Policymaking by Other Means

Abstract

While politicians wage legislative battles about welfare benefits, bureaucratic procedures represent a less visible means to shape access to those benefits. By constructing complex and time-consuming application procedures, the state can effectively create administrative barriers that limit access to benefits. What explains the variation in the level of administrative burden that individuals face? We address this question by examining Medicaid procedures across the 50 U.S. states, identifying different types of administrative barriers. We find that such barriers tend to be lower in states with unified Democratic control of political institutions.

Keywords

administrative rules social policy citizen–state interaction

Introduction

While politicians wage ongoing legislative battles over which social benefits the poor should receive, bureaucratic procedures represent an alternative—and less visible—means of shaping access to those benefits. By constructing complex, confusing, and time-consuming application procedures, the state can effectively thwart an individual from accessing benefits, even if eligible by law. We examine variation in the level of administrative burden that individuals face when they seek to access social benefits, as well as how political factors may shape the level of administrative burden states impose on their citizens. Better understanding of this variation has broad relevance to studies of the political economy of poverty and inequality (Jacobs & Soss, 2010), social policy (Brodkin & Majmundar, 2010), public administration (Moynihan & Herd, 2010), and citizen–state interactions (Mettler & Soss, 2004).

Specifically, this article documents, compares, and explains variation in formal application requirements (i.e., compliance administrative burden) experienced by individuals accessing a standard federal program across U.S. state governments: Medicaid. Medicaid is a means-tested health insurance program for individuals who meet certain income and asset guidelines. The costs of the program are shared between the state and federal governments, and states have a great deal of autonomy in structuring their Medicaid programs—particularly their administrative procedures. Indeed, we find significant variation in administrative burden across states and that the level of burden differs in states governed by Democrats versus Republicans in the previous decade. This illustrates how political factors may play a role in structuring the administrative burden faced by applicants for Medicaid, as well as how states use administrative burdens as an additional (often overlooked) policy tool, or what Lineberry (1977) refers to as “policymaking by other means” (p. 71).

When Do Burdens Become Red Tape?

In this article, we conceptualize our dependent variable in terms of administrative burdens placed on citizens. We define administrative burden as an individual’s experience of policy implementation as onerous (Burden, Canon, Mayer, & Moynihan, 2012). A broad stream of literature focuses on the discretion of street-level government actors in implementing government policies (Lipsky, 1980, 1984) and how citizens are classified by various policies (Schneider & Ingram, 1997), but relatively little research has considered the origins of systematic administrative compliance burdens that citizens face in their interactions with the state.

Perhaps the closest concept to administrative burden, coming from public administration scholarship, is that of red tape. Even in this research area, the definition and study of this concept has been relatively narrow. Red tape is defined by Bozeman (2000) as “rules, regulations, and procedures that remain in force and entail a compliance burden, but do not advance the legitimate purposes the rules were intended to serve” (p. 12). This definition suggests that any rule that advances a legitimate purpose cannot be classified as red tape. A recent reevaluation of the definition concluded that

it is difficult to find rules that had absolutely no functionality or no positive outcomes. Rules that provided some functionality but also disproportionate burdens would, in the classic definition, not be considered red tape. For example, many rules that serve their purpose of limiting fraud in the claiming of public benefits also create significant barriers to access among legitimate claimants. When do the negative impacts of such a rule lead it to being classified as red tape? (Feeney, Moynihan, & Walker, 2010, p. 4)

Perhaps because of the restrictive nature of the definition, little attention has been paid to the actual rules that citizens face; instead, the extant research has been centered on managerial perceptions of red tape (Bozeman & Feeney, 2011).

To explain administrative burden, we focus on the political economy of rules affecting actors outside the organization. Theory on red tape emphasizes the potential for rules that once served a purpose to evolve into red tape over time (rule-evolved red tape). This insight is important, because natural bureaucratic processes can potentially allow problematic rules to continue. Bad rules are not necessarily designed to be bad, but simply have become less useful over time. Red tape research, however, has not explored the potential for such rules to be used as a deliberate policy tool or the possibility that they could have partisan origins. Both the rule-evolved and political economy perspectives can help explain why citizens face onerous rules. But from the citizen’s point of view, conceptual debates about the appropriate dividing line between red tape and administrative burden are beside the point. While citizens might recognize and understand the basic need for procedural fairness (Tyler & Lind, 1992), their perceptions are driven largely by the nature of their experiences with bureaucratic processes, as the next section describes.

How Administrative Burden Matters

The impact of administrative burdens on citizens has been documented in two distinct research areas. Welfare studies show that rules and their implementation significantly affect the ability of eligible claimants to access entitlements. Policy feedback research indicates that citizens who experience negative policy interactions with the state have reduced political efficacy. We examine each in turn.

Several social-welfare studies have examined how program structure alters the access citizens have to benefits and the rate of uptake among legitimate claimants. Much of this research depends on comparisons between means-tested and universal programs, showing that means-tested policies that target poorer populations are more complex, requiring demanding eligibility forms and procedures (Korpi & Palme, 1998), despite offering fewer benefits.

Empirical research has shown that the compliance burdens associated with means-tested programs have a real impact in limiting access to benefits. While almost all individuals eligible for universal programs (such as Medicare and Social Security) receive benefits, only 40% to 60% of those eligible for means-tested Supplemental Social Insurance and about 25% of those who qualify for Medicaid receive benefits (Elder & Powers, 2006; Shore-Sheppard, 2008). Just over a third of those eligible for food stamps do not receive them (Food & Nutrition Service, 2007).

Some percentage of nonparticipation in such programs may be truly voluntary, but the administrative complexity of means-tested programs has been found to explain a large portion of nonuse of benefits (for a review, see Currie, 2006). For example, Bennett (1995) shows that more than a quarter of welfare case closings (removing the claimant from the public rolls) in her sample were attributable to problems the claimants had with documentation rather than to changes in eligibility. Brodkin and Majmundar (2010) further find that procedural barriers faced by claimants explain a significant amount of the decline in welfare caseloads, and that these barriers systematically affect some groups—poorer and less educated claimants—more than others.

