Urban segregation and economic inequality in Sweden (1991

Abstract

Chinese

This article bridges the gap between inequality and segregation research by introducing a method to decompose income segregation across spatial units, income groups, and income sources, applied to detailed full-population register data for Sweden from 1991 to 2017. Sweden’s sharp rise in income inequality—driven largely by capital income gains among top earners—has coincided with intensifying spatial sorting, especially within cities. We find that segregation is most pronounced at the neighborhood level, where income gaps exceed those between regions; regional segregation has risen by 5%, while neighborhood segregation has increased by 15%, entirely due to capital income segregation. Low-income families, disproportionately foreign-born, have become increasingly concentrated in the poorest neighborhoods, and high-income households have moved away—a pattern shaped by urbanization, immigration, and a shrinking rental housing supply. These trends appear driven less by changes in labor income taxation or welfare benefits than by lower taxation of capital income, wealth, and property, propelling the growing wealth concentration found in the inequality literature.

Keywords

economic segregation inequality Sweden variance decomposition

Introduction

The World Inequality Lab has documented a global rise in income inequality, driven largely by gains among the top percentiles and especially by capital income (Piketty and Zucman, 2014). At the same time, residential segregation by income—the spatial sorting of residents along the income distribution—has produced distinct economic geographies. Within cities, segregation across neighborhoods, and between regions, polarization between “superstar cities” and rural hinterlands, have become defining features of recent decades (Rosés and Wolf, 2018; Storper, 2018; Van Ham et al., 2021), with clear links to political polarization (Adler and Ansell, 2020). Yet, the spatial consequences of rising macro inequality remain underexplored (Cottineau-Mugadza, 2025). The economics-of-inequality literature, rooted in national accounts approaches informing the World Inequality Database (Alvaredo et al., 2024; Kuznets, 1955), has focused on nationally aggregated trends, while segregation research has documented spatial patterns without fully capturing the geographical impact of inequality across all income types and groups.

Despite classic theories (Schelling, 1969; Tiebout, 1956) predicting a close connection between inequality and segregation, the two literatures remain surprisingly disconnected.¹ These theories suggest that larger income gaps push people with similar means to cluster together, due either to shared housing and amenity preferences or to a desire for socially similar neighbors, potentially then reinforced by “cycles of segregation” (Krysan and Crowder, 2017). Typically applied to ethnic groups, they can be adapted to income or status groups as well (Malmberg and Clark, 2021). Sweden provides a compelling case: in recent decades, it has experienced both rising income inequality and spatial sorting by disposable income (Figure 1), despite ongoing urbanization that Kuznets expected to eventually lower inequality in the long run (Kuznets, 1955).² Among OECD countries, Sweden has seen one of the largest increases in national income inequality (OECD, 2011; Therborn, 2020), driven especially by gains among top income groups (Korpi and Tåhlin, 2011) and the growing share of capital income (Roine and Waldenström, 2012). Simultaneously, Swedish cities record some of the highest levels of segregation of the income-poor in Europe (Haandrikman et al., 2023; Tammaru et al., 2015),³ alongside widening urban–rural disparities since the 1980s (Henning et al., 2023).

Figure 1.

Inequality and segregation over time.

This development is particularly noteworthy given Sweden’s post-Second World War establishment of a comprehensive welfare state, characterized by robust public services and high taxation. However, the economic crises of the 1970s and 1990s led to major reforms, including reduced tax rates on labor and capital income and less generous welfare provisions. Simultaneously, the abolition of government construction loans led to a decline in housing supply, while urban development increasingly focused on privately built, owner-occupied housing in profitable areas. Ongoing urbanization and successive waves of migration further intensified housing demand in cities and raised housing costs.

We seek to bridge the gap between the inequality and segregation literatures by developing a method to decompose total segregation into contributions from different spatial units, income percentiles, and income sources at different geographical scales. We apply this approach in a comprehensive empirical analysis using Sweden’s detailed full-population register data from 1991 to 2017, which includes all geo-referenced income recipients and detailed income components. This highly granular analysis allows us to go beyond existing survey-based inequality research which stops short at the regional level (Milanovic, 2005; Milanović, 2016). Beyond offering a rich descriptive account that is valuable for policymakers seeking to locate the most important income groups, sources, and geographical levels for segregation, the Swedish context is particularly well-suited for examining how growing disparities in the ability to pay for housing, the decreasing supply of new housing, and immigration contribute to the concentration of low-income residents in less desirable areas. We also investigate whether segregation is further intensified by the out-migration of higher-income households and assess the extent to which place-based equalization policies at various spatial scales can mitigate these developments.

We find that segregation is generally more pronounced at smaller spatial scales, with income differences between neighborhoods within municipalities exceeding those between regions. It is driven largely by labor income disparities and the extremes of both the income and area distributions, with the lowest- and highest-income residents concentrating in the poorest and richest areas, respectively. Over time, segregation has risen by 5% at the regional level, remained unchanged at the municipal level, and increased by 15% at the neighborhood level—entirely due to capital income segregation. Families in the bottom 20% of the income distribution have become more concentrated in the poorest 20% of areas, while those in the top 20% have moved away, making poverty segregation considerably higher than affluent segregation. Foreign-born residents have contributed significantly to these patterns.

Our results indicate that although urbanization has widened the urban–rural divide, segregation has increased even more rapidly within cities. Together with immigration and a shrinking supply of rental housing, urbanization has deepened the concentration of low-income families in poor neighborhoods and prompted affluent households to avoid these areas, a trend that policy has failed to reverse and which contradicts Kuznets’ classical prediction of late urbanization leading to lowering inequality levels. Weakened redistribution has further eroded low-income households’ ability to compete for housing. However, rather than changes in labor income taxes or welfare benefits, our interpretation is that the lower taxation of capital income, wealth, and property, propelling the growing concentration of wealth found in the inequality literature, is the primary driver of rising segregation.

The article proceeds as follows. We begin with a review of the literature on segregation, polarization, and inequality, followed by an institutional background and a discussion of segregation theories. Next, we present the data and introduce a decomposable segregation index that is independent of mechanical increases in inequality. The results are presented in three decompositions—by income group, income source, and geographical scale—after which we discuss the findings and conclude.

