Abstract
In this chapter, an overarching model of implementation will be formulated on the basis of the theories discussed in the previous chapter. This model contains the variables considered to be crucial for the course of implementation in the theories discussed: (a) the implementation agencies’ policy preferences, which include the salience they attach to issues; (b) the possibilities of monitoring and sanctioning implementation agencies’ activities; (c) the political control exercised on the policy implementation in the culture of the policy system; and (d) the level of consensus among political decision makers regarding the proposed decision.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
The possibility that implementation agencies are strategically dependent is not included in the model presented in this book. This dependency is undoubtedly relevant. However, the analysis of this dependency creates such a large amount of complexity that this detracts from the insight provided by the model presented here. Strategic dependencies exist not only between implementation agencies, but also between implementers and policy makers. As a result, the complexity of such a game theoretic model of implementation would be much greater than that of current multi-actor models of strategic decision making, such as for example, Bueno de Mesquita’s model.
In order to do this a number of key informants, or subject area specialists, are consulted. These individuals have a large amount of substantive knowledge concerning the policy area being investigated. Chapter four discusses key-informants in more detail.
For example, the decentralization of an organization leads often to an improvement in the provision of services. At the same time, however, the authority and the budget of the central organization (the management) is reduced. the organization must weigh these conflicting consequences against each other.
The assumption that loss functions decline monotonically is identical to the assumption that preference functions are single peaked. Whether the loss functions of all actors decline monotonically is, of course, dependent on the ordering in which the policy alternatives are represented on the policy scale (Coombs and Avrunin, 1988: 13). Another important assumption is that the preference loss increases monotonically. That monotonicity means that the preference loss will not decrease, but rather will increase for policy alternatives located further from the agency’s policy position on the policy scale.
A more general loss function, (xd — Dd)r, is used in modern decision making theory. If r is smaller than one, it is assumed that the actor is risk seeking. In the model of policy implementation developed in this book, however, it is assumed that agencies are risk averse (the value of r = 2).
The sensitivity to control is a weighting factor which indicates the extent to which a given policy deviation will lead to reputation loss. An agency’s with a greater sensitivity to control will suffer a larger reputation loss given a certain policy deviation than an agency will a smaller sensitivity to control. Sensitivity to control is therefore conceived of as the extent to which implementation agencies themselves perceive the likelihood of sanctions. Given that many different factors determine an agency’s sensitivity to control, this aspect of the model will be elaborated on in the future.
It seems plausible that an implementation agency’s sensitivity to control is also related to the substance of the political decision to be implemented. However, this variation is excluded because it would increase considerably the number of degrees of freedom in the model. In addition, it can be argued theoretically that it is highly costly for political decision makers to establish a separate sanctioning and monitoring instrument for every policy decision (cf. Robinson, 1989).
In chapter two, it was argued that different assumptions regarding political control result in different models of bureaucratic behavior: the implemented preference models from the neo-classical tradition, the political decision models from the neo-institutional approach, or the mixed models in which policy freedom and sensitivity to control are central.
The collective nature of political control within the (implementation) culture of a policy terrain should be distinguished clearly from the individual character which sensitivity to control has. Sensitivity to control originates in organizational characteristics which make it possible for the implementation agency to withdraw itself individually from political control. Political control originates in procedures which are generally binding and which compel adherence to the norm of loyalty (see chapter two).
Therefore, this involves an assumption regarding the value which implementation agencies attach to their reputation, credibility and reliability in relation to the value which they attach to their own preferences. the existence of strong political control will lead to implementation agencies evaluating their own activities on the basis of their reputations. the conditions under which this will occur were discussed in chapter two.
If = 0 and a= 0, then the weighting factor is equal to zero.
As ei approaches zero, approaches infinity. If ei is equal to zero, then is equal to one.
The simple model of implementation can be redefined in terms of a standardized policy deviation: in other words, the fraction of the distance between the agency’s policy position and the political decision which constitutes the policy deviation. the standardized policy deviation () is described by the following function
Therefore, it is extremely important to devote attention to the’ support’ for or acceptance of proposed political decisions by implementation agencies when preparing policy.
