Sept. 10, 2001, morning. Introduction into problems of forest technology / risk management in mountaineous terrain

Administrating Risk – A Social Science Perspective on Natural Hazard Prevention Based on an Austrian Case Study

Gerhard Weiss

Gerhard Weiss: Institute of Forest Sector Policy and Economics, Agricultural University Vienna. - This article is based on a 4-year research study on "Evaluation of policy instruments for natural hazards prevention in Austria" (Weiss, 1998a; 1999). This national project was financed by the Ministry of Agriculture and Forestry and was part of the European COST Action E3 "Forestry in the context of rural development" (Koch/Rasmussen, 1998). The author participates in the EU concerted action MUFOMA "Multifunctional forest management in mountain regions - Evaluation of silviculture and policy means" (http://www.forst.uni-muenchen.de/LST/WALDBAU/mufoma.html)

ABSTRACT

This article discusses how the bodies of public administration deal with natural hazards in theory, and presents the results of an empirical study on policy instruments for natural hazards prevention in Austria. It is argued that different groups have different concepts and conceptions of risk. In Austria, two institutions are active in maintaining forests which protect against natural hazards. They apply different routines: the forest authority follows a vague concept of multifunctional forest management. The torrent and avalanche control service, in contrast, focuses more precisely on the reduction of certain hazards. While the first is primarily oriented at the forest owners' interests, the latter concentrates on the interests of the beneficiaries of protective forests. The applied risk policies are strongly determined by the different stakeholders' interests, who participate in the corresponding policy-networks. Within different institutional settings, official experts develop different risk concepts and consequently different solutions.

Keywords: protective forests, risk policies, public administration, political analysis.

Published in: Gadow, K.v., (ed., 2001): Risk Analysis in Forest Management, this being part of a Kluwer book series on Managing Forest Ecosystems.

Introduction: The role of forests in natural hazards protection

The relation of forests and natural hazards is discussed in the literature concerning two different aspects: The damage that natural hazards may cause in forests; and the contribution of forests in natural hazards protection for settlements, infrastructures, cultivated land, etc. The first is a management task concerning primarily the forest owner. The latter refers to a relation of the forest (owners) to other parties. These two tasks are basically different problems but still they are interrelated: Natural disturbances in forests may provoke dangerous processes starting from the forest. If a forest is struck by wind-throw and/or a bark beetle attack both timber production and the protection of neighbouring properties are possibly affected. Foresters around the world have to deal with the protective effects of forests for neighbouring properties or for society on the whole; they call them "protective functions". In some countries the administration of the policy field "natural hazards protection" is assigned to the forestry administration.

In my article I am dealing with the role forests play in protecting mountain areas against natural hazards. As the protective effects of forests concern different parties – the forest owners and the beneficiaries – this is a task of public policy (1). For this reason, public administration plays an important role in this field, and I will especially focus on this. The key question of my article is: How do public agencies deal with natural hazards? For finding answers I will use theories from interdisciplinary and sociological risk research as well as from policy and bureaucracy studies, and I will use findings of detailed case studies from Austrian policies concerning mountain forest management and natural hazards protection.

About nature: Are natural hazards different from technical hazards?

When the mass media report about natural hazards they ask "who is responsible?" (2)Usually no case of corruption can be found and nobody has acted illegally. In spite of this, people are injured or even killed and houses and infrastructure are regularly damaged by elementary forces like avalanches, land-slides, torrents or debris flows. Often it is simply concluded that life in the mountains is dangerous.(3) Even the most sophisticated technology is limited, they say, when facing nature and its unpredictability. Nature can never be controlled 100% and therefore catastrophes by elemental forces will always occur.

Is this true? Are natural hazards different from technological hazards because of the unpredictability of nature? No, there is basically no difference between technological risks and risks caused by elementary forces. Risks always comprise of natural processes in combination with the utilisation of nature. In the case of atomic engineering, nuclear physics is utilised and it is not possible to control these forces 100%. Accidents occur because the behaviour of water, concrete, heat and radioactivity cannot be predicted precisely. Technological failures always involve human and physical aspects. This is the same with natural hazards: People develop mountain areas, but the natural processes of the landscape cannot be 100% controlled. The specific moment when an avalanche starts and the detailed area it will cover cannot be predicted for each event – but the hazard is usually known. The causes for damage are in both cases neither only human intervention nor only natural processes. (4)

Mountain areas are developed with the awareness of natural hazards i.e. with a perception of risk. The risk is dealt with by regulations and public agencies. Instruments exist to assess risks, and procedures exist to control development in hazard-prone areas. It is not nature threatening the powerless and unaware population but people utilising nature and endangering themselves, or – to put it differently – they take the risk. It is a lame excuse to state that damage may always occur. It is true that nature cannot be controlled 100%, but it is also true that in principle the risk is – more or less – known and regulated. This leads to the following questions: "Who knows about the risks and who takes decisions? How is risk knowledge managed and how is it considered in decision-making?"

(1) The protective effects of forests are in many cases characterised as external effects and public goods which makes private problem-solving difficult. The basic problem structure is described in Glück/Weiss (1997) and Ottitsch/Weiss (1998), more detailed in Schmiderer/Weiss (1999).
(2) E.g. the weekly magazine Der Spiegel, Hamburg, issue 31/97, asks on the cover: "The big flood - natural disaster or work of man?" The cover of the weekly News magazine, Vienna, issue 8/99, is titled: "The avalanche. The disaster of Galtür: Who is responsible?"
(3) E.g. the comment on "Avalanches" in the Frankfurter Allgemeine Zeitung, Frankfurt, issue 25. 2. 1999, p. 1.
(4) This lesson is reflected by Jörg Blech in Die Zeit, Hamburg, issue 32/97, p. 2, who concludes: "The real reasons for the flood disaster sound banal: there was heavy rain, and people settle too close to the river."