Literature

Economic segregation in cities has been studied through the lens of the share of unemployed or average income in neighborhoods, the poor, door-to-door incomes, etc. Income segregation has risen in most countries, as documented by studies in recent decades, for example in the U.S. (Massey et al., 2003; Watson, 2009), in Swedish cities (Hedin et al., 2012), particularly after economic crises (Andersson and Hedman, 2016) and nationwide (Malmberg and Clark, 2021; Mutgan and Mijs, 2023), in Germany (Friedrichs, 2008; Helbig and Jähnen, 2018), in France (Préteceille, 2005), and in European (Musterd et al., 2017) or even global cities (Van Ham et al., 2021). These findings about within-city economic segregation are largely echoed in the literature about inter-regional divergences, which looks at polarization in economic dimensions such as wages, skills, housing (Amaral et al., 2024; Wind and Hedman, 2018), or consumer prices (Diamond and Moretti, 2021). In the U.S., the phenomenon has been referred to as the “Great Divergence” (Moretti, 2012; Storper, 2018). In Sweden, a “double divergence” in terms of inter-regional polarization of population and GDP has taken place since the 1980s (Henning et al., 2023).⁴

At the same time, the economics of inequality literature, most prominently in the wake of Piketty’s work and the World Inequality Lab (Piketty and Zucman, 2014), has largely focused on macro inequality trends across nations using the distributional national accounts approach (Alvaredo et al., 2016). This literature has produced important insights by (1) decomposing the population into income and wealth subgroups and (2) decomposing total income and wealth into their various components. Applied to the case of Sweden, for instance, this research has revealed a U-shaped curve of long-run income inequality across the long 20th century, with a renewed surge of inequality after the 1970s, when top-1% and top-10% income shares started to rise again (Waldenström, 2021). Income inequality increases were strongly driven by the top earners and particularly their capital income component (Roine and Waldenström, 2012), whereas the more moderate wealth inequality growth was strongly affected by the housing and pension wealth of the middle class (Waldenström, 2021). The inequality literature is strong in the inclusion of the entire distribution of income groups as well as all different income sources into the analysis of inequality. Yet, the predominant interest in international comparisons and macro trends and the accompanying lack of harmonized regional data explains the blatant oversight of regional, urban, or any spatial dimension of rising inequality trends, though very recent work has started to remedy this using micro employee establishment surveys (Bauluz et al., 2023).

Hence, few studies have directly linked economic inequality to urban segregation. A recent systematic review study identified 30 empirical core articles scattered across eight disciplines, which mostly find a positive relationship between economic inequality and segregation (Cottineau-Mugadza, 2025). A recent book with evidence from 24 different cities also generally suggests a positive correlation (Van Ham et al., 2021) and within countries, richer and more unequal cities are associated with higher segregation levels (Veneri et al., 2021). Swedish studies show that segregation values are highest for top-income recipients in Stockholm (Andersson and Kährik, 2015), similar to international contexts (Tammaru et al., 2020), but have increased most among the poorest income groups in the three largest municipalities and particularly for households with children and native Swedes (Mutgan and Mijs, 2023). These studies largely document the mechanical relationship that rising income inequality imprints on urban segregation, which, as few studies show, amounts to more than half of total segregation increases (Scarpa, 2016), where Scarpa (2015) claims that sorting is a less important factor in the case of Malmö. Another study comes to similar conclusions in the U.S. for the last four decades (Manduca, 2019).

A handful of studies using primarily data from the U.S. move beyond the mechanical effects in studying income inequality as a driver of residential segregation (Chen et al., 2012; Mayer, 2001; Reardon and Bischoff, 2011; Watson, 2009). These studies find interesting correlations between changes in inequality and segregation across cities. Hu and Liang (2022) provide a deeper causal analysis showing that growing income inequality leads to stronger income sorting across neighborhoods. However, they find that fighting inequality with taxes and transfers cannot mitigate residential segregation by income. Instead, they find evidence indicating that raising the education levels of low-income residents appears effective for mitigating segregation. Overall, the study of how different types of income and income-group inequality are non-mechanically linked to segregation at different geographical scales can still be further developed. This is the gap we want to help to fill.

Institutional background: Changes in the Swedish welfare, taxation, and housing system

Following the Second World War, Sweden developed one of the world’s most comprehensive welfare states, grounded in economic growth, political stability, and social democratic principles. Between 1932 and 1976, successive governments expanded universal social insurance, healthcare, and tuition-free education, while extending public roles in childcare, eldercare, and housing. These cradle-to-grave provisions were funded through high progressive taxes on labor and capital, aiming to reduce poverty and promote equality by combining capitalist growth with strong social protection.

Swedish governance operates at the central, regional, and municipal levels. Most of the 290 municipalities—centered around a principal town or village—were formed during the 1962–1974 mergers to ensure sufficient economic capacity for public service provision, including daycare, education, and elderly care. The 21 regions, based on historical counties, have administered healthcare since the 1920s.

Government loans supported new housing construction, and municipalities were made responsible for ensuring the supply of adequate housing, leading most of them to establish municipal housing companies. Between 1965 and 1974, approximately 1 million homes were built under the Million Homes Program, often characterized by low construction standards and located in suburban blocks. Alongside supply-side measures, demand-side policies aimed to improve affordability for low-income households. Since 1968, rents have been regulated through annual negotiations between tenant and landlord associations, based on apartment size, quality, and standard. Housing allowances, introduced in the 1930s, have played a key role in offsetting housing costs. Initially designed to stimulate demand, these allowances later increasingly targeted economic support for low-income families.

The economic crises of the 1970s and 1990s triggered major welfare reforms, including decentralization and the marketization of education and healthcare through publicly funded mechanisms such as school vouchers. The tax system was overhauled: joint household taxation was abolished, top marginal labor income tax rates were reduced, capital income became separately taxed at a low flat rate, most deductions were eliminated, and the consumption tax base was expanded. Wealth and property taxes were gradually reduced or phased out. On the expenditure side, welfare benefits were curtailed, with housing allowances substantially reduced. After the 1997 reform, the number of recipients fell from about 1 million to under 200,000, primarily renters.

Major housing policy shifts included the abolition of government loans for construction, which reduced new housing starts. Municipal housing companies lost preferential treatment, and many privatized parts of their stock by converting rental units into tenant-owned co-ops. By 2017, 52% of Swedes lived in privately owned detached or semi-detached houses, 16% in co-ops, and 32% in rentals.

Urbanization continued throughout the 20th century and into the present. In response to regional fiscal disparities, Sweden replaced targeted grants (e.g. for maintaining physical infrastructure) with a general equalization system to compensate for differences in tax bases and service costs due to demographic and geographical factors. These transfers are projected to reach 226 billion SEK in 2025, or about 12% of total public sector spending.

Sweden also experienced several waves of refugee immigration, notably from the former Yugoslavia (1994), Iraq (2006), and Syria (2016). Most immigrants settled in urban areas, contributing to rising housing demand alongside continued urbanization and constrained supply. Housing prices rose sharply in larger cities, while the rental market faced excess demand and long waiting lists. Residential segregation increased, as low-income and immigrant households clustered in peripheral neighborhoods dominated by Million Homes-era rental housing. Since the 1990s, the state has implemented place-based initiatives to support such distressed areas.

