state agency discretion in a delegated federal program ...investment (brehm and gates 1997; moe...

State Agency Discretion in a DelegatedFederal Program: Evidence from DrinkingWater Investment

Dorothy M. Daley*, Megan Mulliny andMeghan E. Rubadoy

*University of Kansas; [email protected] University

This article examines the use of discretion by state agencies in the context of multilevel policy.

Research on agency discretion assumes that discretion represents a departure from legislative

intent. However, Congress may delegate authority to promote policy innovation. Using data on

investment in drinking water infrastructure from 2000 to 2008, we examine the relationship

between agency discretion and functional expertise in implementing the Drinking Water State

Revolving Fund program.We focus on two areas where states can exercise discretion: (i) projects

not related to compliance with federal law and (ii) support to small water systems. Our results

indicate that agency expertise influences investment, but problem severity reduces differences

across agencies. Initial choices over agency design affect how states adapt federal programs to

meet state needs.

Fragmentation and multilevel governance remain enduring characteristics in

American government. Across a range of policy areas, devolution of authority has

created networked governance where different levels of government and multiple

agencies share responsibility for policy implementation (Agranoff 2007; Agranoff

and McGuire 2003; Mullin and Daley 2010; Gerber and Teske 2000; Meyers,

Riccucci, and Lurie 2001; O’Toole 1996; Potoski 2001; Scholz and Wei 1986;

Whitford 2007; Woods 2006; Woods and Potoski 2010). Typically, Congress and

the federal bureaucracy set broad policy goals and delegate authority to state

governments to achieve those goals. States then develop rules that transfer

significant responsibility for implementation to bureaucratic agencies.

State agencies play a critical role in the execution of federal public policy, but

our understanding of their behavior remains underdeveloped. Part of the problem

lies in the literature’s inconsistent treatment of state agency discretion in the

context of multilevel governance. On the one hand, scholars of fiscal federalism

treat devolution as an opportunity for states to craft policies that better match the

nature of their problems and the preferences of state populations (Oates 1972).

Publius:TheJournal of Federalism, pp.1^23doi:10.1093/publius/pjt033� TheAuthor 2013. Published by Oxford University Press on behalf of CSFAssociates: Publius, Inc.All rights reserved. For permissions, please email: [email protected]

at Pennsylvania State University on M

ay 10, 2016http://publius.oxfordjournals.org/

Dow

nloaded from

http://publius.oxfordjournals.org/

Devolution can be a mechanism to enhance the efficient provision of public goods,

particularly when faced with a diverse citizenry (Kincaid 1998; Whitaker and Time

2001). It also promotes policy experimentation, which can have benefits for the

entire federal system (Bednar 2011). On the other hand, theoretical work on agency

discretion and empirical research on state bureaucratic decision making typically

depicts the exercise of discretion as a departure from legislative intent and,

therefore, a lapse in democratic representation with unelected civil servants

wielding considerable authority. Factors that contribute to variable outcomes, such

as information asymmetry based on agency expertise, are represented as forces

countering federal control, rather than conditions that ought to influence

implementation of a devolved federal policy (Gailmard and Patty 2007, 2012;

Huber and McCarty 2004).

In this article, we examine the use of discretion by agencies with different types

of functional expertise in the context of a multilevel policy designed to promote

state flexibility and responsiveness to state-level problems. Our analysis capitalizes

on differences across institutional environments as well as problem conditions to

better understand what factors promote or inhibit agency discretion. In the 1996

amendments to the Safe Drinking Water Act (SDWA), Congress established the

Drinking Water State Revolving Fund (DWSRF) program to promote capital

investment in drinking water. Modeled after the Clean Water State Revolving Fund

(SRF) established nine years earlier, the DWSRF creates a system of shared

responsibility for water infrastructure investment. Under the DWSRF program, the

Environmental Protection Agency (EPA) awards capitalization grants to states,

which use these funds to provide low- or no-interest loans to local communities to

invest in drinking water infrastructure. Repayment by loan recipients ensures the

long-term integrity of the funds, allowing future investment in other communities.

Since 1997, DWSRF programs have provided more than $16.2 billion in low

interest loans to maintain, improve, and protect the nation’s drinking water

infrastructure (U.S. EPA 2010). Although the EPA retains authority to set and

enforce drinking water standards, it is up to local actors to design infrastructure

improvements that comply with federal law, and state officials are responsible for

prioritizing among deserving projects.

This type of decentralization can foster innovation and experimentation by

enabling states to design DWSRF programs that meet local needs. However,

resource and information constraints might prevent state agencies from fully taking

advantage of the discretion available to them. The DWSRF program offers a unique

opportunity to systematically examine the interplay between institutional expertise

and agency discretion. There is considerable variation in the functional focus of

agencies administering DWSRF programs, with implementation responsibility

falling under the authority of an environmental, public health, or financial agency,

or some combination of the three.

2 D. M. Daley et al.



Dow

nloaded from

;

,

(DWSRF)


Previous research indicates that institutional structure and administrative

procedures can influence agency decision making and output (Brehm and Gates

1997; Gerber and Gibson 2009; Hammond and Thomas 1989; Moe 1989; Morris

1997; Potoski 2002; Whitford 2002). We expect that an agency’s functional focus,

meaning an environmental, public health or financial focus in agency staffing and

mission, influences how state bureaucrats capitalize on the discretion available to

them in a multilevel program intended to promote state flexibility. Using data on

patterns of drinking water infrastructure investment from 2000 through 2008, we

analyze how these institutional differences influence responsiveness to problem

conditions in a state. We focus on two areas where states have considerable

opportunity to tailor water infrastructure investments: (i) projects not directly

related to compliance with federal law and (ii) support to small water systems that

face significant challenges in providing safe drinking water. We find evidence

indicating that an agency’s functional focus influences investment patterns, but

problem severity reduces differences across agency types.

