Racial Diversity and Public Policy in the States

Abstract

How does racial diversity affect public policy outcomes in the states? The policy backlash hypothesis suggests that the presence of blacks increases antagonistic attitudes toward minorities among whites and thus produces racially conservative policies. The electoral constraint hypothesis argues that the presence of blacks increases the size of population that supports policy liberalism and thus results in more liberal policies. The authors reconcile these competing explanations by arguing that the hypothesized negative impact of minority group size on policy liberalism varies across states in accordance with socioeconomic and institutional conditions that strengthen or weaken the impacts of electoral constraint or backlash.

Keywords

Racial politics,public policy welfare

How does racial diversity affect public policies in the states? Studies suggest that diverse environments alter mass policy preferences, partisan attachments, and rates of political participation (Giles and Hertz 1994; Leighley 2001; Oliver and Wong 2003). Unfortunately, we know far less about the effect of racial diversity on public policy formation at the subnational level. After the passage of the Voting Rights Act in 1965, many expected significant, positive changes in the degree to which minority interests would be represented within the policymaking process. At the state level, scholars theorize that blacks, who prefer liberal social welfare policies, exercise more political power as their population size increases, resulting in the establishment of liberal policy environments. However, a wide range of empirical research finds evidence contrary to this argument. In other words, the presence of a large black population within a state is associated with less liberal welfare policies even after the 1960s. This is true despite a sharp rise in the number of black elected officials over the past few decades. The failure of scholars to find this expected relationship between black population size and liberal racial policy outcomes remains one of the major unresolved issues in the literature on minority representation. How can we explain the inconsistency between scholars’ theoretical expectations and empirical evidence?

In this article, we reexamine the dilemma regarding the relationship between racial context and public policy formation in the states. By contrasting works discussing the effect of racial diversity on mass preferences and the behavior of elected representatives, we develop two conflicting explanations of how minority group size influences the likelihood that states will enact liberal policies that are explicitly or implicitly racialized. The policy backlash hypothesis suggests that states with large minority populations are characterized by negative racial attitudes and are thus more likely to enact racially conservative policies (Key 1949; Blalock 1967; Fellowes and Rowe 2004). The electoral constraint hypothesis, by contrast, argues that minority group size works to limit the extent to which conservative legislators will be willing to pursue policies that are antagonistic to minority interests (Lublin 1997; Yates and Fording 2005).¹

We argue that we cannot identify the exact impact of black group size on policy formation unless we take into consideration the effects of “electoral constraint” and “backlash” simultaneously. We reconcile these competing explanations by arguing that the impact of black population size on policy liberalism varies in accordance with socioeconomic and institutional environments. The backlash effect of black population size on policy liberalism is weak in those environments that strengthen the impact of black electoral constraint or otherwise mitigate the impact of white backlash. In contrast, the negative impact of black population size on policy liberalism is strong in those environments that weaken the impact of black electoral constraint or otherwise intensify the impact of white backlash. These hypotheses are tested using time-series cross-sectional data of minority group size and state welfare spending from 1980 to 2000.

Race and Public Policy in the States

As we noted earlier, previous works offer mixed evidence regarding the relationship between minority group size and public policies that are explicitly or implicitly racialized in the states.² The dominant explanation for the relationship is the power-threat or backlash hypothesis (Key 1949; Blalock 1967). Put simply, the backlash hypothesis states that a “superordinate group (e.g., whites) becomes more racially hostile as the size of the proximate subordinate group increases, which punitively threatens the former’s economic and social privilege” (Oliver and Wong 2003, 568). Racial hostility may manifest itself in numerous ways, such as expressions of racial prejudice and an increased tendency to oppose racially liberal public policies (Giles and Hertz 1994; Glaser 2003; Quillian 1996; Taylor 1998). In short, whites may be more likely to oppose redistributive policies in states with a large black population, moving the position of the median voter among whites in a conservative direction. As a result, a large minority population has potential to produce more racially conservative policies.

Previous works examining the relationship between black population size and representation at the state level lend strong support to the backlash argument. A number of studies (Fellowes and Rowe 2004; Soss et al. 2001) show that the presence of a large black population in a state is associated with lower welfare spending and more restrictive welfare provisions. A recent study by Preuhs (2007) finds that states with larger Latino populations also tend to have less generous welfare policies. Similarly, research in economics demonstrates that higher levels of racial diversity reduce the amount of government redistribution (Alesina, Baqir, and Easterly 1999; Lind 2007).

The relationship between racial diversity and racially conservative policy outcomes may be especially acute at the state level, where residential patterns mean that increased minority group size does not result necessarily in numerous racially diverse legislative districts. Douzet’s (2008) analysis of politics and racial segregation in California argues that California has become a “patchwork state,” with pockets of diversity and ethnic homogeneity. As a result, elected officials see themselves as representing particular, often racially homogenous, constituencies. Representatives from homogenously white districts have little incentive to represent minority interests; rather, they may prefer racially conservative policies that reflect the opinions of their constituents who reside in the vicinity of minority populations but not among them (Rocha and Espino 2009).

This argument may also explain the null relationship Nelson (1991) finds between the descriptive and the substantive representation of minorities at the state level. In accordance with studies of minority representation in Congress (Canon 1999), minority state legislators are significantly more likely to sponsor minority-interest legislation (Bratton and Haynie 1999; Haynie 2001). Thus, a null relationship is unlikely to be a product of minority representatives choosing not to advocate minority interests; rather, it is likely to be a product of nonminority representatives choosing to oppose minority-interest legislation more strongly.

The policy backlash hypothesis, however, is limited because it does not take account of the policy preferences of blacks. Blacks who have distinctive policy preferences impose an electoral constraint on policy outcomes. Yates and Fording (2005) find that Republican control of state government is associated with inequitable and less favorable policy outcomes for blacks. The magnitude of this association is lessened in states with large black populations, implying that blacks in those states serve as an electoral constraint on both Democratic and Republican representatives. Hero’s (1998, 2007) work also shows that more diverse states are characterized by greater levels of racial policy equity.

The notion of electoral constraint corresponds well with what we know about how representatives respond to different racial environments. Lublin (1997) demonstrates that the voting patterns of House members become increasingly liberal as blacks make up a larger proportion of their constituency. Although Canon (1999) finds less consistent evidence regarding the relationship between minority group size and the roll call voting behavior of House members, he does find that representatives of districts with large black populations are more likely to cosponsor legislation and give floor speeches on racial issues. These works present a convincing case that large minority populations serve as an electoral constraint on representatives regardless of partisanship. Representatives stand as advocates for the policy preferences held by their constituencies, and representatives are apt to adopt racially liberal policy positions as minorities make up a larger share of their constituency.³

The electoral constraint hypothesis can be also drawn from Meltzer and Richard’s (1981) model of government redistribution. Meltzer and Richard suggest that the expansion of voter enfranchisement increases the number of voters from the lower end of income distribution and shifts the position of the median voter in a liberal direction. Consequently, any exogenous shock that brings more voters with low income into the electorate increases government redistribution. Husted and Kenny (1997) offer a test of this hypothesis, finding that the elimination of poll taxes and literacy tests in the 1950s and 1960s led to higher welfare spending by state and local governments.⁴ Blacks hold distinctive policy preferences on explicitly and implicitly radicalized issues and are more supportive of redistributive policies than are whites (Kinder and Winter 2001), suggesting that we can expand Meltzer and Richard’s argument to the relationship between the racial context and policy formation in the states. In short, states with larger black populations will be characterized by a more liberal median voter, possibly resulting in more government redistribution.