Theories of segregation

If all individuals had identical incomes, income-based residential segregation would not arise. However, income inequality does not automatically lead to spatial segregation, as households with different incomes can reside in mixed neighborhoods. As Tiebout (1956) demonstrated, differences in preferences and ability to pay for housing quality and local public goods—such as schools or recreational space—can lead to systematic residential sorting. In this framework, disposable income and savings are central to housing choices, implying that redistribution policies shape spatial outcomes.

Schelling (1969) offered a complementary perspective, emphasizing that residential sorting can be driven by social preferences. Highly educated individuals may prefer to live among like-minded peers or place greater importance on neighborhood status. They may also avoid areas with more mixed populations. Since they typically earn more, spatial sorting can occur by gross income even if redistribution equalizes disposable incomes. Both models highlight how individual preferences across the income distribution interact with housing markets. A complementary view, the “consumer city” (Glaeser et al., 2001), sees sorting as driven not only by income-based preferences but also by the amenity endowment attached to cities or different neighborhoods.

While these theories put an emphasis on demand-driven segregation by sorting, others have rather emphasized the availability of (high-skilled) jobs (Storper, 2018), particularly when explaining the rise of the U.S. Sunbelt and the increasing divergence of superstar-cities from other places (Le Galès and Pierson, 2019; Moretti, 2012). The supply of career-advancing jobs can produce “escalator regions,” attracting young talents interested in climbing up the job ladder more quickly and driving the segregation of skills and salaries (Moretti, 2014), and creating a mismatch between the location of jobs and job seekers (Wilson, 1987). Trends toward “superstar cities” may be facilitated by supply-side inelasticities on their housing markets (Gyourko et al., 2013) and further pushed by increasing corporate concentration, where top-company income inequalities drive total income inequality most (Autor et al., 2020).

Once income-segregated neighborhoods are in place, they may themselves have neighborhood effects and feed back into potentially “vicious” circles of segregation (Tammaru et al., 2021) or segregation cycles (Krysan and Crowder, 2017), where disadvantages in one life domain can spill over into others and become concentrated disadvantages (Wilson, 1987). Negative feedback loops from differential education investment can be particularly relevant (Bischoff and Owens, 2019) and segregation can become intergenerationally transmitted (Chetty and Hendren, 2018). Crime can be part of a negative feedback loop (Hipp, 2011; Kang, 2016).

While market structures themselves can have effects on the income distribution both generally and spatially, the tax and transfer policy system can have indirect spatial effects. Given this theoretical link between income inequality and spatial sorting, Sweden’s shift away from progressive redistribution since the 1990s—via less progressive labor income taxes, lower capital taxes, and reduced transfers—may have contributed to increased income segregation. Such segregation would increase average income differences between areas to a greater extent than the corresponding increase in average income differences between individuals. However, the magnitude and importance of different mechanisms remain empirical questions. Rising income inequality may mechanically widen area-level income gaps, but relocation decisions and neighborhood preferences ultimately determine whether segregation intensifies. Liberalization also expanded opportunities for private firms, likely amplifying labor income vis-à-vis capital and business income inequality and wealth concentration—raising the question of how different income sources contribute to spatial segregation.

Beyond taxation, there are deliberately place-based policies, which further influence the spatial allocation of infrastructure, services, amenities, and housing. On the housing side (Beaubrun-Diant and Maury, 2022), reduced rental supply and the expansion of owner-occupied housing have made housing affordability and access to attractive areas increasingly dependent on income, wealth, and savings. Low-income and immigrant households are increasingly concentrated in the remaining rental stock, raising the risk of reaching tipping points where others avoid or exit such neighborhoods. Measuring spatial sorting by income and income source at the neighborhood level is therefore critical for evaluating past place-based policies and informing future interventions to counteract segregation.

Data

We utilize annual data sourced from the GeoSweden database spanning from 1991 to 2017. This dataset encompasses the entire Swedish population, which stood at approximately 10 million inhabitants in 2017. It comprises income-related variables obtained from tax authorities. We link individual observations over time using social security numbers. Furthermore, these observations are linked to supplementary administrative records, which encompass demographic background information. Notably, GeoSweden offers a distinctive feature by geographically associating individuals with their actual residential properties through registered addresses. Consequently, researchers have leveraged the database to investigate neighborhood influences, as well as to analyze residential relocation patterns.

Our examination of income inequality is conducted at the regional, municipal, and neighborhood levels, as exemplified in Figure 2, which displays the average income levels for the 21 Swedish regions in the upper panel, for the 26 municipalities in the Stockholm region in the middle panel, and for the 544 neighborhoods in Stockholm municipality in the lower panel. Each panel’s color code shows the dispersion of income levels at the different geographical levels, which clearly reveals Stockholm to be the richest region and richest municipality, with between-neighborhood variation appearing as the highest.

Figure 2.

Total annual average disposable income (2017 price level) at different geographical levels in Swedish Krona, 2017; produced by authors.

Like in prior studies on segregation, we focus on family income. In our data, a family consists of a married couple with or without children living in the same property, an unmarried cohabiting couple with joint children living in the same property, or a single household. Thus, unmarried cohabiting couples without children are not registered as one family. Our sample consists of individuals above the age of 20 years, and we assign to each individual the mean family income among the adult family members.

We use disposable income, which is defined as the sum of wage, capital, and business incomes plus transfers. Wage income is reported by the employer and includes fringe benefits. Capital income includes interest, payments on savings, stock dividends and incomes derived from other financial assets, rental income from subletting private homes, lottery prizes, and various forms of realized capital gains with losses deductible by 70% normally. Business income covers incomes from active business activities not reported as wage or capital incomes. Transfers include welfare benefits minus taxes, with the housing and child allowances being the most important welfare components. These definitions are used by the Swedish tax authorities. We add the minimum recorded income of 100 SEK to our disposable income measure to enable working with the logarithm of income. Since most families without gross income get welfare benefits, the share of residents with no income is small (approximately 1% most years).

Method

For the purposes of decomposing income inequality of different income components and for different income groups (e.g. percentiles) across different districts in both urban and rural Sweden, we require a statistical measure that meets various criteria at once. It should (1) be able to capture dispersion, (2) be computable at different geographical scales, (3) be separable between spatial units, (4) be decomposable into different income components, (5) be computable for multiple income groups, and (6) be immune to purely mechanic inequality increases. Most segregation measures, for example the dissimilarity index, do not fulfill all these six criteria. However, we note that several segregation measures, for example the neighborhood sorting index, are based on variances that are fairly straightforward to decompose along several dimensions, especially between geographical units. We will develop and apply a variance decomposition method fulfilling all the listed criteria.

