16Th Nispacee Annual Conference

16th NISPAcee Annual Conference

“Public Policy and Administration: Challenges and Synergies”

May 15 – 17, 2008, Bratislava, SlovakRepublic

Policy failures in Hungary – in search of new failure mechanisms

Draft version!

György Hajnal PhD

Head of Department

Department of Public Administration

ECOSTAT Government Institute for Strategic Research of Economy and Society

e-mail:

1Introduction

This paper summarises the results of a research focusing on problems of public policymaking in Hungary. My main, and emphatically subjective, motif triggering this research is a perseverating astonishment over seemingly trivial yet incomprehensible failures of government, observable in the most diverse fields of governmental activity – practically irrespective of governmental tier (central or local government) or policy sector.

The present research intends to contribute to bridging the gap between availeble scholarly results on the one hand, and the acute yet intriguing field of policy failures, on the other, by (i) identifying, and giving a thick empirical description of, policy failure mechanisms not yet – or not so much – identified and described, and (ii) developing theoretical concepts enabling one to desribe and explain those factors of policy failure. The empirical material of the research comes from Hungary, a post-communist central/eastern European country, presumably sharing many features with other countries of the region while exhibiting a historical development path distinctly different from those of many “older” EU member states and of other developed countries. Therefore, while the generalisability of the findings and conclusions is, clearly, a matter of further research and analysis, they should be viewed in the light of the geographical and historical context, in which the research took place.

The next (2.) section gives an introduction into the conceptual and theoretical underpinnings of the research. The section first outlines the concept of policy failure, building both on insights and results of earlier research, and on a critque and subsequent conceptual refinement of those results. The section then goes on to delineate the concept of originality/novelty as used in the context of this research.

The third section describes the data, which is case studies of policy failures in three policy fields, and the method used – that is, qualitative analysis of (mostly) textual data having emerged from field research. The fourth section presents the findings, and the final, fifth section elaborates some of the conclusions of the research.

2Theory

2.1The concept of policy failure

Spectacular and/or enduring phenomena carrying damaging effects and attributed, at least in part, to insufficient or faulty decisions or operations of government have been a frequent subject of research by students of public policy/administration, organizational science, and related disciplines throughout the past decades. The numerous examples include single tragic events as the explosion of the Challenge spaceship, industrial disasters as the Chernobyl or the Three Mile Island atomic power plants, or the riot in the Heysel stadium; just as the more lasting failures presented long-term policies such as military engagements (Vietnam, Iraq) or various sectoral policies, such as the anomalies around the US Department of Housing and Urban Development or the Chicago area electric supply system in the eighties.

It seems however that the many mosaic pieces of such individual research efforts – providing a huge amount of empirical material – did not result in a proportionately well-developed conceptual framework of policy failure. One of the very few such – indeed ambitious – attempts is the work of Bovens and t’Hart (2004), which will be used as a starting point of the conceptual clarification that follows.

It is instructive that the authors, admittedly having spent a significant amount of – theoretical as well as empirical – research effort on elaborating some “hard” concept of policy failure (pp. i-ii) ended up with the recognition that the notion of policy failure is, inevitably, a social construction impossible to be anchored to “objective” (that is, temporally and geographically transferable) criteria. According to their definition, policy failure is an event or a circumstance (i) perceived by a socially and politically significant group as seriously negative (ii) though avoidable, (iii) the occurrence of which is attributed, at least to a significant extent, to the blameworthy conduct of public policymakers (ibid. p. 15).

Labelling an event as “failure” or “scandal” can be perceived, in the final analysis, as a mechanism sustaining the equilibrium of the social – and within that: the political – system. It is a symbolic act sanctioning the violation of some essential societal norm or interest: “The very label of `fiasco` or `failure` entails a political statement. Events are perceived and labelled as `policy fiascos` when important values and issues are at stake, when societal taboos or sacrosanct values have been violated. […] If this occurs, there tends to be an impetus to maintain and restore social order. […] The identification of failures and the analysis of their causes tend to be part of aprocess of accounting andblaming.” (Bovens–t’Hart 2004 p. 129). In a similar vein – nevertheless avoiding definitional details – Hood (2000 p. 23. ff) stresses the inherently culture-based nature of how policy disasters are constructed.