What is risk? Defining and evaluating risk

Decisions, actors and values: the subjective qualities of risk

Risk – this simple word causes ardent public debates and has initiated extensive interdisciplinary studies over the last decades. What is risk? The term itself does not cause much controversy. It is a technical expression like e.g. "duration" or "costs". Describing a measure as risky is no more exciting or threatening than the attributions "long" or "expensive". Risk is normal: each and every decision contains a certain risk because every decision concerns the uncertain future. The term risk does not evoke anything as long as it is not connected with a matter. Mountain climbers or stock-brokers may take high or low risks. Their decisions are only understandable when considering the (subjective) benefits they expect. Discussions about high or low risks are only relevant if it is known what risk and what benefits are concerned. Risk management is usually taken as aiming at the reduction of risk. However, the decision of taking a measure or not is always risky: if the option is taken there is the risk of not reaching the goal; if the option is neglected there is the risk of having failed a chance (Evers/Novotny, 1987: 32ff, 59ff; Wildavsky, 1989). It is not possible to escape this dilemma that is inherent in every decision. So what is the risk debate all about? In order to answer this question, I want to explain the following basic features which are essential to understand "risk":

• Risk is connected with decisions: Risk has to be distinguished from danger (Evers/Novotny, 1987: 32-46; Luhmann, 1993). The concept of risk goes back to medieval trade and is formalised in decision theory. Only the possibility to choose between options sets up a risk situation; if I accept a risk I also take a chance. The choice constitutes the probability of profit or loss (Müller, 1993: 3813f). If I cannot choose I am endangered.
• Risk regards actors: The fact that risk is connected with a decision points to the fact that there is an actor who decides. In formal decision theory this actor is abstracted; it presupposes a "rational decision-maker". In real life decisions, however, the "subjectivity" of all actors is crucial (Evers/Novotny, 1987: 42). There is more than one rationality – it is crucial who decides.
• Risk includes values: According to decision theory, decisions on risks weigh present and potential future benefits and costs and their probability. In real life decisions, benefits and costs are weighed by actors with interests and values. Decision models are constructions that can either meet the real situation of the deciding person or not. Risk concepts are not neutral but include normative assumptions about the decision situation. Economists state that it is impossible to define an "objective" rationality that can generally be applied (Fritsch/Wein/Ewers, 1993: 148).

Decisions about the safety of technical projects are usually based on risk analyses or risk assessments. Risk management tries to find a rational approach to safety decisions by structuring the problem and making the decision transparent.

These rationalist risk models, which are developed by technicians, are often not accepted by the public. Prominent examples for conflicts between experts and laypersons are the discussions about nuclear reactors or genetic engineering. The concepts of risk analysis and risk management were applied in the field of natural hazards as well (Kienholz, 1993; Hollenstein, 1995). These authors recognise the problem of deciding "What is acceptable risk?" They propose to include risk preferences in the assessment process (compare Starr, 1969), but these attempts under the psychometric paradigm were not successful either. In varying situations people accept differing risk levels. The risk preference of one person in a certain situation cannot be transferred to another person in the same situation, nor can conclusions be drawn from the risk preference of one person in a certain situation to other situations. The same is true for different social or cultural groups (Johnson/Covello, 1987; Jungerman/Slovic, 1993).

Technicians may recognise that the question "How safe is safe enough?" cannot be answered by expertise alone, but has to be answered by politics or society. However, it is common that politicians fail to give clear directions about an acceptable level of risk. Kienholz and Hollenstein mention the problem of "political" or "subjective" factors in risk evaluation, but up to now conceptual approaches to cope with these problems have been lacking in natural hazards regulation. Risk managers try to deal with these questions within their technical concepts, e.g. by including risk preferences into the decision model. The questions concerned, however, go beyond the technical concepts. They are questions of political procedures. The bureaucratic decision and the use of risk models is only one possible procedure. These existing procedures put value decisions into the hand of technicians who normally are not aware of the political implications. In many risk debates this technocratic procedural approach is questioned (Nelkin/Pollack, 1980: 243-246).

The traditional bureaucratic approach applied by public agencies has deficiencies in terms of democratic principles because the interests of different groups are not considered evenly in the implementation process. In spite of their formal rationality the concepts applied are often biased towards powerful interests (Mazur, 1985). In a study of the practical application of risk assessment the questions of power distribution and the role of political actors have to be included (Compare, 1995). In practice, risk models may not be adequate in certain situations at all or they may be manipulated or instrumentalised by certain groups. These problems are also found in natural hazards protection (Weiss, 1999). Alternative procedures – e.g. public participation – have not yet been studied thoroughly. (5)

Actors, values and interests: conceptions and politics of risk

Risk is technically defined as "the probability and magnitude of an adverse effect" (Lowrance, 1980: 6) or similar. The objectivist paradigm in risk research believes that risks may be assessed objectively. Technicians work with instruments like risk analysis and they usually assume that "rational actors" would follow their concept. In contrast to this, risk researchers who approach the problem with a constructivist view state that risk is constructed in the mind of people – and is thus constructed differently by different people. From their point of view, an objective definition and assessment of risks is not possible (Douglas/Wildavsky, 1982).

(5)Recently a study about the options of public participation in torrents and avalanche control projects in Austria has been carried out (Ottitsch/Pregernig/Rappold, 1998). This topic is considered in ongoing studies in Europe ("Public participation in mountainous ecosystem management", EFI; EU concerted action "Multifunctional forest management in mountain regions - Evaluation of silviculture and policy means", MUFOMA).

(6) In the literature different risk conceptions as well as different concepts of risk are debated; Shrader-Frechette (1995, 1997) distinguishes positivist and relativist conceptions of risk, whereas the framework developed by Thompson and Dean (1996) proposes a dichotomy of probabilist and contextualist risk concepts. For the purpose of this article it is sufficient to recognise that "risk" does not mean the same for different people.

A possible answer to the conflicting objectivist and constructivist theories is that both views are complementary (Rosa, 1997). It can be assumed that real hazards exist, in the perception of people, however, they are always constructed (Johnson/Covello, 1987). In such a way also the technical, objectivist concept of risk is one cultural form of risk perception (Krohn/Krücken, 1993: 13). Not only the evaluation of risks is based on "cultural specifics" (Kienholz, 1993: 18) but also its conception. (6) The attempt to understand risk debates leads to the necessity of a broader view on the subject. To understand the risk debates the subject has to be framed in a social context. It is useful to take a look at risk concepts from a meta-analytical level: to look how risk concepts are formed. In such a view it can be asked: "Who defines risks in what way?" The probabilistic risk model is just one possible conception of risk among others. It is connected with our modern society, originally being generated in course of the individualisation process that started at the beginning of modern times (Elias, 1976).

What makes risk politically relevant is that different people may be affected by risk decisions. It is the actor assessing the risks who is crucial! The subjective character of risk assessment becomes clear when considering that risk decisions do not only include the "risk factor" but also potential costs and benefits. It is plausible that different actors may value things differently because of personal preferences. It is even more obvious that people judge risks differently depending on how they are affected by the decision: Do they benefit from the measures? Do they have to bear the costs? In the cases of nuclear power plants and genetic engineering in the first place the companies profit but the public bears the risk or the potential costs. A similar situation is found in the case of developing hazard-prone areas in mountainous areas: The property owner makes profit from hotels or shops but if torrents or avalanches destroy the building other people are injured or killed and parts of the costs are covered by the public. The dissent between laypersons and experts is different in the case of the power plant and the case of natural hazards: In the first situation, the official experts typically share the view of the power agency and laypersons oppose. In the latter the experts represent public interest in restricting the development of hazard-prone areas but laypersons are interested in development as property owners. The main reason is the relative power of the interest groups.