Burdens may not just affect access to benefits but also the experience that the citizen has of the state and how he or she responds to that experience (Moynihan & Herd, 2010). Through policy rules, citizens are taught lessons regarding their worth and standing (Schneider & Ingram, 1997). Studies that examine the policy feedback effect of programs on citizens often compare citizen experiences associated with means-tested versus universal programs, while attempting to deal with selection bias by controlling for citizen characteristics. Such research shows that citizens with experiences related to means-tested programs express a lower sense of political influence (Soss, 1999), are less likely to engage in a variety of political actions such as voting or campaigning (Verba et al., 1995), and have lower levels of social trust (Kumlin & Rothstein, 2005). Using a model that controls for the potential selection bias issues that had limited prior analyses, Bruch, Marx-Freere, and Soss (2010) find that citizens who interacted with the Temporary Assistance to Needy Families (TANF) program had lower voting, political participation, civic participation and political engagement relative to those in other means-tested programs. They also determine that these effects were largest in states that were the most paternalistic (in terms of being more active in regulating the behavior of program recipients). They conclude that “social-welfare program designs can influence levels of civic and political engagement among the poor” (Bruch et al., 2010, p. 219).

Application Compliance Burden

Administrative burdens come in a variety of forms. For example, it can be difficult to learn about a program, or whether one is eligible. In addition, applying for and receiving benefits can confer a stigma and a psychological burden. But, here, we study application compliance burden, which we define as the costs borne by individuals when completing application and recertification processes, including required documentation and testing, and responding to or avoiding the discretionary demands of program staff. While the empirical aspect of our analysis focuses only on formal documentation and process requirements, we do not mean to imply that bureaucrats are unquestioning implementers. Any understanding of application compliance burden would be incomplete without some consideration of bureaucratic discretion. Lipsky (1984) characterizes efforts to use administrative burden to reduce take-up of welfare benefits as one form of “bureaucratic disentitlement,” and he finds that this is more likely to occur when resources are limited (see also Keiser & Soss, 1998). Such tools have been used in welfare programs as a deliberate means of social control and disciplining of the poor, especially minorities (Soss, Fording, and Schram, 2011), and a central theme of street-level bureaucracy research is how the characteristics of claimants shape the decisions of bureaucrats. For example, Scott (1997) uses an experimental design to show that the level of organizational control and client characteristics matter to bureaucratic use of discretion in assigning benefits.

In practice, it is difficult to separate the effects of bureaucratic discretion in assigning benefits from the policies and rules they implement. Bureaucrats can alter application compliance burdens through their choices about what information to share with claimants, for example, or by the amount of documentation they demand, or how helpful they are when a claimant is completing forms. Previous studies demonstrate both how much application enrollment burdens limit access and just how important a helpful bureaucratic advocate can be in overcoming these burdens. For example, Schanzenbach (2009) describes an experiment in which a sample of taxpayers was provided assistance in completing Medicaid forms in California. She finds that those who received such assistance were almost 80% more likely to actually apply for benefits than those who were informed they were eligible but given no special assistance. A similar experiment among people applying for federal financial aid for education finds that those given assistance in completing forms were more likely to send in the applications, leading to a 29% increase in actual college enrollment (Bettinger, Long, Oreopoulos, & Sanbonmatsu, 2012).

The formal rules that create administrative burden is an understudied topic that falls between two often-studied streams of literature: policy formation and administrative behavior. For example, we know a good deal about why welfare policies are created to be more or less generous (e.g., Klarner, Mao, & Buchanan, 2007). The exercise of bureaucratic discretion in policy implementation is also well-documented enough to feature a robust literature on street-level bureaucracy (Loyens & Maesschalck, 2010), as well other approaches to studying bureaucratic discretion. But in between policy formation and implementation, blurring both categories, are the administrative rules and procedures intended to reflect policy, constrain bureaucrats, and shape the experiences of citizens in their encounters with the state. These burdens may be mediated or exacerbated by bureaucratic action and citizen capacities, but they are worth studying because they serve as the framework for bureaucratic action and have demonstrable effects in their own right.

Bureaucratic implementation of program eligibility requirements can vary—affecting take-up and making it difficult to separate bureaucratic discretion from other aspects of the administrative burden placed on program claimants. However, the nature of formal documentation requirements has been shown to matter. For example, natural experiments have shown that the introduction of requirements for income documentation results in considerable decline in program participation among eligible participants across time (Brien & Swann, 2001). Requiring face-to-face interviews and failing to provide applicants with assistance also significantly decrease participation (Wolfe & Scrivner, 2005). In the case of food stamps, requiring less frequent recertification procedures (Klerman & Danielson, 2011) and simplifying the reporting procedures results in more successful claimants (Hanratty, 2006). On the other hand, fingerprinting of applicants leads to lower rates of Food Stamp application completion (Bartlett, Burstein, & Hamilton, 2004), and in surveys of nonparticipants, respondents emphasize their desire to avoid the burdens and intrusiveness of application processes (Bartlett & Burstein, 2004).

Administrative burden can be found across a range of different social-welfare programs, but we focus on Medicaid for three reasons. First, the program has substantial variation in administrative rules and eligibility procedures across states, allowing us to examine the factors that shape this variation. Programs like Medicare and food stamps, which are largely controlled by the federal government, lack this kind of variation. Second, the program is large and important. Medicaid provides health insurance to nearly 20% of all Americans and a substantially larger percentage of Americans have relied on the program at some point in their lives. Finally, the passage of the Patient Protection and Affordable Care Act includes a large expansion of Medicaid, but also provides states with implementation flexibility for these expansions.

Administrative Burden as Policymaking by Other Means

Even within relatively similar policy requirements, public actors often retain a good deal of autonomy in structuring and interpreting administrative details in ways that matter to citizens. This is not simply a matter of discretion about implementing rules; rather, agencies pursuing common policies have distinct formal rule structures that alter citizens’ experiences of the state.

