Is urbanization sustainable? A longitudinal study of developing nations, 1990-2015

ABSTRACT

Sustainability scholars have long asked whether urbanization fosters sustainable development. To stimulate progress on this question for the cross-national quantitative literature, we draw on theories from modernization and political economy and address two empirical issues: the lack of a comprehensive metric on sustainable development as well as a need to differentiate between the multiple dimensions of urbanization. Covering the years 1990–2015, first in models with listwise deletion (n = 88) and then using full information maximum likelihood (n = 156), we regress change in the Sustainable Development Index (SDI) and its component parts on changes in the basic percentage urban variable as well as on independent measures for country-level density, urban primacy, the size of urban agglomerations, and slum prevalence, controlling for unit fixed effects. For developing nations, results from these models indicate that the multiple dimensions of urbanization exert countervailing pressures on the social and environmental components of sustainable development. These results highlight competing claims from urban-ecological theories of modernization and political economy.

KEYWORDS:

• Urbanization
• Sustainability
• Cross-national
• Structural equation modeling

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. We focus on developing nations for two reasons: they continue not only to face substantial sustainability challenges but also to experience dramatic urban transformations, in terms of the pace of rural-urban population shifts as well as the size and quality of city life (United Nations 2018a). Also, for developing nations, we use the nations in the ‘Emerging and Developing Economies’ category of the World Economic Outlook published by the International Monetary Fund (2022).

2. This statement from Davis (2006) clarifies an argument made earlier in Planet of Slums, where he wrote, ‘The chaotic form of so many Third World cities - “urban mandelbrots” according to urban theorist Matthew Gandy – annuls much of the environmental efficiency of city life’ (129).

3. As a single variable, basic urbanization percentages do not sufficiently capture the state and extent of urbanization and city life in a country. Thus, based solely on the percentage urban variable, territories can appear similar in their level of urbanization even though they are dramatically different in terms of the extent and condition of city size and urban life (see discussion in Satterthwaite 2010). For example, Djibouti, Greenland, and the United States are roughly 78%, 87%, and 83% urban, respectively; yet, 65% of Djibouti’s urban population is living in slum conditions, and the biggest settlement in Greenland does not exceed 20,000 residents (World Bank 2022). While we do not use developed countries in our analysis, these examples clearly demonstrate the limitations of the basic urbanization variable in capturing cross-national variation in the experience of urban residents.

4. We clarify here that the specific agglomeration variable used in Zhang, Yu, and Chen (2017) is used as a sensitivity check and measured as a percentage of the total population, not as a percentage of the urban population. As a sensitivity check, the reported model does not also control for the level of basic urbanization, thereby potentially confounding the association between agglomeration size and environmental impact.

5. We make two points of clarification here. First, density and slum prevalence are significantly associated with the expected years of education, but not the mean years of education (in Model 4 of Tables 3–4), even though the slope estimates are in the same hypothesized directions. We speculate that the difference in significance levels between expected and mean years of education is an artifact of measuring the former for children and the latter only for adults 25 years and older. Second, in terms of GNI per capita (in Model 5 in Tables 3–4), we speculate that the negative slope estimate for density is an artifact of the dependent variable being a per capita measure and the density variable being measured at the country-level; in other words, economic prosperity per capita decreases with density while the cumulative total of economic prosperity would increase with density. Future research can illuminate better the mechanisms involved in these two findings.

Additional information

Notes on contributors

Matthew Thomas Clement

Matthew Thomas Clement is an associate professor in the department of sociology at Texas State University.

Nathan W. Pino

Nathan W. Pino is a professor in the department of sociology at Texas State University.