Fiscal Consequences of Concentrated Poverty in a Metropolitan Region

Abstract

Poverty concentration has a significant negative effect on the fiscal health of cities in that it increases spending on antipoverty programs and also raises the cost of providing more general public services such as police and fire protection. Spending patterns among Southern California cities over the last two decades show that poverty strongly influences local public expenditures after controlling for demographic, institutional, and fiscal characteristics of cities. Moreover, the fiscal burden of poverty has increased more rapidly for general expenditures than it has for antipoverty programs per se. Because the general service impacts of poverty are typically uncompensated under state and federal fiscal policies, poor cities experience significantly higher uncompensated costs and face growing fiscal pressures. These pressures reinforce patterns of poverty and inequality within the Southern California region and likely fuel exurbanization and environmental degradation.

Key Words:

• fiscal disparity
• urban poverty concentration
• local public expenditure
• uncompensated poverty costs
• Southern California

Acknowledgments

Financial support from the National Science Foundation Program in Geography and Regional Science is gratefully acknowledged (award 9905866). The authors would also like to thank Alejandro Alonso, George Kivork, Julie Park, Lydia Thornton, Lili Wang, and Brooke Zobrist for their research assistance; Thomas W. Lester, Joseph Persky, Haydar Kurban, and Jerry Keffer for their advice on using CFFR data; David Deis for his excellent cartographic support; and Michael Dear, Enrico Marcelli, Manuel Pastor, Chris Weare, and participants at the Southern California Studies Center's Building the Sustainable Metropolis Project for comments and suggestions. Lastly, comments from our anonymous reviewers were invaluable. All errors and omissions, however, are ours alone.

Notes

Note: Poverty categories are based on three quantiles computed separately for each period.

Note: standard errors in italic

****significant within 99% confidence intervals

***95% confidence intervals

Note: standard errors in italic

****significant within 99% confidence intervals

***95% confidence intervals

**90% confidence intervals

Note: standard errors in parentheses.

****significant within 99% confidence intervals

***95% confidence intervals

**90% confidence intervals

*80% confidence intervals

Poverty categories for both periods are based on thresholds obtained from 1980 poverty quantiles.

1. Central cities are designated and defined by the Office of Management and Budget (OMB). They consist of the one (and in some cases two) largest cities contained in Metropolitan Statistical Areas (MSA).

2. Jargowsky (2003) shows that, in the 1990s, poverty concentration, measured by the proportion or number of poor people living in extreme poverty tracts (i.e., tracts with more than 40 percent of the population below the official poverty threshold), had increased in Los Angeles while decreasing in almost every metropolitan area in the United States.

3. A number of studies argue that local expenditures are determined by a jurisdiction's “need” for services (see Yinger 1986). Although the concept of fiscal need is often difficult to define in operational terms, some indicators of environmental costs, such as poverty concentration, housing or population density, age of infrastructure, or other risk factors, can be used to reflect the impact of variations in needs on public expenditures.

4. This adjustment is based on the observation that official poverty thresholds do not fully capture poverty given the changes in consumer expenditure pattern since the 1960s (when the poverty line was first devised) and the especially high cost of living in the Southern California region. For a detailed comparison of poverty measures in Southern California, see Joassart-Marcelli (2005).

5. One of the most common measures of fiscal capacity is computed using a representative tax burden approach. Under this method, city level fiscal capacity is estimated by multiplying a city's aggregate household income by a “representative” tax rate derived from averaging tax rates across all cities in the region (Ladd and Yinger 1989). The city's “tax rate” itself is typically estimated as the ratio of city revenues to aggregate household income. However, cities in California have limited control over intergovernmental aid, sales tax, and property tax revenues (especially since the 1978 Proposition 13); hence, from a standpoint of local fiscal policy, these types of revenues are fixed. To take into account these constraints, we compute a representative rate based on the discretionary component of income (i.e., all revenues minus IGR and sales and property taxes) to calculate local discretionary capacity. Local capacity is then computed by adding actual sales and property taxes over which there is little control. Finally, total capacity is obtained by adding IGR to local capacity.

6. This measure of fiscal burden is only partially valid in Southern California where the proportion of local revenues raised from traditional property taxes has declined from 13 to 11 percent during the 1980s and 1990s (Musso 2000). In this context, cities rely on a wide range of other instruments, including fees for services such as sanitation and other taxes like utility taxes, to raise revenues. Such instruments are highly regressive, and as a consequence, the fiscal burden is more likely to be shared by all residents rather than just the nonpoor.

