From Parity to Polarization: Iowa Income Inequality

Transcription

1 From Parity to Polarization: Iowa Income Inequality The past four decades have been one of tremendous economic change in the United States and the American Midwest. The relative postwar prosperity of the 1960s and 1970s gave way to recessions and farm crises in the 1980s, which hit the Midwest particularly hard. The economic boom of the 1990s increased incomes and lowered inequality and poverty across much of the nation, driven in part by information technology. The pendulum swung back during the decade of the 2000s as the dot-com and housing bubbles precipitated recessions that wiped away the gains of the 1990s. However, there is evidence suggesting that past economic expansions and recessions have largely affected the distribution of income rather than the aggregate amount of income (DeNavas-Walt et al. 2011). This has led some social scientists to theorize that the United States is experiencing a Great U-Turn of inequality (Moller et al. 2009; Nielsen and Alderson 2001; Peters 2013; Piketty 2014). This body of literature concludes that although certain regions of the Midwest and United States are very prosperous, they also have some of the highest levels of relative deprivation. Why does income inequality matter? Previous research in the United States has found that more unequal states tend to have worse health and mental health outcomes, lower educational outcomes especially for children and teens, more crime and drug use, less social cohesion and trust, and less social mobility (Wilkinson and Pickett 2009a). These negative social outcomes partly gave rise to the Occupy Movement in hundreds of cities across the United States, most notably the Occupy Wall Street protests in New York City. Therefore, it is important for both the public and policymakers to better understand the current inequality trends in Iowa. This publication seeks to answer two questions: Is inequality a problem in Iowa, and does it have a negative impact on community well-being? To answer these questions, we examine inequality trends in Iowa across time and place; and use statistical techniques to determine whether inequality leads to poorer socioeconomic outcomes in Iowa s communities. David J. Peters, Ph.D. PM 3068 September 2014

2 Iowa has more equal incomes compared to the rest of the nation. Inequality is measured using the Gini coefficient (G), where values close to zero indicate equal income distributions and values close to one indicate unequal distributions. For example, perfect equality is where both the bottomearning 10% and top-earning 10% of households each own 10% of the community s income. By contrast, perfect inequality is where the bottom 10% owns none of the community s income and the top 10% owns it all. Data for this analysis is taken from the Decennial Censuses and the American Community Survey (hereafter referred as 2012 data), both produced by the U.S. Census Bureau. Historical Gini scores at the county-level are not available from Census, and had to be computed from income distribution data using a specific methodology. Thus, Gini scores reported here are similar but not exact to current ACS-Census estimates. Further information on the data and statistical methods is presented in the appendix. Inequality Today Currently, income inequality in Iowa is low compared to the rest of the nation. In 2012, Iowa s Gini coefficient was compared to the national rate of Although rates are low in the Hawkeye State, inequality has grown over the last several decades. Since 1970, Gini scores in Iowa grew by or 11.7%, but this is much slower than growth in national inequality over the same period (0.079 or 20.5%). Over the decade of the 2000s, however, Iowa inequality has grown slightly faster than the 2 From Parity to Polarization: Iowa Income Inequality PM 3068 U.S. rate, posting gains of or 3.1% versus or 2.1% for the nation. In short, Iowa has more equal income distributions compared to the rest of the U.S., but recently incomes are polarizing slightly faster than the national rate. During the 1970s Iowa experienced a small inequality gain due in part to a strong agricultural and manufacturing economy (a continuation 1960s economic expansion) that increased income shares of the top 20% of earners slightly faster than the bottom 20%. The prosperity of the 1970s gave way to farm and manufacturing crises of the 1980s that contracted incomes across the board, causing inequality to remain flat. After the lean years of the 1980s, inequality jumped as the economy rebounded during the 1990s. Incomes of poor and middle class Iowans stabilized, causing poverty rates in the state to fall. Top-earning Iowans, however, saw their incomes grow during the 1990s, causing inequality to rise. The decade of the 2000s has ushered in two recessions that have hit lowerincome households hardest, but have largely benefitted higherincome ones. Over this decade incomes for lower and middle income Iowans shrunk, while at the same time upper income households saw theirs grow. As a result, the 2000s has seen fairly sharp rises in both Gini scores and poverty rates reversing gains of the 1990s. Within Iowa there are some key rural-urban differences (see figure 1). In 1970 and 1980, inequality rates were generally similar across Iowa s