According to Bovens and t’Hart – and to their definition provided above –, identifying an event or circumstance as “policy failure” involves (at least) three, rather intricate steps.

-Firstly, the event has to be evaluated in terms of its societal impact: is it – as measured by the standards of “a socially and politically significant societal group” – negative, and if yes: sufficiently negative to qualify as a policy failure?

-The second step can be taken if the first question resulted in a positive answer. In this case one has to position the event in the “misfortune vs. mismanagement” continuum; the question is, in other words, whether the harmful consequences at hand were realistically avoidable as opposed to being inevitable.

-Finally, if both of the previous two questions were answered with a “yes”, the third step consists of confirming whether the blame of the ensuing societal damage can be put, at least to a significant extent, on some public policymaker(s), as opposed to other agents outside the realm of government.

In the subsequent paragraphs (i) I will – although admitting that the social constructivist approach offered above is the only workable way – argue that some other, key definitional elements offered above suffer from conceptual inconsistency, and (ii) on the basis of this critique I will present an alternative, modified concept of policy failure.Specifically, I will argue that

(a)the only viable concept of “consequence” (of a course of action by a public policymaker) inevitably implies that, provided that the relevant policymakers possessed/exercised a sufficient level of policy management skills, the given (negative) consequence was indeed avoidable; therefore

(b)the definitional elements of “avoidability” and “blameful misconduct of policymakers” are tautological and thus needless.

Before proceeding to presenting the arguments it is noted that the above propositions are not mere “l’art pour l’art” exercises. Although – if well-founded – they have a theoretical significance in and by themselves, they also form a necessary foundation for the design of the present research.

At the heart of the concept of policy failure one finds the notion of consequence.Surely, a necessary feature of any event or circumstance aspiring to be qualify as “policy failure” is that it must bring about specific (that is, socially damaging) consequences. But what does the statement “Y is the consequence of Policy X” exactly mean? This question is asked neither by Bovens and t’Hart, nor by a number of other scholars researching the topic. Rather, it is treated in a – if not epistemologically but at least ontologically/conceptually –straightforward and unambiguous manner. Looking for the causes of given, specific consequences is assumed to take the form of a criminal investigation; lack of information or factual knowledge might, in specific cases, hinder one in singling out of a relationship between some specific consequence and some agents (and their courses of action), but all such ambiguity can be eliminated, provided that sufficient level of detail is available on who did what, when and why.

Contrary to this view, my proposition is that the “criminal investigation” metaphor is not an adequate one in the world of policy failures. The reason for this is a key difference between investigating criminal acts vs. policy failures. Namely, in the former case a relatively clear-cut catalogue of non-permissible – and thus potentially blameful – acts is defined by the Penal Code. The investigator’s task in this case is “only” establishing whether any one element of this list was carried out by any actor. Here, the very notion of crime already involves the notion of consequence; there is no homicide without a casualty, no robbery without goods taken away, and so on.The situation, however, is different with investigating the causes of policy failures (or, vice versa, investigating the consequences of specific policies). Here, the requirements of (i) social damage, (ii) governmental misconduct and (iii) a causal relationship between these two exist independently from one another[1];and if any one of these preconditions is not met one cannot talk about a policy failure.

Traditionally, the problem of identifying the consequences of X is tackled in the social science by the notion of the counterfactual. According to Marshall (1998, pp. 123-124) it is “a proposition which states what would have followed had the actual sequence of events or circumstances been different. Thus, to claim that the Battle of Alamein altered the outcome of the Second World War is to imply the counterfactual that, had victory in the desert actually gone to the defeated German forces, thn the Allies would have lost the war”.