One shortcoming of the electoral constraint hypothesis is its assumption that the policy preferences of nonminorities do not change with the size of the minority population, an assumption that conflicts with the expectations of the backlash hypothesis. While a larger minority population produces more antagonistic attitudes among whites toward issues such as redistributive policies, it also increases the number of voters who have more supportive views of such policies, as discussed earlier.

Figure 1 offers an example of how the overall distribution of preferences on redistributive policies changes within a state as the size of the black population increases from 10 percent to 40 percent with and without considering the impact of electoral constraint.⁵ The bold vertical lines in the figure represent the position of the median voter, while the dotted vertical lines represent the middle point of the horizontal scale. The left side of the horizontal scale represents more supportive and liberal views of government redistribution, while the right side of the scale represents less supportive or more conservative views. We assume that blacks are more supportive of redistribution than whites. The top panel illustrates the location of the median voter when we ignore the impact of black electoral constraint. The location of the median voter is exactly equivalent to the location of the white median voter. As the size of the black population increases, the median voter moves to the more conservative direction because of the backlash effect. The bottom panel illustrates the location of the median voter when we take into consideration black electoral constraint. The backlash effect is weaker in the bottom panel simply because blacks who are supportive of redistributive policies also affect the location of the median voter.

Figure 1.

The position of the median voter within a state

This example suggests that the negative relationship between black group size and policy liberalism does not necessarily mean that black group size has no electoral constraint on the policy formation. In other words, the negative relationship is not necessarily equivalent to the total impact of white backlash on policy liberalism because the effect of white backlash may be counteracted by the impact of black electoral constraint. In short, Figure 1 indicates that we cannot identify the exact impact of black size on policy formation unless we take into consideration the electoral constraint and backlash effects simultaneously.

Hypotheses

We develop several hypotheses to reconcile electoral constraint and policy backlash theory. Our explanation is based on the notion that the position of the median voter has the decisive impact on policy formation (Meltzer and Richard 1981; Erikson, Wright, and McIver 1993).

In line with the predictions of the backlash argument, we begin by suggesting that the presence of blacks results in a conservative shift within the electorate. Attitudes on issues such as welfare are likely to become increasingly negative among whites, who make up a larger share of population in most of the states. In short, we hypothesize that the greater presence of blacks within a state is associated with less liberal policies for domains that are racialized such as welfare (Hypothesis 1).

To identify the relative effects of black electoral constraint or white backlash, we next consider the role of socioeconomic and institutional conditions within states, which determine the relative impacts of backlash and electoral constraint on the location of the median voter. We argue that the effect of black group size on racial policy outcomes is conditional on the socioeconomic and institutional characteristics of states because they affect the relative level of political participation between blacks and whites and the level of hostility among whites toward blacks.

We expect that the relative influence of any backlash effect is conditional upon the disparity of socioeconomic resources between blacks and whites, voter registration laws, and the socioeconomic status of whites within a state. The socioeconomic disparity between blacks and whites and voter registration laws are expected to affect the relative level of electoral involvement between blacks and whites and thus determine the degree to which minorities affect the position of the median voter and thus constrain the behavior of elected officials. The socioeconomic status of whites is expected to affect the extent to which whites negatively react to the presence of blacks and thus determines the size of any policy backlash effects.

Our second hypothesis focuses on these socioeconomic and institutional conditions. Scholars have long noted that all individuals are not equally likely to turn out or otherwise participate in politics. Specifically, Verba et al. (1993) demonstrate that blacks are significantly less likely to participate in politics than are whites. This may limit the extent to which black group size affects the position of the median voter through the mechanism suggested by the electoral constraint hypothesis, thereby increasing the effect of backlash on policy formation.⁶

Lower rates of participation among blacks are primarily the product of class, as socioeconomic status is an equally important determinant of participation for blacks and whites (Leighley and Vedlitz 1999). Controlling for socioeconomic status, blacks are no less likely to participate in most forms of electoral activities, including voting (Verba, Schlozman, and Brady 1995). Based on this evidence, we expect the electoral influence of blacks to be stronger in states where blacks enjoy more socioeconomic resources relative to whites. Therefore, we hypothesize that the negative impact of the black population size on policy liberalism is weakened by greater socioeconomic equality between blacks and whites (Hypothesis 2). That is, the backlash effect of the black group size on policy liberalism is strong when the difference in the socioeconomic status between blacks and whites is large, while the backlash effect is weak when the difference in the socioeconomic status is small. This hypothesis is consistent with previous works that suggest a relationship between class or racial bias in voter turnout and political representation (Griffin and Newman 2008; Hill, Leighley, and Anderson 1995).

We also expect that the relative political involvement of blacks and whites is also shaped by institutional settings. Specifically, we focus on the role of voter registration laws because previous studies suggest that voter registration laws play a role as a determinant of the relative level of political participation between blacks and whites.⁷ Wolfinger and Rosenstone (1980) argue that the dissuading impact of restrictive registration laws on voter turnout is more pervasive among the less educated and affluent because institutional barriers increase the cost of voting above the threshold of those who have fewer resources. Hanmer (2009) likewise confirms that less restrictive registration laws are more beneficial for less educated people, though the difference in impacts is quite small. This line of research leads us to expect that less restrictive registration laws boost participation among the disadvantaged and in turn produce a more liberal electorate. Because blacks are more likely to be socially and economically disadvantaged, we expect that less restrictive registration laws also encourage their participation and enhance their electoral influence.⁸ Thus, we hypothesize that the negative impact of black population size on policy liberalism is moderated in states with less restrictive registration laws (Hypothesis 3a).