3 U.S. Iowa Metro Iowa Micro Iowa Rural Iowa Gini Since 1970, Iowa inequality grew much slower than the national rate (11.7% vs. 20.5%) * Figure 1. Inequality (Gini) in Iowa and the United States metropolitan, micropolitan, and rural areas (using 1993 definitions). Inequality between these population groups began to diverge in the 1980s as rural and micropolitan inequality fell due to farm and manufacturing crises. The 1990s saw a sharp rise in disparities in urban (especially mircopolitan) areas as the economy rebounded. However, by 2012 the situation reversed as polarization is now higher in rural places due to a booming farm economy, while metro and micro inequality is lower due to recessions. Over the past 40 years inequality has grown fast across most urban and rural places in Iowa (Gini increases between 6 and 8%). However, over the 2000s economic disparities in the countryside (7.7% growth) far outpaced that of the state s metros and micros (-1.3 and -0.9% declines, respectively). State averages mask important differences across Iowa s counties. Income polarization tends to be spatially clustered in four main areas of the state (see table 1 and figure 2). First, inequality is highest in north-central Iowa. This includes rural agricultural counties of Emmet, Kossuth, and O Brien; natural amenity Metros have an urban population of 50,000 or more. Micros have an urban population of 20,000-50,000. Table 1. Highest and lowest inequality counties in Iowa, 2012* Most Unequal 2012 Most Equal 2012 County Gini Major City County Gini Major City Emmet Estherville Cedar Johnson Iowa City Benton suburban Cedar Rapids Decatur Mitchell O Brien Mills suburban Council Bluffs Jefferson Fairfield Hancock Henry Mount Pleasant Jones suburban Cedar Rapids Worth suburban Mason City Lyon Adams Delaware Story Ames Butler Kossuth Algona Allamakee *Average between From Parity to Polarization: Iowa Income Inequality PM

4 Figure 2. Inequality (Gini) in Iowa 2012* Since 2000, however, Iowa incomes polarized slightly faster than the U.S. (3.1% vs. 2.1%). areas including Spirit Lake and the other Iowa Great Lakes; and the micropolitan centers of Fort Dodge, Mason City, and Spencer. The second cluster is in university towns of Ames and Iowa City, where lower income college students and higher income professionals live. Third, another rural inequality cluster is located in southwest Adams, Decatur, and Montgomery counties. Lastly, inequality is prevalent in southeast Iowa communities that are home to smaller colleges and correctional centers (especially Burlington, Fairfield, Mount Pleasant, and Ottumwa). By contrast, economic disparities are very low in the northeast part of the Hawkeye State. Another way to look at inequality is by how much each county contributes to Iowa s overall inequality rate (see figure 3). This is measured using a different inequality measure called Theil T (or the generalized entropy measure GE1), permitting decomposition of state-level inequality to each county based on population and polarization levels. Des Moines (Polk County), Figure 3. Inequality contribution (Theil T) in Iowa 2012* 4 From Parity to Polarization: Iowa Income Inequality PM 3068