The procedural compliance burdens placed on citizens give rise to some of the most common complaints about the public sector: that it is bureaucratic, slow, unresponsive, and rule oriented. When rules leave citizens dissatisfied or substantively restrict citizenship rights, these rules have negative impacts—not just for individuals but for governance legitimacy more broadly. (Moynihan & Herd, 2010, p. 656)

The primary hypothesis of this article is that administrative rules in social policy implementation represent an extension of political preferences. As such, variation in these rules would reflect state political context, rather than economic or other factors that could provide less-politicized rationales for expanding or limiting access to Medicaid. One might think that such an idea is already well-documented, but that is not the case. Certainly, empirical evidence shows that politics impacts welfare choices (Hacker, 2002). For example, Klarner et al. (2007) demonstrate that the power of business interests explains variation in the generosity of TANF policies across states. Keiser and Soss (1998) find that partisan control of government matters to bureaucratic discretion in terms of granting benefits, but not in terms of how vigorously bureaucrats restrict claims.

Analogous debates take place in other policy areas in ways that reflect partisan preferences. For instance, Republicans generally argue that voter identification laws are necessary to prevent fraud, while Democrats argue that such rules serve mainly as a burden that reduces turnout among eligible voters (Moynihan & Herd, 2010). As a result, voter identification laws are more likely to be adopted in states where Republicans control policymaking institutions (Hale & McNeal, 2010). Though such research generally overlooks administrative rules, a few types of rules have garnered attention. For example, Yackee (2006) identifies how business interests influence the federal rulemaking process. But unlike the rulemaking process, which creates formalized structures through which interests can directly influence the process, the politics of state-created administrative rules for welfare programs is less clear.

Theoretical Expectations

Why do some states have complex administrative requirements that seem designed to discourage citizens from accessing benefits, while other states take care to make the process administratively simple? In general, those who lack political power or who are seen as undeserving tend to be less successful in winning benefits from the policy process (Schneider & Ingram, 1997). Some circumstantial evidence indicates that administrative burden is deliberately used to limit those rights. Brodkin (1987) notes that as far back as the Nixon administration, welfare state programs have not been designed to balance take-up by eligible claimants with mistaken payments to ineligible beneficiaries; instead, administrative procedures have been used to reduce the former to protect against the latter. Historically, federal quality control guidelines offered states much stronger incentives to avoid overpayment rather than to enroll eligible participants (Hanratty, 2006). Such a tendency remains in contemporary performance evaluations of welfare programs, where reducing fraud is often an important goal, but beneficiary take-up is neglected (Wichowsky & Moynihan, 2008).

To some degree, federal policies are shifting toward making administrative barriers less burdensome through a mixture of flexibility, incentives, and restrictions. For example, since 1999, the federal government has reduced sanctions on states that have pursued more streamlined application procedures in food stamps programs. The 2009 Children’s Health Insurance Program Reauthorization Act (CHIPRA) authorizes states to use new enrollment policy options designed to facilitate greater enrollment. The 2010 Patient Protection and Affordable Care Act prohibits states from making their medical assistance “eligibility standards, methodologies, or procedures” more restrictive under the threat of losing federal matching funds. Even if a tide-change is occurring in how the federal government thinks about the trade-off between enrollment and fraud in welfare programs, the intergovernmental nature of the U.S. system will still give states significant discretion, and we expect that political conditions at the state level will shape how these programs are implemented.

If politics shapes the nature of administrative rules, we expect that in the case of welfare programs, states with more homogeneous populations, more liberal political cultures, and more Democratic control of political institutions will produce policies more aligned with the interests of Medicaid clients, and will be less likely to establish administrative barriers to services. Our expectation regarding racial minorities stems from research demonstrating how racial stereotypes affect perceptions of welfare programs (Gilens, 1999; Rigby, Soss, Booske, Rohan, & Robert, 2009) and tolerance for inequality (Glazer, 2003). Where citizens are inclined to believe that poverty is due to lack of effort or that social programs are exploited by minority groups, they become less sympathetic to these programs and are more likely to find ways to prevent a perceived exploitation of public dollars. In terms of political ideology, we expect a willingness to expand social benefits among more liberal states. Similarly, a central plank of the Democratic Party ideology is the value of social programs, while Republican political ideology portrays such programs as unhelpful to claimants and a poor use of taxpayer dollars (Hacker, 2002; Keiser & Soss, 1998). If these parties are using administrative rules to pursue their goals, then we expect that states that feature strong Democratic control will feature less administrative barriers, while states with strong Republican control will be more likely to establish barriers.¹

In the following section, we describe our data and methodology for documenting variation in states’ administrative burdens and for examining their political roots.

Data and Method

The key measures of administrative burden came from two primary sources. First, 2010 Medicaid application forms for all 50 states and the District of Columbia were content-analyzed by two graduate students in public affairs using a coding instrument developed by the authors. This process produced data on the number of questions on each state’s form and the requirements for reporting income and expenditures, an approach similar to Slemrod’s (2005) method for assessing variation in the complexity of state tax systems. We developed a pilot coding instrument, which revealed that many states did not use a single Medicaid form but instead had multiple versions, indicating that the administrative burden an individual faced depended upon their characteristics. To ensure consistency between the two coders, we asked them to assume a detailed profile of a hypothetical individual applying for Medicaid benefits (a 35-year-old, single white female, previously married, with two children—a relatively common set of characteristics among applicants). Coders were trained about the basic concepts examined in the coding tool, and they raised questions based on issues they encountered during the pilot phase. To ensure intercoder reliability, both coders compared their results after coding the same six states. Each coding instrument had 306 items for coders to complete (a total of 1,836 items across the six states). Items had up to four responses (see the appendix), not including “not applicable” responses. Intercoder reliability was calculated based on percent agreement, where agreement was counted only if coders scored exactly the same on an item, which occurred just under 95% of the time across the six test states.

A secondary source of data about the Medicaid administrative burden came from the Kaiser Family Foundation, which collects data on Medicaid practices for all 50 states.² This source provided information on states’ interview requirements and on reapplication policies in 2010. Using these two data sources, we identified the following range of indicators of administrative burden that applicants to Medicaid face (see the appendix for additional information on the coding scheme):³

Length of application: How long (measured in number of words) is the state Medicaid application?

Number of questions: How many questions does a hypothetical applicant have to answer to complete the process?

Income reporting burden: Do individuals have to report income sources in the application process? Do they have to document income sources? How many different types of income have to be specified?

