Theories of the Policy Process

by Christopher M. Weible

location-choice competition is relevant in the international setting for policies designed to influence a variety of choices of individuals or firms, including business location, individual and corporate purchasing decisions, and migration (Oates 2001; Genschel 2002).

In spillover-induced competition, an adoption by government B has a (positive or negative) externality effect on government A that changes A’s expected net benefit from adopting.

Coercion. Government A is coerced into adopting a policy when a more powerful government, B, takes action that increases A’s incentive to adopt or, in the extreme case, forces A to adopt.

horizontal coercion, in which a powerful country encourages a weaker country to adopt a policy, sometimes by threatening action if the weaker nation does not capitulate (Simmons, Dobbin, and Garrett 2006).

In other cases, the national government uses a “carrot” rather than a “stick,” by creating an incentive for a state to adopt a policy.

international organizations such as the International Monetary Fund (IMF) and the World Bank can play a similar coercive role by requiring a country to adopt some policy as a condition for financial aid (e.g., Weyland 2007; Simmons and Elkins 2004; Barrett and Tsui 1999).

Complicating matters, multiple mechanisms may underlie a policy’s diffusion.

each model relies on one or more of the five mechanisms described above to justify why governments emulate other governments when making public policy.

three models that dominated early scholarship on policy innovation. One—the national interaction model—assumes that policy diffuses because of learning. The others—the regional diffusion model and the leader-laggard model—are consistent with multiple mechanisms for diffusion.

The National Interaction Model This learning model was developed and formalized by communication theorists analyzing the diffusion of an innovation through a social system (assumed to be of ...

Early in the diffusion process, adoptions occur relatively infrequently. The rate of adoptions then increases dramatically but begins to taper off again as the pool of potential adopters becomes small.

First, the model assumes that, during any time period, all potential adopters that have not yet adopted are equally likely to do so; the only variable influencing the probability that a potential adopter will adopt during any time period is the cumulative number of adopters prior to that period.

Thus, the theory is well suited for when the social system is a large society of individuals and the scholarly interest is in a macrolevel description of the diffusion process.

But when studying the diffusion of a policy through the fifty states, it seems less reasonable to treat the states as undifferentiated units;

It is also likely that contacts between officials from different states are patterned rather than random.

The regional diffusion model posits that governments are influenced primarily by other governments that are geographically proximate.

Most of these models assume that governments are influenced exclusively by those jurisdictions with which they share a border; as such, we call them neighbor models.

Other models, which we term fixed-region models, assume that some collectivity of jurisdictions (e.g., the states in the United States, the countries in Latin America) is divided into multiple regions and that governments within the same region tend to emulate each other’s policies (e.g., Mooney and Lee 1995).

Fixed-region models presume (if only implicitly) that all governments within the same region experience the same channels of influence. In contrast, neighbor models—by avoiding fixed regional groupings of jurisdictions and instead pointing to the influence of all bordering jurisdictions—assume that each government has a unique set of reference jurisdictions for cues on public sector innovations.

neither pure model is entirely realistic.

a weakness of the regional diffusion model is that multiple mechanisms for diffusion can be the basis for an expectation that diffusion channels are regional.

The leader-laggard model assumes that certain jurisdictions are pioneers in the adoption of a policy and that other jurisdictions emulate these leaders (Walker 1969, 893).

This model can be modified easily, however, to reflect the notion of a national leader among US states or an international leader among countries: a jurisdiction that, when it adopts a new program, increases the likelihood that other jurisdictions, regardless of their geographical location, will adopt.

Like the regional diffusion model, the leader-laggard model can be motivated by multiple mechanisms for diffusion.

such models are often flawed by their failure to identify a priori (1) the jurisdictions (or even types of jurisdictions) that are expected to be pioneers, and (2) the predicted order of adoption of the governments expected to follow.

Such a hypothesis specifies (in a testable fashion) the characteristics of leaders (high economic development) and a clear ordering of successive adoptions (from most- to least-developed countries).

Internal determinants models presume that the factors causing a government to adopt a new program or policy are political, economic, and social characteristics of the jurisdiction.

these models preclude diffusion effects in which a government is influenced by the actions of other governments.

internal determinants models assume that once a jurisdiction’s policymakers are aware of a new policy, the jurisdiction’s internal characteristics, rather than pressure created by other governments’ adoptions or explicit evaluations of the impacts of the policy in earlier-adopting jurisdictions, determine its probability of adoption.

hypothesis that larger, wealthier, and more economically developed American states are more innovative.

