ated under the Telecommunications Act of 1996 and was set up as a $2.25
billion annual fund to provide discounts (between 20 percent and 90 per-
cent) to schools and libraries for connectivity costs for the Internet (Carvin,
Conte, & Gilbert, 2001).
Programs meant to expand Internet access were downsized or eliminated
under the George W. Bush administration. During his tenure, the problem
of disparities in Internet usage was recast in terms of technological literacy.
His administration worked to reduce inequalities in usage by improving
computer skills through the No Child Left Behind Act (NCLB). This act
expanded the defi
finition of literacy to include technical competence and
information literacy and provided money to schools for a broader variety
of Internet resources such as teacher training, support staff, and software
(Dickard, 2003).
State Response to Obama’s Broadband Access Policy 245
The Obama administration has redirected the federal government
response to Internet inequalities to a strategy more closely resembling Bill
Clinton’s by focusing on increasing Internet access but with an emphasis of
addressing infrastructure barriers to residential areas. Under the American
Recovery and Reinvestment Act of 2009, $7.2 billion was provided to states
and local governments to extend high-speed broadband access to rural areas
that have not been served by existing broadband providers. The money was
distributed through the Broadband Technology Opportunities Program
(BTOP) and the Broadband Initiatives Program (BIP). The two programs
provide grants and loans to state and local communities to either update
existing telecommunication infrastructure or to put into place necessary
infrastructure to provide broadband service to areas that do not currently
have broadband access or are underserved. The Obama administration has
also enlisted the aid of the private sector through awarding $100 million
dollars under the Recovery Act to four satellite companies to help broaden-
ing access to broadband in rural area (Recovery Accountability and Trans-
parency Board, 2010). Additionally, these programs are complemented by
the State Broadband Initiative. This program was created in 2009 as a joint
venture between the Recovery Act and the Broadband Data Improvement
Act. Currently, it has awarded $293 million in grants to support the use of
broadband for projects that help the states compete in the digital economy,
including expansion of Internet access (NTIA, 2011a).
3 COMMUNICATIONS MODEL
While each of the three administrations took a diff
fferent approach to
addressing the issue of inequality in Internet usage, the success of each of
the policies relied heavily on the cooperation from state and local govern-
ments. In exploring variation to state compliance with the current adminis-
tration’s policy, this study turns to Goggin et al.’s (1990) Communications
Model designed to frame intergovernmental policy. The goal of this model
is to depict implementation over time and determine why there is variation
in how states implement federal laws. The dependent variable under this
framework is state implementation. Specifi
fically, the dependent variables
include outputs and outcomes. Outputs can be characterized as agency
eff
fforts. Outcomes involve the impact that the law had on society (1990).
The intervening variables are state organizational and ecological capaci-
ties or resources that allow the state to ignore messages from other politi-
cal actors. State organizational capacity refers to items such as a state’s
administrative effi
fficiency and competency. State ecological capacity refers
to factors such as the partisan make-up of the governor’s offi
ffice and the
state legislature (1990). The independent variables are federal-level and
state-level inducements and constraints. An example of a federal level
inducement is the allocation of resources to implement a law. Conversely, a
246 Ramona
McNeal
restraint would include sanctions against states that fail to implement a law
as directed (Goggin et al., 1990).
Goggin et al. (1990) argue that communications takes center stage in imple-
mentation. The message and content of the policy, in addition to the level of
communication federal agencies have with state and local implementation
agencies, is also likely to aff e
ff ct the success or failure of the implementation
of a law. If state and local implementers regard the message and content as
credible, their execution of the law is more likely to mirror its original intent.
Typically, higher levels of communication facilitate better implementation.
Also, the less communication there is, the more likely that competing mes-
sages from other political actors will result in the implementation of policy
that deviates from its original design (Goggin et al., 1990).
A review of the Communication Models suggests a number of factors
that can infl uenc
fl
e state compliance with federal policy. In the next section,
their infl u
fl ence—including that of state organizational and ecological capac-
ity on the state-level response to one such U.S. telecommunication policy,
the Obama administration’s programs to extend Internet access—will be
examined. Multivariate logistic regression analysis of fi
fifty state data will
be used to test rival factors for state response as of December 2010.