Federalism and State Decision Making

Much of the literature on agency discretion assumes that discretion represents a

departure from program designers’ intent. For some scholars, conflict between

political principals and bureaucratic agents is intrinsic to the concept of discretion;

Calvert, McCubbins, and Weingast (1989) define discretion as ‘‘the departure of

agency decisions from the positions agreed upon by the executive and legislature at

the time of delegation and appointment’’ (589). A large literature on political

control of the bureaucracy examines the tools that elected officials use to increase

their leverage over agency behavior (Hammond and Knott 1996; McCubbins, Noll,

and Weingast 1987; Moe 1989, 1990; Whitford 2005; Wood 1990; Wood and

Waterman 1991, 1994). Other scholars who focus on the role of street-level

bureaucrats in delivering public services similarly treat bureaucrats’ use of

discretion in applying rules to particular cases as shifting policy outcomes away

from those defined in law (Jones et al. 1978; Lipsky 1980; Maynard-Moody and

Musheno 2003).

If delegation necessarily involves agency drift, the problem should be particularly

acute for complex, multilevel programs (Pressman and Wildavsky 1973). Indeed,

empirical studies of state implementation of federal policies reveal considerable

variation in state-level outcomes (Atlas 2007; Barrilleaux and Miller 1988; Gerber

and Teske 2000; Konisky 2007, 2009; Konisky and Woods 2010; Sigman 2003;

Potoski 2001; Whitford 2007). In some policy contexts, however, exercising

political control over state bureaucracies may not be the primary goal of federal

policy makers. Devolution instead can be an intentional effort to encourage

innovation and responsiveness to state problems and needs.

State Agency Discretion in a Delegated Federal Program 3



Dow

nloaded from

1

2

;

;

;


For example, in creating the Temporary Aid to Needy Families (TANF)

program, the federal government allowed states to set welfare eligibility

requirements and benefit levels with the explicit intent to foster states’ creativity

in designing effective welfare systems. Studies of TANF programs reveal important

interstate differences in implementation, but unlike the case with the DWSRF

program, these differences are the product of political negotiations in state

legislative and executive branches (Soss et al. 2001; Fellowes and Rowe 2004). The

TANF experience demonstrates that state lawmakers will respond to a variety of

factors including political and economic conditions, demographics, problem

severity, and the policies of neighboring states when they pass legislation to

implement federal law, just as they do when enacting laws autonomously (Berry

and Berry 1999; Boushey and Luedtke 2011; Daley and Garand 2005; Gerber and

Teske 2000; Moreland-Russell et al. 2013; Ringquist 1993; 1994; Shipan and Volden

2006, 2008; Volden 2006). We know less about how state agencies use discretion

that may be granted under federal law.

Our research focuses on multilevel policies where the federal government seeks

to promote state flexibility. States implementing these policies have some discretion

to design programs to meet state needs and political demands, but they are far

more constrained than in the settings for most state policy adoption research. The

multilevel context not only limits the autonomy of state decision makers, but it

also complicates questions about political control of the bureaucracy by creating

two sets of principals delegating authority to state agencies. State agencies receiving

competing signals from federal and state political actors might seek to minimize

conflict by hewing closely to federal standards and other states’ policies. Another

factor that may reduce variation in state policies are the information demands

associated with state-level policy making. State bureaucrats might not have the

time, expertise, or data necessary to craft policies that address specific state needs,

and therefore rely instead on professional norms and agency expertise in making

implementation decisions.

We look specifically at the relationship between agency expertise and the exercise

of discretion, and examine how expertise interacts with state-level problem

conditions. An agency’s functional focus may help agencies navigate the two

challenges to exercising discretion outlined above. Bureaucrats facing competing

signals from multiple principals have an incentive to acquire further policy

expertise (Gailmard and Patty 2007). They will then utilize this expertise to

overcome information constraints in decision making. Implementation research

points to the critical role of organizational characteristics, including agency

mission, culture, and professionalism in predicting agency behavior (Bardach 1998;

Brehm and Gates 1997; Downs 1964; Kaufman 2006; Soss, Fording, and Schram

2011; Thomas 2003; Wilson 1989; Whitford 2010). If reliance on mission and

professional norms is a mechanism that agencies use to navigate complex decision




Dow

nloaded from

;


making, we should expect agencies with different functional responsibilities, such as

environmental protection or financial management, to focus on different aspects of

drinking water infrastructure investment.

State Discretion in PrioritizingWater Infrastructure Spending

Congress created the DWSRF program to assist community water systems in

meeting rising federal standards for drinking water protection. By design, the

DWSRF program is meant to give state and local governments considerable latitude

in managing investments in water infrastructure. However, federal DWSRF policy

creates incentive for state agencies to pursue multiple, competing goals: to foster

innovative environmental investment and to maintain the fiscal integrity of the

loan funds. Implementing agents are responsible to political principals at both the

state and federal levels, but they may lack essential data about drinking water

problems and needs. In this complex policy environment, the functional focus of

an agency is likely to direct attention to some problems and issues over others, and

bureaucrats will rely on their expertise and preferences in directing drinking water

investment (Brehm and Gates 1997; Moe 1990; Watkins-Hayes 2011; Wilson 1989;

Whitford 2008). For example, DWSRF agency staff may have different levels of

knowledge about issues such as ecosystem health and source water protection, the

integration of drinking water with stormwater and wastewater management, or

loan portfolio analysis and cash flow modeling.

Drinking water protection is a goal clearly shared by environmental and public

health professionals. However, the longstanding organizational separation between

environmental and public health professionals has resulted in disengagement

between the fields and little mutual understanding (Gordon 1995; Kotchian 1997).

Previous work has reported that public health professionals have a tendency to

ignore environmental data and that pressures from interest groups and the political

environment contribute to environmental agencies focusing on their regulatory

responsibilities over their role in public health protection (Black 2000; Kotchian

1997). Financial agencies bring an entirely different set of tools to program

management. These agencies are staffed with individuals who have been trained to

promote fiscal responsibility and understand financial risk. The organizational

culture and expertise of staff working in financial agencies lends itself more to the

fund management aspects of DWSRF implementation than to the policy’s

substantive goals.

Substantial differences in agency mission and staff expertise exist between the

three types of agencies involved in DWSRF decision making. Past research

highlights the critical nature of organizational factors in shaping agency discretion

(Brehm and Gates 1997; Soss, Fording, and Schram 2011; Wilson 1989). Even when

public health and environmental agencies diversify by hiring financial professionals




Dow

nloaded from

governments


as agency staff, these financial experts are likely to be profoundly influenced by

established organizational norms (Soss, Fording, and Schram 2011; Watkins-Hayes

2011) and they face considerable hurdles in changing agency culture (Bardach 1998;

Kaufman 2006; Wilson 1989). Our research examines the relationship between

agency focus and outcomes for a program operated by different agency types.