Lawrence Mohr (1969, 114) proposes that the probability that an organization will innovate is inversely related to the strength of obstacles to innovation and directly related to (1) the motivation to innovate, and (2) the availability of resources for overcoming obstacles.

Factors reflecting the motivation to innovate. Numerous scholars have hypothesized that problem severity is an important determinant of the motivation to innovate.

Problem severity can influence the motivation of government officials to adopt a policy directly, by clarifying the need for the policy, or indirectly, by stimulating demand for the policy by societal groups.

elected officials should be responsive to public opinion when deciding whether to adopt a new policy, their responses should vary with their level of electoral security. The more insecure they feel, the more likely they will be to adopt new policies that are popular with the electorate, and the less likely they will be to adopt new policies that are widely unpopular, or at least sufficiently unpopular with some segment of the electorate to be deemed controversial.

politicians anticipating closely contested elections are especially likely to embrace new programs to try to broaden their electoral support. Implicit in this hypothesis is that the new programs are popular with the public. In the case of unpopular programs (like the imposition of a new tax), electoral competition is likely to reduce the probability that politicians will support the program.

Obstacles to innovation and the resources available to overcome them. Theories of individual and organizational innovation have stressed the importance of financial resources (i.e., wealth and income levels for individuals and slack resources for organizations) and other characteristics (e.g., a high level of education for an individual and large size for an organization) reflecting the capability of the potential adopter to innovate.

Walker (1969), Sigelman and Smith (1980), Andrews (2000), and McLendon, Heller, and Young (2005) maintain that American states with legislatures that give their members generous staff support and extensive research facilities should be more likely to adopt new policies than states with less professionalized legislatures,

Brooks (2005) posits that the level of party fragmentation in a country is inversely related to the country’s likelihood of innovation.

it can be argued that the capacity of a jurisdiction’s economy to finance extensive public services is the ultimate determinant of the jurisdiction’s prop...

Indeed, several theorists, recognizing the rarity of government innovation, have argued that innovation can be expected to occur only in the unusual case wherein various independent conditions happen to occur simultaneously.

The unit of analysis for this equation is the jurisdiction i eligible to adopt a policy (i.e., the jurisdiction that has not yet adopted the policy) in a particular year t. The dependent variable—ADOPTi,t—is the probability that jurisdiction i will adopt the policy in year t. EXTERNALi,t denotes variables reflecting diffusion effects on jurisdiction i at time t; thus, these variables would measure the behavior of other jurisdictions at time t, or in the recent past.

MOTIVATIONi,t represents variables indicating the motivation of public officials in jurisdiction i at time t to adopt the policy; these variables would include the severity of the problem motivating consideration of the policy, the character of public opinion and electoral competition in the jurisdiction, and other ad hoc motivation factors.

RESOURCES/OBSTACLESi,t denotes variables reflecting obstacles to innovation and the resources available for overcoming them. For many policies, the government’s level of economic development and the professionalism of its legislature would be among the variables included. Factors indicating the presence (and skill) of interested policy entrepreneurs, or the strength of advocacy coalitions, in a jurisdiction could also be included.

OTHERPOLICIESi,t is a set of variables indicating the presence or absence in jurisdiction i of other policies that have implications for the likelihood that ...

Mahajan and Peterson (1985, 39–40) identify four types of “innovation interrelationships”: innovations may be (1) independent, (2) complementary, (3) contingent, or (4) substitutes.

If we are seeking to explain the adoption of policy B, and policy A is largely independent of B (in the sense that a jurisdiction’s probability of adopting B is unaffected by whether it has already adopted A), obviously we need not concern ourselves at all with policy A.

Sometimes two policies are complementary: the adoption of policy A increases the probability that a government will adopt policy B.

Another possibility is that policy B’s adoption is contingent on the previous adoption of policy A, in which case the probability that a jurisdiction will adopt B is zero until the jurisdiction adopts A.

A final alternative is that policy A is a substitute for policy B. When A is an exact substitute for B, A’s adoption completely precludes the possibility of adopting B. However, exact policy substitutes are rare; partial substitutes are more likely. In the case of partial substitutes, the adoption of A does not preclude the adoption of B; it only reduces its likelihood.

Berry’s results, which we summarize here, paint a very pessimistic picture of the ability of the traditional methodologies to help us understand government innovation.