5 Cedar Rapids (Linn County), Davenport (Scott County), and Iowa City (Johnson County) contribute most to Iowa s overall inequality. Although Polk County only has above average inequality, its large population drives state rates. By comparison, Johnson County has a much smaller population base, but much greater income polarization in the county drives its contribution. In general, most of the state s large population centers contribute to state inequality. Importantly, although north Iowa has the most polarized income distributions in the state, low population sizes in those places minimizes the impact on Iowa rates. Although there are some very unequal places in Iowa, they are still relatively equal compared to other parts of the United States (see figure 4). Economic disparities are highest in the large global cities such as Atlanta, Boston, Chicago, Los Angeles, Miami, New York, San Francisco, and Washington, D.C. Disparities are also high in much of the American South (especially Florida and Texas), in the Dakotas on Native American reservations and near mining areas, and across most of California. On the other hand, disparities are smallest in parts of the Midwest, the Great Lakes states, and the Mountain West. Inequality Trends Inequality since the 1970s Unlike much of the nation, Iowa inequality has grown more slowly over the past four decades, rising by only 9.1% versus 17.6% for the nation (see table 2 and figure 5). Incomes polarized fastest in rural areas of north Iowa (Emmet, Kossuth, and O Brien counties), in the state s micropolitan areas with large institutions like community colleges or prisons (Burlington, Fort Dodge, Mason City, and Ottumwa), and in smaller metropolitans areas (Iowa City, Davenport, and Council Bluffs). Despite being the largest urbanized center in the state, Des Moines only posted above average growth, although inequality spiked in suburban Dallas During the 2000s economic disparities in the rural Iowa far outpaced (7.7% growth) that of the state s metro (-1.3% decline) and micro (-0.9% decline) areas. Figure 4. Inequality (Gini) in the United States 2012* From Parity to Polarization: Iowa Income Inequality PM

6 Table 2. Growing and declining inequality counties in Iowa, Top Gains Top Declines County Gini Major City County Gini Major City Emmet Estherville Winneshiek Decorah Des Moines Burlington Howard Scott Davenport Allamakee Kossuth Algona Mills suburban Council Bluffs O'Brien Benton suburban Cedar Rapids Decatur Cedar Henry Mount Pleasant Lyon Wapello Ottumwa Mitchell Webster Fort Dodge Delaware Linn Cedar Rapids Wright Many rural Iowa counties are among the 10 fastest growing inequality places in the nation out of over 3,100 counties placing them in the top 1 percent. Figure 5. Inequality (Gini) in Iowa County. Counter to the general trend of rising inequality, some places in Iowa saw an inequality turnaround. A cluster of counties in northeast Iowa became more equal over the past 40 years (Allamakee, Howard, Mitchell, and Winneshiek counties). The only metro less polarized over this period was Ames. Iowa and its surrounding states in the Midwest and Great Plains stand out as having slower inequality gains compared to the rest of the nation. Income became much more polarized over the past 40 years across most counties 6 From Parity to Polarization: Iowa Income Inequality PM 3068 in the United States, especially in the New England, Rust Belt, Southern, and Pacific coast states (see figure 6). Inequality since the 2000s However, Iowa fared less well over the past decade, as Gini scores grew at a slightly faster pace than the nation as a whole (3.1% versus 2.1%). In the 2000s, most places in Iowa saw their incomes distributions become more unequal, and this was especially apparent in rural Iowa (see table 3 and figure 7). Most

7 Figure 6. Inequality change (Gini) in the United States Table 3. Growing and declining inequality counties in Iowa, Top Gains Top Declines County Gini Major City County Gini Major City Emmet Estherville Fayette Oelwein O Brien Guthrie suburban Des Moines Worth suburban Mason City Buchanan Independence Adams Mitchell Montgomery Red Oak Delaware Kossuth Algona Keokuk Henry Mount Pleasant Allamakee Shelby Mills suburban Council Bluffs Hamilton Webster City Cedar Tama Toledo Audubon Inequality is concentrated in rural north-central Iowa and micropolitans with large public institutions. Figure 7. Inequality change (Gini) in Iowa From Parity to Polarization: Iowa Income Inequality PM