The dissent of laypersons and experts is frequently explained by limited information and insight of the laypersons. Experts tend to believe that their view on risks is objective and the public perceives risks differently – i.e. in a distorted way (Starr, 1980: 4). Slovic/Fischhoff/Lichtenstein (1980: 212), however, show that experts are not immune to biases in risk perception. Laypersons’ perceptions are often considered "irrational" but their behaviour can also be called "rational in a different sense". On the contrary, one could even say that laypersons behave more rational than experts in certain cases. In the example of nuclear power plants Perrow (1984) shows that experts failed to recognise the specific susceptibility of the technology to accidents but laypersons did. In this case, relying on limited calculations and working with probabilities by experts can be considered irrational; the sceptic attitudes of laypersons with regard to unpredictable consequences of accidents seems rational. The dissent between laypersons and experts can partly be explained by clarifying what is understood by risk. The risk model of experts is limited to probabilities of a certain damage. Studies show that laypersons often assess probabilities of different hazards quite accurately (Lichtenstein et al., 1978; Renn, 1984). Different results are gained in surveys if not death rates are asked but "how risky" a certain activity is. People call sources of risk more or less risky quite independent from their own assessment of death rates or damage (Renn, 1984: 152). That means that in their risk assessment qualitative aspects are included additionally to quantitative information. It is important to recognise that risks are not only evaluated differently but also constructed differently! There are not just different preferences but different models behind the decision. In the case of natural hazards it will make a difference if the property owners who are interested in development are asked about their opinion or inhabitants who strive for maximum safety. The land owners would criticise other aspects of official risk assessments than other persons.

Finally it has to be stated that the public risk debates were about more than just risk: they were about technology, about democracy, about equity, about the environment (e.g. Nelkin/Pollack, 1980: 243; Beck, 1988: 20). In this respect the risk debate can also be interpreted as a ritualised confrontation between social or political groups (Conrad, 1989: 183-190). The attractive topic of risk may hide other topics like the criticism of technologies, but it also hides other latent conflicts concerning the application of these technologies (Conrad, 1989: 190-194). If problems are concerned where the concepts of risk and insurance are no longer adequate the symbolic risk debate distracts from the real problem. In natural hazards prevention such cases may be found concerning "close-to-nature" prevention concepts. Conservationists’ criticism of e.g. a check-dam may go beyond the impact of this single construction but may question the whole prevention concept.

At the beginning of this chapter I stated that it is the context which makes discussing risks interesting. It makes a difference what kind of risk, what kind of activity is concerned, and it makes a difference who assesses the risk. Three important conclusions may be drawn from this chapter for analysing risk:

• Risk assessments always involve values and interests of groups. Therefore, the question of power may not be excluded when studying risk debates. Risk decisions include (potential) costs and benefits and therefore it is crucial which groups will have to bear the costs and who profits.
• Different social or cultural groups not only evaluate risks differently but also have different perceptions and conceptions of risks. For instance, dissents between laypersons and experts may not only occur because of varying assessments of the probability or a varying evaluation of damage but also because of varying basic assumptions about and different aspects considered in the assessment.
• Public debates about risk may explicitly or implicitly concern additionally other topics connected with the activity in question, e.g. the use of technologies, impacts on the ecology, democratic processes, etc.

Forests, natural hazards and risk assessment

Who assesses the risk in the field of forests and natural hazards? Important groups of actors are: forest owners, direct beneficiaries, the general public, and public administration. The owners of properties from which dangerous processes may start (e.g. forest owners) have direct influence on the risk situation by their regular management decisions. The question here is: "In which way, if ever, do they consider the risk situation in their decisions and how can this consideration be retained?" Answers can be found in their value system and their interests. Appealing to their solidarity may make them consider the safety of people living below their forest. Forest owners may feel responsible for possible damage; however, they try to reduce costs for protective measures because they do not benefit from the safety (unless they own the properties below themselves). Their liability in case of a damage possibly provoked by their management measures is an interest-related factor determining management decisions of forest owners. The owners of properties potentially endangered by natural hazards, or benefiting from protective effects of forests (direct beneficiaries), are directly interested in reducing the hazard. They are interested in the highest possible safety for the lowest possible costs. The general public is interested in many positive effects of the forest but latent interests of big groups are difficult to organise (Olson, 1965). The public administration tries to ensure the public interest, which is sustainable forest management including the maintenance of the protective effects of the forest. Public bodies formulate and implement regulations for forest management and natural hazards protection. One important part of this regulation is the definition of property rights in protective forests: "Is the forest owner obliged to maintain the protective effects of the forest or not?" And another question is: "How is the regulation implemented?" I want to discuss this question in the following section.(7)

Property rights and the public administration form the institutional background for the policy of natural hazards protection. The basic determinants for the behaviour of the forest owners was shortly explained above. The following chapter especially focuses on the behaviour of the bureaucracy.

(7) The question of property rights is not elaborated further on in this article. Detailed results on the behaviour of owners of protective forests, beneficiaries and public agencies can be found in Schmiderer/Weiss (1999).

To administer risk: The behaviour of bureaucracy in risk regulation

Public agencies as political actors

Public agencies play a major role in risk policy because they are main actors in both the formulation and the implementation of norms, concepts and procedures for risk regulation. Policy formulation is usually not done by parliament independently from executive bodies. It is common that the administration elaborates proposals for sector laws. By doing so, not only their knowledge, but also their value system and interests are incorporated in the regulation (Peters, 1984). Furthermore, they are the main implementing actors. Laws are formulated on a general level. Very often policy goals are formulated in the law only vaguely or even contradicting. Conflicts between interest groups that have not been solved in the formulation process lead to vague formulations, which hide underlying conflicts. They then have to be solved in the implementation process. If strong interest groups still oppose regulation at this stage, laws cannot be implemented. Implementation deficits are the result (non-decision problems, Wolf, 1996: 145, 151).

In the implementation process, general norms are applied by the administration which thus strongly influences the interpretation of the laws. However, they are not independent in this process: actually, there is an implementation network in which other actors are involved as well. Through this network, values and interests of parties also affected by the regulations become effective in implementation. Each group pursues their interests, and the result of the implementation process depends on the power of the actors involved in the network. As a result, risk concepts and risk evaluation of experts and official agencies are biased. Risks are conceptualised within the value system of the profession or the policy community, respectively (Dietz/Rycroft, 1987: 77). Technicians, for instance, tend to describe risks mathematically. Depending on influential interest groups different criteria may be included in the evaluation: economic value of damage, magnitude of possible catastrophes, kind of damage expected, number of deaths expected, impairments for quality of life, short-term and/or long-term effects, ecological damage, etc.