Expense reporting burden: Do individuals have to report expenses in the application process? Do they have to document expenses? How many different types of expenses have to be specified?

Resident: Do individuals have to provide documentation proving their residency in the state?

Enrollment process: Does the process require an in-person interview to qualify for the program, and/or does an applicant have to renew their status every 6 months as opposed to every year?

We considered these to be individual indicators of administrative burden, as well as elements of an underlying latent construct tapping the ease or difficulty involved in claiming Medicaid benefits. Together, these indicators tell us about the overall administrative burden imposed on Medicaid applicants in the state. For example, we assumed that the more information that was required to be provided and documented, the greater the burden on the applicant. Similarly, the coding mechanism also assumed that vague documentation requirements that provided little detail to claimants on how to satisfy the requests were more burdensome than relatively specific guidelines.

As an example, consider the application package for New York State. Applicants must provide day care costs, documented by a written statement from a day care provider and cancelled checks. Income documentation requires 4 weeks of consecutive paycheck stubs, a dated and signed letter from an employer documenting employment, and income tax returns. Child support information requirements include a court-signed letter, documents from the provider of funds, and check stubs. Locating such documentation is a daunting task for anyone, but even more so for individuals with limited resources. If one lacks a bank account, a car, or easy access to a photocopier or the Internet, these tasks become even more challenging. Such requirements are by no means unique to the Empire state.

Requiring applicants to face additional rounds of verification via interviews and repeated regular renewal processes is another step that allows applicants to fail (Wolfe & Scrivner, 2005). In particular, some of the requirements seem redundant. A state may have an interest in asking for detailed documentation of primary income sources, but most such sources can be collected via administrative data the state already has access to, and reporting requirements on certain types of income (such as income from panhandling, required in five states; or reparations as a result of being interned in prison camps during World War II, required in three states) seem both unlikely and excessive.

To capture state-level factors that might help explain variation in the level of administrative burden placed on applicants for Medicaid, we drew from a range of data sources to generate six measures: program generosity, state income, the federal Medicaid matching rate, the proportion of racial minorities in the state, political ideology in the state, and party control in the state. We focused on these, which were most central to our hypotheses and the Medicaid program, because we were limited by the small population size (N = 51) of these cross-sectional models. The first three measures—program generosity, state income, and federal Medicaid matching rate—reflect less political factors that could shift the financial incentives for states to make enrollment in Medicaid easier or harder (shown in Model 1 below) to contain the costs of this very substantial state-level budget item.

1. Y Admin Burden = a + B1 (Program Generosity) + B2 (State Income) + B3 (FMAP rate)

The other three variables—racial minorities, political ideology, and party control—tap more politicized factors. We hypothesize that these factors will play a role in structuring administrative burden (as tested in Model 2), even after controlling for the economic factors (as tested in Model 3).

2. Y Admin Burden = a + B1 (Proportion Minority) + B2 (Political Ideology) + B3 (Party Control)

3. Y Admin Burden = a + B1 (Program Generosity) + B2 (State Income) + B3 (FMAP rate) + B4 (Proportion Minority) + B5 (Political Ideology) + B6 (Party Control)

Because we wanted to predict the state administrative burden as of 2010 (the year of the application forms we content-analyzed), we operationalized our independent variables as an average across the previous decade: 2001-2008.⁴ We began by calculating state Medicaid generosity—measured as the maximum eligibility threshold to access a Medicaid benefit plan that meets federal guidelines—as a percentage of the federal poverty line. These data were drawn from the Kaiser Family Foundation (2011) and averaged 89.23 (SD = 56.50). We expected that states with more generous Medicaid programs would require more administrative barriers to enrollment. Therefore, we were most interested in the association between political (and other) factors and administrative burden, after controlling for the generosity of the state Medicaid program. The measure of generosity also acts as an additional measure of political ideology, since states where elected officials are less sympathetic to welfare claimants are expected to have lower rates.

We used two measures to model additional fiscal considerations that could shape states’ choices to erect administrative barriers to Medicaid enrollment. The first was the average personal income, presented in US$1,000s of 2010 constant dollars and gathered from the Bureau of Economic Advisors (M = 33.66, SD = 5.51). The second factor was the average state Federal Medical Assistance Percentage (FMAP) rate, which determines the amount of federal matching funds for state expenditures for Medicaid. These rates were gathered from the Federal Register for each year (M = 60.50, SD = 8.32). A higher FMAP rate implies that the federal government is covering a higher proportion of the state’s Medicaid bill. We expected to find a lower administrative burden in states with greater economic resources and/or those receiving greater assistance from the federal government in paying for Medicaid.

We focused on three measures of state political context: racial context, ideology, and party control. First, we expected greater administrative burden in states that have a higher proportion of racial minorities. We tested this expectation with a variable capturing the proportion of minorities (both non-Whites and/or people of Hispanic origin)⁵ in the state, which averaged 0.27 (SD = 0.16). These data were gathered from U.S. Census Bureau demographic profiles. We used the proportion of the popular vote for the Democrat in the 2008 presidential election (drawn from the Office of the Clerk, U.S. House of Representatives) to proxy for citizen ideology. This measure varied from 0.33 to 0.92, with an average of 0.51 (SD = 0.11). This measure is highly correlated with standard measures of political ideology, but has the additional advantage of allowing the inclusion of Washington, D.C.

To capture partisan control, we generated two variables. The first, a measure of unified Democratic Party control, captures the proportion of years between 2000 and 2009 in which the Democratic Party controlled the state’s upper and lower chambers, as well as the governor’s office. This measure was calculated from data provided by Klarner (2006) and averaged 0.21 (SD = 0.27). Similarly, we generated a measure of unified Republican control (M = 0.23, SD = 0.33) and divided party control (M = 0.56, SD = 0.33). Nebraska’s nonpartisan legislature was coded as continuous divided control, and Washington, DC, was coded as unified Democratic control for the entire time due to the Democratic party affiliation of the DC mayor and the majority of the DC city council during our time period.