8 Unequal places tend to be more rural, older, and more diverse in terms of population. of northern and central Iowa polarized during the 2000sas the farm economy boomed late in the decade. Inequality also grew in rural southwest Iowa home to community colleges and correctional facilities. The state s largest metropolitans had almost no change in inequality over the last 10 years, as recessions of the 2000s likely stagnated upper incomes in urban centers. Countering this trend, most of northeastern Iowa had falling or stable Gini scores. For the United States, the Great Recession slowed income polarization across most communities. The exception to this trend is found in Iowa and other states of the Farm Belt and Great Plains, where many communities posted some of the highest income polarization gains (see figure 8). For example, the top three Iowa counties are among the 10 fastest growing inequality places in the nation out of over 3,100 counties placing them in the top 1 percent. In other words, several rural places in Iowa are experiencing unprecedented polarization. Inequality and Socioeconomic Conditions Does inequality matter for Iowa communities, specifically is it associated with poorer socioeconomic outcomes? To answer this question, key demographic and economic variables are used to predict current Gini scores and their change using spatial regression models, presented in table 4 (see appendix for details). In terms of demographics, more unequal places in 2012 tend to be more rural and have fewer urban centers (higher urban-rural continuum codes). Communities with more elder populations (age 65 years and older) also have higher Gini scores. The two are related to some degree as a larger share of the rural population is past the traditional retirement age. The statistical model affirms what is shown in the maps (see figure 2), that inequality is widespread across rural Iowa. Unequal places are apt to have more minorities (non-whites and 8 From Parity to Polarization: Iowa Income Inequality PM 3068 Figure 8. Inequality change (Gini) in the United States

9 Hispanics). In most of Iowa this typically means Hispanics, both from domestic and international migration, and to a lesser extent immigrants from Africa and Asia. Immigrants to Iowa may have limited English ability, limited skills or unrecognized educational qualifications, and larger families. These factors contribute to lower family incomes compared to native born Americans, resulting in greater income polarization. By all accounts, this demographic will continue to grow in the future and will become an important part of Iowa s culture and economy, especially in the state s urbanized areas. One negative social outcome links inequality to larger numbers of single-headed families with children. The economic hardship of providing and caring for children as a single parent is well documented in the social inequality literature (Partridge and Rickman 2006). Compared to single-parent families, the presence of two adults significantly increases wage income and allows for greater time sharing of family responsibilities. It is this wage and time differential that likely drives income polarization. This aspect of inequality is already well understood by most social service agencies, and there is an array of programs focusing on child care access, parenting and household skills, and how to manage work-life balance. However, addressing this issue has been problematic because of limited funding, differences in program availability across places (especially child care), and wavering public support based on different sets of social values. In terms of economic conditions, we find current inequality is highest in places where most labor income is from jobs in low-skill and low-wage services sectors (retail and leisure services like entertainment, food, and accommodation). Further, communities that saw growth in these services also experienced growing rates of inequality. Retail and leisure services jobs contribute to polarization by depressing incomes for the lowest-earning households because such jobs tend to be part-time, temporary, low paid with limited benefits, and offer few advancement opportunities. Highlighting this, in this industry over 90% of jobs contain occupations requiring minimal skills, and this is reflected in the low average wage of $27,000 per job (OES-BLS 2013). On the other hand, inequality is also high in Iowa communities with high-skill high-wage professional services. Professional jobs drive inequality by increasing earnings at the top of the income distribution, since wage and benefit levels are high and jobs are more secure. For example, in this industry skilled professions account for over 60% of employment earning on average $92,000 per job, while the remaining one-third are semiskilled earning far lower wages of $42,000 (OES-BLS 2013). Inequality also appears to be driven by growth in farm and agricultural incomes, but not by base levels. Despite the fact that many high inequality counties are located in rural areas, income earned in farming and agricultural jobs has no impact Unequal places are more dominated by the dual-services economy, consisting of both low-skill (leisure and retail) and high-skill (professional) services. From Parity to Polarization: Iowa Income Inequality PM