An alternative expectation regarding the effect of voter registration laws can be drawn from research following Wolfinger and Rosenstone’s (1980) seminal work. Nagler (1991) finds no difference in the impact of registration laws on people with different levels of education. Highton and Wolfinger (1998) argue that those with moderate amounts of education (e.g., a high school degree or some college education) are the most likely to be affected by the removal of restrictive registration laws, while Wolfinger and Hoffman (2001) report that blacks and Latinos are less likely to take advantage of the voter motor registration act. Similarly, Hill and Leighley’s (1996) state-level analysis indicates that the positive impact of less restrictive voter registration laws is greater for middle- to upper-class individuals (see also Brians and Grofman 2001). Oliver (1996) shows that the introduction of mail voting has the largest impact on participation among advantaged individuals. In short, some works suggest that less restrictive registration laws are more likely to benefit the socially and economically advantaged and in turn result in a more conservative electorate. Less restrictive registration laws may encourage the majority to vote while having little impact on the participation of minority groups. As a result, we consider the possibility that the negative effect of minority population size on policy liberalism is stronger in states with a less restrictive registration system (Hypothesis 3b).⁹

Finally, any backlash effect may be conditional on the socioeconomic status of whites. Previous studies show that negative reactions to out-group members are stronger among low-status individuals (e.g., Sniderman, Brody, and Tetlock 1991), in low-status environments (Branton and Jones 2005; Gay 2006; Oliver and Mendelberg 2000), or in worsening economic situations (Quillian 1995). These findings lead us to expect that whites will hold increasingly negative views of blacks and racialized policies, such as welfare, when whites enjoy fewer economic resources. In other words, whites in states with a large minority population and fewer economic resources reside in a “maximum-threat context” (Branton and Jones 2005, 360), which in turn enhances their negative reactions to black group size. Accordingly, we would expect the negative reactions of whites to be weaker in more affluent contexts. Extending this literature, we argue that less affluent whites will hold more antagonistic attitudes when confronted with a large black population and therefore express conservative preferences on racialized policy issues. When whites hold such strong negative attitudes, the position of the median voter will move further in a conservative direction. Thus, we hypothesize that the backlash effect will be stronger in states where whites enjoy fewer socioeconomic resources (Hypothesis 4).

Data and Method

We employ a pooled time-series cross-sectional design to test our hypotheses regarding the relationship between race and welfare policies in forty-nine states for the years 1980-2000. Nebraska is excluded from our analysis because of its nonpartisan legislature. We selected these years due to the availability of data for our key independent and dependent variables. The maximum number of observations is 1,029.

In line with previous research, we develop two measures of state support for welfare programs as our dependent variables. Earlier studies have employed several alternative measures of state welfare policy: the benefit levels of Aid to Families with Dependent Children (AFDC) or Temporary Assistance for Needy Families (TANF) (Brown 1995; Fording 1997, 2001; Preuhs 2006; Tweedie 1994), eligibility and flexibility rules (Fellowes and Rowe 2004; Soss et al. 2001), and per capita spending on public welfare (Lind 2007; Radcliff and Saiz 1995; Smith 1997) or on AFDC programs (Barrilleaux, Holbrook, and Langer 2002). Some works (Fellowes and Rowe 2004; Hill and Leighley 1992; Preuhs 2007) use two or all of these measures to enhance the robustness of their findings. Typically, the benefit levels or the eligibility rules are used as a measure of state welfare generosity, while per capita spending on welfare is used as a measure of state welfare effort.

To incorporate both of these aspects of state welfare programs, we use per capita spending on public welfare and AFDC benefit levels for our regression analysis. Per capita spending for welfare equals the total amount of expenditures on public welfare programs in constant dollars divided by the total population.¹⁰ Public welfare programs include medical and cash assistance, payments to private vendors providing welfare services, welfare-related community action programs, construction for welfare institutions, and children services.¹¹ Note that we take a log of per capita welfare spending for our regression analysis to account for the percentage changes (i.e., the growth rate) rather than changes in the absolute amount of spending.¹²

AFDC benefit levels are measured as the ratio of the AFDC maximum benefit for a family of three to the median family income for a family of four.¹³ We use the ratio to take into account differences in the cost of living across states.¹⁴ AFDC benefit levels are available only from 1980 to 1996 because of the major policy shift in 1996, although the results reported here hold even after TANF maximum benefit levels are included for the years from 1997 to 2000.¹⁵

These two measures of state support for welfare programs are positively correlated, but only weakly so.¹⁶ AFDC benefit levels are a popular measure among scholars because they are controlled by the state, while other measures are influenced by a variety of external factors. As Alesina, Baqir, and Easterly (2000) note, however, pro-welfare politicians in state legislatures may avoid increasing AFDC benefit levels because cash assistance is an explicit redistributive policy and consequently may trigger strong opposition from anti-welfare politicians. Pro-welfare politicians may exploit a less explicitly redistributive policy for providing welfare services to satisfy their constituency and circumvent the opposition. Accordingly, it is also important to direct our attention to the total amount of expenditures spent for a variety of welfare programs. The robustness of our findings increases if our hypotheses receive evidence from both of these measures, but we have no a priori expectations regarding how the effects of our key independent variables will differ across welfare expenditures and AFDC benefit levels.

We also chose to rely on welfare expenditure and benefit-based measures because they represent a policy output rather than an outcome. Policy outcomes, such as the percentage of minorities living in poverty, are undoubtedly of great substantive importance to minority communities and merit examination. However, outcomes are not directly manipulable by voters and political institutions in the same way as outputs.

Our main expectation is that state support for welfare programs will be a function of the size of the black population. We account for the proportion of blacks within a state with data obtained from the Census Bureau’s U.S. Population Estimates. We further argue that the relationship between black group size and policy outputs is contingent upon the socioeconomic disparity between blacks and whites, the absence of restrictive voter registration laws, and socioeconomic status of whites within a state. We measure socioeconomic inequality by creating a black/white income ratio. The income ratio is based on the black and white median family incomes in constant dollars as reported in the 1980, 1990, and 2000 decennial censuses. The ratios are then computed by dividing the median black family income by the median white family income. Thus, a larger score denotes higher income equality. Noncensus years are imputed using a weighted moving average. The average black/white income ratio is 0.65.

To test for the possible effect of liberal registration laws, we use voter registration information found in the Book of the States to create a variable that equals 0 when a state requires voters to be registered before election day and equals 1 when a state has no such requirement and voters can register on election day.¹⁷ North Dakota has no registration requirement and is included with the states that offer voter registration on election day. We chose to categorize states by the requirement of voter registration before election day rather than closing days of voter registration because we have clear evidence that allowing registration on election day or requiring no registration has a significant positive impact on the rate of voter turnout (Besley and Case 2003; Brians and Grofman 2001), while the findings on the closing days are mixed (Wolfinger and Rosenstone 1980; Nagler 1991). We also tested the hypothesis using the closing day of voter registration. The substantive result reported in the following holds.

The socioeconomic status of whites is measured with two different indicators. The first is the proportion of whites in a state living in poverty. Our second indicator is the proportion of whites over 25 who have graduated from college. We use both of these measures to test Hypothesis 4 since some research suggests that economic distress is a predictor of whites’ racial attitudes (Oliver and Mendelberg 2000), and other research suggests that educational attainment is more influential (Branton and Jones 2005).