Informal interests, professional values, and safe decisions

To understand public bureaucracies one has to be aware that not only the formal goals (contents of laws, general principles of bureaucracy like legality, economy, expediency) but also the informal goals determine their behaviour (e.g. Downs, 1967; Mayntz, 1985; Scharpf, 1997). Such informal goals which are significant for every organisation, are for instance: maximising their budget or safeguard their autonomy, competence field and legitimacy. Public agencies are often characterised by a clientelistic behaviour: In order to minimise conflicts they try to regulate a sector rather in correspondence with than against the interests and values of their main clientele (selective clientelism). Measures are negotiated between the administration and the addressee (bargaining process). The result of this negotiation process depends on the power of the groups concerned: If the private actors concerned by a regulation are powerful (e.g. if they have a significant economic status, a high number of employees, etc.) the result may be that the public agency serves the interests of the private group rather than controlling them (capturing, Wolf 1986; "state failure", Jänicke, 1986). Clientelistic behaviour like in industry is also found in the forestry sector (Glück, 1976; Krott, 1990; Pregernig/Weiss, 1998). In environmental policy within the industrial sector clientelism leads to "end-of-pipe" solutions (Wolf 1986: 158). A change of structures and processes is not achieved because of the resistance of established actors which benefit from the existing situation. Typical environmental problems are externalisations of costs; an internalisation is defended by the system (Jänicke, 1986). In natural hazards policy this means that, for instance, constructions to defend damage by torrents and avalanches are preferred to afforestations (Weiss, 1999).

The realisation of measures depends not only on the "objective" problem pressure but also on the problem solving capacity of the administration. As the 'catastrophe paradox' Prittwitz (1990) describes a phenomenon that if the problem solving capacity is restricted, small problems are solved whereas bigger problems are not. In the work of a bureaucracy even a reversal of the rational principle of choosing means according to certain goals might occur: The administration may have certain instruments at hand and, with the informal goal of keeping its territory and legitimacy, they seek goals to legitimise their work (Luhmann, 1966; Downs, 1967; Edelmann, 1976). By applying their routines technicians would look for technical solutions, foresters contribute by silvicultural measures. Eventually measures are taken in a way and at locations that can be realised without great troubles (aversion of conflicts). In their problem solving efforts they apply their professional values and routines and change policies only under high pressure (Sabatier, 1986, 1993).

For public agencies the problem of risky decisions is not a new situation: all decisions include risk because decisions are always taken under uncertainty. For the administration, however, decisions have to be certain. Procedures serve to transform the uncertainty that lies in the problem into certainty of the decision (Luhmann, 1993: 144f). In the case of environmental risk, setting standards of maximum emissions, measuring emissions, and mathematical procedures rationalise the highly uncertain and value-laden decision on "what is a health hazard" or "what is acceptable risk" (Wolf, 1986). Formal procedures make the decision apparently value-free. Also in the assessment of natural hazards procedures exist that are ostensibly "objective". Yet, the underlying normative assumptions of what risk is referred to as acceptable is hidden behind expertise.

Two agencies – two concepts: Case studies on risk regulation in Austria

The forestry administration in Austria deals with classical forestry tasks as well as the protection against torrents and avalanches. Besides the forest authority an agency was established under the ministry of agriculture, forestry department, especially for the task of natural hazards protection (torrent and avalanche control service, TACS). (8)Partly, the same policy means are applied by both institutions. This allows a comparative analysis of two institutions active in the same field. By this comparison the importance of the institutional design in regulating a policy field can be illustrated. Following I present the application of exemplary policy instruments by these two institutions and draw conclusions with regard to the determinants of their behaviour in risk regulation. All findings are the result of a research study which I have conducted in the years 1994-1998. (9)

(8) Although there is a federal law and not provincial laws on forestry, forest authorities exist for each province. To simplify matters, I will speak of the forest authority in singular. By this I mean the provincial authorities and the ministry of agriculture, forestry department. The TACS on the contrary, is directly subordinated to the ministry.
(9) The examples used originate from the research project "Evaluation of policy instruments for natural hazards prevention in Austria" (Weiss, 1998a). Aspects on the administration of risk are partly included in the dissertation which is based on this material (Weiss, 1999). In this project, data was gathered at federal and provincial governments (interviews, statistics, documents) as well as at the local level. In 20 districts the local officers of the forest authority and the TACS were interviewed (representing about ¼ to ¾ of the districts in the mountainous part of Austria) and 50 projects were chosen for detailed analysis (interviews, document analysis and visits at site). The results can be regarded representative for the administration of torrents and avalanches in Austria.

The forest authority

Protected vs. protective forests: the politics of definitions

The formal institutional goals of the forest authority are the public interests in the forest, which are in Austria legally defined as the effects of the forests on wood production, natural hazards protection, protection of environmental goods, and recreation. The fact that 80% of all forests in Austria are privately owned has a strong effect on forest policy, because of strong forest owners’ interest groups. For instance, most subsidies are granted for a better utilisation of the forest by the owner (so-called "production function"). Over the last three decades, however, activities of the forest authority in "protection forests" has increased gradually (so-called "protection function").

Traditionally, Austria's mountain forests are regarded as "protection forests" (Schutzwald) implying a special protected status for these areas. The term "protection forest" comprises two aspects: (a) the protection of the forest against over-use or improper management, and (b) the importance of the forest in protecting the soil against erosion. Going back to the Austrian Empire’s Forest Act of 1852, also the Austrian Forest Act of 1975 defines "protection forests" as site-protective forests. They do not necessarily have direct protective effects for settlements or infrastructures against natural hazards. Protection forest/Schutzwald is an ambiguous term: While the official definition refers to site protection, in colloquial use (also by foresters) it is rather connected with the protection against natural hazards. This ambiguity is used politically by the forest administration: Statistics are based on the legal definition which results in large areas of protection forests (most parts of mountain forests); in public relations, however, it is argued that these forests are necessary for protecting against natural hazards. In fact, only few forests directly protect against rock-fall or avalanches. In their mind Austrian foresters typically do not distinct between forests at risk and forests as part of a risk reduction scheme. This undifferentiated connection of protection forests and protective effects of the forest in their colloquial use influences the practical work of the forest agencies: They are hardly aware of different protective effects of the forest and have no corresponding management concepts at their disposal.

