Chapter IV Establishing the concept of institutional-learning
F. Institutional-learning Framework and Policy Change
While the aim of this thesis is to analyse IL by the European Union in its state-building policies, the “institutional-learning Framework” (IL Framework) on which the following analysis is based also allows for and incorporates other elements which might have influenced policy change by the EU – besides that of IL itself. This inclusion should prevent any contentions of post hoc ergo propter hoc argumentation or “logical fallacy”, in the sense that one occurrence follows another and therefore the resulting occurrence must necessarily have been caused by the first, or that because the thesis sets out to find evidence of IL it is bound to find it.
Policy change occurs for numerous reasons and in a complex environment. Change does not occur in a vacuum and the more actors that are involved in the policy the more difficult it may be to change it. The different triggers initiating a policy change are sometimes difficult to distinguish, and in some instances it may be more than one influential trigger. Triggers are also not mutually exclusive. In some instances it may not be possible to prioritise which trigger is the “strongest” or most influential. What may be said with any sense of certainty is that the force of these triggers must have been strong enough for the actor to act upon them; change is not always simple to initiate without being faced with opposition (to different degrees).
Goldmann534 lists three dimensions influencing the extent to which a country’s foreign
policy is likely to change, dimensions which nicely summarise the triggers and incentives for governments to change and which can with little hesitation be translated into an EU setting: 1) ‘The degree of institutionalisation, or roughly the extent to which the government
has become committed to the policy.
2) The degree of support, or roughly the extent to which the various actors in domestic politics support, are indifferent to, or oppose the policy.
3) The degree of salience or roughly the significance of the issue in the domestic power struggle.’
The actors involved in this vary from elected officials, political parties, bureaucrats and civil servants, pressure groups, policy entrepreneurs and experts to supra-national institutions535.
534 Goldmann (1988), p. 44
136 When Etheredge introduces his overview of government learning he first analyses executive-branch learning as a ‘self-reflective study for the professional social scientist’536.
For him, learning has to take place on a presidential or high government official level, but this restriction is questionable. In most cases it can be agreed in the words of Etheredge that: ‘learning by government [sic] is often a function of current active political conflict’537.
This conflict can be internally based or may be ignited by external factors, which in turn trigger the government to change course. Therefore Levy is correct in stating that ‘the relationship between politics and learning is usually reciprocal rather than unidirectional’538.
The policy change with which this thesis is concerned is planned change, not policy change which occurs accidentally, as such accidental changes are unlikely to have been affected by IL and the origin of such changes will be difficult to trace to the accidental source. Policies, as understood in this thesis, consist of a range of actions (and inactions), taking the form of laws, treaties, policy statements, programmes, statements of principle, and processes. It has been established that the European Union as such represents a unique form of policy-making system, since its problem-solving mechanism is characterised by consensus and the decision-making process takes place in an international political environment. As a result, one can detect a conflict between obtaining the most effective policy and the most consensual policy539. A multi-level organisation like the EU is particularly prone to
“garbage-can” politics which derive from political activities and effects on different levels (national and EU) and in different institutional bodies. The unique structure of the EU therefore invites lobbying by state and non-state interest groups, and the policy window is therefore relatively small. In the EU, ‘the policy agenda is formally the preserve of the supranational, non-party European Commission. This body is granted the sole right of [policy] initiation’540.