My next proposition is that it is exactly this notion of the counterfactual scenario that is being employed, albeit often implicitly and/or unconsciously, by many analyses of policy failures. For better illustrating this let’s take the example of the following statement: “The high unemployment in Hungary is consequence of policy mismanagement”. This statement actually involves an implicit reference a counterfactual policy scenario, in which the unemployment rate is less than the actual one. If the current unemployment is 12% and the counterfactual unemployment rate is 0%, then one can say that the consequence of this policy failure is 12 percentage points unemployment; if one accepts that some level – say, 6% – of unemployment would have been unavoidable anyway, then the actual damage caused by this policy failure is 6 percentage points unemployment. However if one holds that achieving a rate of unemployment less than 12% would have been impossible anyway than (s)he cannot attribute the attached societal damage to government policy.

It is important to notice here the judgmental – indeed, socially constructed – nature of the “counterfactual policy”. Identifying some hypothetical policy as a realistic and feasible alternative of the actual one involves a number of assumptions, culturally determined predispositions/expectations, and value judgments varying strongly in time, space, and even within individual political systems. For example, natural disasters such as typhoons or earthquakes have been for ages – and in many countries still are – accepted as unavoidable events of misfortune (“vis maior”). However nowadays governments, especially in developed countries, are expected to effectively counteract these events (via early warning systems, construction regulations, catastrophe relief measures and so on). Likewise, the extent, to which governments are expected (or assumed to be able, or to be permitted) to intervene into societal problems like workplace stress, aging, migration or global warming – to name but a few areas – is constantly changing.

To sum up, talking about a policy failure inevitable involves the presumption of an alternative, counterfactual policy – and a corresponding state of the world – that (i) could have been, according to the speaker, realistically achieved and (ii) would have brought about better societal outcomes (either higher benefits or lower costs) than what was actually achieved. Now let’s suppose that, in a given, specific policy case this was the case indeed; that is, a feasible policy alternative was available for the policymakers that would have resulted in societal outcomes better than those having occurred. In this case we can say that the unwanted consequences were avoidable, and that it was the fault of the policymaker that the actual policy was implemented instead of the counterfactual one (otherwise we couldn’t say that the counterfactual policy option was a realistic one). My initial statement – i.e. that the definitional elements of avoidability and blameful misconduct are already included in the very concept of policy failure and thus are tautological – is therefore demonstrated.

What remains then as the core concept of policy failure? For the purposes of this study, policy failure is present if the societal outcomes of a given policy could have been substantially better – as measured by the value standards of a socially and politically significant societal group –, provided that an alternative, feasible course of action had been taken; “better societal outcomes” can refer either to higher benefits or to less damage to society.

2.2Causes of failure

The central aim of this paper is not identify policy failures per se; rather, it is to identify certain types of causes leading to policy failure. Specifically: I am looking for factors that in some significant respect stand out from the “usual” variety of similar factors. In order to further specify this task it is necessary to introduce some additional conceptual elements and distinctions. It is however beyond the ambition and the scope of this paper to give an ontologically and epistemologically exhaustive account of the ways, in which such terms as explanation, cause or causation are understood and used. Therefore – as a kind of golden mean – only certain conceptual cornerstones will be laid down as follows.

“Explanation” should, at least ideally, refer to the set of preconditions– relevant from the investigator’s perspective – under which the given explanandum is likely to occur, but which is unlikely to occur if any one of those preconditions is missing. It is however clear that this empiricist view of explanation and causation is not a viable one when it comes to explaining policy failures. Policy failures, like a number of social phenomena, are often historically unique; moreover, they are characterised by the presence of a vast number of human, social and technical variables intertwined in a complex structure. Therefore a more interpretivist stance has to be taken.Namely, instead of looking for “objectively verifiable” causal relationships the focus is on key stakeholders’ causal attributions and, not the least, on the process of sense-making by the researcher in the process of qualitative analysis.