Let total income $Y_{i}$ of individual $i$ consist of labor income $Y_{i}^{l}$ , capital income $Y_{i}^{c}$ , business income $Y_{i}^{b}$ , and transfers minus taxes $Y_{i}^{t}$ . To limit the influence of outliers in the top-income end, we will work with the logarithm of total income:

y_{i} = \ln (Y_{i}), Y_{i} = Y_{i}^{l} + Y_{i}^{c} + Y_{i}^{b} + Y_{i}^{t} .

(1)

A possible measure of inequality is the total variance of log total income in Sweden, $\sum_{i} \frac{1}{I} {(y_{i} - \bar{y})}^{2}$ , where $\bar{y} = \sum_{i} y_{i} / \bar{I}$ is the national mean log income and $\bar{I}$ is the national population.

Sweden has $R = 21$ counties or regions, $M = 290$ municipalities, and $N = 3363$ RegSO-areas, our measures of neighborhoods, and we will use subscript $r$ , $m$ , and $n$ to denote individual areas. By the law of total variance, we can decompose total variance into a between-area part and a remainder within-area part. As an example, $\sum_{i} \frac{1}{\bar{I}} {(y_{i} - \bar{y})}^{2} = \sum_{r} \frac{I_{r}}{I} {(y_{r} - \bar{y})}^{2} + \sum_{i} \frac{I_{i}}{\bar{I}} {(y_{i} - y_{r})}^{2}$ , where for region $r$ , $y_{r}$ is the mean log income and $I_{r}$ denotes its population. Applying the decomposition formula iteratively on the within-area part yields the multi-level spatial decomposition:

\begin{matrix} \sum_{i} \frac{1}{I} {(y_{i} - \bar{y})}^{2} = \sum_{r} \frac{I_{r}}{\bar{I}} {(y_{r} - \bar{y})}^{2} + \sum_{m} \frac{I_{m}}{\bar{I}} {(y_{m} - y_{r})}^{2} \\ + \sum_{n} \frac{I_{n}}{\bar{I}} {(y_{n} - y_{m})}^{2} + \sum_{i} \frac{1}{\bar{I}} {(y_{i} - y_{n})}^{2} . \end{matrix}

(2)

The between-variances measure spatial inequality between geographical units at the same level, that is, segregation between regions, municipalities, and neighborhoods, respectively. We will work with the standard deviations, that is, the square root of variances, of each of these components. To avoid a widening of the income distribution mechanically raising the income differences between spatial units even when nobody relocates, Jargowsky (1996) constructed the widely used neighborhood sorting index (nsi) as the between-neighborhood standard deviation divided by the total within-municipal standard deviation. We will adopt this way of accounting for general changes in the income distribution and define sorting indices according to the following:

\begin{matrix} R S I = \frac{\sqrt{\sum_{r} \frac{I_{r}}{\bar{I}} {(y_{r} - \bar{y})}^{2}}}{\sqrt{\sum_{i} \frac{1}{\bar{I}} {(y_{i} - \bar{y})}^{2}}} * 100, \\ M S I = \sum_{r} \frac{I_{r}}{\bar{I}} {m s i}_{r}, {m s i}_{r} = \frac{\sqrt{\sum_{m_{r}} \frac{I_{m}}{I_{r}} {(y_{m} - y_{r})}^{2}}}{\sqrt{\sum_{i_{r}} \frac{1}{I_{r}} {(y_{i} - y_{r})}^{2}}} * 100, \\ N S I = \sum_{m} \frac{I_{m}}{I_{r}} {n s i}_{m}, {n s i}_{m} = \frac{\sqrt{\sum_{n_{m}} \frac{I_{n}}{I_{m}} {(y_{n} - y_{m})}^{2}}}{\sqrt{\sum_{i_{m}} \frac{1}{I_{m}} {(y_{i} - y_{m})}^{2}}} * 100, \end{matrix}

(3)

${m s i}_{r}$ is the municipal sorting index for region $r$ , and $M S I$ quantifies the country-level importance of municipal sorting within the Swedish regions. For municipalities, we analogously define ${n s i}_{r}$ and $N S I$ . A comparison of $R S I$ , $M S I$ , and $N S I$ discloses the scale, at which spatial sorting is the strongest.

Moving on to separating the contribution from a spatial unit, we note that the between-variance consists of population-weighted averages of each unit’s mean per-capita square deviation; for example, ${(y_{n} - y_{m})}^{2}$ are the building stones of ${n s i}_{m}$ . We will take advantage of this additively separable property and construct each neighborhood’s per capita contribution to segregation according to the following:

{n s i}_{m}^{n} = \frac{{(y_{n} - y_{m})}^{2}}{\sum_{n_{m}} \frac{I_{n}}{I_{m}} {(y_{n} - y_{m})}^{2}} {n s i}_{m} .

(4)

Equation (1) enables further additive decomposition into the contribution of each income source $s = l, c, b$ by its share of the total income deviation in the neighborhood:

{n s i}_{m}^{n s} = \frac{Y_{n}^{s} - Y_{m}^{s}}{\sum_{s} Y_{n}^{s} - Y_{m}^{s}} {n s i}_{m}^{n} .

(5)

One could also ask the extent to which households from different parts of the income distribution contribute to segregation. We decompose the between-neighborhood variance into contributions from different income quintiles. Let $D^{q} (y_{i})$ be indicator variables of whether household $i$ belongs to the national income quintile $q = 1, 2, 3, 4, 5$ . We start from the fact that the non-squared deviation is additively separable in quintile contributions: $y_{n} - y_{m} = \sum_{q} {\tilde{w}}_{n}^{q}$ with ${\tilde{w}}_{n}^{q} = d_{n}^{q} (y_{n}^{q} - y_{m})$ , where $d_{n}^{q} = \sum_{i_{n}} \frac{1}{I_{n}} D^{q} (y_{i_{n}})$ is the share of residents in neighborhood $n$ from quintile $q$ and $y_{n}^{q} = \sum_{i_{n}} \frac{1}{I_{n}^{q}} D^{q} (y_{i_{n}}) y_{i_{n}}$ is the mean log income of these residents. The term $d_{n}^{q}$ measures how concentrated residents from quintile $q$ are in neighborhood $n$ , and $y_{n}^{q} - y_{m}$ measures their per-capita contribution to the income deviation. Segregation arises when residents from a quintile are disproportionally represented, that is, over- or under-concentrated in a neighborhood. When the residents are evenly sorted across neighborhoods with the same population share in every neighborhood $d_{n}^{q} = d_{m}^{q} = \sum_{i} \frac{1}{I_{m}} D^{q} (y_{i})$ , $y_{n}^{q} = y_{m}^{q} = \sum_{i} \frac{1}{I_{m}^{q}} D^{q} (y_{i}) y_{i}$ , and ${\tilde{w}}_{n}^{q} = \tilde{w_{m}^{q}} = d_{m}^{q} (y_{m}^{q} - y_{m})$ which serve as an even-sorting benchmark. We decompose the income deviation into quintile components each measuring how the over- (under-) concentration of residents from a quintile raises (lowers) the neighborhood income deviation (and its square) relative to the benchmark and this yields the following:

\begin{matrix} {n s i}_{m}^{n q} = \frac{w_{n}^{q}}{\sum_{q} w_{n}^{q}} {n s i}_{m}^{n}, w_{n}^{q} = {\tilde{w}}_{n}^{q} - {\tilde{w}}_{m}^{q} \\ = d_{n}^{q} (y_{n}^{q} - y_{m}) - d_{m}^{q} (y_{m}^{q} - y_{m}) . \end{matrix}

(6)

Since $\sum_{q} {\tilde{w}}_{m}^{q} = 0$ , our weights are intuitive as they sum to the income deviation: $\sum_{q} w_{n}^{q} = y_{n} - y_{m}$ .