In many states DWSRF programs are operated jointly by more than one agency.

Thirty-nine states have relied upon environmental agencies to manage the revolving

funds alongside other environmental programs. In twenty-two states, health

agencies have provided substantive policy leadership for the state funds, typically in

place of the environmental agency or occasionally in cooperation. In thirty-one

states, some type of financial agency has played a role in DWSRF decision making.

While these implementation arrangements are complex, they allow us to examine

how agency expertise influences investment decisions, and we expect the functional

focus of agencies to influence the allocation of funds even in cases of shared agency

authority.

Analyzing the relationship between the functional focus of an agency tasked with

implementation and programmatic outcomes begs the question of whether the

choice of an implementing agency is exogenous to the outcome of interest. We find

no evidence that states designate lead substantive agencies for the DWSRF program

in order to pursue a certain type of outcome. Delegation to the substantive agency

is dictated by primacy, or the agency that has primary enforcement responsibility

over public water systems in the state. In every state where either an environmental

or public health agency has at least partial responsibility for the revolving fund,

program management corresponds to the agency (or to at least one of the agencies,

in states where primacy is shared) for SDWA implementation. In the majority of

states, designation of SDWA enforcement authority was granted prior to 1980,

nearly two decades prior to establishment of the DWSRF program (U.S. EPA

1994). Moreover, recent work by Sinclair and Whitford (2012) suggests that

organizational structure, at least with respect to state environmental and health

agencies, may be better explained by historical factors than contemporary political

conditions.

Endogeneity is a bigger concern for finance agency involvement in DWSRF

implementation. In most cases, states delegate fund management to a financial

agency if the agency already manages the Clean Water SRF or some other

infrastructure financing program. Of the thirty-one states where finance agencies

have had at least partial responsibility over the DWSRF program, in all but three

states the agency already had established a role with the state’s Clean Water fund.

Financial agency involvement also might be an indication that the state intends to

issue leveraging bonds: finance agencies play a role in program management in

seventeen of the twenty-one (81 percent) leveraged states but in only thirty-one

states overall (62 percent). In general, the evidence suggests that assignment of




Dow

nloaded from


DWSRF program responsibility to one or more agencies is largely a function of

historical responsibilities over water policy and choices about whether to leverage

federal funds. Case studies of SRF programs in three Southern states (Morris 2010)

indicate that interagency competition and administrative capacity also contributed

to state decisions about the location of program authority. Because the factors

contributing to agency assignment are complex and may be state-specific, we have

some confidence that the relationships we observe between an agency’s functional

expertise and loan allocation decisions are not spurious. Without strict exogeneity

in agency assignment, we cannot assert causal relationships, yet our results reveal

important patterns in how problem conditions shape the relationship between an

agency’s functional authority and decision outcomes.

Hypotheses, Models, and Data

We focus on two areas of spending in which states have some discretion: assistance

for projects not related to compliance with federal water quality law and assistance

for small water systems. In the EPA DWSRF spending data we use, states report the

amount of loan assistance they provide that helps currently noncompliant systems

meet compliance, the amount that helps compliant systems maintain compliance,

the amount that aids compliant systems to meet anticipated future standards, and

the amount not related to compliance. It is among the latter projects where state

agencies exercise discretion to provide assistance based on their own criteria,

including factors related to infrastructure improvement, affordability, the

applicant’s financial health, population served, and green projects. Our review of

the formulas states use to prioritize among projects reveals that the most influential

of these criteria is infrastructure improvement; these projects often take a broader

approach to water infrastructure development, aiming to advance green

infrastructure, promote watershed management and source water protection,

improve flow, or minimize service disruptions rather than address some particular

contaminant.

Given their expertise, environmental agencies should be most supportive of this

broad view of infrastructure development and therefore to dedicate a greater

portion of loan funds to projects not directly related to compliance [H1].

Professionals in these agencies have more background in the relationship between

water supply and ecosystem health. However, we expect that the severity of water

quality challenges within a state will moderate this response. High rates of water

quality violations will focus the attention of environmental agencies on compliance

goals, minimizing the influence of agency functional focus [H2].

The second area of spending we examine is assistance for small water systems.

Providing support for public water systems serving 10,000 or fewer people is an

important priority for the U.S. EPA. The 1996 SDWA Amendments placed




Dow

nloaded from

non-


emphasis on improving the technical, managerial, and financial capacity of small

systems and created several mechanisms for helping small systems comply with

drinking water regulations, including an optional set-aside fund that states may

establish within their DWSRFs. DWSRF officials can exercise discretion and invest

in small water systems. However, small systems entail substantial financial risk;

most small water systems have no credit history and remain vulnerable to local

economic conditions (Copeland 2010). States face competing pressures to design

DWSRF programs that both respond to their particular water quality challenges

and at the same time emphasize the fiscal health of applicant communities to

ensure the long-term integrity of the fund. The SRF approach creates a powerful

incentive for fiscal discipline. Enacting legislation for the state loan programs

envisioned that they would operate without ongoing federal support after initial

appropriations to capitalize the funds.1 The potential consequences for taking on

financial risk are substantial: without sound management of its loan programs, a

state may drain its funds and prevent its localities from accessing financial

assistance for infrastructure projects for the long term.

We hypothesize that DWSRF programs located in financial agencies will

emphasize the long-term fiscal integrity of loan funds, and therefore be more

cautious in assisting small water systems [H3]. As with noncompliance spending,

however, we expect that problem factors, such as the number of small systems

within a state, will outweigh the influence of agency type. Therefore the

relationship between institutional design and outcomes will be smaller in states that

are more reliant on small water systems [H4]. Furthermore, given their expertise

and training in financial management, we predict that financial agencies will be

more responsive than other management structures to the financial health of the

fund in deciding whether to award loans to small systems [H5].

Models and Data

Data on revolving fund spending patterns come from the DWSRF National

Information Management Systems (DWNIMS), compiled by the EPA for

accountability purposes. Our panel begins in 2000, by which time all states had

begun providing DWSRF loan assistance, and runs through 2008.2 We merged the

DWNIMS data with data from a variety of secondary sources to estimate the

influence of political and problem factors on state discretion in the implementation

of a multilevel policy program, and to assess how agency expertise conditions

responsiveness to problem conditions.