Several frameworks were considered as a basis for analysis in this thesis – but none seemed to fit properly. Adapting the Organisational Behaviour Model (Model II) developed by Allison541 provides a detailed description of the problem at hand, while allowing for
535 Dolowitz and Marsh (1996), p. 345
536 Etheredge (1981), p. 74 537 Ibid., p. 135
538 Levy (1994), p. 302
539 A clear indicator for why the RCt was rejected at the beginning of this chapter.
540 Hague and Harrop (2007); p. 379. For the corresponding provisions in the treaty cf. Treaty of Lisbon( TEU, Art 17) European Union (2008)
137 bounded rationality and taking the analysis away from the individual decision maker (state leader), which was also one of the main reasons for ruling out Model I (the Rational Actor Model). Model II supports the notion of IL (although Allison refers to organisations, this distinction may be ignored at this stage). Allison even notes that ‘organisations do change. Learning occurs gradually, over time’542. Another element in favour of this model is that
Allison is not concerned with ‘named individuals or entire governments; rather the subjects in Model II explanations are organisations, and their behaviour is explained in terms of organisational purposes and practices’543. On the other hand, Model II was developed for a
very specific case (the Cuban Missile Crisis) and in a very different setting (Cold War), and is therefore prone to a very “black and white” approach. It furthermore deals with a limited time-span of 13 days, whereas this thesis considers a much longer time-frame (for the Afghanistan case study alone, it covers 9 years). Additionally, as Bendor and Hammond have pointed out in their criticism of the Allison Models, Model II suggests that ‘simple rules, or standard operating procedures, sharply limit and constrain behaviour, that is, that simple rules generate simple, predictable behaviour’544 and it is doubtful whether any scholar of EU
policies would describe them as “simple” or “predictable”.
Using Model II, an adaption of the organisational behaviour paradigm545 would be similar to
the following:
1) Basic Unit of Analysis: Government Action as Organisational Output: This is the European Union’s state-building policy.
2) Organisational Concepts:
a. Organisational actors: the EU as an entity is considered as one singular actor. b. Factored Problems and Fractionated Power: while the EU is considered as
an entity, the thesis is fully aware of the incumbent power struggles within it, in particular struggles between the Commission and the Council. Additionally, the differing national interests and bargaining powers of the Member States are recognised.
c. Organisational Mission: the mission of the EU is clearly stated in its treaties and its official communications; these form the basis for tracking any policy changes.
3) Action as Organisation Output:
a. The EU has a very distinct set of rules of procedure for making decisions or implementing programmes.
542 Ibid., p. 144
543 Ibid., p. 144
544 Bendor and Hammond (1992), p. 309 (emphasis as in original) 545 Cf. Allison and Zelikow (1999), pp. 163-185
138 b. As the EU has evolved as an institution, so has its search to assimilate differing circumstances (e.g. the shift from a Cold War to a Post-Cold War setting); it has had to contend with changes in its self-perception, which has also led to an evolution in its policies.
4) Decision of leaders: the EU in particular is prone to shifts in emphasis occurring due to change in national governments and to different emphases placed by different Commissions – in this context, it is worthwhile recalling the controversy surrounding the Prodi Commission.
5) Existing Organised Capabilities Influence Institution’s Choice:
a. The existence of certain departments and capacities (or lack thereof) to implement certain actions increases the probability that a specific action will be chosen.
b. Differing priorities – as communicated in the reformed treaties –leads to the emphasising of certain goals over others. Sometimes goals may be conflicting, as analysis will show with respect to economic goals over liberal and human rights.
6) Implementation:
a. Activities following previously implemented programmes are not to be considered as far-sighted and flexible adaptation to “the issue at hand”. Details and nuances can easily be determined by organisational routines and not by a specific direction taken.
b. Decision makers (leaders) neglect calculations of administrative feasibility – there is a tendency to fall back on blueprints to deal with situations arising. c. Flexibility to change may be limited by constraints such as budget limitations
or organisational priorities. The implementation of new programmes exerts more pressure on an institution’s operations than the adaptation of an established programme.
While Model II cannot be operationalised for analysis in this thesis, it has nevertheless provided a clear and structured overview of the problem at hand and in particular the constraints of the actor (EU).
i. The Framework
Instead of Allison’s model, a more simplified but more tailored and purpose-driven framework will be used in the analysis of the policies and to guide fieldwork. The Framework used is depicted in Figure 4.4.