Several common segregation measures are based on $d_{n}^{q}$ . For example, we could quantify the segregation between a quantile and all other quantiles across neighborhoods in a municipality using the dissimilarity index ${d i}_{m}^{q}$ . Rearranging terms gives:

{d i}_{m}^{q} = \frac{1}{2 d_{m}^{q} (1 - d_{m}^{q})} \sum d \underset{n}{\overset{q}{i}}, {d i}_{n}^{q} = | d_{n}^{q} - d_{m}^{q} | .

(7)

The generalized dissimilarity index (Reardon and Firebaugh, 2002), a multi-group version of the dissimilarity index, also consists of weighted averages of $| d_{n}^{q} - d_{m}^{q} |$ .⁵

It is straightforward to aggregate the components in equations (5)–(7) at the national level according to the following:

{N S I}^{s} = \sum_{n} \frac{I_{n}}{I} {n s i}_{n}^{s}, {N S I}^{q} = \sum_{n} \frac{I_{n}}{I} {n s i}_{n}^{q} .

(8)

Our decomposition of the neighborhood sorting index can be done at other geographical levels. We will provide results for analogous decompositions of $M S I$ and $R S I$ .

Results

The findings of our variance decomposition analysis of sorting indices are reported in Table 1, which offers a synopsis of the sorting index by region, municipality, and neighborhood levels over time and its main contributing components. We provide decompositions by (sub)areas in columns (2) and (3), income sources in columns (4)–(6), and income groups in columns (7) and (8). The decompositions are made at the regional level in panel A, municipal level in panel B, and neighborhood level in panel C, and we provide results for different years across rows. The first line, for instance, reads: the regional sorting index is at 5.7 in 1991, of which the 20% poorest regions contributed 1.5, labor income contributed 7.0, and the poorest 20% of residents contributed −0.6.

Table 1.

Decomposition of segregation by income groups, sources, and areas.

		I. Areas		II. Sources			III. Groups
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Year	Index	Q1-areas	Q5-areas	Labor	Capital	Transfer	Q1-ind	Q5-ind
	RSI	A. Regions (some RSI components)
1991	5.7	1.5	3.6	7.0	0.5	−1.9	−0.6	5.9
1996	3.9	1.4	2.0	2.0	1.1	−0.5	−1.2	4.7
2001	5.0	1.8	2.7	6.8	0.8	−3.1	−0.7	5.6
2007	6.2	3.9	1.9	9.6	0.1	−4.2	3.1	2.9
2012	6.5	4.2	2.0	7.0	1.6	−2.4	3.3	3.0
2017	5.9	3.2	2.2	5.6	2.1	−1.9	1.7	4.0
Period diff	0.3	1.7	−1.4	−1.4	1.6	−0.1	2.3	−1.9
% change	4.9	30.7	−25.3	−24.7	27.5	−1.3	41.0	−34.3
	MSI	B. Municipalities (some MSI components)
1991	8.9	3.4	4.5	16.2	−2.4	−4.4	3.8	4.1
1996	8.2	3.0	4.3	17.2	−4.0	−5.7	3.7	3.6
2001	8.3	3.1	4.1	11.9	−1.3	−3.8	4.2	3.3
2007	9.0	4.0	4.2	10.6	−2.1	-1.0	5.2	3.0
2012	9.2	4.1	4.2	7.9	−1.4	1.3	5.6	3.0
2017	9.0	4.0	4.2	7.7	0.6	−0.8	5.0	3.4
Period diff	0.1	0.7	−0.4	−8.5	3.0	3.6	1.1	−0.7
% change	0.7	7.4	−4.1	−95.3	34.2	41.1	12.9	−7.9
	NSI	C. Neighborhoods (some NSI components)
1991	15.7	7.0	6.8	29.9	−7.9	−4.8	8.1	5.3
1996	15.7	7.7	6.3	39.6	−9.8	−12.9	8.2	5.3
2001	15.8	8.3	5.8	29.2	−6.2	−5.0	8.3	5.3
2007	17.4	9.9	5.6	28.6	−3.9	−10.1	9.7	5.5
2012	18.4	10.8	5.6	21.9	−1.3	−4.5	10.7	5.5
2017	18.0	10.5	5.6	22.1	0.8	−6.9	10.0	5.9
Period diff	2.2	3.4	−1.2	−7.8	8.7	−2.1	1.9	0.6
% change	14.1	21.9	−7.9	−49.8	54.9	−13.5	12.0	3.7

See the previous methods section for the decomposition formulas. % change refers to period diff divided by the RSI, MSI, or NSI value in 1991, that is, the same denominator across columns within each panel.

In Figure 3, we visualize the contribution of different subareas and income groups to total spatial inequality at different geographical levels in 2017 and the change in contributions from 1991 to 2017. In each figure, we draw the contribution of each subarea (thick solid lines), from the one with the lowest to the highest income (population-weighting the length of the line). For instance, the Stockholm region is the richest region in Sweden, and the right end of the thick solid line in Panel A shows that it has a high per-capita contribution to the regional segregation index (RSI), but not as high as the poorest regions in the left end. In Figure 4, we instead visualize the contributions of each income quintile of the population.

Figure 3.

Segregation and its change by sources and areas, produced by authors.

Figure 4.

Segregation and its change by quintiles and areas.

Starting with the spatial decompositions, column 1 of Table 1 shows that segregation is higher at smaller geographical scales (regional < municipal < neighborhood), with RSI = 5.9, MSI = 9.0, and NSI = 18.0 in 2017. In Figure 3 panels A–C, this is reflected in a stronger U-shaped pattern for total segregation (thick lines) at a lower spatial level. The strong U-shape (rather than a more moderate V-shape) indicates that segregation is almost fully driven by the poorest and richest areas. This pattern holds over time (Table 1, columns 2 and 3). Bottom-end segregation has been reinforced over time, with increases of 30.7%, 7.4%, and 21.9% at the regional, municipal, and neighborhood levels, respectively (column 2), which has led to RSI rising by 4.9%, MSI by 0.7%, and NSI by 14.1% (column 1).