Our dependent variables are percentages of total loan funds awarded. Spending

on noncompliance-related projects is a percentage of spending across all

compliance categories (loans that help currently noncompliant systems meet

compliance, loans that help compliant systems maintain compliance, loans that aids




Dow

nloaded from

non-compliance

non-


compliant systems to meet anticipated future standards, and loans not related to

compliance), and assistance for small systems is the percentage of spending

dedicated to systems serving 10,000 or fewer people. We estimate expenditures

using Tobit, with censoring at 0 and 100.3 The premise of the Tobit model is that a

latent dependent variable is equal to the observed dependent variable within some

range, but censoring creates inequality between the latent and observed values

outside of that range. It is an appropriate estimation strategy for percentages that

are only observed between 0 and 100.

Coding of agency type comes from an EPA report to Congress reporting agency

leadership as of 2001 and information from DWNIMS about agency leadership as

of 2010. Where agency leadership changed, we identified the date of change using

state program reports and news articles. The SRF implementation environment is

complex. In 68 percent of state-years, program management is shared among two

or more agencies. Consequently, our agency variables measure whether an

environmental or financial agency has any formal decision making authority in

program management. Both are treated as dichotomous variables, because agency

type categories are not mutually exclusive in cases with shared program

management. Appendix 1, available as supplementary data at Publius online,

reports DWSRF agency leadership by state for the last year in our analysis, 2008.

To test how problem conditions might moderate the influence of an agency’s

functional focus, we interact agency type with problem condition variables

appropriate for our two models: (i) water quality violations to examine discretion

in supporting projects not related to SDWA compliance; and (ii) the percentage of

small systems and fiscal integrity of the loan fund to examine discretion in

supporting small water systems. To measure compliance, we used data from the

Safe Drinking Water Information System (SDWIS) on the percentage of water

systems in a state reporting water quality violations. This variable is lagged by one

year to give a realistic portrayal of when program officials might learn about their

state’s performance in compliance. Data on system size also come from the SDWIS.

To measure a loan fund’s fiscal status, we calculated the fund’s cumulative balance

using data and a formula from DWNIMS.

Drinking water infrastructure investments are likely shaped by a range of factors

in addition to the functional expertise of implementing agencies and problem

severity. Therefore, we include a number of control variables in our models. State

demographic characteristics may help direct infrastructure investment.

Infrastructure spending may be partially predicated on population density and

community wealth. Our analysis includes controls for population size, per capita

income, and percentage urban. Drinking water infrastructure investments tend

to be relatively low profile and low salience decisions. But, political factors

may influence investment patterns (McCubbins, Noll, and Weingast 1987; Wood

1990). Therefore we control for state political composition by including two




Dow

nloaded from

http://publius.oxfordjournals.org/lookup/suppl/doi:10.1093/publius/pjt033/-/DC1

1

2


variables: party control of the governor and the legislature and public opinion

liberalism. Given the sheer need for widespread investment in water infrastructure

(U.S. EPA 2002) and the human capital needed to oversee this investment, we

include a measure to control for bureaucratic capacity: number of state employees

per 1,000 residents. Similarly, infrastructure investment is likely to be controlled by

overall financial constraints. Therefore, we include a measure indicating the total

level of assistance awarded by a DWSRF each year and the cumulative funds

available.4 Because the agency variables that interest us vary little over time, we

omit state fixed effects but include fixed effects for year and Census region and

calculate standard errors with observations clustered by state.

Results and Discussion

Table 1 presents results of the two Tobit models predicting state discretion in SRF

programs. Summary statistics for all variables are given in Appendix 2 available as

supplementary data at Publius online. Omitted from the analysis are two state-years

with no DWSRF spending and, in the compliance model, six observations with

missing data.

Investment Outside of SDWA Compliance

Our first model examines the exercise of discretion when states invest in water

infrastructure that does not specifically advance SDWA compliance. This type of

investment can include projects that advance green planning, protect source water,

integrate technological innovations in drinking water systems, and improve

drinking water flow. We predicted that environmental agencies responsible for SRF

programs would invest more in noncompliance-related projects because these

professionals are in the best position to view the broad environmental and public

health benefits from targeting green infrastructure and technological innovation,

the type of projects that would typically fall into the noncompliance category (H1).

We also predicted that differences between agency types would be smaller where

water quality problems are more severe (H2). Model I in table 1 displays estimates

from our Tobit models.

To interpret the interaction between an agency’s functional focus and problem

conditions, we used the Tobit estimates to calculate average predicted differences in

DWSRF loan allocations between agencies of different types. Holding all other

variables at their actual values, we generated predicted loan allocations for each

observation with the agency variable set either to 0 (in this case indicating no

environmental agency management) or to 1 (in this case indicating environmental

agency management), accounting for censoring of the dependent variable.

Calculating the difference between these probabilities and averaging across all

observations produces the average predicted difference based on agency type.




Dow

nloaded from

available on-line


non-compliance

non-


Testing the first hypothesis, we estimate that environmental agencies dedicate 2.03

percentage points more of their loan funds to noncompliance-related projects, but

this difference is not significant.5

We find more support for the hypothesis that differences between agency types

will vary across problem conditions (H2). The left panel of figure 1 depicts the

average predicted values of funding for projects not compliance related by agency

type across the range of values of compliance. The right panel shows the average

predicted difference between agency types. As seen in the right panel,

environmental agency management has a positive and significant relationship

Table 1 State discretion in DWSRF spending, 2000–08

I: Loans for projects

not compliance related

II: Loans for

small systems

Functional focus

Environment agency 57.425** (24.028) –

Finance agency – –414.942*** (131.017)

Problem conditions

Percent systems in violation, lagged 3.203 (1.980) 1.273** (0.523)

Percent small systems 1.589 (0.978) –0.175 (0.710)

Ln(funds available) –3.561 (15.912) 7.142 (6.763)

Functional focus * Problem conditions

Environment * percent systems in

violation, lagged

–4.987*** (1.834) –

Finance * percent small systems – 2.025*** (0.786)

Finance * Ln(funds available) – 12.415** (4.879)

Control variables

State employees per 1,000 population 2.781 (2.010) –1.330*** (0.446)

Governor party, lagged 3.694 (4.634) 2.224 (1.947)

Legislature party, lagged –0.477 (8.338) –4.918 (4.228)

Opinion liberalism 0.216 (1.147) 1.210*** (0.409)

Ln(population) 12.850 (13.722) –0.729 (5.766)

Ln(per capita income) 25.820 (23.250) –9.124 (11.893)

Percent urban –0.421 (0.695) –0.453** (0.214)

Ln(DWSRF assistance) 2.212 (3.441) –24.446*** (2.262)

Constant –568.477 (407.552) 509.055*** (185.577)

N 442 448

Pseudo R2 0.036 0.064

Note. Models estimated using Tobit with censoring at 0 and 100. Year and census region fixed

effects not shown. Standard errors clustered by state.