139 Figure 4.04 Institutional-learning Framework
The diagram emphasises the fact that changes in (EU state-building) policies can be influenced by two main elements: IL and adaptation. Each category has several sub- categories. IL is divided into three different forms of how learning occurs or from what sources lessons may be drawn: lessons from history, lesson-drawing and learning by doing. Adaptation is divided into two main blocks: EU internal pressures and pressures external to the EU. EU internal pressures account for changes due to Member-State pressure, pressures from institutional bureaucracy (inertia is included to account for any lack of change due to bureaucratic pressure) and budget constraints. The pressures external to t he EU can take the form of a shock event or may be due to obligations to comply with international law or interdependence. In this setting, policy change is the dependent variable, as any policy change depends on IL and Adaptation (or other). Only intentional (planned) changes in policy can be accounted for in this setting. In conjunction with this framework, the main questions governing the analysis are as follows:
Is it institutional-learning?
What kind of institutional-learning is present and from what sources are the lessons drawn?
140
Is the change in policy due to adaptation?
Is the adaptation pressure situated within the EU or outside?
Which adaptive element accounts for the change in policy (with possible reasons)? These questions will guide the analysis in the form of a decision tree. Thus, for each noted event of change, the above questions will be posed succinctly in the analysis section in order to determine which category is present. In other words, the framework will be applied whenever the question of IL/adaptation arises. Converting this framework into a matrix thus provides us with the following Table:
Institutional-learning Adaptation Other Lesson
from history
Lesson-
drawing Learning by doing
EU internal pressure Pressure external to EU EU action/policy
The following section will deal in detail with the IL classifications and adaptation. The IL triggers (a-c) develop endogenously in the institution and in contrast to adaptation (d) result in a change in the belief-system (double-loop learning).
a) Lessons from History
Do lessons from history influence political leaders in their decisions? Levy suggests that some influential American foreign policy decisions were based on lessons from the past, in an attempt to avoid failures from the past; such lessons were “lessons of Munich”, “lessons of Korea” and “lessons of Vietnam”546. Why then did the US still insist on invading
Vietnam (which was later agreed to have been disastrous) and then engage in the invasion and occupation of Iraq since 2003 (which is already being considered a disaster by many academics and policymakers alike)? Fear therefore claims that learning from history either does not take place, or if it does, that the knowledge gained is rarely a guide for state action547. Levy rejects this claim by propounding that ‘policymakers reconsider their goals
and objectives only after repeated strategic failures’548. Speaking in terms of a learning
curve, one would therefore have to assume that this is very shallow in the case of governments. But why should this be so? Or might the general perception of when or whether learning has taken place need adjustment? Are there maybe only certain instances in which lessons are drawn from the past? Possibly the analysis will shed light on these questions.
546 Levy (1994)
547 Cf. Fear (2001) 548 Levy (1994), p. 286
141 Often, “learning from history” is equated with “learning by failure”; cases which involve policy failure are useful to consider ‘since failure serves as a trigger for considering policy redesign and as a potential occasion for policy learning’549. It is the dissatisfaction with the
outcome which serves as a stimulus to explore alternative policies: most importantly it can ‘stimulate a reconsideration of the existing dominant causal reasoning about policy’550.
Learning from success is less frequent, but could supply a basis by allowing the tracing of conditions necessary for policy success. History lessons – particularly the negative ones – are often drawn from previous dealings with unexpected crises, either crises threatening the viability of the organisation itself or external events551. Rose argues that ‘lessons are
different from the analogies that constitute the uses and abuses of history’; he continues by indicating that ‘lessons must identify circumstances that are different as well as those that are the same’552. But the main reason for policymakers to revert to lessons of experience
‘reflects their concern with feasibility’553. Policymakers assume that this will reduce the
possibility of policy failure. By drawing on lessons from the past, policymakers also have the opportunity not only to learn what has worked, but also what has not. Dolowitz and Marsh claim that ‘only when routines stop providing “solutions” is it necessary to search for lessons’554. This thesis claims that “lessons from history” are important and can be useful
tools in ameliorating policies, but the circumstances and conditions of the situation, with respect to the two policies to be compared, must be compatible.