10 Growth in farm and agricultural income during since the 1970s is linked to greater income disparities. on current Gini scores in However, counties experiencing faster income polarization since the 1970s and 2000s also have fast growth in farm incomes. This contributes to inequality by sizably increasing incomes of a small number of farm households, especially in rural Iowa where the population base is low. Some economic sectors actually slowed income disparities during the last decade. Labor income growth in manufacturing, transportation, public utilities, and communications resulted in slower inequality gains during the 2000s. One reason why these sectors tend to promote equality is they hire many middle-skill and middle-wage workers. These occupations require some education and training beyond high school (trades or two-year degrees), generally have good wages and benefits (unionization or public utilities), and tend to be more stable (full-time and full-year jobs). In manufacturing, for example, production occupations account for over half of employment and pay middle wages of around $40,000 per job (OES-BLS 2013). Ancillary occupations also pay middle wages, such as office and production clerks (9% of jobs that pay $38,000) and transportation and warehousing workers (8% of jobs paying $32,000). These findings are consistent with what is called the polarization thesis, which argues that a globalizing post-industrial economy has increased inequality (Hamnett, 2003; Sassen, 1991). The polarization thesis acknowledges that the shift 10 From Parity to Polarization: Iowa Income Inequality PM 3068 towards globalization and postindustrialism has increased the number of higher skill and higher wage jobs in the financial, business, and professional services sectors. However, they also argue there has been parallel growth in relatively lower skilled and lower wage services jobs, which supports professional and managerial functions and the people working within them. The dual growth of both high-paid and low-paid services, along with declines in industrial goods-producing sectors, is argued to have reduced middle-skill and middle-wage jobs and resulted in the growing polarization of incomes. Summary and Implications for Iowa Is inequality a problem in Iowa, and does it have a negative impact of community well-being? In terms of major trends, Iowa has more equal income distributions than the rest of the nation, but incomes are polarizing slightly faster than the rest of the United States. In the 1980s and 1990s, inequality grew much faster in urban Iowa than in rural parts of the state. However, the trend reversed during the 2000s as economic disparities in the countryside outpaced that or urban Iowa. In short, inequality is an emerging problem in the state, especially in rural Iowa. There is also a clear spatial clustering of unequal counties in Iowa. To start, inequality is highest and fastest growing in a cluster of north Iowa counties whose economies are primarily dominated by agriculture

11 Table 4. Predicting inequality (Gini) across Iowa counties, Base Change Change y = Gini & Change (0-100 scale) β β β Demographics Economics Intercept (β0) Spatial Error (λ) ** ** Urban-Rural Continuum * Population Density (per sq.mi.) Minority Population (%) 0.121*** Age 17 and Under (%) Age 65 and Older (%) 0.500*** ** Single-Headed Families with Children (%) 0.077** Bachelor s Degree or Higher (%) College Student Population (%) Labor Force Participation (%) *** *** Previous Gini (G) * *** *** Agriculture & Natural Resouces $ (%) ** 0.059** Construction $ (%) Manufacturing $ (%) * Transport, Utilities & Communication $ (%) ** Finance, Insurance & Real Estate $ (%) Retail & Leisure Services $ (%) 0.470*** 0.030** 0.038** Health & Education Services $ (%) n.a. Professional Services $ (%) 0.379** n.a. Professional, Health & Educ Services $ (%) n.a. n.a Model Fit *p<.10, **p<.05, ***p<.01 R2 / PRE * AIC Heteroscedasticity Test (B-P) * Spatial Dependence Test (LL) Income polarization during the 2000s was kept in check by industries employing middleskill and middle-wage workers (such as manufacturing, transportation, and utilities). (notably Emmet, Kossuth, O Brien, and Worth counties). Further, inequality gains in these communities were so large that many ranked in the top 10 fastest growing places (out of over 3,100) in the nation over the last decade. Next, disparities are also high and growing in smaller micropolitan areas dominated by public institutions, such as community colleges or correctional facilities (such as Burlington, Fort Dodge, Mason City, Mount Pleasant, Ottumwa, and Spencer). Lastly, income polarization is a problem in Iowa s university towns of Ames and Iowa City. The statistical correlates of inequality show that more unequal places in Iowa tend to be more rural, older, and more diverse in terms of population. There are also more single-headed families, a negative social outcome related to higher poverty. In terms of economics, unequal places are more dominated by the dualservices economy, consisting of both low-skill (leisure and retail) and high-skill (professional) services. This is consistent From Parity to Polarization: Iowa Income Inequality PM