***p5.01, **p5.05, *p5.10 (two-tailed).




Dow

nloaded from

non-compliance


(p5.05) with the percentage of loans dedicated to projects that are not compliance

related where compliance rates are high overall. Where just 1 percent of a state’s

water systems are in violation of drinking water quality laws, environmental

agencies award 13 percentage points more in noncompliance-related funding.

Environmental agencies scale back and behave more like other agency types when

violation rates are more severe.6 When levels of noncompliance are high, we detect

no difference between agency types.7

The model explains little of the overall variation in support for projects not

related to SDWA compliance. Political control does not predict patterns of SRF

spending; the estimated effects of the party of the governor and the legislature are

small and insignificant. Given the relatively low salience of drinking water

infrastructure investment, it is not surprising that political variables have no

consistent influence in our model. The only control variable that has a nearly

significant relationship with investment in projects that are not compliance related

is the percentage of water systems serving 10,000 or fewer customers (p¼ .101).

Both state and federal officials charged with drinking water protection have

particular concern about small water systems, which often lack the resources

Figure 1 Agency type and loans for projects not compliance-related, by violations

Note. Estimates calculated from Tobit coefficients shown in table 1. Estimates are expected values,

accounting for censoring. Shaded area shows a 95 percent confidence interval.




Dow

nloaded from

one

non-compliance

non-


necessary to maintain complex infrastructure and comply with water quality

regulations. Consequently, a number of states give funding priority to projects that

might not address specific compliance problems but instead consolidate small

systems or help small systems improve their operating capacity.

Investment in Small Water Systems

Loans to small water systems support the EPA’s goal of improving the performance

of these systems, but they are higher risk investments that offer no greater reward

than loans to larger, more financially sound water systems. We predicted that loan

funds managed in part by financial agencies would give greater attention to a

fund’s risk profile (H3). The pattern that emerges in this analysis is similar to that

in noncompliance-related spending: agency type does not have a direct relationship

with patterns of DWSRF investment when examined across the full range of

observations, but it appears to shape how program administrators respond to

problem conditions and fund status.

Using average predicted differences to interpret the interaction depicted in

Model II of table 1, DWSRFs managed by financial agencies do not dedicate a

significantly smaller percentage of their loan awards to small systems on average.

However, we find a significant difference between agency types in states where

small water systems are less prevalent (H4). Figure 2 shows model predictions

across the range of values of small system prevalence. The majority of community

water systems—92 percent in 2008—serve fewer than 10,000 people, although these

small systems serve just 18 percent of the total population receiving water from

community water systems (U.S. EPA 2011). As figure 2 demonstrates, in states with

a low relative percentage of small systems, DWSRFs managed by financial agencies

provide a significantly lower percentage of total funds to these small systems than

funds managed without financial agency involvement. The difference is as much as

32 percentage points at the lowest rate of small systems (69 percent). As the

percentage of small systems in a state rises, financial agencies appear to respond to

their need and commit significantly more funds to them. Thus the differences

across agency types dissipate with a higher percentage of small systems.

Results of the analysis of fund availability appear in figure 3. For all SRFs, the

proportion of funds made available to small systems rises as the overall SRF fund

balance rises. Consistent with H5, this relationship is stronger among financial

agencies than among other agency types and only significant among financial

agencies. By our measure of funds’ fiscal integrity—calculated using an EPA

formula—financial agencies appear to be more responsive than other agency types

to fund scarcity when making loan decisions. The result is that financial agencies

are more generous than other agency types to small systems where the status of

their funds is healthy, a finding that we had not anticipated. However, it may be




Dow

nloaded from

non-compliance


that financial agencies are in a better position to evaluate the fiscal health and

potential financial risk of small systems compared to their public health or

environmental counterparts. This expertise may foster investment while public

health and environmental officials may adopt a conservative investment approach,

even as their fund balance increases, to avoid jeopardizing infrastructure investment

over the long term.

Several control variables are related to small system assistance. States with a large

percentage urban area dedicate significantly less funding to small systems; urban

states have a lower percentage of small systems, and loans to large urban water

utilities can easily crowd out other SRF loan activity. Holding constant fund

availability, the amount of assistance an SRF awards in the current year also has a

negative relationship with small system support, likely for the same reason that

large loans can dominate fund activity. Moreover, our measure of state capacity,

the number of state employees per 1,000 residents, is associated with lower rates of

small system funding. As further evidence that DWSRFs are responding to state

need, assistance for small systems is positively related to compliance with the

SDWA, with a one-point increase in the percentage of systems in violation yielding

Figure 2 Agency type and loans for small systems, by small systems






Dow

nloaded from


a 0.98-point increase in percentage of funds to small systems. Considering that the

vast majority of SDWA violations occur in small systems (92 percent in our

dataset), increasing support for small systems is a sensible policy response to

noncompliance. The only political variable showing any association with SRF

funding patterns in either of our analyses is citizen ideology. Here, a shift from the

most conservative to the most liberal state is associated with a 34-point increase in

the percentage of funds going to small systems.

Conclusion

The devolution of policy authority to the states sometimes involves explicit

delegation of decision making responsibility with the hope that states will design

programs that are responsive to state-level conditions. However, the exercise of

discretion by state agencies implementing devolved policies typically is treated in

the literature as a departure from policy makers’ goals. We begin our analysis from

a different perspective: recognizing that states might be asked to exercise discretion

in policy implementation and to produce variable outcomes, but that state

Figure 3 Agency type and loans for small systems, by fund level






Dow

nloaded from

-


bureaucrats operate in a murky environment with multiple political principals and

information constraints that can limit their ability to direct program implemen-

tation toward addressing their own state’s needs. Our goal was to explore the extent

to which states utilize their discretion in a multilevel program designed to foster

divergent approaches and to examine how the functional focus of an agency affects

responsiveness to problem conditions.