In our income-source decompositions (Table 1 columns 4–6 and Figure 3), we see that labor income is the largest contributor to total income segregation. However, its contribution has dropped over time by between 24.7% and 95.3% across the different geographical scales (column 4), although the contribution increased in the poorest regions and neighborhoods (Figure 3 panels D and F labor graphs). For most years, capital income increases total segregation at the regional level but counteracts it at the municipal and neighborhood levels. Over time, increasing capital income segregation has driven the total segregation increase, most notably contributing 8.7 (column 5) to the rise of total neighborhood segregation (of 2.2 in column 1), or a 54.9% increase in neighborhood capital segregation, with contribution both due to capital flight from poor areas and concentration in rich areas (Figure 3 panel F capital graph). The (tax and) transfer system has spatially equalizing effects that are strong at the regional and neighborhood levels. Column 6 shows that in 2017, transfers moderate RSI by −1.9 (total RSI is 5.9) and NSI by −6.9 (total NSI is 18.0). Between 1991 and 2017, the regional redistributive power of transfers has remained fairly constant, with some improvements for the poorest regions (Figure 3 panel D transfer graphs). It has worsened at the municipal level (changed MSI by 3.6) and improved at the neighborhood level (changed NSI by −2.1). Business income makes generally low contributions to both segregation levels and changes when compared to capital or labor income.

From the income-group decompositions (Table 1 columns 7 and 8 and Figure 4), we find that the bottom and top 20% of households make up most of the segregation. In 2017, they contribute 5.7 (1.7+4.0), 8.4 (5.0+3.4), and 15.9 (10.0+5.9) to RSI, MSI, and NSI, respectively; which amounts to 80%–95% of total segregation (total RSI = 5.9, MSI = 9.0, and NSI = 18.0). Whereas affluence segregation of top-20% residents is more substantial across regions, poverty segregation of bottom-20% residents is worse across neighborhoods. Figure 4 panels A–C reveal that both components add to segregation in poor and rich areas. While the clustering of low-income residents is more important in poor areas and the clustering of high-income residents is more important in rich areas, the low concentration of high-income residents in poor areas and of low-income residents in rich areas is also substantial at the municipal and neighborhood levels. Over our sample period, poverty segregation has deteriorated across geographical scales by 41% at the regional level, and 12%–13% at the lower scales, driving the total segregation increase (column 7). Although affluence segregation has not changed significantly overall, the dilution of high-income residents in poor neighborhoods contributes to their declining income levels (Figure 4 panel F Q5-graph).

In the Appendix (Figure A1), we provide an alternative way to quantify the segregation of different income groups and how this changed over time using the commonly used dissimilarity index. The main conclusion from our decomposition analysis remains: the extreme percentiles have concentrated in space, and this pattern has been reinforced over time. However, affluence segregation now plays a greater role.

In Figure 5, we plot segregation and its change without foreign-born residents. Both segregation levels and changes due to low incomes in poor areas are less pronounced for people born in Sweden. The sorting pattern of immigrants thus accounts for significant parts of the declining incomes in poor areas, although the concentration of low-income natives accounts for most of the rising segregation of poor neighborhoods.

Figure 5.

Segregation and its change without immigrants, produced by authors.

Concluding discussion

We have bridged the gap between inequality and segregation research by introducing a method to decompose income segregation across spatial units, income groups, and income sources. We applied it to detailed full-population register data for Sweden from 1991 to 2017. Sweden’s sharp rise in income inequality—driven largely by capital income gains among top earners (Waldenström, 2021)—has coincided with intensifying spatial sorting, especially within cities. Our empirical findings show that segregation is more pronounced at smaller spatial scales, with income differences between neighborhoods within municipalities exceeding those between regions. It is driven largely by labor income disparities and the extremes of both the income and area distributions, with the lowest- and highest-income residents concentrating in the poorest and richest areas, respectively. Over time, segregation has risen by 5% at the regional level, remained unchanged at the municipal level, and increased by 15% at the neighborhood level—entirely due to capital income segregation. This new finding may also spark further research into the difficult territory of wealth segregation by measuring the flows from wealth (capital gains, interest, dividends) rather than the stock (Suss et al., 2024). Surprisingly, direct-business income—which has decreased in general importance but is generally unequally distributed—hardly makes any segregation contributions. Families in the bottom 20% of the income distribution have become more concentrated in the poorest 20% of areas, while those in the top 20% have moved away. Foreign-born residents have contributed significantly to these patterns. While we decomposed income by groups and sources, future studies could combine these with different incomes by sectors, particularly business-service (Wessel, 2022) or financial-sector incomes (Godechot et al., 2024), as potential drivers in Sweden.

Our empirical findings show that urbanization has not only driven population outflows from poorer to wealthier regions but also resulted in a concentration of remaining low-income residents in disadvantaged regions, a finding in line with recent work in the U.S. (Massey, 2020), North-Western Europe (Andersson et al., 2018), and OECD countries (OECD, 2018). Contrary to studies in most of these countries, we find affluence segregation to be considerably lower than poverty segregation in neighborhoods and municipalities (though not regions). While our dissimilarity-index findings (Appendix Figure A1) confirm prior findings about the rich tending to cluster in few enclaves (Haandrikman et al., 2023; Reardon and Bischoff, 2011), the sorting indices also show that these wealthy enclaves still contain a certain high mix of lower-income residents, balancing the average income better than in their lower-income-area counterparts. However, government transfers have helped moderate the decline in disposable incomes in the poorest regions, smoothening not only the income distribution but also its spatial implications. Importantly, regional polarization and its changes are relatively minor compared to the much greater income segregation observed within municipalities, where the poorest neighborhoods are increasingly falling behind. In the Swedish context, the challenges associated with growing cities and declining neighborhoods are therefore more pressing than those related to the urban–rural divide.

Alongside urbanization, waves of immigration have increased housing demand in urban areas. This trend has coincided with reduced government support for housing construction and a more market-driven supply of privately owned homes in desirable locations—an unfortunate combination. As theory predicts, we observe a rise in segregation. Our key finding is the concentration of the lowest-income families in the poorest neighborhoods, to some extent fueled by the spatial clustering of immigrants, but also native poor families, a finding resonating with prior studies (Malmberg and Clark, 2021). At the same time, affluent families are increasingly avoiding these areas, suggesting tipping points where multiple downward spirals reinforce one another in cycles of segregation—something that place-based policies in Sweden have so far failed to mitigate. While we do not probe into potential channels driving these trends, prior research suggests that the housing market may be a key one (Dwyer, 2007; Gordon and Bruch, 2020).