"Multifunctional" protection forest restoration

The traditional protection forest policy was a "do-not-touch" policy, which is understandable when considering the "classical" foresters’ fears of over-exploitation. In the last decades a new problem definition and a new paradigm in protection forest management evolved: As timber prices fall, revenues from wood production decrease. At many sites especially in mountain areas wood production is no longer profitable. Instead of over-use, "under-use" or "lacking management" of mountain forests is considered a problem today. This view serves the interests of the authority which – facing the decreasing economic significance of the forestry sector – searches for new tasks to maintain its own legitimacy. The forest authority initiated new subsidy programmes for "protection forest restoration". Although these programmes were officially argued with the protective effects of forests against natural hazards the measures focused on protection forests according to the Forest Act. Only a small part of these forests actually provide protection against natural hazards as well. When granting subsidies under these programmes the forest authority does not primarily consider direct protective effects against rock-fall, avalanches, etc. Practically all forest management measures carried out by forest owners are subsidised, and measures specifically oriented at an enhancement of protective effects are hardly considered. The forest authority assumes that all "forest functions" are enhanced in an indirect way by regular forest management. Furthermore they hope that protective effects are safeguarded by supporting economically oriented forest management, which is carried out by the owner. Multifunctional forest management is not operationalised according to the different "functions". (10)Thinking in the tradition of the "wake theory" it is believed that orderly forest management automatically fulfils all forest functions. (11)

Perceptions of risk

In subsidised projects for protection forest restoration in Austria, wind-throw areas are usually treated traditionally, which includes the harvesting of the timber followed by afforestation. This treatment produces bare ground like clear-cuts, thus increasing the threat of avalanches and erosion. In many cases leaving the thrown timber at the site would have been cheaper and more favourable in terms of protective effects. Why is the clearing of the wind-throw sites promoted by the forest officials? Clear-cuts are traditionally regarded as normal treatments, and therefore accepted by forest officials as well, although they are unfavourable in terms of natural hazard protection. Clear-cuts are assessed positively because they are intentional actions; wind-throw on the other hand is regarded as damaging - as nature disturbing planned forest management. The language chosen by forest officials shows how management measures and natural 'catastrophes' are perceived: wind-throw areas are often called areas 'cleared' by storms, although the storm alone did not actually clear the ground - this is eventually done by the forest workers who remove the timber from the site. It was not the natural incident but human intervention that produced bare ground. Foresters traditionally fear risks if wind-throw sites or even healthy forests were left untreated.

Forest officials and silviculture scientists in mountainous countries in Europe usually appeal for more active management of mountain forests (e.g. Mayer, 1967a, b, 1976; BMLF, 1993). They argue that the natural development of forest stands includes unstable 'stages' such as the 'decomposition stage' - when old trees die and regeneration starts - susceptible to natural threats. Another view - that natural mountain forests are particularly stable ecosystems (e.g. Meister, 1985) - is shared by some foresters and environmentalists. They consider forest management destructive to mountain forests, for example producing unnatural and unstable stands susceptible to wind-throw, insect attacks and other threats. Both views have firm arguments; both are based on scientific knowledge and practical examples may be found for both positions; the generalisations, however, are guided by ideological beliefs that can be interpreted as different views of nature. Nature may be seen as harmony or as chaos (Sieferle, 1997: 19-21). In the first view human interventions cause risks, in the latter natural forest dynamics do (Weiss, in press). In their "silvacentrist" view (Glück/Pleschberger, 1982) foresters do not differentiate between risks for the forest and from the forest. For them risks for the forest are at the same time risks for the society.

Recent research indicates that many cases need less intensive interventions in mountain forests than foresters tend to believe – even forests damaged by wind-throw and bark beetles (Frey et al., 1995). In certain cases, however, treatment seems to be necessary to avoid unacceptable levels of risk, e.g. on very steep slopes directly above settlements or infrastructures. The actual policy applied is highly determined by the stake-holders involved in the policy-making process: A low-intervention strategy where timber has been left on wind-throw sites was practically applied in certain cases in the Austrian province of Vorarlberg and in the neighbouring countries Switzerland and Bavaria. In these countries the share of municipal forest property is high which gives other interests than timber production more weight in management decisions. Especially in Switzerland nature protection groups are explicitly included in forest policy-making (SBN, 1994).

(10)The forest policy doctrine of forest functions goes back to Dieterich (1953). In practice, the term "forest function" is rather to be understood as an ideological than a scientific expression (Heeg, 1973; Glück/Pleschberger, 1982; Krott, 1985).
(11) The ideology of the so-called "wake theory" seemingly harmonises competing interests in the forest (Glück/Pleschberger 1982: 652).

The torrent and avalanche control service (TACS)

The "forest-technical system"

The institution of the forest-technical service for torrent and avalanche control dates back to the late 19^th century when the Act for Torrent Control came into force after severe floods that occurred in large parts of the Austrian Empire in 1882. Since that time the service is assigned to the forestry jurisdiction and implements what is called the forest-technical system of torrent and avalanche control. This concept aims at combining biological and construction measures for preventing natural hazards. Technical means should not only be applied to prevent damage; the land should be used in a way that hazards are avoided as well (precautionary principle). For instance, alpine pastures should not be overgrazed so that no erosion or land-slides occur; alpine meadows should be afforested if this can prevent avalanches; in forest management, clear-cuts that might cause avalanches, rock-fall, or land-slides should be avoided. In recent time another instrument, following a precautionary approach as well, was developed by the TACS: hazard zones mapping. Land-use should not only be controlled in the catchment areas of torrents and avalanches but also in the hazard-prone areas themselves. While primarily agriculture and forestry are concerned in the catchments, the land threatened by hazards are often settlement areas. In prospering times many new buildings were constructed in such areas, thus causing additional needs for prevention measures against hazards. Hazard zones planning is now striving to prevent the development of such areas.

Conflicts in land-use regulation

The TACS carries out biological and construction measures and elaborates hazard zones plans. A large amount of money is invested in prevention measures every year (over 1 thousand million ATS). On average 80% of the project costs are subsidised. Beneficiaries, e.g. municipalities, have to apply for projects; they are the clientele of the TACS. Although the forest-technical system comprises theoretically both land-use regulations and construction measures, actually mainly technical measures are carried out.

The preference for technical constructions is not strictly dependent on weighing system performance or by considering their return on investment. One important reason why constructions are preferred over land-use regulations is that the restriction of land-use causes more conflicts with local users (e.g. afforestation of alpine pastures). Furthermore, technical constructions like check-dams have immediate effects which have a higher political value. Politicians at municipal level prefer immediate effects and are not only interested in physical impacts (less future damage) but also economic (increased property values) and symbolic benefits (proof that they "care" about the safety of their municipality). The symbolic value of constructions is higher because their protective effects are more obvious. Although afforestations would provide a long-term enhancement of the communal safety, they are not promoted adequately. Why – if the TACS considers land-use regulations a central part of watershed amelioration – are these measures not safeguarded more strongly?

shows the typical policy network in planning an avalanche prevention project. The acceptance of constructions and biological measures (afforestation) in such a project is indicated with regard to the type of political actors usually involved. Actors participate in the policy network formally or informally. As a rule, not only the village/town (applicant) has high interests in realising technical constructions but also the county authority which informally represents influential regional interests. The constructions themselves are opposed only by nature protection concerns. On the other hand, there are rather powerful actors opposing the afforestation of alpine pastures in such a project (farmers, hunters, agrarian officials, tourist managers) while only a few actors support this land-use regulation (representatives of the ministry, the local forest office and the nature protection office). According to the principles of their "forest-technical system" the local TACS officer usually proposes afforestations wherever ecologically feasible but only few of them pursue this goal if they face strong opposition from local actors.