Findings: Administrative Burden in State Medicaid Programs

We find a good deal of variation in each of our indicators of administrative burden. In regard to the application itself, state applications varied from 669 words to 9,085 words (M = 4,026.55, SD = 2,247.30) and from 49 to 248 questions (M = 107.14, SD = 45.84). The income reporting burden measure (coding described in the appendix) ranged from 7 to 84 (M = 27.33, SD = 17.19); the expense reporting burden ranged from 0 to 15 (M = 3.78, SD = 3.42); and the residency reporting requirement ranged from 0 to 2 (M = 0.75, SD = 0.66). Finally, 24% of states required more onerous enrollment procedures—either an in-person interview or frequent recertification.

To capture an overall measure of states’ administrative burdens, we created a composite score by taking the average of the standardized z scores of each indicator variable. That value was then restandardized with a mean of 0 and standard deviation of 1. This burden measure is presented in Figure 1, which maps states based on this value. This measure of administrative burden ranges from more than two standard deviations above the mean in Nevada (3.06) and Nebraska (2.31) to the lowest administrative burdens in Connecticut (−1.28) and Arkansas (−1.35). Table 1 provides a full listing of state values on our overall measure of burden, as well as on each of the indicators.

Figure 1.

Distribution of administrative burden among the states.

Table 1.

Indicators of Medicaid Administrative Burden in the 50 U.S. States.

	Burden score	Number of words	Number of questions	Income reporting	Expense reporting	Residency documentation	Enrollment process
Nevada	3.06	7,083	214	61	11	2	1
Nebraska	2.31	9,085	223	40	15	1	0
Tennessee	1.76	3,118	102	70	7	2	1
Pennsylvania	1.72	5,973	157	46	8	1	1
Michigan	1.63	6,119	248	15	6	1	1
Alaska	1.56	5,196	147	64	7	2	0
Virginia	1.48	5,754	138	84	7	1	0
New York	1.30	7,453	103	42	8	2	0
Texas	1.09	4,191	141	35	6	1	1
Minnesota	0.85	7,770	184	36	2	1	0
Florida	0.79	6,262	148	23	6	0	1
Montana	0.67	4,238	116	42	9	1	0
Indiana	0.57	3,883	77	24	12	0	1
North Dakota	0.24	4,384	74	29	9	1	0
New Hampshire	0.08	4,043	70	8	4	1	1
South Dakota	0.08	3,588	122	22	6	1	0
Washington	0.07	2,249	83	16	4	1	1
South Carolina	0.04	6,109	67	36	3	1	0
Missouri	0.00	8,625	180	13	0	0	0
Kentucky	0.00	5,467	72	36	3	1	0
Massachusetts	−0.02	7,759	124	39	0	0	0
Oklahoma	−0.04	7,532	103	21	0	1	0
Oregon	−0.05	7,342	96	25	0	1	0
Maryland	−0.10	2,900	110	21	3	0	1
Colorado	−0.12	6,796	148	21	2	0	0
North Carolina	−0.22	2,538	59	9	3	1	1
California	−0.29	1,198	104	17	2	2	0
Utah	−0.33	1,303	104	20	6	1	0
Vermont	−0.38	3,242	101	18	3	1	0
Rhode Island	−0.40	3,379	110	22	1	1	0
Wisconsin	−0.42	2,785	88	19	4	1	0
Hawaii	−0.44	4,520	89	45	1	0	0
Delaware	−0.49	3,498	69	20	1	1	0
Arizona	−0.51	5,019	146	15	1	0	0
New Jersey	−0.52	1,606	60	27	5	1	0
West Virginia	−0.55	950	66	48	1	1	0
Alabama	−0.55	2,493	107	52	0	0	0
D.C.	−0.66	2,196	57	9	3	2	0
Ohio	−0.72	2,366	55	10	5	1	0
Georgia	−0.72	669	92	15	1	0	1
Idaho	−0.76	2,988	166	7	1	0	0
Louisiana	−0.79	2,631	99	28	2	0	0
Illinois	−0.80	3,557	74	20	4	0	0
Wyoming	−0.98	2,379	58	12	1	1	0
New Mexico	−1.03	2,255	49	8	2	1	0
Iowa	−1.10	2,447	74	24	1	0	0
Maine	−1.16	2,123	73	18	2	0	0
Kansas	−1.24	1,508	80	20	1	0	0
Mississippi	−1.28	1,417	78	19	1	0	0
Connecticut	−1.28	1,896	100	7	1	0	0
Arkansas	−1.35	1,472	59	16	2	0	0

Table 2 shows that the composite measure of administrative burden is positively correlated with each of the individual indicators (.45 < r < .76). In addition, additional exploratory analyses (not shown)⁶ indicate that administrative burden is associated with actual enrollment in Medicaid—even after controlling for standard measures of need/poverty that we know to drive take-up. Specifically, we found higher rates of take-up among both the general population and among children when state applications had fewer questions, required lower expense reporting burden, and did not require an interview. These tests provide further face validity to our measures of administrative burden.

Table 2.

Correlations Among Medicaid Administrative Burden Indicators.

		(1)	(2)	(3)	(4)	(5)	(6)	(7)
(1)	Burden score	1.00
(2)	Number words	.65*	1.00
(3)	Number of questions	.69*	.64*	1.00
(4)	Income reporting	.65*	.35*	.29*	1.00
(5)	Expenses reporting	.76*	.29*	.36*	.41*	1.00
(6)	Residency reporting	.54*	.13	.06	.34*	.41*	1.00
(7)	Enrollment procedures	.45*	.01	.22	.04	.35	.08	1.00

Results: The Politics of Administrative Burden

Table 3 presents results from ordinary least squares (OLS) models regressing the composite measure of administrative burden on the six factors specified in Models 1 to 3 presented earlier. Model 1 includes the variables tapping the three less politicized, mostly fiscal considerations that could lead a state to use administrative burden to limit access to its state Medicaid program. Notably, the generosity of states’ Medicaid program (captured by their income eligibility level) was not associated with the level of red tape. We also obtained this null finding in bivariate models without additional controls and in the fully controlled models. In short, we did not find the expected trade-off between expanding eligibility and easing enrollment procedures.

Table 3.

Relationship Between State Context and Administrative Burden.

