miller, J. www.unequalscenes.com.
It is important to consider that this interpretation is based primarily on the definition of the environmental indicators as measuring environmental burden (i.e. higher values indicate worse quality or more negative conditions). If this definition were to change, so would the interpretation of the relative position of the Concentration Curves. It is also important to mention that this approach is not intended to compare income versus environmental distributions - in other words, it should not be used to judge whether the distribution of an environmental outcome is more or less equitable than that of income for the same city, or between the same curves for different cities.
Hsu, A., Chakraborty, T., Thomas, R., Manya, D., Weinfurter, A., Chin, N. J. W., ... & Feierman, A. (2020). Measuring what matters, where it matters: A spatially explicit urban environment and social inclusion index for the sustainable development goals. Frontiers in Sustainable Cities, 2, 556484.
Wu, S., & Chen, Z. M. (2023). Carbon inequality in China: Evidence from city-level data. China Economic Review, 78, 101940.
Sanders, B. F., Brady, D., Schubert, J. E., Martin, E. M. H., Davis, S. J., & Mach, K. J. (2024). Quantifying Social Inequalities in Flood Risk. ASCE OPEN: Multidisciplinary Journal of Civil Engineering, 2(1), 04024004.
Kakwani, N., Wagstaff, A., & Van Doorslaer, E. (1997). Socioeconomic inequalities in health: measurement, computation, and statistical inference. Journal of econometrics, 87-103.
Kaufman, J. S. (2017). Methods in social epidemiology (Vol. 16). John Wiley & Sons.
Elgar, F. J., McKinnon, B., Torsheim, T., Schnohr, C. W., Mazur, J., Cavallo, F., & Currie, C. (2016). Patterns of socioeconomic inequality in adolescent health differ according to the measure of socioeconomic position. Social indicators research, 127(3), 1169-1180.
Costa-Font, J., Hernandez-Quevedo, C., & Sato, A. (2018). A Health ‘Kuznets’ Curve’? Cross-Sectional and Longitudinal Evidence on Concentration Indices’. Social indicators research, 136(2), 439-452.
The trapezoidal rule is a method for calculating the integral of a function based on the construction of trapezoids for sequential sections of an interval, for this case the sections are build by the pairs of sequential cartesian points that define the Concentration Curve and the Lorenz Curve. The areas of the trapezoids are added and the sum is equal to the integral of the function, which by definition is the area under the curve (AUC).
Maguire, K., & Sheriff, G. (2011). Quantifying the Distribution of Environmental Outcomes for Regulatory Environmental Justice Analysis (No. 201102). National Center for Environmental Economics, US Environmental Protection Agency.
By definition, the Concentration Index includes the assumption that the inequality in the distribution of the outcomes is equivalent for income groups. As a result, if there are similar levels of inequality for the poorest and richest groups - represented by a curve that is above and below the 45 degree line - the results of the ECI will be closer to zero, because both inequalities would mathematically compensate each other.
Average Gini coefficient values for all UESI cities.
Das, M., & Basu, S. R. (2022). Understanding the relationship between income inequality and pollution: A fresh perspective with cross-country evidence. World Development Perspectives, 26, 100410.
Bonds, M. H., Keenan, D. C., Rohani, P., & Sachs, J. D. (2010). Poverty trap formed by the ecology of infectious diseases. Proceedings of the Royal Society B: Biological Sciences, 277(1685), 1185-1192.
Auffhammer, M. (2022). Climate Adaptive Response Estimation: Short and long run impacts of climate change on residential electricity and natural gas consumption. Journal of Environmental Economics and Management, 114, 102669.
Yu, Y., & Kittner, N. (2024). Hot or cold temperature disproportionately impacts US energy burdens. Environmental Research Letters, 19(1), 014079.