Our regression models include several control variables that account for political, economic, and demographic differences across states. Following previous research that examines the formation of welfare policy at the state level (Fellowes and Rowe 2004; Tweedie 1994), we select three sets of time-varying variables that may distort the relationship between racial context and state support for welfare programs. The first set of variables represents the influence of constituents. These variables include the proportion of Latinos, the proportion of school-age children, the proportion of the population over 65 years old, the proportion of a state’s population that is unemployed, and citizen ideology. Although the literature thus far emphasizes connections between welfare policy preferences and racial attitudes toward blacks, the presence of Latinos within a state may also impact welfare generosity and effort. Previous research reports that Latino population size decreases state welfare provisions (Fellowes and Rowe 2004; Hero and Preuhs 2007; Preuhs 2007). We control for the impact of Latino population size using the U.S. Population Estimates generated by the U.S. Census Bureau. The proportions of school-age children, senior citizens, and unemployed workers determine the demand for welfare services. These data are from various years of the Statistical Abstract of the United States. Erikson, Wright, and McIver (1993) and Berry et al. (1998) contend that liberal public opinion results in more liberal state policies. Accordingly, we use Berry et al.’s citizen ideology measure as a control. The measures range from 0 to 100, where higher values indicate higher levels of liberalism.

The second set of variables we employ concerns institutional characteristics. The racial and partisan composition of state legislatures and the partisanship of the governor are likely to be a significant determinant of state expenditures. Preuhs (2006, 2007) shows that black and Latino representation in state legislatures is associated with more provisions for welfare services. Data regarding the racial composition of state legislatures come from the National Association of Latino Elected Officials and the Joint Center for Political and Economic Studies. Research (Barrilleaux, Holbrook, and Langer 2002; Brown 1995; Besley and Case 2003; Smith 1997) also shows that Democratic state government is associated with higher support for welfare programs. Although Lublin (1997) points out instances where the voting patterns of black representatives are clearly distinguishable from white Democrats, we expect Democratic control of the state legislature and the presence of a Democratic governor to be associated with greater levels of spending. Our data on the partisan makeup of state governments come from the National Conference of State Legislatures.

The third set of variables concerns state economic resources for welfare programs. Our model accounts for the total population size of a state, the total amount of federal aid to a state, per capita gross state product (GSP), and state median income for a four-person family. We take a log of the total population size, federal aid, GSP, and median income. These data are from various years of the Statistical Abstract of the United States and U.S. Bureau of the Census.¹⁸ Descriptive statistics of all variables are reported in Table 1.

Table 1.

Summary Statistics

	M	SD	Minimum	Maximum
Log per capita welfare expenditure	5.691	0.458	4.162	6.876
Aid to Families with Dependent Children (AFDC) benefit levels (N = 833)	0.117	0.041	0.037	0.271
Proportion Black	0.097	0.093	0.002	0.365
Proportion Latino	0.056	0.077	0.005	0.427
Black–white income ratio	0.653	0.103	0.445	1.122
Voter registration	0.110	0.313	0.000	1.000
Proportion of whites in poverty	0.097	0.027	0.045	0.191
Proportion of whites with a college degree	0.213	0.051	0.103	0.373
Proportion black legislators	0.060	0.058	0.000	0.265
Proportion Latino legislators	0.018	0.053	0.000	0.402
Proportion Democratic representatives	0.570	0.175	0.129	0.952
Proportion Democratic senators	0.575	0.181	0.086	1.000
Democratic governor	0.533	0.494	0.000	1.000
Proportion unemployed	0.062	0.022	0.020	0.180
Log gross state product per capita	2.789	0.216	2.310	4.043
Log total federal aid per capita	6.166	0.312	4.606	7.811
Citizen ideology	0.480	0.150	0.093	0.939
Log personal income per capita	2.625	0.177	2.150	3.163
Log total population size	8.057	1.014	5.996	10.434
Proportion school-age population	0.192	0.018	0.154	0.268
Proportion 65 years old and over	0.121	0.022	0.029	0.182

Note: N of observations = 1,029. Data cover forty-nine states from 1980 to 2000. Nebraska is excluded because of its nonpartisan legislature.

For examining the impacts of these independent variables on the welfare policies, we use a linear regression model with state and year fixed effects.¹⁹ State fixed effects are expected to control for time-invariant differences across states. For example, cultural differences across states are accounted for by state dummies. Year fixed effects are included for controlling differences across time. For example, the major policy shift from AFDC to TANF in 1996 is accounted for by the year fixed effects. We exploit variation across time within a state for estimation. All standard errors are clustered by states to take account of nonspherical error structures.²⁰

Findings

We begin by examining the effect of black population size on per capita welfare expenditures in the states without considering any possible interactive effects. Our first hypothesis (Hypothesis 1) predicts that black population size is negatively related to welfare spending. The first column in Table 2 reports the results for Hypothesis 1. As expected, the proportion of a state’s population that is black has a negative and statistically significant impact on per capita welfare spending. State governments allocate less money to welfare programs as black group size increases.

Table 2.

The Conditional Impact of Black Group Size on Per Capita Welfare Spending, 1980-2000

	(1)		(2)		(3)		(4)		(5)
	Hypothesis 1		Hypothesis 2		Hypothesis 3		Hypothesis 4		Hypothesis 5
	Regression estimate	SE	Regression estimate	SE	Regression estimate	SE	Regression estimate	SE	Regression estimate	SE
Proportion black	−5.668*	2.100	−12.129*	3.925	−4.917*	1.990	−4.025*	2.139	−5.859*	2.853
Proportion Black × Black–white Income Ratio			9.565*	4.304
Black–white income ratio			0.079	0.151
Proportion Black × No Voter Registration					−9.190	5.716
No voter registration					0.173*	0.087
Proportion Black × Proportion White in Poverty							−25.085*	10.892
Proportion white in poverty							2.348	2.333
Proportion Black × Proportion White College Graduate									1.709	4.769
Proportion white college graduate									−1.281	1.414
Proportion Latino	−3.396*	1.451	−3.015*	1.502	−3.491*	1.408	−3.433*	1.470	−3.019*	1.556
Proportion black legislator	0.623	0.467	0.651	0.426	0.623	0.481	0.503	0.452	0.522	0.559
Proportion Latino legislator	0.564	0.407	0.551	0.417	0.587	0.408	0.529	0.383	0.607	0.416
Proportion Democratic representatives	0.101	0.152	0.124	0.155	0.061	0.149	0.160	0.152	0.105	0.168
Proportion Democratic senators	−0.050	0.122	−0.082	0.114	−0.031	0.120	−0.032	0.115	−0.022	0.125
Democratic governor	0.025	0.015	0.026*	0.015	0.028*	0.015	0.027*	0.015	0.024	0.015
Proportion unemployed	0.606	0.890	0.627	0.870	0.471	0.856	0.694	0.875	0.782	0.878
Log gross state product	0.010	0.132	0.022	0.127	0.027	0.118	−0.015	0.133	0.009	0.126
Log total federal aid	0.451*	0.147	0.454*	0.148	0.439*	0.143	0.466*	0.152	0.449*	0.148
Citizen ideology	0.170	0.122	0.186	0.121	0.179	0.121	0.152	0.113	0.167	0.128
Log income	−0.268	0.214	−0.267	0.219	−0.269	0.211	−0.233	0.216	−0.220	0.212
Log total population	0.853*	0.258	0.732*	0.268	0.827*	0.257	0.876*	0.260	0.844*	0.257
Proportion school-age population	−5.410*	1.715	−4.773*	1.580	−4.932*	1.741	−5.346*	1.717	−5.317*	1.658
Proportion 65 and older population	0.473	2.848	0.371	2.870	−0.145	2.695	−0.083	2.976	0.249	2.821
Constant	−1.654	2.850	−0.575	3.038	−1.611	2.842	−1.937	2.861	−1.591	2.868
R ²	0.928		0.929		0.929		0.929		0.928