12 Economic participation is key for reducing disparities. with the polarization thesis, where disparities are driven by growth in the services jobs and corresponding declines in goods producing ones. Of particular importance for Iowa, growth in farm and agricultural income during since the 1970s is also linked to greater disparities. On the other hand, income polarization during the 2000s was kept down by industries that employ many middle-skill and middle-wage workers (such as manufacturing, transportation, and utilities). More broadly, growing rates of labor force participation is linked to falling rates of inequality, suggesting economic participation is key for reducing disparities. While these demographic and economic factors that are important, more research is clearly needed to better understand the causes and consequences of income polarization in Iowa. What can be done to address economic inequalities? The most conclusive policy recommendations demonstrate that inequality can be held at bay through public investment in quality education and jobs, and a tax system to support such investments (Wilkinson and Pickett 2009b). The best way to reduce economic disparities is to provide quality employment opportunities. However, not all jobs are created equal in terms of their pay, benefits, and stability. State and local economic development programs in Iowa can help reduce inequality by focusing their efforts at expanding and attracting jobs that meet certain criteria. First, jobs should be at least moderately skilled so they are accessible to persons with some education and training beyond high school. However, this requires a quality K-12 education system that is both adequately funded and accountable for student learning outcomes. Second, jobs should pay above average wages and offer benefits, which increase the wealth and health of the household. Jobs should be full-time and full-year so families have stable incomes. Lastly, development efforts should be targeted at industries that are growing nationally, or to those in which the state or region has a competitive advantage. To pay for all of this takes large public investments in a time of state budget austerity. The most common approach to raise the needed revenues is to expand the progressive income tax structure to increase the rates on top earners; and to tax capital gains income at a higher rate than wage income. In most of Iowa, inequality can be reduced through programs that increase the share of income owned by the bottom-earning households. The challenge, of course, is creating these quality jobs and providing quality educations in areas that have experienced population and economic declines. However, the alternative is for Iowa to lose its reputation of being a place of economic opportunity and social mobility for everyone regardless of the family circumstances. This is not just a cultural value but economic fact. A recent project by Harvard University (www. equality-of-opportunity.org) found that Iowa has the fourth highest upward social mobility in the nation, behind North Dakota and Nebraska but tied with 12 From Parity to Polarization: Iowa Income Inequality PM 3068

13 Minnesota. In other words, the real concern is that rising income inequality will slow economic mobility and career aspirations for Iowa s children (especially the poorest), which over time is likely to erode the Hawkeye State s high quality of life. From Parity to Polarization: Iowa Income Inequality PM

14 References DeNavas-Walt, C., Proctor, B., & Smith, J. (2011) Income, Poverty, and Health Insurance Coverage in the United States: P60 239, Census Bureau. Washington, DC: U.S. Department of Commerce. Hamnett, C. (2003) Unequal City: London in the Global Arena. London: Routledge. Moller, S., Alderson, A., & Nielsen, F. (2009) Changing Patterns of Income Inequality in U.S. Counties, American Journal of Sociology 114: Nielsen, F. & Alderson, A. (2001) Trends in Income Inequality in the United States. Pp in Sourcebook on Labor Markets: Evolving Structures and Processes, Berg, I. & Kalleberg, A. (Eds.), New York: Plenum Publishers. Partridge, M. & Rickman, D. (2006) The Geography of American Poverty: Is There a Need for Place-Based Policies? Kalamazoo, MI: Upjohn Institute. OES-BLS (2013) Occupational Employment Statistics: National Industry-Occupation Estimates. U.S. Bureau of Labor Statistics. Peters, D. (2013) American Income Inequality Across Economic and Geographic Space, Social Science Research 42: Piketty, T. (2014) Capital in the Twenty-First Century. Cambridge, MA: Belknap Press of Harvard U. Sassen, S. (1991) The Global City: New York, London and Tokyo. Princeton, NJ: Princeton U. Press. Wilkinson, R. & Pickett, K. (2009a) Income Inequality and Social Dysfunction. Annual Review of Sociology 35: Wilkinson, R. & Pickett, K. (2009b) The Spirit Level: Why Greater Equality Makes Societies Stronger. New York: Bloomsbury Press. 14 From Parity to Polarization: Iowa Income Inequality PM 3068