In an ex ante evaluation paper555 the EU identifies “lessons from the past” as a source of
EU learning; the guiding principle they imply in this context is one of ‘don’t repeat old mistakes, make new ones!’ The paper admits that most programme proposals follow on from existing programmes; hence the need not to repeat previous failures, particularly with respect to relevance, effectiveness and any other problems encountered.
“Learning from history” as understood by this thesis is very similar to “learning through failure” as previous policy failures provide the trigger for an evaluation of the causal reasoning behind the policy; by thus questioning these failures, alternative strategies are devised, leading to institutional-learning.
549 May (1992), p. 341
550 Ibid.
551 Cf. Rerup (2009) 552 Rose (1991), p. 9
553 Ibid., p. 5. The concern for feasibility also incorporates the threat of free-riders; meaning that some countries might not want to “waste” resources on a certain problem and profit from another countries’ efforts. (Cf. Ibid, p. 19)
554 Dolowitz and Marsh (1996), p. 346
142 b) Learning by doing
Learning by doing is sometimes also called “experimental learning”; it is often achieved by group or trial-and-error discussions556. Institutional “learning by doing” is – of course -
very akin to learning processes conducted by an individual and constitutes probably the most pragmatic and a posteriori form of learning experienced within systems of education. As such, it is the form of learning generally analysed most by educational psychologists. This type of learning tends to be less intentional and less systematic than the other two types of learning discussed here557. It is greatly enhanced by the availability and analysis of
feedback. Huber rightly states that organisational learning is facilitated by the increase in accuracy of feedback about cause-effect relationships determining institutional actions and outcomes. Learning by doing mainly occurs while the action is still being carried out. There are no benchmark points as such with which to compare it – other institutions’ policies constitute analogies to be drawn on from previous attempts. In the case of learning by doing, the sequence between action and policy feedback tends to be shorter; as a result, the double-loop of the learning process is less established than in e.g. learning from history, where institutionalisation of the feedback from the mismatch of outcomes can take several years to be analysed and implemented. It could even be argued that, due to the shortened timeframe, learning by doing is mainly limited to single-loop learning. In a way, learning by doing is the simplest form in which an institution can learn, and the lessons drawn tend to be more “quick fixes”. The main concern of learning by doing is whether the knowledge acquired through this type of learning depreciates or whether it is fed into the institution’s memory and is thereby institutionalised558.
The EU makes brief reference to this learning category by commenting on the benefit of internal evaluations to promote learning by doing, since managing services themselves are closely involved in questioning their activities on “why” and “how”of559.
Learning by doing occurs while the action is being carried out and is mainly limited to single -loop learning due to its shortened timeframe. It is mainly achieved by feeding new information into the decision -making process without a significant change in the belief system. It is therefore prone to “short-term” fix approaches.
556 Zuber-Skerritt (2002)
557 Huber (1991), p. 91
558 Epple, Argote and Devadas ibid., p. 64
143 c) Lesson Drawing and Policy Transfer
This is the most complex form of IL. Even though every state has its own problems, certain concerns are similar to all citizens in whatever nation they may live: national security, education, economy, unemployment, social security - to mention just a few. Therefore there is a certain vested interest among the policymakers of all nations (all of whom would like to be re-elected) – and particularly among its citizens – to obtain the best system possible for their national environment. As a result, most policymakers “shop around” and try to incorporate systems which seem to work in countries other than their own. With reference to this phenomenon, Rose makes the valid point that ‘it is easier to see similarities between the same policy area in different countries than to find similarities between social security and environmental or defence policies within a country’560. But
Dolowitz and Marsh have examined this proviso and noted that it is not so applicable as it