Note: Table entries are regression estimates with heteroskedasticity and serial correlation consistent standard errors. The dependent variable is logged per capita expenditures on welfare. Data included forty-nine states from 1980 to 2000. Nebraska was excluded for its nonpartisan legislature. The number of observations is 1,029. Fixed effects (not reported here) are included for states and years.

p < .05 (one-tailed tests).

All columns in Table 2 report that the size of the Latino population is also negatively associated with welfare spending. This result is consistent with Preuhs (2007), indicating that the presence of a large Latino population generates a backlash effect rather than an electoral constraint effect on welfare spending. The results for our other control variables in columns 1 through 5 are mixed. Most are estimated to have an insignificant impact on welfare expenditures, though their coefficients are generally in the expected direction. The proportions of Democratic representatives and senators have no statistically significant impact, yet the presence of Democratic governor has a positive and significant impact on the level of welfare spending. In addition, Table 2 does show that more federal aid, greater population size, and a small school-age population are associated with greater per capita expenditures on welfare programs.²¹

The remaining columns in Table 2 report the results for our conditional hypotheses. We test Hypothesis 2 through Hypothesis 4 by including a series of interaction terms between black group size and socioeconomic and institutional contexts within the states. Column 2 examines Hypothesis 2, which argues that the negative impact of black population size on welfare spending is moderated by greater socioeconomic equality between blacks and whites. The coefficient associated with black group size remains negative and statistically significant after the interaction term between black group size and the black/white income ratio is included in the model. The interaction term, however, is positive. This result suggests that the negative impact of black group size decreases as income equality increases. Table 3 presents the marginal effect of black group size when the income ratio is set at the value of the 5th percentile and the 95th percentile. When income equality is high (ratio = .872), the marginal effect of black group size on welfare spending is negative but much smaller. In sum, these results offer evidence in support of our argument that the negative impact of black group size is moderated by greater socioeconomic equality between blacks and whites.

Table 3.

The Marginal Effect of Black Group Size on per Capita Welfare Spending and Aid to Families with Dependent Children (AFDC) Benefit Levels

	Marginal effect	SE
Per capita welfare spending
Conditions
Hypothesis 2: Black–white income ratio
5th percentile = 0.502	−7.324	2.335
95th percentile = 0.872	−3.787	2.024
Hypothesis 3: Voter registration on election day
No = 0	−4.917	1.990
Yes = 1	−14.107	5.852
Hypothesis 4: Proportion of whites in poverty
5th percentile = 0.057	−5.464	2.113
95th percentile = 0.145	−7.655	2.407
Hypothesis 4: Proportion of whites with a college degree
5th percentile = 0.141	−5.618	2.479
95th percentile = 0.309	−5.331	2.223
AFDC benefit levels
Conditions
Hypothesis 2: Black–white income ratio
5th percentile = 0.502	−0.172	0.189
95th percentile = 0.872	−0.263	0.163
Hypothesis 3: Voter registration on election day
No = 0	−0.134	0.161
Yes = 1	−0.999	0.275
Hypothesis 4: Proportion of whites in poverty
5th percentile = 0.057	−0.158	0.166
95th percentile = 0.145	−0.376	0.170
Hypothesis 4: Proportion of whites with a college degree
5th percentile = 0.141	−0.560	0.226
95th percentile = 0.309	−0.405	0.188
Conditions
Hypothesis 2: Black–white income ratio
5th percentile = 0.502	−0.172	0.189
95th percentile = 0.872	−0.263	0.163
Hypothesis 3: Voter registration on election day
No = 0	−0.134	0.161
Yes = 1	−0.999	0.275
Hypothesis 4: Proportion of whites in poverty
5th percentile = 0.057	−0.158	0.166
95th percentile = 0.145	−0.376	0.170
Hypothesis 4: Proportion of whites with a college degree
5th percentile = 0.141	−0.560	0.226
95th percentile = 0.309	−0.405	0.188

Note: Table entries are the marginal effects of black group size on logged per capita welfare spending and AFDC benefit levels with standard errors. The marginal effects are computed using the estimated results in Tables 2 and 4.

We next turn to the competing expectations we discussed earlier regarding the possible conditioning impact of voter registration laws. Hypothesis 3a predicts that the negative impact of black population size on policy liberalism is moderated in states with less restrictive registration laws. Conversely, Hypothesis 3b predicts that the negative effect of black population size on policy liberalism is stronger in states with less restrictive registration laws. In column 3 of Table 2 we see that the coefficient associated with black group size is negative, as is the interaction term between the black group size and the presence of same day voter registration. The coefficient of the interaction term is statistically significant at the 0.10 level. These results indicate that the negative impact of black group size is stronger in states that offer same-day voter registration. Table 3 reports the marginal effects of black group size in states with and without same day voter registration systems. The negative impact of black group size is strongly enhanced by same day registration. These findings suggest that less restrictive registration systems increase electoral involvement among whites, resulting in a more conservative electorate. Thus, our analysis supports Hypothesis 3b. Note that states that allow voter registration on election day are characterized by a relatively small black population.²² If states with a large black population allowed voter registration on election day and the pattern in Table 2 held, we might see a stronger negative effect of black group size on state welfare support.

Columns 4 and 5 in Table 2 examine our fourth hypothesis (Hypothesis 4), which holds that the negative effect of black group size is stronger in states where whites enjoy fewer socioeconomic resources. We measure socioeconomic resources among whites using two indicators: the proportions of whites living in poverty and the proportion of whites with a college education. The results presented in column 4 find that the coefficients associated with black group size and the interaction term between black group size and the proportion of whites living in poverty are both negative. Both coefficients are also statistically significant. Table 3 reports the marginal effect of black group size when the poverty level is set at the 5th and 95th percentiles (5.7 percent and 14.5 percent, respectively). As expected, the negative impact of black group size is much stronger in states with a greater proportion of poor whites. When whites enjoy fewer socioeconomic resources, they tend to express more negative attitudes toward blacks, which enhance the negative effect of policy backlash. This result supports Hypothesis 4.