Rising segregation in Sweden also appears to be linked to weakened redistribution, which has reduced low-income families’ ability to afford housing. However, labor income and transfers are not the primary drivers, despite lower income taxes and less generous welfare benefits. Our results instead underscore the pivotal role of capital income. Individuals without capital income have been concentrated in poor areas, while those with substantial capital income tend to avoid them. Our interpretation is that this pattern reflects a growing reliance on accumulated wealth to secure desirable housing, reinforced by intensified competition and a shrinking supply of rental homes in cities. Consequently, lower capital, wealth, and property taxation, propelling the concurrent concentration of wealth found in the inequality literature, have likely played a significant role in driving segregation. Future studies could try to further assess the spatial implications of different major tax and welfare reforms which have occurred in Sweden.

Footnotes

Appendix

Our income-group decomposition measures how the concentration of rich or poor residents contributes to income differences across areas. The concentration of the extremely rich or poor contributes more to such differences. One might instead be interested in how unevenly different income groups are distributed across areas, ignoring how much that contributes to area income differences. The dissimilarity index is a popular measure of the uneven distribution of a group of individuals in space. In Figure A1, we report the dissimilarity index for different income percentiles across subareas at different geographical scales and how the indices changed during our sample period. The main conclusion from our main decomposition analysis remains that the extreme percentiles are more concentrated in space and that this pattern has been reinforced over time. The figures also show that the most extreme top percentiles are much more concentrated in space than the rest of the population.

ORCID iDs

Matz Dahlberg

Sebastian Kohl

Che-Yuan Liang

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We would like to thank the Jan Wallander and Tom Hedelius Foundation (P23-0083), the Swedish Research Council (VR, 2023-01296), and the Swedish Research Council for Health, Working Life, and Welfare (FORTE, 2023-00527) for their financial support.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Notes

References

Adler

Ansell

(2020) Housing and populism. West European Politics 43(2): 344–365.

Alvaredo

Atkinson

Piketty

, et al. (2016) WID – The World Wealth and Income Database. Springer.

Alvaredo

Atkinson

Piketty

, et al. (2024) World inequality database. Available at: http://wid.world/data (accessed 1 January 2025).

Amaral

Dohmen

Kohl

, et al. (2024) Interest rates and the spatial polarization of housing markets. American Economic Review: Insights 6(1): 89–104.

Andersson

Hedman

(2016) Economic decline and residential segregation: A Swedish study with focus on Malmö. Urban Geography 37(5): 748–768.

Andersson

Kährik

(2015) Widening Gaps: Segregation Dynamics During Two Decades of Economic and Institutional Change in Stockholm. Routledge, pp.110–131.

Andersson

Janssen

Ham

, et al. (2023) Contextual poverty and obtained educational level and income in Sweden and the Netherlands: A multi-scale and longitudinal study. Urban Studies 60(5): 885–903.

Andersson

Malmberg

Costa

, et al. (2018) A comparative study of segregation patterns in Belgium, Denmark, The Netherlands and Sweden: Neighbourhood concentration and representation of non-European migrants. European Journal of Population 34(2): 251–275.

Autor

Dorn

Katz

, et al. (2020) The fall of the labor share and the rise of superstar firms. Quarterly Journal of Economics 135(2): 645–709.

10.

Barro

(2012) Convergence and modernization revisited. Report, National Bureau of Economic Research.

11.

Bauluz

Bukowski

Fransham

, et al. (2023) Spatial wage inequality in north America and western Europe: Changes between and within local labour markets 1975–2019. Technical report, Kiel Working Paper.

12.

Beaubrun-Diant

Maury

T-P

(2022) On the impact of public housing on income segregation in France. Demography 59(2): 685–706.

13.

Bischoff

Owens

(2019) The segregation of opportunity: Social and financial resources in the educational contexts of lower- and higher-income children, 1990–2014. Demography 56(5): 1635–1664.

14.

Chen

W-H

Myles

Picot

(2012) Why have poorer neighbourhoods stagnated economically while the richer have flourished? Neighbourhood income inequality in Canadian cities. Urban Studies 49(4): 877–896.

15.

Chetty

Hendren

(2018) The impacts of neighborhoods on intergenerational mobility ii: County-level estimates. Quarterly Journal of Economics 133(3): 1163–1228.

16.

Cottineau-Mugadza

(2025) Economic inequality and economic segregation: A systematic review of causal pathways. Social Forces 2025.

17.

Diamond

Moretti

(2021) Where is standard of living the highest? Local prices and the geography of consumption. Technical report, National Bureau of Economic Research.

18.

Dwyer

(2007) Expanding homes and increasing inequalities: U.S. housing development and the residential segregation of the affluent. Social Problems 54(1): 23–46.

19.

Friedrichs

(2008) Gespaltene Städte? Soziale und Ethnische Segregation in Deutschen Großstädten. VS Verlag für Sozialwissenschaften.

20.

Gennaioli

La Porta

Lopez De Silanes

, et al. (2014) Growth in regions. Journal of Economic Growth 19(3): 259–309.

21.

Glaeser

Kolko

Saiz

(2001) Consumer city. Journal of Economic Geography 1(1): 27–50.

22.

Godechot

Tomaskovic-Devey

Boza

, et al. (2024) The great separation: Top earner segregation at work in advanced capitalist economies. American Journal of Sociology 130(2): 439–495.

23.

Gordon

Bruch

(2020) Home inequity: Race, wealth, and housing in St. Louis since 1940. Housing Studies 35(7): 1285–1308.

24.

Gyourko

Mayer

Sinai

(2013) Superstar cities. American Economic Journal Economic Policy 5(4): 167–199.

25.

Haandrikman

Costa

Malmberg

, et al. (2023) Socio-economic segregation in European cities: A comparative study of Brussels, Copenhagen, Amsterdam, Oslo and Stockholm. Urban Geography 44(1): 1–36.

26.

Hedin

Clark

Lundholm

, et al. (2012) Neoliberalization of housing in Sweden: Gentrification, filtering, and social polarization. Annals of the Association of American Geographers 102(2): 443–463.

27.

Helbig

Jähnen

(2018) Wie brüchig ist die soziale architektur unserer städte? Trends und analysen der segregation in 74 deutschen städten. Report, WZB Discussion Paper.

28.

Henning

Westlund

Enflo

(2023) Urban–rural population changes and spatial inequalities in Sweden. Regional Science Policy & Practice 15(4): 878–893.

29.

Hipp

(2011) Spreading the wealth: The effect of the distribution of income and race/ethnicity across households and neighborhoods on city crime trajectories. Criminology 49(3): 631–665.

30.

Liang

C-Y

(2022) Does income redistribution prevent residential segregation? Journal of Economic Behavior & Organization 193: 519–542.

31.

Jargowsky

(1996) Take the money and run: Economic segregation in U.S. metropolitan areas. American Sociological Review 61(6): 984–998.

32.