*) Interest groups are ranked according to the significance and power in the local implementation process.

**) The actors’ interests in implementing the stated measures are valued as: support (+), neutral (0), resistance (-); the question mark (?) means that the position of the actor is "not homogeneous".

Table : Typical attitudes of political actors towards avalanche control projects with and without afforestations (source: WEISS, 1998b: 123).

As the study underlying this article indicates, the result of this distribution of interests is that "technical" projects are usually preferred over afforestation projects, only minor afforestation areas are included in projects, and the success of afforestations is often poor because they are not properly protected against damage by browsing etc. Other interest groups or the public that possibly would criticise unsustainable land-uses and thus could strengthen the position of the TACS are not included in the decision-making process.

Bias in risk-benefit analysis

As a prerequisite for granting subsidies for natural hazards prevention projects in Austria, cost-benefit analyses (CBA) have to show a positive result. This tool goes back to an initiative of the ministry of finance. The guidelines were elaborated by scientists under the head of the ministry of agriculture. A cost-benefit analysis is a management tool of the central organisation – it is not in the very own interest of the street-level bureaucrats because it restricts their competence of decision. The policy evaluation shows that the TACS does not accept this instrument as a decision tool at all. Decisions what projects will be realised are not based on a cost-benefit analysis; this analysis tends to be elaborated when the decision has already been taken unofficially by the TACS regional officer and the measures are already planned in detail. The CBA is just used for legitimising the functionality and efficiency of the project proposed. Decisions are mainly taken in the regional offices of the TACS that choose, project and realise the projects. Effective factors for the practical decision besides of the estimated hazard are (a) the interest and the financial capacity of the applying community, (b) public attention, (c) informal interests of the agency, and (d) the regional economic development or political priorities on provincial level. The impacts of the formal CBA in torrent and avalanche control can be seen in (a) a certain influence on the thinking of the decision makers in terms of efficiency when projecting measures and (b) in sorting out the very least efficient projects.

The shortfalls of the CBA as applied in practice are manifold. There is a wide range in assessing the benefits of such projects. For most projects supported by the agency the CBA can be calculated in a way that the result supports its realisation. The formal need is only to prove the single project "positive"; projects are not ranked nation-wide. The CBA as applied within the TACS reflects neither the economic goals of the ministry of finance nor the official principles of the TACS, neither in formulation nor in implementation.

Although being claimed an "objective" tool the CBA falls short this pretension with regard to both theoretical and pragmatic aspects. The CBA as applied is not scientifically sound with respect to economic theory because different value categories are simply aggregated. (12)

It does also not pragmatically reproduce the formal goals of the TACS. The positions considered within the CBA framework do not adequately reflect the official principles of the TACS. Principles formulated by the agency state that (a) the protection of existing facilities should be prioritised over developing new ones, (b) the protection of lives is more important than the protection of material values, and (c) developments should not be supported that result in future needs for increasing prevention measures. (13)There are positions included in the CBA that support projects which are not favourable according to these principles. Land that might be developed due to the project in question is evaluated with its expected future value (estimated rise of the property price). This value is usually several times higher than the estimated damage of existing houses (the damage is devaluated and reduced according to the probability of a damage). Furthermore, the CBA itself does not consider human lives. Representatives of the TACS on the one hand call this a big drawback of the analysis, on the other hand they believe that human lives should not be expressed in monetary terms. They use this situation to argue that the threat of lives is to be included in the decision additionally to the CBA. The hazard zones plans that indicate the number of houses possibly affected by natural disasters serve as a database for the consideration of possible deaths. However, the consideration of potential fatalities does not take place formally. As a result, there is a wide spectrum of possible decisions. The agency does not want their discretion to be restricted by a standardised procedure.

(12) Guidelines ("Kosten-Nutzen-Untersuchung"), BMLF (1978).
(13) E.g., guidelines "Hinderungsgründe", BMLF (1980).

The political system undermines the CBA in another way: by using regional economic arguments, even projects might be realised that do not produce a positive cost-benefit relation. The project is then called a contribution to the economic development of disadvantaged regions. Most projects, however, are realised in prospering regions, e.g. regions with tourism. As a result the natural hazards protection policy does not restrict but support economic development in mountain regions. Growing economic activities in these regions mean growing threats and increasing needs for protective measures.

Operative deficits of hazard zones planning

The main policy means of the TACS – prevention measures and hazard zones mapping – are not implemented independently from each other. There are several relations between these instruments. Formally (according to the Forest Act), there are two purposes of the plans: (a) to prevent the development of hazard-prone areas for housing etc., and (b) to provide comprehensive data on the hazardous processes as a basis for systematic planning of prevention concepts. Informally, hazard zones plans also examine the "objective" necessity of prevention measures as a first step. The TACS uses the plans to argue for the elaboration of prevention projects in villages/towns if buildings are situated in "red zones" (high hazard zones). In a second step, the plans serve for evaluating the achievements of the service for the safety of the community. The municipalities, who have to cover a share of the project costs, want to "see results". A higher standard of safety is visible in technical constructions and furthermore in changed hazard zones: after the completion of a project, "red zones" are reduced in the maps. What politically counts is not the real reduction of hazards, but the reduction of hazard zones on the map. With the elimination of "red zones" these areas are free for economic development. The value of these properties may increase dramatically, especially if situated in tourist regions.

The problem with this procedure is that risks are partly ignored. The hazard zones maps provide only restricted information; the hazard assessment refers to a "standard or reference event". This reference event is a disaster of a magnitude that is statistically expected every 100 or 150 years. The remaining risk of events with higher impact, the uncertainties of the assessment, as well as the uncertainties of the construction measures are not considered in the assessment and are not shown in the map. Also constructions are laid out for preventing the "standard event" only. If new houses are erected in formerly "red zones", now protected by technical constructions, they are still threatened by remaining risks and uncertainties. These remaining risks are ignored (Jäggi, 1996); the prevention concepts provide "false" or "relative" safety (Weinmeister, 1988). There are examples in Austria where newly built houses were damaged by events that were more dramatic than "expected". In other cases the prevention measures failed (Aulitzky, 1968, 1970, 1996). 100 years ago the reference event for prevention constructions was an event with a 30 years probability of recurrence; 50 years ago this standard was 50 years. What safety will be expected in the future? In history, villages and towns were built in the protection of constructions, the protective effects of which are insufficient according to today’s safety standards. Today’s measures probably do not meet future expectations. The development of areas protected by technical constructions inevitably results in future damage. The protection gained by the constructions is given away if new houses are erected. Such a policy enlarges the developable area but does not reduce hazards on the whole.