We do not find strong support in Table 2 for Hypothesis 4 when we consider the level of education among whites. The interaction term between black group size and proportion of whites with college degree shows a positive but statistically insignificant impact. This result is reflected in the marginal effect in Table 3. When the proportion of whites who are college educated increases from the value of the 5th percentile to the 95th percentile (14.1 percent and 30.9 percent, respectively), the negative effect of black population size decreases to some extent. Yet, the marginal effect of black group size does not differ dramatically between the two conditions. In sum, we find that the negative reaction of whites to the increased presence of blacks is largely determined by a lack of economic resources among whites.

We next replicate the previous analysis using our other measure of state welfare policy: the maximum level of AFDC benefits. Table 4 reports the estimated results where the dependent variable is the ratio of annual maximum AFDC benefit levels to median family income. The same set of the independent variables are included in our regression models. Column 1 of Table 4 reports that black group size has a negative effect, but the estimate is not statistically significant at the conventional level. However, the interaction terms between black group size and socioeconomic and institutional contexts within the states are statistically significant for the most part and consistent with the findings in Table 2. The only insignificant result is found in column 2 for the interaction between black group size and the black/white income ratio.

Table 4.

The Conditional Impact of Black Group Size on Aid to Families with Dependent Children (AFDC) Benefit Levels, 1980-1996

	(1)		(2)		(3)		(4)		(5)
	Hypothesis 1		Hypothesis 2		Hypothesis 3		Hypothesis 4		Hypothesis 5
	Regression estimate	SE	Regression estimate	SE	Regression estimate	SE	Regression estimate	SE	Regression estimate	SE
Proportion black	−0.196	0.180	−0.048	0.355	−0.134	0.161	−0.015	0.189	−0.689*	0.273
Proportion Black × Black–white Income Ratio			−0.247	0.421
Black–white income ratio			0.072*	0.026
Proportion Black × No Voter Registration					−0.865*	0.239
No voter registration					0.008	0.008
Proportion Black × Proportion White in Poverty							−2.495*	1.051
Proportion white in poverty							0.082	0.203
Proportion Black × Proportion White College Graduate									0.920*	0.486
Proportion white college graduate									0.136	0.158
Proportion Latino	−0.276*	0.109	−0.310*	0.101	−0.298*	0.106	−0.272*	0.104	−0.318*	0.116
Proportion Black legislator	0.045	0.038	0.034	0.037	0.041	0.038	0.046	0.031	0.018	0.040
Proportion Latino legislator	0.025	0.016	0.026	0.017	0.026	0.017	0.024	0.015	0.024	0.015
Proportion Democratic representatives	−0.022*	0.012	−0.018*	0.010	−0.022*	0.012	−0.017	0.012	−0.011	0.012
Proportion Democratic senators	0.001	0.009	−0.001	0.008	0.002	0.009	0.000	0.008	0.001	0.009
Democratic governor	−0.001	0.001	−0.000	0.001	−0.001	0.001	−0.000	0.001	−0.000	0.001
Proportion unemployed	0.068	0.066	0.082	0.062	0.061	0.064	0.060	0.065	0.048	0.064
Log gross state product	−0.015	0.014	−0.015	0.014	−0.016	0.014	−0.020	0.013	−0.018*	0.010
Log total federal aid	0.018*	0.007	0.017*	0.007	0.017*	0.007	0.020*	0.008	0.018*	0.007
Citizen ideology	−0.002	0.010	−0.006	0.010	−0.002	0.010	−0.001	0.010	−0.006	0.011
Log income	−0.112*	0.011	−0.108*	0.009	−0.112*	0.011	−0.111*	0.012	−0.114*	0.013
Log total population	0.095*	0.018	0.096*	0.016	0.094*	0.017	0.091*	0.016	0.095*	0.015
Proportion school-age population	−0.501*	0.098	−0.538*	0.091	−0.488*	0.099	−0.457*	0.099	−0.444*	0.093
Proportion 65 and over population	0.092	0.148	−0.066	0.128	0.045	0.139	0.075	0.139	0.161	0.144
Constant	−0.263	0.175	−0.297*	0.165	−0.252	0.170	−0.224	0.179	−0.198	0.170
R ²	0.959		0.961		0.960		0.960		0.961

Note: Table entries are regression estimates with heteroscedasticity and serial correlation consistent standard errors. The dependent variable is the ratio of annual AFDC benefit levels for a family of three to the median family income. Data included forty-nine states from 1980 to 1996. Nebraska was excluded for its nonpartisan legislature. The number of observations is 833. Fixed effects not reported here) are included for states and years.

p < .05 (one-tailed tests).

The results for our other control variables in Table 4 are similar to those in Table 2. The size of the Latino population has a negative and statistically significant impact on AFDC benefit levels in all columns. Unlike the results in Table 2, Table 4 reports that the proportion of Democratic representatives has a negative and significant impact on AFDC benefit levels. The presence of a Democratic governor has no significant impact. More federal aid, lower personal income, greater population size, and a small school-age population are associated with greater benefits.

Column 3 of Table 4 also shows that the interaction term between black group size and the presence of same day voter registration is negative and statistically significant. Table 3 reports the marginal effects of black group size in states with and without same day voter registration systems. We see no significant effect of black group size on the benefit levels in states without same day voter registration. However, the effect of black group size is negative and statistically significant in states with same day voter registration. Columns 4 and 5 report that the interaction term between black group size and the proportion of whites living in poverty is negative and significant, while the interaction term between black group size and the proportion of whites with a college education is positive and significant. Table 3 indicates that the negative effect of black group size on AFDC benefit levels is enhanced in states with less affluent or less educated whites. That is, the effect of black group size is conditional on the socioeconomic environments of whites. In short, these results offer evidence for Hypothesis 3, Hypothesis 4, and Hypothesis 5.

Conclusion

In their seminal work, Erikson, Wright, and McIver (1993) show that public policies within the states are remarkably responsive to public opinion. Working from the assumption that political actors react to changes in the policy preferences of citizens, some scholars argue that states with large minority populations are more likely to produce racially liberal public policies. Others, as we note, suggest the opposite. According to this view, large minority populations produce a sense of racial threat among whites. The net result is less liberal policies in states with large minority populations.

The current literature offers empirical evidence for each perspective; however, both tend to oversimplify the literature on how racial/ethnic context influences racial attitudes and policy preferences. We argue that preferences on racialized policies within a state are determined by the interaction of black group size with institutional and socioeconomic conditions.