	(1)	(2)	(3)
Income eligibility	−0.003 (0.00)		−0.003 (0.00)
Personal income	−0.029 (0.04)		−0.010 (0.04)
FMAP rate	−0.057* (0.02)		−0.037 (0.02)
Proportion minority		0.799 (0.93)	0.647 (0.97)
Vote for Dem in 2008		0.599 (1.42)	0.371 (2.10)
Proportion Dem party control		−1.583* (0.60)	−1.274* (0.62)
Proportion divided control		0.157 (0.47)	0.100 (0.48)
Constant	4.668* (2.25)	−0.277 (0.66)	2.636 (2.57)
R ²	0.146	0.176	0.242

Note. N = 51. Coefficients from OLS models regressing the composite measure of administrative burden on measures of state economic/fiscal and political context. Standard errors are listed in parentheses below each coefficient.

p < .05.

We also found no independent association between administrative burden and state income. We, however, did find a significant association between administrative burden and the amount of assistance the state received from the federal government, measured as the state FMAP (matching) rate. States with higher FMAP rates had fewer administrative barriers to enrollment in Medicaid (b = −0.057, SE = .02). If we consider a one-standard deviation increase in a state FMAP rate (8.31), this model predicts a decrease in the state administrative burden by 0.47 (nearly half a standard-deviation of the administrative burden measure). Another way to consider this association is to note that these three fiscally focused variables together explain approximately 15% of the variation in state administrative burden.

Model 2 presents results from OLS models regressing the measure of administrative burden on the variables capturing more political factors: the proportion of racial minorities in the state, political ideology, and party control. We find only one factor to be significantly associated with administrative burden: Democratic Party control. States that experienced more years of unified Democratic Party control had erected fewer barriers to enrollment in Medicaid (b = −1.58, SE = .60). This coefficient can be interpreted directly: A state that experienced continuous Democratic Party control would have an administrative burden score that is 1.58 (one-and-a-half standard deviations) lower than would a state that experienced continuous Republican Party control. This model explains 18% of the variation in state administrative burdens.

Model 3 includes all the variables used in Models 1 and 2 to consider their influence over and above one another. The inclusion of the political factors reduced the FMAP coefficient to nonsignificance (although it remained quite similar in direction and magnitude). However, the inclusion of the economic factors does little to mediate the influence of Democratic Party control (with the coefficient reduced only 20% to 1.27). Democratic Party control remains the only significant factor related to administrative burden. This relationship is depicted in Figure 2, which presents the predicted value of administrative burden estimated from Model 3 in Table 3. Each bar represents the predicted value under each level of Democratic Party control (vs. Republican Party control). These estimates assume that divided control is set to zero and all other variables are set at their mean level.

Figure 2.

Administrative burden, by Democratic party control.

Finally, to examine the robustness of these findings across each indicator of administrative burden, we reestimated Model 3 from Table 3 for each of the six indicators. To ease comparison across models, we standardized each dependent variable to have a mean of zero and standard deviation of one. These models are presented in Table 4. The first four columns present coefficients from OLS regression models, and the final two employ ordered logit models. The models in Table 4 offer two primary take-aways. First, Democratic Party control remains quite important in understanding administrative burden. States with more years of united Democratic Party control are expected to require shorter applications (b = −1.46, SE = .61 for the number of words; b = −1.52, SE = .62 for number of questions) and require less burdensome documentation of expenses (b = −1.46, SE = .65). This variable is also negative (although not significant) for income and residency reporting requirements, two other indicators of administrative burden. The second key finding is the distinct predictors of more burdensome enrollment procedures (requiring either an interview or more frequent recertification/reapplication periods). Unlike the other indicators, our measures of party control was not significant (or even in the expected direction). Instead, all three of the economic (or less political) variables were significant in the model. In short, states were less likely to require these burdensome enrollment procedures when they had more generous eligibility levels (b = −0.05, SE = .02), greater personal income (b = −0.93, SE = .46), or higher federal matching rates (b = −0.62, SE = .25).

Table 4.

Relationship Between State Context and Indicators of Administrative Burden.

	(1)	(2)	(3)	(4)	(5)	(8)
	Number of words	Number of questions	Income reporting	Expenses reporting	Residency reporting	Enrollment procedures
Income eligibility	0.000 (0.00)	0.000 (0.00)	−0.001 (0.00)	−0.005 (0.00)	0.007 (0.01)	−0.046* (0.02)
Personal income	−0.005 (0.04)	−0.021 (0.04)	−0.001 (0.04)	0.007 (0.04)	0.087 (0.09)	−0.933* (0.46)
FMAP rate	−0.018 (0.02)	−0.034 (0.02)	−0.018 (0.03)	−0.010 (0.03)	0.021 (0.05)	−0.624* (0.25)
Proportion minority	0.267 (0.95)	0.696 (0.96)	1.173 (1.03)	−0.473 (1.01)	1.131 (2.09)	0.303 (3.21)
Vote for Dem in 2008	0.785 (2.07)	−0.274 (2.08)	−2.298 (2.23)	1.895 (2.19)	−1.358 (4.33)	1.332 (9.79)
Proportion Dem control	−1.455* (0.61)	−1.522* (0.62)	−0.180 (0.66)	−1.458* (0.65)	−0.930 (1.28)	2.485 (2.17)
Proportion divided control	0.470 (0.48)	−0.030 (0.48)	0.769 (0.51)	−0.438 (0.51)	−0.192 (0.99)	−1.049 (1.73)
Constant	0.844 (2.53)	3.068 (2.55)	1.715 (2.73)	0.457 (2.68)
Cut 1					3.528 (5.65)	−69.330* (31.26)
Cut 2					6.238 (5.71)
R ²	0.268	0.255	0.146	0.175
Pseudo R²					0.040	0.410

Note. N = 51. Coefficients from models regressing each indicator of administrative burden on measures of state economic/fiscal and political context. Models 1 to 4 are estimated from OLS models, while Models 5 and 6 are estimated using ordered logistic models. Standard errors are listed in parentheses below each coefficient.

p < .05.