“Cause” or “causal motif” is understood as one element of the above mentioned set of preconditions. Causal motifsare a central concept of this study as the analytical focus is, throughout, on identifying individual causal motifs rather than elaborating complex explanations involving the various possible networks or constellations of preconditions/causes.

There is a vast amount of scholarship on the different mechanisms – hereinafter referred to as government dysfunction –, by which government operations might deviate from the “correct”, expected, or prescribed track. It is emphasised that government dysfunction – be it of a political or a more administrative nature – does not necessarily lead to policy failure (the latter necessarily involving damaging societal consequences). Rather, it may or may not do so; as a matter of fact it may in some cases even promote societal conditions and/or prevent policy failure from taking place.

One large class of these factors are related to the realm of governmental apparatuses, such as various forms of bureaucratic dysfunction (March–Simon 1993) originating from e.g. dysfunctional learning of organisational members or adverse effects of increasing demand for control at the top of the organisation, or the structural framework and processes within/by which implementation takes place (Pressman–Wildavsky 1984). Lacking financial and human resources, expertise and/or institutional infrastructure has been another major source of government dysfunction both in the onetime Third Worldas well as in the context of transition countries. Regulatory capture (Stigler 1971) and administrative corruption, or various rational choice concepts, such as the “budget-maximizing bureaucrat” (Niskanen 1971) or, more recently, the bureau-shaping model (Dunleavy 1991/2001) are other, frequently appearing elements of this cluster.

Another major class of government dysfunction originate from the upper echelons of the policy arena. These manifest themselves mostly in the policy decision, which are (or are not) made by the policymakers. Besides models relying on the concept of the utility-maximizing politician (e.g. Buchanan-Tullock 1962), other forms of legislative failure (Weimer-Vining 1999), the overwhelming effect of vested (economic) interests, or group dynamics leading to wrong decisions in high-level decision making bodies (cf. “groupthink”; Janis 1982) can be mentioned here as examples.

More empirically oriented research (i) focussing specifically on the dysfunctions/limitations of governments’ public policymaking capacity and (ii) geographically covering much or all of the post-communist region is – contrary to efforts aiming at the “higher”, political system level (Offe 1997, Greskovits 1998, Henderson 1998) – relatively scarce.

On the one hand, there are important and interesting, predominantly empirical (case) studies – mosaic-like pieces mostly focussing on one or a few countries and pursuing a limited ambition (Barnes 2003, Cummings-Norgaard 2004, Mungiu-Pippidi 2003, Vandeveer-Dableko 2001). On the other one finds a number of more conceptually oriented pieces (Fritz 2003, Sotiropoulos 2002), which are however less rooted in empirical evidence. Most of this type of research deal with the appearance and effects of already well-known problems of public policymaking, such as political culture or corruption). Finally, attempts to identify/describe „genuine” types of policy failures is much more infrequent (Hajnal 2005, Ionita 2003 and 2005, Kovryga-Nickel 2004). However, conceptual-theoretical and/or empirical limitations – the latter being a particularly severe constraint in the overwhelmingly secretive and thus data-poor worlds of post-communist policymaking – this direction of research still has a long way to go.

Of course it is not possible to give even a rough classification of the extensive terrain of government dysfunctions as models, concepts and schools largely overlap with one another; nor is it my ambition to make an attempt at drawing up an exhaustivecatalogue of relevant factors, concepts and models explaining all possible type of government failure in the Hungarian – or, possibly, post-communist – context. Rather, the role of this very brief overview is to put the ambition of this research into context: it is this vast universe of (possible) causal motifs and explanations of policy failures, which this research intends to supplement with some additional elements. I see no sound method (i.e., research design) ex ante guaranteeing that once the proposed path is followed carefully it will lead to such a result; nor is it possible to devise a method capable of substantiating a claim ex post that such a relatively new element has been found indeed. Rather, the outputs of this research are to be treated as propositions, which are to be tested, and – if necessary – modified or rejected in response to future arguments.