We predicted that initial decisions over allocating program responsibility to an

agency might have long-term effects on program implementation. State revolving

funds provide a rare opportunity to examine bureaucratic discretion and agency

expertise, because the same program is managed in different states by agencies with

distinct professional expertise and organizational norms. Looking across eight years

of data and two areas of water infrastructure expenditure, we find no direct

relationship between agency type and the exercise of state discretion. However, our

analysis indicates that agencies do use their discretion to respond to problem

conditions in their state, and that patterns of responsiveness vary across agency

types.

We find evidence that agencies’ functional expertise is associated with decision

making in areas where the state does not face a pressing problem, but that problem

severity diminishes differences between agency types. This is a particularly

interesting finding that highlights how problem conditions constrain agency

discretion. While some scholars equate discretion with agency drift, our results

emphasize the interplay of agency discretion and problem conditions. When

problems are severe, state agencies—regardless of their functional expertise—focus

their attention toward the problems in their state. For example, where a state has

high levels of compliance with drinking water quality regulations, environmental

agencies that manage DWSRF funds are more likely to engage in innovative

investment, that is, investment that is not directly aimed toward specific drinking

water contaminants. Differences between environmental and other agencies in

compliance-related spending disappear where a higher percentage of water systems

are in violation of water quality laws. Similarly, financial agencies demonstrate less

support for the EPA priority of funding small systems in states with relatively fewer

small systems, but where small systems are more prevalent, their support level rises

and they behave more like other agency types. Financial agencies also dedicate a

smaller percentage of loans to small systems where state fund balances are low,

behaving more like other agency types. Consistent with research by Mullin (2009),

we find that problem conditions moderate the relationship between institutional

design and outcomes, such that institutional factors have less influence where

problems are more severe. For federalism scholars, our findings lend support to one

of the underlying tenants of decentralization: subnational governments are in a

strong position to identify and respond to local problem conditions, regardless of

specific details of agency and program design.




Dow

nloaded from

--

--

i.e.


These results also reinforce the lessons from public administration research

highlighting the importance of organizational identity (Bardach 1998; Wilson

1989). Although locating an SRF program in a particular type of agency does not

directly produce a certain outcome, our results suggest that institutional location

influences how implementing officials perceive and respond to problem conditions

within a state. This creates opportunity for state political actors to exert ex ante

political control over a multilevel program by situating the program in an agency

that will be most sympathetic to the problems or goals the politician cares about

(McCubbins, Noll, and Weingast 1987; Moe 1989).

Ultimately, our research highlights the importance of considering the intent

behind devolution, delegation, and political control. State agency behavior in the

context of multilevel programs, even while those programs become increasingly

common, remains understudied. Our results suggest that state agencies exercise

discretion when substantive problems are not severe. Federal actors cede

responsibility to subnational units for a multitude of reasons and create complex

multilevel governance networks where state agencies must operate. Yet, research to

date tends to treat delegation as a necessary evil that will result in bureaucratic drift

with agencies deliberately pursuing their own goals. Given the near constant

drumbeat of devolution across multiple substantive areas, it is important to

consider devolution and delegation that is designed to promote state discretion and

determine the ways in which that discretion operates.

Supplementary Data

Supplementary data can be found at www.publius.oxfordjournals.org.

Notes

This project was supported by a Faculty Senate Seed Money Grant from Temple

University. The authors thank Heather Bosak and Arnold Kim for their help with

data collection.

1. The original goal was for SRFs to become self-sustaining, but implementation problems

combined with ongoing needs have led to continued annual Congressional appropriations.

The magnitude of federal appropriations for both Clean Water and Drinking Water SRFs

has declined over time, until the 2009 American Recovery and Reinvestment Act (ARRA)

dedicated $6 billion in emergency appropriations for SRF funds. Detailed information on

capitalization grants is available at: http://water.epa.gov/grants_funding/dwsrf/dwnims.cfm

2. We end our analysis in 2008 because new rules under the ARRA affected SRF decision

rules in many states even for non-ARRA spending.

3. Because states may report negative spending for accounting purposes, a total of nine

observations across the two analyses have values either less than 0 or greater than 100.




Dow

nloaded from


www.publius.oxfordjournals.org

http://water.epa.gov/grants_funding/dwsrf/dwnims.cfm


Results are robust to estimating the models using ordinary least squares or to omitting

the outlying observations, and both strategies narrow the confidence intervals in the

tails. We present the more conservative estimates here acknowledging the censoring that

applies for most observations.

4. All variables but one are annual data or, in the case of the urban variable, linearly

interpolated from decennial Census data. The exception is the measure of public opinion

liberalism, a time-invariant indicator calculated by Erikson, Wright, and McIver from

public opinion polls conducted from 1976 to 2003. The authors have demonstrated that

state-level ideology has shown considerable stability over several decades (Erikson,

Wright, and McIver 2006).

5. Because we interact agency type with variables measuring problem conditions, the

coefficient in table 1 for environment agency indicates the difference in spending on

loans for projects that are not compliance related where the lagged percentage of systems

in violation is 0. As shown in the right panel of figure 1, environmental agencies do

spending significantly more on noncompliance-related spending in these high

compliance states, but the difference associated with agency type is not significant

when analyzed across the full range of observations at all levels of compliance.

6. These results are even stronger when we restrict our analysis to focus on cases where a

single agency implements the state’s DWSRF program. Under these conditions, an

agency’s functional focus should have a stronger influence on investment decisions

compared to situations where implementation is shared among agencies. When

environmental agencies implement DWSRF programs on their own, they dedicate a far

greater percentage of their funds to projects not related to compliance (endeavors like

source water protection, ecosystem management, and green infrastructure). The

functional focus of an agency has a greater impact where agencies act alone then where

they operate in cooperation with other agencies. The interaction with compliance is also

stronger among single agencies, so that agency type becomes insignificant where problem

severity is high.

7. Similarly, we find that agency type has no relationship with loan allocations that should

allow less agency discretion. In a placebo test, we regressed agency type on the

percentage of loan funds dedicated to projects for noncompliant systems to meet

compliance (rather than projects for maintaining compliance or meeting future

standards, or those not compliance-related). Helping systems that are in violation to

meet current regulatory standards should be the top priority for any funding program,

and indeed we find no difference between agency types regardless of statewide

compliance rates.