Ringeling (1978) conduced research on the factors which increase and decrease the policy freedom of bureaucrats. These factors encompass more than those referred to by the concept of room for maneuver in the simple model of implementation. the policy freedom is, according to Ringeling (1978: 69—70), dependent on five factors: (a) the information advantage; (b) the level of hierarchical control over the implementation agencies; (c) the programming of the policy, including the level of consensus among policy makers, which will be discussed extensively later in this chapter; (d) the extent to which bureaucrats strive toward their own goals, which are distinct form those of political decision makers (this is included in the incentive to deviate in the implementation model presented in this book); (e) the democratic control compelled by administrative procedures (this is described as normative control in the model of implementation presented here).
Of course, a special situation occurs when the effective control is equal to zero. In that case, the room for maneuver is not defined. Then it is assumed that the room for maneuver is infinitely large.
This can be illustrated by two extreme situations. If an implementation agency has no room for maneuver whatsoever (= 0), then, in formula (6), zero is divided by a. the preference weight is then zero: the implementation agency then attaches no weight whatsoever to its own policy position, and full weight to the political decision. It will not deviate. Alternatively, if an implementation agency has an infinitely large room for maneuver (→∞), then, in formula (6), an infinitely large number is divided by itself plus a the preference weight will then approach one. In that case, the implementation agency attaches full weight to its own policy position. It will then realize its own policy position.
A useful metaphor for understanding the working of the norm of bureaucratic loyalty is the following. Implementation agencies sometimes want to deviate from the political decision (incentive), and have the power to do so (room for maneuver). the norm works as a magnet which pulls all implementers toward the political decision, but with a different force because the room for maneuver differs between implementers.
This does not detract from the fact that the model of implementation is also applicable in political systems which are not democracies. In a dictatorship, the political decision will also be the guideline for the activities of implementation agencies.
The two-stage model of collective decision making is based on this principle (Stokman and Van den Bos, 1992; Berveling, 1994a).
The term overarching model of implementation is used because specifications can be derived from it which represent each of the theoretical traditions discussed in chapter two.
The preference weight consists of the implementation agency’s room for maneuver on the policy decision concerned, and the reputation sensitivity of the implementation agencies in the policy system.
This is the maximum deviation toward which an implementation agency strives. This maximum deviation is restricted to the limit of the toleration interval.
A comparable derivation can be represented for an agency i, which is highly sensitive to control, and an agency j, which has a low sensitivity to control.
Mathematically, it is true that the predicted policy performance of implementation agency j, which attaches a high level of salience to the decision, is located further away from the political decision than that of agency i:ôdi > ôdj.
In order for there to be a difference between the models’ predictions, the implementation agency must have an incentive to deviate. If this incentive is absent (if the political decision reflects the policy position of the implementation agency), all the models will predict that no deviation will occur.
We assume that the actual policy performance can be represented on the policy scale; this is the same assumption as that regarding the policy alternatives on the issue. This will be illustrated extensively in section 4.3.
The Mean Squared Error is a regular statistical measure used to determine how efficient an estimator is, when the extent to which this estimator is biased systematically is unknown (Wonnacott and Wonnacott, 1985: 197). a good estimator shows no bias and has a small spread.
Here, direction expresses nothing more than the position on the policy scale with respect to the political decision: left of the political decision (negative sign), or right of the political decision (positive sign).
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Torenvlied, R. (2000). Overarching Model of Policy Implementation. In: Political Decisions and Agency Performance. Library of Public Policy and Public Administration, vol 2. Springer, Dordrecht. https://doi.org/10.1007/978-94-011-4285-4_3
Download citation
DOI: https://doi.org/10.1007/978-94-011-4285-4_3
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-010-5857-5
Online ISBN: 978-94-011-4285-4
eBook Packages: Springer Book Archive