4 DATA AND MEASUREMENT
The dependent variable is constructed to measure the extent of broadband/
high-speed Internet access policy in a state. It measures whether the state
has taken action that is in compliance with the Obama’s administration
policy to expand broadband access either through authorizing fi
financial
support to provide infrastructure to facilitate broadband access or working
with local government to provide municipal owned broadband service. It
is coded 1 if such actions were taken and 0 otherwise. The dependent vari-
able was created from a summary of current state broadband laws avail-
able through the National Council of State Legislatures (2010a). Because
of problems related to multicollinearity, two models will be presented. The
dependent variable remains the same but several independent and control
variables change between models.
The main independent variables under the Communication Model are
federal-level and state-level inducements and constraints. The federal gov-
ernment can compel the states to act through inducements such as grants,
constraints such as sanctions, or a combination of both. The Obama
administration is relying entirely on inducements in the form or grants
and loans to encourage state compliance. In Model 1, two variables are
included to measures these inducements. The fi
first measure is the amount of
grants in 100’s of millions of dollars awarded to each state under the State
Broadband Initiative program (NTIA, 2011a). The second is the amou
nt of
grants rewarded to the states for infrastructure under the BTOP program
State Response to Obama’s Broadband Access Policy 247
in millions of dollars (NTIA, 2011b). This second measure is not included
in Model 2.
Actors at the state and local level (interest groups, local offi
fficials and
agencies) can shape the implementation of legislation. Depending on how
legislation impacts local groups, they may act either to boost or to hinder
implementation. Because e-government may have the ability to increase
political engagement and facilitate a more participatory democracy (Pardo,
2000), it is expected that good government groups would play an impor-
tant role in supporting the extension of broadband. The number of good
government groups in a state was included as a control for interest group
strength (Project Vote Smart, 2009). One group proven to influenc
fl
e state
Internet access policy is the telecommunication service providers. Acting as
a proxy for the strength of telecommunication service providers is whether
or not a state has passed laws that either restrict or prohibit municipal
owned broadband service (a policy that was hard fought for by service pro-
viders). The measure is coded 1 if the state has such a law and 0 otherwise
(Baller Herbst Law Group, 2011).
Although actors at the federal, state, and local levels may attempt to influ-
fl
ence state policy, states may still disregard these players and enact its own
preferences. This can occur if the “messages” sent by these actors are not
considered credible. Credibility is based on a number of factors including clar-
ity of message, accompanying resources to implement a policy and whether
the “message” came from an actor who is perceived to be credible and legiti-
mate. How much leeway a state has in disregarding such messages is based on
state resources. The ability of states to discount outside messages is defi ne
fi d by
Goggin et al. (1990, p. 119) as state capacity. This capacity falls into two cat-
egories: ecological capacity and organizational capacity. Ecological capacity
concerns the “contextual environment in which state government operates”
(Goggin et al. 1990, p. 911). The state operates within three environments:
economical, situational and political. Economical capacity concerns the avail-
ability of monetary resources. The ability of a state to decline federal grants
and loans for Internet access depends on state wealth. Following Walker
(1969) in Model 2, educational attainment measured as the percent of the
state population over the age of 25 with a bachelor’s degree or higher (U.S.
Census Bureau, 2012) is included as a measure of societal resources.
The political environment includes both the attitudes of the citizens as well
as public offic
ffi ials. A number of factors can infl uenc
fl
e the opinion of policymak-
ers. The fi r
fi st is partisanship. A measure of party control of the government is
included, coded 1 if the Republican Party controls both houses of the legis-
lature and the governorship, 0 if control is divided between the two parties,
and–1 if the Democratic Party controls both houses of the state legislature and
the governorship (National Council of State Legislatures, 2010b). Research
on partisanship and e-government (McNeal et al., 2003; Tolbert, Mossberger,
& McNeal, 2008) found a positive relationship between Republican con-
trolled legislatures and implementation of e-government policies. Both studies
248 Ramona
McNeal
concluded that states with Republican controlled legislatures were more likely
to be innovators in e-government because of the belief that e-government would
increase both effi
c
ffi iency and cost savings. To control for the possible response
to citizen concerns, included in the models is the voting age turnout in the state
for the 2010 midterm election (United States Election Project, 2011).