7. It is important to note that while state and county grants to cities are included in this analysis, direct state and county expenditures to cities are not. Unfortunately, there is no systematic way to estimate the contribution of such expenditures in California cities. However, given that state direct spending from its own revenue is less than half the amount of federal spending and is allocated according to similar formulas, it is unlikely that it would counteract existing trends and promote equalization.

8. Under the California Constitution, Charter cities have more flexibility in their governmental structure and greater home rule powers than General Law cities. For instance, the former often have district elections while the latter have a council-manager form of government with five city council members elected at large.

9. While the nontransformed versions of the model show some level of heteroscedasticity, natural logarithmic transformation of the dependent variable partially corrected for this problem as suggested by Cook-Weisberg tests. Plots of residuals against fitted values indicate similar variances across observations with slightly larger residuals for cities with low levels of spending. Correlation analysis also reveals no strong correlation between any two of the independent variables, with a maximum of 0.5, suggesting low multicollinearity. Furthermore, tolerances were sufficiently high to rule out biases introduced by multicollinearity.

10. Standardized beta coefficients are the estimated coefficients for the equation in which all variables have been standardized by subtracting their means from them and by dividing them by their own standard deviations. A high beta indicates that the associated independent variable may be important in explaining variations in the dependent variable.

11. To avoid losing a large number of observations for cities that had no poverty-related expenditure, the natural log was computed using per capita expenditure+1.

12. Recent Census data on southern California counties show that the increase in poverty of the 1980s noted above accelerated in the 1990s. While poverty in the region increased from 21 percent to 22 percent of the population in the first decade, it rose up to 27 percent in 2000, affecting 4,266,882 people. Recent studies by Jargowsky (2003) and McConville and Ong (2003) more specifically highlight the increase in poverty concentration in the Los Angeles Metropolitan Area and the Southern California region.

13. The analysis of fiscal capacity excluded school district revenues, which in California are districts independent of cities. Moreover, school finance in California is highly centralized due to the joint effects of Proposition 13 (the property tax limit) and the California Supreme Court's 1976 decision in Serrano v. Priest, which required fiscal equalization of school finance on the grounds that fiscal inequities violated the California Constitution's equal protection protections. Hence, while, in other states, school finance might substantially contribute to fiscal disparities, in California, differences in school revenues would not be expected to contribute substantially to urban fiscal disparities.

14. It should be noted that because this model is in reduced form, the empirical work may not adequately differentiate the effects of immigration in the case where immigrants are in poor communities. While there is an absence of collinearity between poverty and immigration, the indirect effects of immigration may nonetheless be carried through the poverty variable. Our thanks to an anonymous reviewer for pointing this out.

15. As in Table 2 above, the sample of cities is divided into three groups of equal size based on their level of poverty. We chose not to adopt any pre-set definitions of high poverty places (such as Jargowsky [1997]'s definition of high poverty neighborhood as those with poverty rates above 40 percent) because these tend to be arbitrary and are typically applied to smaller geographical areas (e.g., census tracts). Moreover, the ranking of cities according to their poverty rate did not reveal any clear break points leading to specific categorization.

16. This result would not necessarily hold in states where cities are responsible for the provision of education, given that there is likely an association between concentration of immigrants and concentration of school-age children.

17. Some studies have questioned the extent to which poorer individuals migrate in response to welfare differentials (Schram, Nitz, and Krueger 1998). It should be noted that most studies of migration in response to welfare differentials or fiscal differences are based on aggregate data or focus on state-to-state differences. As such, they miss finely textured movements within a metropolitan area. The incentives for intrametropolitan movement, and corresponding disincentives for localities within a metropolitan area to provide substantial antipoverty services, are clearly much greater as the costs of moving within the area are relatively low. Moreover, as Gramlich and Laren (1984) point out, the actual migration behavior is less important than the beliefs of policy makers about the extent to which such migration will occur.

18. In California, the revenue capacity of cities has been limited since 1978 when Proposition 13 effectively capped property taxes. The state has also used other mechanisms such as the creation of special funds (e.g., the Educational Revenue Augmentation Fund) to appropriate funds originally tagged for local governments.