A joint restoration programme for protective forests

Initiation and formulation

In the late 1980’s a joint programme of the forest authority and the TACS was launched to support mountain forest restoration. The ministry – the initiator of this subsidy programme – observed that the TACS prioritised technical constructions over measures aiming at maintaining the protective effects of forests. The forest authority, in contrast to this, was interested in carrying out such forest management measures but the budget of the protection forest restoration programme existing at that time was insufficient. With the new programme money from the Disaster Relief Fund administered by the TACS was made available for the forest authority. For formal reasons restoration projects had to be borne jointly by both institutions. In the following a dispute between the two institutions arose about the objectives of the programme. Officially, the programme was called "restoration of damaged forests in catchments of torrents and avalanches" because as a part of the disaster relief programme it was restricted to natural hazards. Practically, the measures were called "protection forest restoration" by the forest authority because they wanted to implement their restoration concept in "protection forests" according to the Forest Act (see above). In the understanding of the TACS, restoration projects should have the character of watershed management measures oriented at enhancing the protective effects of the forests against natural hazards ("protective forests"). Both agencies applied their routine concepts: While the forest authority wanted to be active in all "protection forests", the TACS considered only "protective forests" targets of the programme.

The conflict between "protection forest restoration" and "watershed management" was in fact a conflict over competencies and resources of these two institutions. The forest authority insisted on the term "protection forests" according to the Forest Act and the TACS adhered to their definition of "protective forests" in terms of natural hazards. Defining the programme goal with reference to "protection forests" would have granted the forest authority a larger field of activity. Focusing on "protective functions" would have strengthened the TACS. Eventually, a compromise was reached by not generally excluding measures on a larger scale but giving priority to forests with direct protective effects.

Implementation

The obligation to joint projects of the two agencies also caused difficulties because of unclear responsibilities. This problem was solved by defining a main responsible agency for each project. Practically, projects carried out by the forest authority or the TACS differ (see Fehler! Unbekanntes Schalterargument.): The forest authority includes larger areas in the projects and encourages the forest owners to carry out the measures. Measures are planned in such a way that they also meet the interests of the forest owners. The TACS, in contrast to this, focuses on areas directly protecting settlements or traffic lines, and carries out the work on its own. Measures of the TACS are designed for immediate and maximum protective effects.

The main clienteles’ interests influence the decisions of the public agencies. This results in a timber production orientation of the forest authority additionally to the protection goal (preferences of the forest owners) and in a short-term orientation of the TACS (preferences of the beneficiaries). While the management concept of the forest authority includes forest management measures also oriented at timber production, the TACS focuses more on the specific protective effects desired.

Two risk concepts

The two institutions operate with different risk and management concepts

The forest authority aims at a multifunctional forest. They assume that "close-to-nature management" will result in a "stable" forest, i.e. the best provider of all "functions" of the forest. Forest owners are supposed to play an important role in keeping a stable forest. The forest authority’s subsidy approach is oriented at the forest owner. Subsidies should be incentives for them to consider public interests in their management decisions. Measures subsidised by the forest authority do not derive from specific protective effects of the forest because they require that mountain forests generally provide a long line of important protective effects for society. As they consider indirect effects just as important as direct effects, in their view the specific protective effects do not have to be proven in each situation. They have a holistic but vague approach. In the view of the forest authority the risk of forests losing their potential to carry out this "function" may be a result of natural disturbances. They look at risks for the forest. If the forest is at risk, its "functions" are endangered as well. Furthermore, the forest they generally have in mind is rather a "timber-forest" than a protective forest because this is the traditional image of the ideal forest.

The TACS, in contrast to this, applies a monofunctional management concept. They start with certain protective effects forests have for settlements or infrastructures. Measures derive from specific protection goals related to certain threats. In their view they subsidise measures in the interest of the beneficiaries of the protective effects. Their risk concept is directly connected with the security of the properties below the forest. Their protection concept is derived from certain natural processes, which potentially threaten human constructions, and the related protective effects of the forests. Using a causal model they search for the risks which the forest exerts on other properties.

Table : Institutional background, risk perception and management concepts of two agencies active in mountain forest restoration

Discussion and conclusions

In land management it is a typical problem that decisions of certain actors may produce risks for other parties. As the effects of land management for the safety of other parties are often characterised as external effects and public goods, a problem solution strategy using private instruments, e.g. contracts, is difficult and regulations are necessary. The management of forests, for instance, may have negative effects on properties below the forest; the management of watersheds has effects on all properties situated along the river. In the first case the circle of beneficiaries is small which makes it basically possible for them to organise their interests themselves; in the latter a big group benefits from appropriate land management strategies which makes it necessary for the State to intervene, e.g. by taking preventive measures. These public activities again may result in the need for accompanying regulations: As the State covers the costs for hazards prevention and disaster relief, the development of hazard-prone areas for housing etc. has to be restricted. As the property owners’ liability for their decisions is limited, they would take higher risks at the cost of others (e.g. hotel guests) or the public (the erection of prevention constructions is subsidised). Furthermore, as areas for development are officially zoned the State also takes responsibility for their safety.

We see, there are several reasons for the State to become active in natural hazards protection. How do public agencies deal with this task? The examples quoted confirm the statement that the perception of risks is strongly determined by the interests and values of the political actors involved in policy-making. The two agencies have different approaches to their work: While the forest authority strives for multifunctional forest management, the TACS applies a monofunctional concept in managing protective forests. General findings about the behaviour of public administration are true for risk regulation as well: Public agencies do not only pursue formal but also informal goals. They are oriented at their main clientele; they want to safeguard their autonomy, territory and legitimacy; they strive for safe decisions; they apply their routine and change policy only under high pressure. Different combinations of main competence field and clientele of agencies on the one hand and certain tasks of risk management on the other produce different problem solving strategies.

One leading principle of bureaucratic behaviour is the need for safe decisions. In the case of the forest authority it was sufficient to define the areas of protection forests with need for restoration only roughly. The criterion of "protection forest" is laid down in the Forest Act, the main legal basis for their activity. The criterion "need for restoration" was assessed by silvicultural expertise and in three priority levels. The assessment was practically oriented at the models of the traditional "timber forest" and "close-to-nature" management; for "protective forests", no specific model was developed.