Kang

(2016) Inequality and crime revisited: Effects of local inequality and economic segregation on crime. Journal of Population Economics 29(2): 593–626.

33.

Korpi

Tåhlin

(2011) Changing work–life inequality in Sweden: Globalization and other causes. In: Blossfeld

Buchholz

Hofäcker

, et al. (eds) Globalized Labour Markets and Social Inequality in Europe. Springer, pp.177–208.

34.

Krysan

Crowder

(2017) Cycle of Segregation: Social Processes and Residential Stratification. Russell Sage Foundation.

35.

Kuznets

(1955) Economic growth and income inequality. American Economic Review 45(1): 1–28.

36.

Le Galès

Pierson

(2019) “Superstar cities” & the generation of durable inequality. Daedalus 148(3): 46–72.

37.

Malmberg

Clark

WAV

(2021) Migration and neighborhood change in Sweden: The interaction of ethnic choice and income constraints. Geographical Analysis 53(2): 259–282.

38.

Manduca

(2019) The contribution of national income inequality to regional economic divergence. Social Forces 98(2): 622–648.

39.

Massey

(2020) Still the linchpin: Segregation and stratification in the USA. Race and Social Problems 12(1): 1–12.

40.

Massey

Fischer

Dickens

, et al. (2003) The geography of inequality in the United States, 1950–2000. Brookings-Wharton Papers on Urban Affairs 2003: 1–40.

41.

Mayer

(2001) How the growth in income inequality increased economic segregation. Report, JCPR Working Papers 230, Northwestern University/University of Chicago Joint Center for Poverty Research.

42.

Milanovic

(2005) Half a world: Regional inequality in five great federations. Journal of the Asia Pacific Economy 10(4): 408–445.

43.

Milanović

(2016) Global Inequality: A New Approach for the Age of Globalization. The Belknap Press of Harvard University Press.

44.

Moretti

(2012) The New Geography of Jobs. Houghton Mifflin Harcourt.

45.

Moretti

(2014) Are cities the new growth escalator? World Bank Policy Research Working Paper 6881.

46.

Musterd

Marcińczak

van Ham

, et al. (2017) Socioeconomic segregation in European capital cities: Increasing separation between poor and rich. Urban Geography 38(7): 1062–1083.

47.

Mutgan

Mijs

(2023) Income inequality and residential segregation in ‘egalitarian’ Sweden: Lessons from a least likely case. Sociological Science 10(12): 348–373.

48.

OECD (2018) Divided Cities: Understanding Intra-Urban Inequalities. OECD Publishing.

49.

OECD (2011) Divided we stand: Why inequality keeps rising. Organisation for Economic Co-operation and Development Publishing. Available at: http://www.i-red.eu/resources/publications-files/oecd-divided-we-stand.pdf (accessed 22 June 2026).

50.

Piketty

Zucman

(2014) Capital is back: Wealth-income ratios in rich countries 1700–2010. Quarterly Journal of Economics 129: 1255–1310.

51.

Préteceille

(2005) La ségrégation sociale a-t-elle augmenté? La métropole parisienne entre polarisation et mixité. Sociétés contemporaines 62: 69–93.

52.

Reardon

Bischoff

(2011) Income inequality and income segregation. American Journal of Sociology 116(4): 1092–1153.

53.

Reardon

Firebaugh

(2002) 2. Measures of multigroup segregation. Sociological Methodology 32(1): 33–67.

54.

Roine

Waldenström

(2012) On the role of capital gains in Swedish income inequality. Review of Income and Wealth 58(3): 569–587.

55.

Rosés

Wolf

(2018) Regional economic development in Europe, 1900–2010: A description of the patterns. In: Ramón Rosés

Wolf

(eds) The Economic Development of Europe’s Regions. Routledge, pp.3–41.

56.

Scarpa

(2015) The impact of income inequality on economic residential segregation: The case of Malmö, 1991–2010. Urban Studies 52(5): 906–922.

57.

Scarpa

(2016) Looking beyond the neighbourhood: Income inequality and residential segregation in Swedish metropolitan areas, 1991–2010. Urban Geography 37(7): 963–984.

58.

Schelling

(1969) Models of segregation. American Economic Review 59(2): 488–493.

59.

Stand

Rising

WIK

(2011) An overview of growing income inequalities in OECD countries: Main findings. In: OECD (ed) Divided We Stand: Why Inequality Keeps Rising, pp.21–45.

60.

Storper

(2018) Separate worlds? Explaining the current wave of regional economic polarization. Spatial Transformations 136: 247–270.

61.

Suss

Kemeny

Connor

(2024) Geowealth-us: Spatial wealth inequality data for the United States, 1960–2020. Scientific Data 11(1): 253.

62.

Tammaru

Knapp

Silm

, et al. (2021) Spatial underpinnings of social inequalities: A vicious circles of segregation approach. Social Inclusion 9(2): 65–76.

63.

Tammaru

Marcińczak

Aunap

, et al. (2020) Relationship between income inequality and residential segregation of socioeconomic groups. Regional Studies 54(4): 450–461.

64.

Tammaru

Van Ham

Marcińczak

, et al. (2015) Socio-Economic Segregation in European Capital Cities: East Meets West. Routledge.

65.

Therborn

(2020) Sweden’s turn to economic inequality, 1982–2019. Structural Change and Economic Dynamics 52: 159–166.

66.

Tiebout

(1956) A pure theory of local expenditures. Journal of Political Economy 64(5): 416–424.

67.

Van Ham

Tammaru

Ubarevičienė

, et al. (2021) Urban Socio-Economic Segregation and Income Inequality: A Global Perspective. Springer Nature.

68.

Veneri

Comandon

Garcia-López

, et al. (2021) What do divided cities have in common? An international comparison of income segregation. Journal of Regional Science 61(1): 162–188.

69.

Waldenström

(2021) Wealth and history: An update. CESifo Working Paper.

70.

Watson

(2009) Inequality and the measurement of residential segregation by income in American neighborhoods. Review of Income and Wealth 55(3): 820–844.

71.

Wessel

(2022) Business services, income inequality, and income segregation in metropolitan areas: Direct and indirect links. Economic Geography 98(5): 464–486.

72.

Wilson

(1987) The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. University of Chicago Press.

73.

Wind

Hedman

(2018) The uneven distribution of capital gains in times of socio-spatial inequality: Evidence from Swedish housing pathways between 1995 and 2010. Urban Studies 55(12): 2721–2742.

Urban segregation and economic inequality in Sweden (1991–2017): A decomposition analysis

Abstract

Keywords

Introduction

Literature

Institutional background: Changes in the Swedish welfare, taxation, and housing system

Theories of segregation

Data

Method

Results

Concluding discussion

Footnotes

Appendix

ORCID iDs

Funding

Declaration of conflicting interests

Notes

References