A large black population is not necessarily associated with conservative policies when the income disparity between blacks and whites is relatively low. We also find that the presence of blacks has an enhanced negative effect on policy liberalism when whites enjoy fewer socioeconomic resources. These findings suggest that the relationship between racial context and policy outputs is not as simple as other research thinks it to be. Early scholars (Allport [1954] 1979) and recent empirical work (Branton and Jones 2005) argue that the relationship between minority group size and racial attitudes in contingent upon socioeconomic context. Our study offers the first attempt to apply this argument to the literature on minority representation at the state level.

In a departure from previous works, we maintain that the presence of lax voter registration laws does not allow minorities to more easily translate their group size into favorable political outcomes. States with large minority populations enacted racially conservative policies before the passage of the Voting Rights Act as minorities had few means to hold elected officials accountable. Since the mid-1960s, the increased accessibility of the ballot has been thought to reverse the effect of racial context on politics at the state level (see Fording 1997). Recent studies, however, show that less restrictive voter registration laws benefit economically advantaged groups. Applying this finding to the study of state racial policy outputs, our empirical results show that the effect of black group size is exacerbated in states with no or same-day voter registration.

The descriptive representation of minorities has increased in recent years along with minority voter enfranchisement. Still, the ability of minorities to translate access to the ballot into substantive outcomes remains unclear. Our research suggests that the political influence of minorities varies considerably in the states. Moreover, forces that scholars might expect to have a liberalizing effect on racial policy outputs at the subnational level may produce unintended consequences. Lax voter registration laws do not enhance the political power of blacks, at least in the domain of state welfare policies. The presence of minorities can affect the political discourse surrounding racialized policy issues and this can in turn affect public opinion all groups.

The dynamics of racial politics in the United States are seldom easy to decipher. Relationships tend to be complicated and are often conditional. The result is often a series of mixed findings within the literature. In this article, we point to the relationship between racial context and public policy outputs in the states as one example. It is essential that future research consider the role of socioeconomic conditions and electoral institutions in shaping the effects of racial context.

Footnotes

Declaration of Conflicting Interests

The authors declared that they had no conflicts of interests with respect to their authorship or the publication of this article.

Funding

The authors received no financial support for the research and/or authorship of this article.

Notes

References

Alesina

Alberto

Baqir

Reza

Easterly

William

. 1999. “Public Goods and Ethnic Divisions.” Quarterly Journal of Economics 114:1243-284.

Alesina

Alberto

Baqir

Reza

Easterly

William

. 2000. “Redistributive Public Employment.” Journal of Urban Economics 48:219-41.

Arellano

1987. “Computing Robust Standard Errors for within-Groups Estimators.” Oxford Bulletin of Economics and Statistics 49:431-34.

Barrilleaux

Charles

Holbrook

Thomas

Langer

Laura

. 2002. “Electoral Competition, Legislative Balance, and American State Welfare Policy.” American Journal of Political Science 46:415-27.

Berry

William D.

Fording

Richard C.

Hanson

Russell L.

2000. “An Annual Cost of Living Index for the American States, 1960–1995.” Journal of Politics 62:550-67.

Berry

William D.

Ringquist

Evan J.

Fording

Richard C.

Hanson

Russell L.

1998. “Measuring Citizen and Government Ideology in the American States, 1960-93.” American Journal of Political Science 42:327-48.

Besley

Timothy

Case

Anne

. 2003. “Political Institutions and Policy Choices: Evidence from the United States.” Journal of Economic Literature 52:7-73.

Blalock

Hubert M.

1967. Toward a Theory of Minority-Group Relations. New York: John Wiley.

Branton

Regina P.

Jones

Bradford S.

2005. “Reexamining Racial Attitudes: The Conditional Relationship between Diversity and Socioeconomic Environment.” American Journal of Political Science 49:359-72.

10.

Bratton

Kathleen A.

Haynie

Kerry L.

1999. “Agenda Setting and Legislative Success in State Legislatures: The Effects of Gender and Race.” Journal of Politics 61:658-79.

11.

Brians

Craig Lernard

Grofman

Bernard

. 2001. “Election Day Registration’s Effect on U.S. Voter Turnout.” Social Science Quarterly 82:170-83.

12.

Brown

Robert D.

1995. “Party Cleavages and Welfare Effort in the American States.” American Political Science Review 89:23-33.

13.

Canon

David T.

1999. Race, Redistricting, and Representation: The Unintended Consequences of Black Majority Districts. Chicago: University of Chicago Press.

14.

Douzet

Frederick

. 2008. “Residential Segregation and Political Balkanization.” In The New Political Geography of California, eds. Douzet

Frederick

Kousser

Thad

Miller

Kenneth P.

(pp 45-70). Berkley Public Policy Press:Berkley, CA.

15.

Erikson

Robert S.

Wright

Gerald C.

McIver

John P.

1993. Statehouse Democracy. New York: Cambridge University Press.

16.

Fellowes

Matthew C.

Rowe

Gretchen

. 2004. “Politics and the New American Welfare States.” American Journal of Political Science 48:362-73.

17.

Fording

Richard C.

1997. “The Conditional Effect of Violence as a Political Tactic: Mass Insurgency, Welfare Generosity, and Electoral Context in the American States.” American Journal of Political Science 41:1-29.

18.

Fording

Richard C.

2001. “The Political Response to Black Insurgency: A Critical Test of Competing Theories of the State.” American Political Science Review 95:115-30.

19.

Gay

Claudine

. 2006. “Seeing Difference: The Effect of Economic Disparity on Black Attitudes toward Latinos.” American Journal of Political Science 50:982-97.

20.

Gilens

Martin

. 1999. Why Americans Hate Welfare? Chicago: University of Chicago Press.

21.

Giles

Michael W.

Hertz

Kaenan

. 1994. “Racial Threat and Partisan Identification.” American Political Science Review 88:317-26.

22.

Glaser

James M.

2003. “Social Context and Inter-Group Political Attitudes: Experiments in Group Conflict Theory.” British Journal of Political Science 33:607-20.

23.

Goren

Paul

. 2008. “The Two Faces of Government Spending.” Political Research Quarterly 61:147-57.

24.

Griffin

John David

Newman

Brian

. 2008. Minority Report. Chicago: University of Chicago Press.

25.

Hanmer

Michael J.

2009. Discount Voting: Voter Registration Reforms and their Effects. New York: Cambridge University Press.

26.

Haynie

Kerry Lee

. 2001. African American Legislators in the American States. New York: Columbia University Press.

27.

Hero

Rodney E.

1998. Faces of Inequality. New York: Oxford University Press.

28.

Hero

Rodney E.

2007. Racial Diversity and Social Capital. New York: Cambridge University Press.

29.

Hero

Rodney E.

Preuhs

Robert R.

2007. “Immigration and the Evolving American Welfare State: Examining Policies in the U.S. State.” American Journal of Political Science 51:498-517.

30.