15 Appendix Gini Coefficients for United States and Iowa, * Estimated Gini * United States Iowa Metro Iowa Micro Iowa Rural Iowa Adair Adams Allamakee Appanoose Audubon Benton Black Hawk Boone Bremer Buchanan Buena Vista Butler Calhoun Carroll Cass Cedar Cerro Gordo Cherokee Chickasaw Clarke Clay Clayton Clinton Crawford Dallas Davis Decatur Delaware Des Moines Dickinson Dubuque Emmet Fayette Floyd Franklin Fremont Estimated Gini Greene Grundy Guthrie Hamilton Hancock Hardin Harrison Henry Howard Humboldt Ida Iowa Jackson Jasper Jefferson Johnson Jones Keokuk Kossuth Lee Linn Louisa Lucas Lyon Madison Mahaska Marion Marshall Mills Mitchell Monona Monroe Montgomery Muscatine O Brien Osceola Page Palo Alto Plymouth Pocahontas Polk Pottawattamie From Parity to Polarization: Iowa Income Inequality PM

16 Appendix Gini Coefficients for United States and Iowa, (continued) Estimated Gini Poweshiek Ringgold Sac Scott Shelby Sioux Story Tama Taylor Union Van Buren Estimated Gini Wapello Warren Washington Wayne Webster Winnebago Winneshiek Woodbury Worth Wright *Average between From Parity to Polarization: Iowa Income Inequality PM 3068

17 Appendix Data and Methods Data Sources This analysis uses a unique set of spatial data from the 1970 to 2000 Decennial Censuses and the American Community Survey (ACS), both produced by the U.S. Census Bureau. Decennial and ACS Census are the only source of income distribution data at the county level. Demographic and economic covariates are also taken from the same Census data. All data are by place of residence. The ACS has replaced the long-form decennial census, and there are some important differences between the two that should be noted. First, ACS data represent average values for each year between 2008 and 2012, rather than point-in-time estimates. Second, income and employment status are for the previous 12-month period, rather than for the previous calendar year. Third, standard errors for the ACS tend to be higher for smaller geographic units than was the case in previous census periods using the long form. However, analysis of the standard errors finds few estimates whose coefficient of variation exceeded 25 percent, indicating adequate data quality. Definitions of metropolitan and micropolitan counties are based on the 1993 Rural-Urban Continuum Codes from USDA s Economic Research Service. Metro definitions change with each decennial Census that complicates analysis over time. To prevent misclassification of counties into rural and urban groups, the 1993 definitions are an ideal midpoint in between the period. For example, 1970 definitions would classify too many established suburban metro counties as rural; and 2012 definitions would classify too many recently rural areas as suburban metros. Estimating Inequality Census reports pre-tax income for all workers age 15 and older in the household. A household may contain a single individual or a group of related individuals such as a family. People living in group quarters, a housing unit managed by an organization, are excluded. Group quarters include college residence halls, skilled nursing facilities, residential treatment centers, group homes, and correctional facilities. Income is counted from all sources including wages and salaries, self-employment income from farms and businesses, capital gains (dividends, interest, and rents), retirement and social security income, cash public assistance, and other miscellaneous income. It excludes the value of assets and noncash benefits. Census does not report inequality. Census only reports income distributions at the county-level, preventing use of most inequality metrics that require household-level data. To solve this issue, county data are manipulated to estimate the household distribution using the following method. Census reports the number of households within predetermined income ranges, typically 16 to 20 categories. For each category, the income midpoint is repeated equal to the number of households. Taken together across all categories, the result is a range of incomes equal to the total number of households in the county. Another issue is the top-coded income category is censored, meaning it has no upper bound (e.g. topcoded income in ACS is $200,000 or more annually). In order to estimate average top-coded income, the sum of aggregated income using category midpoints is calculated excluding the top-coded category. This sum is compared to total aggregate income reported by the Census. If the difference between the two is greater than the product of the top income category by the number of households, then the difference is used. From Parity to Polarization: Iowa Income Inequality PM

