Urbanization Favors High Wage Earners

The study investigates how the distribution of wages within cities varies with city size, extending the well-documented urban wage premium beyond averages to intra-urban inequality. Prior work shows that wages and incomes scale superlinearly with population, implying higher average earnings in larger cities. Yet inequality in resource distribution tends to rise with development, and the relationship between inequality, social trust, innovation, and sustainability is complex. The authors ask whether larger U.S. metropolitan areas exhibit different scaling behavior across the wage distribution and how this relates to inequality within cities.

Urban wage premia and agglomeration economies have been noted since Marshall and studied extensively using urban scaling frameworks, revealing superlinear scaling of wages, income, innovation, and other urban indicators with population. Inequality has been linked to urbanization and macro drivers such as globalization, technological change, and union decline. Prior analyses of within-city inequality often rely on distributional descriptors (e.g., Gini coefficients, percentile ratios) or fixed income bins fitted against population. The latter approach can be confounded by regional cost-of-living differences. Power-law scaling has been effective for aggregate urban measures, but how the intra-urban wage distribution itself scales with city size remained unclear. Related work has examined scaling of income distribution bins and occupational/industrial diversification, suggesting hierarchical emergence of complex, higher-paying activities in larger cities.

Data come from the U.S. Census Bureau’s American Community Survey (ACS) Public Use Microdata Sample (PUMS), 1-year files for 2005–2019 (~1% of the U.S. population). Individuals are geolocated to Public Use Microdata Areas (PUMAs) and mapped to Metropolitan Statistical Areas (MSAs) using the IPUMS PUMA-to-MSA crosswalk. Person weights are adjusted by the crosswalk to synthesize complete MSA populations. The sample is restricted to workforce participants who worked 50–52 weeks in the previous 12 months and at least 35 hours per week (full-time, full-year), excluding those under age 16. For each of 382 MSAs and each year, individuals are placed into wage-ranked population deciles (weighted n-tiles using the R package hutils). For each MSA m and decile d, the authors compute total wages W_md and estimate the power-law scaling of decile aggregate wages with MSA population P_m via log–log regressions of the form W_md = α P_m^β, taking β as the decile-specific urban wage premium. To interpret mechanisms, they also estimate: (1) scaling of employment by 2-digit industry with city size, E_mi = a P_m^{β′}; (2) scaling of employment by 2-digit occupation with city size, E_mo = a P_m^{β′′}; and (3) for each occupation group, scaling of total occupational wages with occupational employment within MSAs, W_mo = a E_mo^{β‴}, to assess whether within-occupation wages rise with city size. They additionally estimate scaling by educational attainment categories. Analyses are repeated annually from 2005 to 2019; results emphasized for 2019.

• Decile scaling is heterogeneous and monotonic by wage rank. In 2019, all deciles exhibit superlinear scaling, but β increases with decile: from β≈1.056 (lowest decile) to β≈1.156 (tenth decile). The aggregate across all earners is β≈1.11. - Over 2005–2019, the pattern is stable. The tenth-decile β rose slightly (1.142→1.156); β for deciles 5–9 also increased modestly, while β for deciles 1–2 decreased slightly. - Decile composition correlates with education and gender: higher deciles have larger shares with college degrees and lower shares of female workers (e.g., decile 10: 78.3% with a degree; 25.8% female; mean wage $232.9k; versus decile 1: 15.6% with a degree; 52.9% female; mean wage $15.4k). - High-scaling occupations are higher paid and concentrate in larger cities. In 2019, 2-digit occupations with high β include Legal (β=1.329; $138.7k), Computer and Mathematical (β=1.327; $95.8k), and Arts/Design/Entertainment (β=1.261; $72.2k). Many lower-wage/lower-skill occupations scale near-linearly or sublinearly (e.g., Production β=0.939; Farming/Fishing/Forestry β=0.906; Military-specific β=0.688). Occupation β correlates positively with mean wages across MSAs (R^2=0.46, excluding military). - Within-occupation wage premiums also rise with city size. Total occupational wages scale superlinearly with occupational employment for legal (β‴=1.14), sales (1.09), management (1.09), and arts (1.08), indicating higher pay for the same occupational groups in larger cities. This β‴ also correlates positively with occupation mean wages (R^2=0.60, excluding military). - Education scales superlinearly with city size, increasing with credential level: Professional degree β=1.260 ($169.8k); PhD 1.196 ($128.6k); Master’s 1.191 ($101.7k); Bachelor’s 1.150 ($82.7k). Some college/associate (β=0.984; $52.7k) and high school or less (β=0.969; $42.1k) scale sublinearly/near-linearly, indicating diminishing representation of less educated workers in larger cities. - Interpretation: Larger cities disproportionately host more (and better paid) high-wage occupations and more highly educated workers, yielding higher average wages alongside increasing intra-urban wage inequality.

The analysis shows that the urban wage premium is not uniform across the wage distribution: higher wage deciles benefit disproportionately as city size increases. This resolves the research question by demonstrating that within-city wage inequality rises with city size due to two reinforcing mechanisms: (1) superlinear concentration of complex, high-wage occupations in larger cities, and (2) higher within-occupation pay in larger cities. The strong superlinear scaling of advanced educational attainment further supports sorting and matching of high-skill workers into large MSAs. While urbanization raises average wages, it also intensifies inequality, complicating its role in promoting human well-being and sustainability. The findings align with theories of hierarchical capability accumulation and labor market matching in larger cities and suggest that policies addressing inequality must consider the structural concentration of lucrative jobs and skills in large urban areas.

Using individual-level microdata and power-law scaling by wage-ranked population deciles, the study demonstrates that wage growth with city size is superlinear across the distribution but steepest for the top deciles, implying increased intra-urban wage inequality in larger U.S. MSAs. The concentration of high-wage occupations and advanced education, along with higher within-occupation pay in large cities, underlies this pattern. Future research should: (1) incorporate part-time and partial-year workers to assess their contribution to inequality; (2) further examine the influence of regional price differences on scaling magnitudes and inequality measures; and (3) extend scaling analyses to skills to identify which skill sets concentrate in larger cities and how they relate to wages.

• Sample excludes part-time and partial-year workers, potentially understating inequality, especially in lower deciles. - Analysis period (2005–2019) is relatively short; historical evidence suggests earlier periods may display different scaling patterns and city-size gradients. - Regional price differences affect the magnitude of scaling coefficients and can alter results in alternative methodologies using fixed income bins, though patterns here appear robust. - Mapping from PUMAs to MSAs introduces approximation error. - Observational, cross-sectional scaling analyses are correlational and cannot alone establish causal mechanisms.