Conclusion

Jacobs and Soss (2010) offer a recent critique of the field of political science, arguing that it has neglected its own traditions and a powerful explanatory tool by turning its back on political economy arguments that can shed light on policy outcomes. In particular, they call for an examination of how different groups have battled over public policies that have resulted in reinforcing, rather than compensating for, a growing inequality in the distribution of resources in the United States. This article follows such an approach, offering a preliminary analysis of the political economy of administrative burden in U.S. state Medicaid policies. Administrative burden matters because it directly affects the rate at which eligible claimants access benefits. Prior research has made clear that administrative burden can negatively affect participation.

One goal of the article is to simply demonstrate the importance of these types of measures as an appropriate dependent variable in social science. Political scientists already examine policy variation in legislated welfare policies (e.g., Klarner et al., 2007), but there is less attention paid to the administrative rules that are also a form of policy. Even though policy feedback research suggests that rules can undermine the political efficacy of clients and social policy research shows that such rules have a significant effect on the capacity of eligible claimants to access services, we know relatively little about the source of such rules. In this article, we identify a series of rules that will plausibly make it more difficult for applicants to claim services, and then we examine whether such rules are associated with political variables.

There are obvious limitations to any cross-sectional analysis with sample size of 51, primarily the risk of endogeneity, the difficulty in asserting causality, and the inability to specify enough variables to develop a complete model. Additional cross-time data would overcome some of these limitations. A qualitative approach would also aid in identifying some of the casual processes we assume here. For example, Herd, DeLeire, Harvey and Moynihan (2013) case-study of Medicaid in the state of Wisconsin illustrates how a Democratic administration sought to reduce burdens. There are limitations in measurement as well. In this article, we examine formal aspects of the enrollment compliance burden. Such rules have demonstrated importance to program take-up, but informal practices also make a difference to the burdens that citizens face, and our analysis should not be interpreted to suggest that such rules are implemented uniformly or that bureaucratic discretion does not matter. Administrators, whether employing their discretion or under the direction of political principals, can use informal means to encourage or discourage access to programs. In addition, our measures of political control and ideology, while standard indicators, are broad and will miss some variation between states that we group together.

While the models are inevitably incomplete, there is enough evidence to suggest that which party is in charge matters to the administrative burden. In particular, in states where Democrats exerted strong political control, claimants were less likely to face burdensome enrollment procedures. Although political ideology did not matter, unified Democratic control predicted lower burdens. Our analysis does not allow us to directly observe the causal processes by which Democratic dominance matters, but it seems plausible that it provides a form of political insulation for political and administrative actors from attacks that such programs are too generous or rife with fraud. This finding adds to accounts of red tape as something that occurs largely because rules that may have begun with a useful purpose evolve to become a nuisance without a benefit (Bozeman & Feeney, 2011). The “rule-evolved” approach is almost certainly present in the formal enrollment burdens we measured—think, for example, of demands for documentation of World War II prison camp reparations. But our findings suggest that red tape does not only arise because of inattention to rules that have outlived their usefulness; there is also a political economy of administrative burden in Medicaid. We see enormous potential for future studies to build upon this finding to further examine why burdens occur in a wide variety of policy contexts.

Footnotes

Appendix

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this research was provided by the Graduate School and the Institute for Research on Poverty at the University of Wisconsin–Madison (no grant numbers).

Notes

Author Biographies

Donald P. Moynihan is a professor at the La Follette School of Public Affairs, University of Wisconsin–Madison. He is a fellow of the National Academy of Public Administration, the author of The Dynamics of Performance Management, and the winner of the ASPA/NASPAA Distinguished Research Award.

Pamela Herd is an associate professor of public affairs and sociology at the University of Wisconsin–Madison. Her research on inequality, aging, and health appears in journals such as Social Forces, and Journals of Gerontology. She directs the Wisconsin Longitudinal Study, and is a member of the Board of Overseers of the General Social Survey and the Panel Study of Income Dynamics, and is a member of the National Academy of Social Insurance.

Elizabeth Rigby is an assistant professor of public policy and public administration in the Trachtenberg School at George Washington University. Her research examines the interplay of politics and inequality in the American policymaking process. She holds a PhD from Columbia University and received postdoctoral training at the University of Wisconsin–Madison as a Robert Wood Johnson Foundation Health and Society Scholar.

References

Bartlett

Burstein

(2004). Food stamp program access study: Eligible nonparticipants (Report No. E-FAN-03-013-2). Retrieved from http://www.ers.usda.gov/publications/efan03013/efan03013-2/efan03013-2.pdf

Bartlett

Burstein

Hamilton

(2004). Food stamp program access study: Final report (Report No. E-FAN-03-013-3, Report to the U.S. Department of Agriculture). Cambridge, MA: ABT Associates. Retrieved from http://www.ers.usda.gov/publications/efan03013/efan03013-3/efan03013-3c.pdf

Bennett

(1995). No relief but upon the terms of coming into the house—Controlled spaces, invisible disentitlements, and homelessness in an urban shelter system. Yale Law Journal, 104, 2157-2212.

Bettinger

E. P.

Long

B. T.

Oreopoulos

Sanbonmatsu

(2012). The role of application assistance and information in college decisions: Results from the H&R Block Fafsa Experiment. Quarterly Journal of Economics, 127, 1205-1242.

Bhargava

Manoli

(2011). Why are benefits left on the table? Assessing the role of information, complexity, and stigma on take-up with an IRS field experiment. Retrieved from http://econweb.umd.edu/~davis/eventpapers/BhargavaBenefits.pdf

Bozeman

(2000). Bureaucracy and red tape. Upper Saddle River, NJ: Prentice Hall.

Bozeman

Feeney

M. K.

(2011). Rules and red tape: A prism for public administration theory and research. Armonk, NY: M.E. Sharpe.

Brien

M. J.

Swann

C. A.

(2001, May). Prenatal WIC participation and infant health: Selection and maternal fixed effects (Working Paper). Charlottesville: Department of Economics, University of Virginia.

Brodkin

E. Z.

(1987). Policy politics: If we can’t govern, can we manage? Political Science Quarterly, 102, 571-587.

10.

Brodkin

E. Z.

Majmundar

(2010). Administrative exclusion: Organizations and the hidden costs of welfare claiming. Journal of Public Administration Research and Theory, 20, 827-848.

11.

Bruch

S. K.

Marx-Freere

Soss

(2010). From policy to polity: Democracy, paternalism, and the incorporation of disadvantaged citizens. American Sociological Review, 75, 205-226.