18 Using the estimated income distributions, standard inequality metrics are calculated including the Gini G coefficient and Theil T index (generalized entropy measure with alpha set to one). Gini is a nondecomposable measure that ranges from 0.0 (perfect equality) to 1.0 (perfect inequality). For example, perfect equality is where both the bottom-earning 10% and top-earning 10% of households each own 10% of the community s income. By contrast, perfect inequality is where the bottom 10% owns none of the community s income and the top 10% owns it all. Theil T belongs to a class of generalized entropy measures that permit decomposition of aggregate inequality into its component parts, in this case individual counties. Theil T is estimated when the sensitivity parameter (α) is set to 1.0, making the calculation sensitive to incomes at the top of the distribution. G = 2 2 N y N i( y i y) i= 1 GE ( α ) N 1 1 y = α 1 ( α 1) N i= 1 y i α Where N number of households i current household y household income α equals 1 for Theil T As a result of these manipulations, Gini scores reported here are similar but not exact to current ACS- Census Gini estimates. Reasons for the differences are: (i) use of income midpoints rather than actual household income from the Census form; and (ii) estimation of top-coded average income. Actual incomes are not available to the public, and must be estimated using aggregated data. Gini scores reported here tend to be lower than actual Census estimates, largely driven by underestimation of topcoded incomes. Statistical Models Spatial regression is used to predict current Gini scores and their change using socioeconomic covariates for the base period (2012) and change over time ( and ). Spatial regression controls for dependence in the residuals caused by the spatial nature of the data; and failure to control for this dependence may lead to unreliable tests of model parameters. Spatial error regression models (SERMs) are used in this analysis because the dependent variable is influenced by some unknown spatial effect that is omitted from the model, which is controlled by spatially correlating the errors using queen contiguity weights. Most assumptions of SERM are met. There is some multicollinearity among predictors, but tolerance statistics are in acceptable ranges. Error variances are constant, normal, and uncorrelated except in the model, where the Breusch-Pagan (B-P) test indicates heteroscedasticity. 18 From Parity to Polarization: Iowa Income Inequality PM 3068

19 y = Xβ + λwu + ε Where y Gini score X socioeconomic covariates b regression parameters λ spatial parameters W spatial weights u spatially correlated residuals ε spatially uncorrelated residuals From Parity to Polarization: Iowa Income Inequality PM

20 For More Information David J. Peters, Ph.D. Department of Sociology 304 East Hall Iowa State University Ames, IA TEL: FAX: Prepared by David J. Peters, Ph.D., associate professor and extension sociologist. Renea Miller provided valuable layout assistance to this report....and justice for all The U.S. Department of Agriculture (USDA) prohibits discrimination in all its programs and activities on the basis of race, color, national origin, age, disability, and where applicable, sex, marital status, familial status, parental status, religion, sexual orientation, genetic information, political beliefs, reprisal, or because all or part of an individual s income is derived from any public assistance program. (Not all prohibited bases apply to all programs.) Persons with disabilities who require alternative means for communication of program information (Braille, large print, audiotape, etc.) should contact USDA s TARGET Center at (voice and TDD). To file a complaint of discrimination, write to USDA, Director, Office of Civil Rights, 1400 Independence Avenue SW, Washington, DC , or call (voice) or (TDD). USDA is an equal opportunity provider and employer. Issued in furtherance of Cooperative Extension work, Acts of May 8 and June 30, 1914, in cooperation with the U.S. Department of Agriculture. Cathann A. Kress, director, Cooperative Extension Service, Iowa State University of Science and Technology, Ames, Iowa. 20 From Parity to Polarization: Iowa Income Inequality PM 3068