The fi
final area of ecological capacity is state situational capacity. Goggin
et al. (1990, pp. 145–6) include in this category such factors as public aware-
ness. States are more likely to respond to an issue if the public believes that
a problem exists. Measures of demand, such as the number of Internet users
in a state or problem severity, may also aff e
ff ct policy adoption and the scope
of implementation (Goggin et al., 1990). Several measures were included for
barriers to Internet access. The fi
first is the average number of computers avail-
able for public use per public library in 2009 (U.S. Census Bureau, 2012).
The greater the access to the Internet in public places such as the library, the
severity of the problem may not seem as acute. The second measure is the per-
centage of rural areas in a state with three or fewer wireless providers (NTIA,
2011c). Citizens living in rural areas often have limited choices for service
providers and therefore connection fees that are considerably higher than in
urban/suburban areas (Stover, 1999). Because the cost of Internet service tends
to be lower in more densely populated areas, a measure of state population
density calculated by the population per square mile was included (U.S. Cen-
sus Bureau, 2011). In states with greater population density, there may be less
of a pressing need to implement these policies. In Model 2, a fourth variable
measuring the percentage of households with Internet access within the state
in 2009 was included (U.S. Census Bureau, 2012). The greater the percent-
age of households with Internet access, the less likely the state will feel that it
needs to implement additional policies to further extend broadband access.
While ecological capacity focuses on the environment in which policy
implementation takes place, organizational capacity concerns the resources
available to the state agencies that oversee policy implementation. Institutional
capacity includes items such as a state’s administrative effi
c
ffi iency and compe-
tency (Goggin et al., 1990). As a measure of ecological capacity, included is
an indicator of whether the state has an existing broadband task force, com-
mission, or authority to oversee state-level broadband initiatives (National
Council of State Legislatures 2010c). It is coded 1 if such an agency exists and
0 otherwise.
Markell (1993) suggests that measures of resources include a strong record
of policy implementation. States that have a history of innovation in an issue
area may be more likely to continue placing new ideas on the table and imple-
ment additional programs. Several measures of innovation in e-government
are included. In Model 2 is included an index that measure state innovation
in electronic commerce (Atkinson & Wilhelm, 2002). The second measure
is West’s (2007) innovation index—a measure of the overall state ranking of
government websites. The fi
final measure is a count of social network sites such
as Faceboo
k and Twitter being utilized by legislative agencies and caucuses in
a state (National Council of State Legislatures, 2011).
State Response to Obama’s Broadband Access Policy 249
5 FINDINGS
AND DISCUSSION
In Table 18.1, the dependent variable is coded so that higher scores are associated with increased likelihood of adopting state policies that are
in compliance with the Obama administration’s policy goals of extend-
ing Internet access to underserved areas. Because the dependent variable
is binary, logistic regression models are used. Although two models were
explored because of muticollinearity concerns, the fi
findings were the same.
The same subset of variables was found to be significant in both mod-
els. The fi
findings suggest that a limited number of variables including State
Broadband Initiative grants, good government interest groups, Republi-
can control of state government, number of computers per library, West’s
Table 18.1 State-Level Broadband Access Policy Implementation
Gov’t Ownership or
Gov’t Ownership or
Financing (Model 1)
Financing (Model 2)
Variables
b (se)
p>|z|
b (se)
p>|z|
Federal-Level Inducements and
Constraints
State Broadband Initiative
.22(0.09)
.019
.27(0.13)
.042
Grants
BTOP Infrastructure Monies
.02(0.14)
.225
—
—
State-Level Inducements and
Constraints
Barrier Laws
4.12(2.76)
.132
2.87(2.02)
.156
Good Government Interest
-1.46(0.68)
.033
-1.28(0.60)
.034
Group Strength
Ecological Capacity
Republican Government
3.97(2.41)
.098
3.03(1.69)
.073
Control
Voter Age Population Turnout
-.45(0.31)
.140
-.35(0.24)
.145
Percent With Bachelor’s Degree
—
—
-.34(0.46)
.457
Rural Broadband Availablity
.06(0.05)
.250
.06(0.05)
.166
Population Density
3.4E-3(3.7E-3)
Public Sector Transformation Through E-Government Page 43