Highton

Benjamin

Wolfinger

Raymond

. 1998. “Estimating the Effects of the National Voter Registration Act of 1993.” Political Behavior 20:79-104.

31.

Hill

Kim Quaile

Leighley

Jan E.

1992. “The Policy Consequences of Class Bias in State Electorates.” American Journal of Political Science 36:351-65.

32.

Hill

Kim Quaile

Leighley

Jan E.

1996. “Political Parties and Class Mobilization in Contemporary United States Elections.” American Journal of Political Science 40:787-804.

33.

Hill

Kim Quaile

Leighley

Jan E.

1999. “Racial Diversity, Voter Turnout, and Mobilizing Institutions in the United States.” American Politics Quarterly 27:275-95.

34.

Hill

Kim Quaile

Leighley

Jan E.

Anderson

Angela Hinton

. 1995. “Lower-Class Mobilization and Policy Linkage in the U.S. States.” American Journal of Political Science 39: 75-86.

35.

Husted

Thomas A.

Kenny

Lawrence W.

1997. “The Effect of the Expansion of the Voting Franchise on the Size of Government.” Journal of Political Economy:54-82.

36.

Johnson

Martin

. 2001. “The Impact of Social Diversity and Racial Attitudes on Social Welfare Policy.” State Politics and Policy Quarterly 1:27-49.

37.

Judson

Ruch A.

Owen

Ann L.

1999. “Estimating Dynamics Panel Data Models: A Guide for Macroeconomists.” Economics Letters 65:9-15.

38.

Kellstedt

Paul M.

2003. The Mass Media and the Dynamics of American Racial Attitudes. New York: Cambridge University Press.

39.

Key

V O.

[1949] 1984. Southern Politics in State and Nation. Knoxville: University of Tennessee Press.

40.

Kinder

Donald R.

Winter

Nicholas

. 2001. “Exploring the Racial Divide: Blacks, Whites, and Opinion on National Policy.” American Journal of Political Science 45:439-56.

41.

Kiviet

Jan F.

1995. “On Bias, Inconsistency, and Efficiency of Various Estimators in Dynamic Panel Data Models.” Journal of Econometrics 68:53-78.

42.

Leighley

Jan E.

2001. Strength in Numbers? The Political Mobilization of Racial and Ethnic Minorities. Princeton, NJ: Princeton University Press.

43.

Leighley

Jan E.

Vedlitz

Arnold

. 1999. “Race, Ethnicity, and Political Participation.” Journal of Politics 61:1092-114.

44.

Levin

Andrew

Lin

Chien-Fu

Chu

Chia-Shang James

. 2002. “Unit Root Tests in Panel Data: Asymptotic and Finite-Sample Properties.” Journal of Econometrics 108:1-24.

45.

Lind

Jo Thori

. 2007. “Fractionalization and the Size of Government.” Journal of Public Economics 91:51-76.

46.

Lublin

David

. 1997. The Paradox of Representation. Princeton, NJ: Princeton University Press.

47.

Meltzer

Allan H.

Richard

Scott F.

1981. “A Rational Theory of the Size of Government.” Journal of Political Economy 89:914-27.

48.

Nagler

Jonathan

. 1991. “The Effect of Registration Laws and Education on U.S. Voter Turnout.” American Political Science Review 85:1393-405.

49.

Nelson

Albert J.

1991. Emerging Influentials in State Legislatures. New York: Praeger.

50.

Oliver

J. Eric

. 1996. “The Effects of Eligibility Restrictions and Party Activity on Absentee Voting and Overall Turnout.” American Journal of Political Science 40:498-513.

51.

Oliver

J. Eric

Mendelberg

Tali

. 2000. “Reconsidering the Environmental Determinants of Racial Attitudes.” American Journal of Political Science 44:574-89.

52.

Oliver

J. Eric

Wong

Janelle

. 2003. “Intergroup Prejudice in Multiethnic Settings.” American Journal of Political Science 47:567-82.

53.

Preuhs

Robert R.

2006. “The Conditional Effects of Minority Descriptive Representation.” Journal of Politics 68:585-99.

54.

Preuhs

Robert R.

2007. “Descriptive Representation as a Mechanism to Mitigate Policy Backlash.” Political Research Quarterly 60:277-92.

55.

Quillian

Lincoln

. 1995. “Prejudice as a Response to Perceived Group Threat.” American Sociological Review 60: 586-611.

56.

Quillian

Lincoln

. 1996. “Group Threat and Regional Change in Attitudes toward African Americans.” American Journal of Sociology 102:816-60.

57.

Radcliff

Benjamin

Saiz

Martin

. 1995. “Race, Turnout, and Public Policy in the American States.” Political Research Quarterly 44:775-94.

58.

Rocha

Rene R.

Espino

Rodolfo

. 2009. “Racial Threat, Residential Segregation, and the Policy Attitudes of Anglos.” Political Research Quarterly 62: 415-426.

59.

Smith

Mark A.

1997. “The Nature of Party Governance: Connecting Conceptualization and Measurement.” American Journal of Political Science 41:1042-056.

60.

Sniderman

Paul M.

Brody

Richard A.

Tetlock

Philip E.

1991. Reasoning and Choice. New York: Cambridge University Press.

61.

Soss

Joe

Schram

Sanford F.

Vartanian

Thomas P.

O’Brien

Erin

. 2001. “Setting the Terms of Relief: Explaining State Policy Choices in the Devoluation Revolution.” American Journal of Political Science 45:378-95.

62.

Squire

Peverill

. 2007. “Measuring State Legislative Professionalism: The Squire Index Revisited.” State Politics and Policy Quarterly 7:211-27.

63.

Swain

Carol M.

1993. Black Faces, Black Interests. Cambridge, MA: Harvard University Press.

64.

Taylor

Marylee C.

1998. “How White Attitudes Vary with the Racial Composition of Local Populations: Numbers Count.” American Sociological Review 63:512-35.

65.

Tweedie

Jack

. 1994. “Resources Rather than Needs: A State-Centered Model of Welfare Policymaking.” American Journal of Political Science 38:651-72.

66.

Verba

Sidney

Schlozman

Kay Lehman

Brady

Henry E.

1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge, MA: Harvard University Press.

67.

Verba

Sidney

Schlozman

Kay Lehman

Brady

Henry

Nie

Norman H.

1993. “Race, Ethnicity, and Poli- tical Resources.” British Journal of Political Science 23: 453-97.

68.

Wolfinger

Raymond E.

Hoffman

Jonathan

. 2001. “Registering and Voting with Motor Voter.” Political Science & Politics 34:85-92.

69.

Wolfinger

Raymond E.

Rosenstone

Steven J.

1980. Who Votes? New Haven, CT: Yale University Press.

70.

Yates

Jeff

Fording

Richard

. 2005. “Politics and State Punitiveness in Black and White.” Journal of Politics 67:1099-121.