References

Agranoff, Robert. 2007. Managing within networks: Adding value to public organizations.

Washington, DC: Georgetown University Press.

Agranoff, Robert, and Michael McGuire. 2003. Collaborative public management: New

strategies for local governments. Washington, DC: Georgetown University Press.




Dow

nloaded from


Atlas, Mark. 2007. Enforcement principles and environmental agencies: Principal-agent

relationships in a delegated environmental program. Law & Society Review 41: 939–80.

Bardach, Eugene. 1998. Getting agencies to work together: The practice and theory of

managerial craftsmanship. Washington, DC: Brookings Institution Press.

Barrilleaux, Charles, and Mark Miller. 1988. The political economy of state Medicaid policy.

American Political Science Review 82: 1089–107.

Bednar, Jenna. 2011. Nudging federalism towards productive experimentation. Regional &

Federal Studies 21: 503–21.

Berry, Frances Stokes, and William D. Berry. 1999. Innovation and diffusion models in

policy research. In Theories of the Policy Process, ed. P. Sabatier. Boulder, CO: Westview

Press.

Black, Harvey. 2000. Environmental and public health: Pulling the pieces together.

Environmental Health Perspectives 108: A513–15.

Boushey, Graeme, and Adam Luedtke. 2011. Immigrants across the U.S. Federal Laboratory:

Explaining state-level innovation in immigration policy. State Politics & Policy Quarterly

11: 390–414.

Brehm, John, and Scott Gates. 1997. Working, shirking, and sabotage: Bureaucratic response to

a democratic public. Ann Arbor: University of Michigan Press.

Calvert, Randall, Mathew McCubbins, and Barry Weingast. 1989. A theory of political

control and agency discretion. American Journal of Political Science 33: 588–611.

Copeland, Claudia. 2010. Rural water supply and sewer systems: Background information.

Congressional Research Service Report RL7-5700. Washington, DC: Congressional

Research Service.

Daley, Dorothy, and James Garand. 2005. Horizontal diffusion, vertical diffusion, and

internal pressure in state environmental policymaking, 1989–1998. American Politics

Research 33: 615–44.

Downs, Anthony. 1964. A theory of bureaucracy. Santa Monica, CA: Rand Corporation.

Erikson, Robert S., Gerald C. Wright, and John P. McIver. 2006. Public opinion in the states:

A quarter century of change and stability. In Public opinion in state politics, ed. Jeffrey

E. Cohen. Palo Alto, CA: Stanford University Press.

Fellowes, Matthew, and Gretchen Rowe. 2004. Politics and the new American welfare states.

American Journal of Political Science 48: 362–73.

Gerber, Brian, and Paul Teske. 2000. Regulatory policymaking in the American states: A

review of theories and evidence. Political Research Quarterly 53: 849–86.

Gerber, Elisabeth, and Clark Gibson. 2009. Balancing regionalism and localism: How

institutions and incentives shape American transportation policy. American Journal of

Political Science 53: 633–48.

Gailmard, Sean, and John Patty. 2007. Slackers and zealots: Civil service, policy discretion,

and bureaucratic expertise. American Journal of Political Science 51 (4): 873–89.

———. 2012. Formal Models of Bureaucracy. Annual Review of Political Science 15: 353–377.




Dow

nloaded from


Gordon, Larry. 1995. Environmental health and protection: Century 21 challenges. Journal of

Environmental Health 57: 28–34.

Hammond, Thomas, and Jack Knott. 1996. Who controls the bureaucracy?: Presidential

power, congressional dominance, legal constraints, and bureaucratic autonomy in a model

of multi-institutional policy-making. Journal of Law Economics & Organization 12: 119–66.

Hammond, Thomas, and Paul Thomas. 1989. The impossibility of a neutral hierarchy.

Journal of Law, Economics, & Organization 5: 155–84.

Huber, John, and Nolan McCarty. 2004. Bureaucratic capacity, delegation, and political

reform. American Political Science Review 98 (3): 481–94.

Jones, Bryan, Saadia Greenberg, Clifford Kaufman, and Joseph Drew. 1978. Service delivery

rules and the distribution of local government services: Three Detroit bureaucracies.

Journal of Politics 40: 332–68.

Kaufman, Herbert. 2006. The forest ranger: a study in administrative behavior. Special reprint

ed. Washington, DC: Resources for the Future.

Kincaid, John. 1998. The devolution tortoise and the centralization hare. New England

Economic Review May: 13–40.

Konisky, David. 2007. Regulatory competition and environmental enforcement: Is there a

race to the bottom? American Journal of Political Science 51: 853–72.

———. 2009. Assessing U.S. state susceptibility to environmental regulatory competition.

State Politics & Policy Quarterly 9: 404–28.

Konisky, David, and Neal Woods. 2010. Exporting air pollution? Regulatory enforce-

ment and environmental free riding in the United States. Political Research Quarterly 63:

771–82.

Kotchian, Sarah. 1997. Perspectives on the place of environmental health and protection in

public health and public health agencies. Annual Review of Public Health 18: 245–59.

Lipsky, Michael. 1980. Street-level bureaucracy: Dilemmas of the individual in public services.

New York: Russell Sage Foundation.

Maynard-Moody, Steven, and Michael Musheno. 2003. Cops, teachers, counselors: Stories from

the front lines of public service. Ann Arbor: University of Michigan Press.

McCubbins, Mathew, Roger Noll, and Barry Weingast. 1987. Administrative proce-

dures as instruments of political control. Journal of Law, Economics, & Organization 3:

243–86.

Meyers, Marcia, Norma Riccucci, and Irene Lurie. 2001. Achieving goal congruence in

complex environments: The case of welfare reform. Journal of Public Administration

Research and Theory 11: 165–201.

Moe, Terry. 1989. The politics of bureaucratic structure. In Can the Government Govern?, ed.

J. E. Chubb and P. E. Peterson, 276–329. Washington, DC: Brookings Institution Press.

———. 1990. The politics of structural choice: Toward a theory of public bureaucracy. In

Organization theory: From Chester Barnard to the present and beyond, ed. O. E. Williamson,

116–53. New York: Oxford University Press.