12.

Burden

Canon

Mayer

Moynihan

(2012). The effect of administrative burden on bureaucratic perception of policies: Evidence from election administration. Public Administration Review, 72, 741-751.

13.

Currie

(2006). The take-up of social benefits. In Auerbach

A. J.

Card

Quigley

J. M.

(Eds.), Public policy and the income distribution (pp. 80-148). New York, NY: Russell Sage.

14.

Elder

Powers

(2006). The incredible shrinking program. Research on Aging, 28, 341-358.

15.

Feeney

M. K.

Moynihan

D. P.

Walker

R. M.

(2010). Rethinking and expanding the study of administrative rules: Report of the 2010 red tape research workshop. Madison: Robert M. La Follette School of Public Affairs, University of Wisconsin–Madison.

16.

Food and Nutrition Service. (2007). Reaching those in need: Food stamp participation rates in 2005. Washington, DC: Author.

17.

Gilens

(1999). Why Americans hate welfare: Race, media and the politics of antipoverty policy. Chicago, IL: University of Chicago Press.

18.

Glazer

(2003). On Americans and inequality. Daedalus, 13, 111-115.

19.

Hacker

J. S

. (2002). The divided welfare state: The battle over public and private social benefits in the United States. New York, NY: Cambridge University Press.

20.

Hale

McNeal

(2010). Election administration reform and state choice: Voter identification requirements and HAVA. Policy Studies Journal, 38, 181-302.

21.

Hanratty

(2006). Has the food stamp program become more accessible? Impacts of recent changes in reporting requirements and asset eligibility limits. Journal of Policy Analysis and Management, 25, 603-621.

22.

Herd

DeLeire

Harvey

Moynihan

D. P.

(2013). Shifting administrative burden to the state: The case of Medicaid take-Up. Public Administration Review. doi: 10.1111/puar.12114

23.

Jacobs

L. R.

Soss

(2010). The politics of inequality in America: A political economy framework. Annual Review of Political Science, 13, 341-364.

24.

Kaiser Family Foundation. (2011). Annual updates on eligibility rules, enrollment and renewal procedures, and cost-sharing practices in Medicaid and SCHIP. Retrieved from http://www.kff.org/medicaid/50StateSurvey.cfm

25.

Keiser

Soss

(1998). With good cause: Bureaucratic discretion and the politics of child support enforcement. American Journal of Political Science, 42, 1133-1156.

26.

Klarner

(2006). Measurement of partisan balance of state government. Retrieved from http://www.indstate.edu/polisci/klarnerpolitics.htm

27.

Klarner

Mao

Buchanan

(2007). Business interest groups and temporary assistance to needy families. Social Science Quarterly, 88, 104-119.

28.

Klerman

Danielson

(2011). The transformation of the Supplemental nutrition assistance program. Journal of Policy Analysis and Management, 30, 863-888.

29.

Korpi

Palme

(1998). The paradox of redistribution and strategies of equality: Welfare state institutions, inequality, and poverty in western countries. American Sociological Review, 63, 661-687.

30.

Kumlin

Rothstein

(2005). Making and breaking social capital: The impact of welfare state institutions. Comparative Political Studies, 38, 339-365.

31.

Lineberry

(1977). American public policy: What government does and what difference it makes. New York, NY: Harper & Row.

32.

Lipsky

(1980). Street-level bureaucracy: Dilemmas of the individual in public service. New York, NY: Russell Sage Foundation.

33.

Lipsky

(1984). Bureaucratic disentitlement in social welfare programs. Social Science Review, 58, 3-27.

34.

Loyens

Maesschalck

(2010). Toward a theoretical framework for ethical decision making of street-level bureaucracy existing models reconsidered. Administration & Society, 42, 66-100.

35.

Mettler

Soss

(2004). The consequences of public policy for democratic citizenship: Bridging policy studies and mass politics. Perspectives on Politics, 2, 55-73.

36.

Moynihan

D. P.

Herd

(2010). Red tape and democracy: How rules affect citizenship rights. American Review of Public Administration, 40, 654-670.

37.

Rigby

Soss

Booske

Rohan

Robert

(2009). Public responses to health disparities: How group cues influence support for government intervention. Social Sciences Quarterly, 90, 1321-1340.

38.

Schanzenbach

D. W.

(2009). Experimental barriers of the barriers to food stamp enrollment (Report No. IRP DP 1367-09). Retrieved from http://www.irp.wisc.edu/publications/dps/pdfs/dp136709.pdf

39.

Schneider

Ingram

(1997). Policy design for democracy. Lawrence, KS: University of Kansas Press.

40.

Scott

P. G.

(1997). Assessing determinants of bureaucratic discretion: An experiment in street-level decision making. Journal of Public Administration Research and Theory, 7, 35-58.

41.

Shore-Sheppard

L. D.

(2008). Stemming the tide? The effect of expanding Medicaid eligibility on health insurance coverage. BE Journal of Economic Analysis & Policy, 8, Article 6.

42.

Slemrod

(2005). The etiology of tax complexity: Evidence from US state income tax systems. Public Finance Review, 33, 279-299.

43.

Soss

(1999). Lessons of welfare: Policy design, political learning, and political action. American Political Science Review, 93, 363-380.

44.

Soss

Fording

Schram

(2011). Disciplining the poor: Neoliberal paternalism and the persistent power of race. Chicago, IL: University of Chicago Press.

45.

Tyler

Lind

E. A.

(1992). A relational model of authority in groups. In Zanna

(ed.), Advances in experimental social psychology (pp. 115-191). New York, NY: Academic Press.

46.

Verba

Schlozman

K. L.

Brady

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

47.

Wichowsky

Moynihan

D. P.

(2008). Measuring how administration shapes citizenship: A policy feedback perspective on performance management. Public Administration Review, 68, 908-920.

48.

Wolfe

Scrivner

(2005). The devil may be in the details. Journal of Policy Analysis and Management, 24, 499-522.

49.

Yackee

S. W.

(2006). Sweet-talking the fourth branch: Assessing the influence of interest group comments on federal agency rulemaking. Journal of Public Administration Research and Theory, 26, 103-124.