Dow

nloaded from


Moreland-Russell, Sarah, Amy Eyler, Colleen Barbero, J. Aaron Hipp, and Heidi Walsh.

2013. Diffusion of complete streets policies across US communities. Journal of Public

Health Management and Practice 19: S89–96.

Morris, John. 1997. The distributional impacts of privatization in national water-quality

policy. Journal of Politics 59: 56–72.

———. 2010. Dirty water, clean water: Infrastructure funding and state discretion in

Southern states. In Speaking Green with a Southern accent: Environmental management and

innovation in the South, ed. Gerald Andrews Emison and John C. Morris, 45–64. Lanham,

MD: Lexington Books.

Mullin, Megan. 2009. Governing the tap: Special district governance and the new local politics

of water. Cambridge, MA: MIT Press.

Mullin, Megan, and Dorothy Daley. 2010. Working with the state: Exploring interagency

collaboration within a federalist system. Journal of Public Administration Research and

Theory 20: 757–78.

Oates, Wallace. 1972. Fiscal federalism. New York: Harcourt Brace Jovanovich.

O’Toole, Laurence. 1996. Hollowing the infrastructure: Revolving loan programs and

network dynamics in the American states. Journal of Public Administration Research and

Theory 6: 225–42.

Potoski, Matthew. 2001. Clean air federalism: Do states race to the bottom? Public

Administration Review 61: 335–42.

———. 2002. Designing bureaucratic performance: administrative procedures and agency

policy choice. State Politics and Policy 2 (1): 1–23.

Pressman, Jeffrey, and Aaron Wildavsky. 1973. Implementation: How great expectations in

Washington are dashed in Oakland; or, why it’s amazing that federal programs work at all,

this being a saga of the Economic Development Administration as told by two sympathetic

observers who seek to build morals on a foundation of ruined hopes. Berkeley: University of

California Press.

Ringquist, Evan. 1993. Environmental protection at the state level: Politics and progress in

controlling pollution. Armonk, NY: Sharpe.

———. 1994. Policy influence and policy responsiveness in state pollution control. Policy

Studies Journal 22: 25–43.

Scholz, John, and Feng Heng Wei. 1986. Regulatory enforcement in a federalist system.

American Political Science Review 80: 1249–70.

Shipan, Charles, and Craig Volden. 2006. Bottom-up federalism: The diffusion of

antismoking policies from U.S. cities to states. American Journal of Political Science 50:

825–43.

———. 2008. The mechanisms of policy diffusion. American Journal of Political Science 52:

840–57.

Sigman, Hilary. 2003. Letting states do the dirty work: State responsibility for federal

environmental regulation. National Tax Journal 56: 107–22.




Dow

nloaded from


Sinclair, Amber, and Andrew Whitford. 2012. Separation and integration in public health:

Evidence from organizational structure in the states. Journal of Public Administration

Research and Theory. Advance Access Published April 18; doi:10.1093/jopart/mus005.

Soss, Joe, Sanford Schram, Thomas Vartanian, and Erin O’Brien. 2001. Setting the terms of

relief: Explaining state policy choices in the devolution revolution. American Journal of


Soss, Joe, Richard Fording, and Sanford Schram. 2011. The organization of discipline: From

performance management to perversity and punishment. Journal of Public Administration

Research and Theory 21(suppl 2): i203–32.

Thomas, Craig. 2003. Bureaucratic landscapes: Interagency cooperation and the preservation of

biodiversity. Cambridge, MA: MIT Press.

U.S. Environmental Protection Agency. 1994. Federal register citations for the Safe Drinking

Water Act: A compilation of references. Office of Water EPA 810-B-94-004. Washington,

DC: U.S. EPA.

———. 1999. Prioritizing drinking water needs: A compilation of state priority systems for the

Drinking Water State Revolving Fund program. Office of Water EPA 816-R-99-001.

Washington, DC: U.S. EPA.

———. 2002. The clean water and drinking water infrastructure gap analysis. Office of Water.

EPA-816-02-020. Washington DC: U.S. Environmental Protection Agency.

———. 2010. Drinking Water State Revolving Fund 2009 annual report. Office of Water EPA

816-R-10-021. Washington, DC: US EPA.

———. 2011. National characteristics of drinking water systems serving 10,000 or fewer people.

Office of Water EPA 816-R-10-022. Washington, DC: U.S. EPA.

Volden, Craig. 2006. States as policy laboratories: Emulating success in the Children’s Health

Insurance Program. American Journal of Political Science 50: 294–312.

Watkins-Hayes, Celeste. 2011. Race, respect, and red tape: Inside the black box of racially

representative bureaucracies. Journal of Public Administration Research and Theory

21(suppl 2): i233–51.

Whitford, Andrew. 2002. Decentralization and political control of the bureaucracy. Journal of

Theoretical Politics 14: 167–93.

———. 2005. The pursuit of political control by multiple principals. Journal of Politics 67:

29–49.

———. 2007. Decentralized policy implementation. Political Research Quarterly 60: 17–30.

———. 2008. A test of the political control of bureaucracies under asymmetric information.

Rationality and Society 20: 445–70.

———. 2010. The hierarchical consequences of reinvention: evidence from the American

bureaucracy. Journal of Management History 16: 59–74.

Whitaker, Ingrid P., and Victoria Time. 2001. Devolution and welfare: The social and legal

implications of state inequalities for welfare reform in the United States. Social Justice 28:

76–90.




Dow

nloaded from


Wilson, James Q. 1989. Bureaucracy: What government agencies do and why they do it. New

York: Basic Books.

Wood, B. Dan. 1990. Does politics make a difference at the EEOC? American Journal of


Wood, B. Dan., and Richard Waterman. 1991. The dynamics of political control of the

bureaucracy. American Political Science Review 85: 801–28.

———. 1994. The bureaucratic dynamics: The role of a bureaucracy in a democracy. Boulder,

CO: Westview Press.

Woods, Neal. 2006. Primacy implementation of environmental policy in the U.S. states.

Publius: The Journal of Federalism 36 (2): 259–76.

Woods, Neal, and Matthew Potoski. 2010. Environmental federalism revisited: Second-order

devolution in air quality regulation. Review of Policy Research 27: 721–39.




Dow

nloaded from


state agency discretion in a delegated federal program ...investment (brehm and gates 1997; moe...

Documents