The TACS defined watershed areas and the protective functions of the forest to legitimise their interventions because their principal field of activity is the protection of settlements and infrastructures against torrents and avalanches. In their main field of activity, which is directly connected with the physical processes of natural hazards, the TACS developed more sophisticated risk assessment models. For assessing hazards they introduced the concept of the "probability of recurrence" describing the frequency of certain dangerous processes at a certain location. Their assessments are oriented at a standard event with a certain frequency (usually 100 or 150 years). Although the available data are so limited that a sound figure for this frequency cannot be given in most of the cases they still work with this concept. In their hazard zones maps, where they are forced to decide on a certain definition of the area potentially affected by a standard event, the weakness of their database is especially challenging. The map may be questioned by people whose properties are zoned as "non-developable" land. For this reason the procedures demand that the hazard zones maps are to be approved by a multi-step evaluation within the TACS hierarchy. The official hazard zones plan is approved by the federal minister of agriculture. The database for the assessment is not improved by this procedure but the decision becomes "safer" symbolically.

Another important informal behaviour of the administration is their orientation at their main clientele. The forest authority's approach is determined by their "multifunctional" forest management concept and their orientation at forest owners’ interests. The multifunctional concept is for a large part a symbolic rhetoric which is used to safeguard the agency’s legitimacy because it considers the many good effects of forests for the public. In practice, their policy is strongly oriented at timber production which is generally the main income for forest owners. Other interests, e.g. nature protection concerns, are not pursued unless NGO’s exercise high pressure and/or the forest owners are interested themselves. This is the case with community forests which, for instance, are typical of neighbouring Switzerland. Hunting, which is often in the interest of the owner, is officially not supported by the forest authority because this is the responsibility of a different agency. The interventions of the Austrian forest authority in protection forest management were caused by their attempt to safeguard or extend their competence field, their resources and their legitimacy. As the necessity of forest restoration measures was not demanded by the benefiting parties and the agency used public money, they were not forced to develop specific management concepts to meet the needs of this group. The beneficiaries of the protective forests were not included in the policy network. Therefore, the measures are actually for the benefit of the forest owners instead of the property owners whose land is situated below the forest.

This is on the other hand the case with measures carried out by the TACS. They implement measures demanded by property owners/municipalities threatened by hazards. Additionally projects are co-financed by state subsidies and the money of the beneficiaries of the protection measures. However, as these groups are usually more interested in immediate and obvious protection, the TACS is more active in building technical constructions than maintaining the forests. The result of the TACS’ orientation at the interests of their clientele is that most of their measures are rather for the benefit of economic development than for reducing hazards in general. The hazard zones plans which follow the precautionary principle but restrict land-use are implemented by offering subsidies for preventive constructions in exchange. Zoning hazard-prone areas is accepted locally because the TACS promises subsidised measures, which reduce the hazard and the hazard zones. Remaining risks and uncertainties are neglected in the design of constructions and zoning. Defining areas that are protected by technical constructions as free for development leads to future damage or even deaths because of the remaining risks.

The most important lesson to learn from these examples is that the risk policies applied are strongly determined by the stakeholders' interests included in the policy network. Different institutional settings lead to different policies. Experts do not necessarily disagree with these policies; on the contrary, official experts develop different risk concepts and thus different solutions depending on the institutional framework.

On the basis of this insight improvements of the political-administrative system may be found by institutional changes. If certain groups shall benefit from the measures, their inclusion by political participation and/or financial involvement promises more effective and efficient implementation. With these groups participating in decision-making, goals might be formulated more specifically; if benefiting groups are additionally involved in financing, cost-benefit relations might be considered more seriously. At the moment, only powerful groups pursue their interests in the implementation process in an informal way. Formal participation of all groups that are affected by policies would make the political practice of public administrations more democratic. Further research is suggested on ways how beneficiaries and the public could be included in policy-making and on the possible effects of such institutional and procedural innovations.

References:

Aulitzky, H., 1968a: Analyse der Schadensursachen von Unwetterkatastrophen zum Zweck der Vorbeugung. Österreichische Wasserwirtschaft 20, H. 5/6 und 7/8.

Aulitzky, H., 1970: Der Enterbach (Inzing in Tirol) am 26. 7. 1969: Versuch einer Analyse eines Murganges als Grundlage für die Neuerstellung einer zerstörten Wildbachverbauung. Wildbach- und Lawinenverbau 34, H. 1: p. 31-66.

Aulitzky, H., 1996: Beispiele vermeidbarer Hochwasser- und Lawinenschäden. In: Naturkatastrophen – hausgemacht? Conference Proceedings, Österreichische Gesellschaft für Ökologie, Wien: p. 25-55.

Bechmann, G. (ed.), 1993: Risiko und Gesellschaft. Grundlagen und Ergebnisse interdisziplinärer Risikoforschung. Westdeutscher Verlag, Opladen.

Beck, U., 1988: Gegengifte. Die organisierte Unverantwortlichkeit. Suhrkamp, Frankfurt/Main.

BMLF (Bundesministerium für Land- und Forstwirtschaft), 1978: Kosten-Nutzen-Untersuchung. Wien.

BMLF (Bundesministerium für Land- und Forstwirtschaft), 1980: Richtlinien Hinderungsgründe. Wien.

BMLF (Bundesministerium für Land- und Forstwirtschaft), 1993: Projekt Schutzwaldverbesserung. Wien.

Compare, Alexander, 1995: Risikokommunikation als Machtproblem. Europäische Hochschulschriften 22. Lang, Frankfurt/Main.

Conrad, J., 1989: Risiko, Ritual und Politik. In Schmidt, M. (ed.), Leben in der Risikogesellschaft. Der Umgang mit modernen Zivilisationsrisiken. Müller, Karlsruhe: p. 179-204.

Dieterich, V., 1953: Fostwirtschaftspolitik. Hamburg – Berlin.

Dietz, T.M. and Rycroft, R.W., 1987: The Risk Professionals. Sage, New York.

Douglas, M. and Wildavsky A., 1982: Risk and Culture. An Essay on the Selection of Technological and Environmental Dangers. University of California Press, Berkeley.

Downs, A., 1967: Inside Bureaucracy. Little and Brown, Boston.

Edelmann, M., 1976: Politik als Ritual. Deutsche Übersetzung. Campus, Frankfurt/Main.

Elias, N., 1976: Über den Prozeß der Zivilisation. 2 Bd., Suhrkamp, Frankfurt/Main.

Evers, A. and Novotny, H., 1987: Über den Umgang mit Unsicherheit. Die Entdeckung der Gestaltbarkeit von Gesellschaft. Suhrkamp, Frankfurt/Main.

Frey, W., Forster, B., Gerber, W., Graf, F., Heininger, U., Kuhn, N. and Thee, P., 1995: Risiken und Naturgefahren auf Windwurfflächen. Schweizer Zeitschrift für Forstwesen 146 (11): p. 863-872.

Fritsch, M., Wein, T. and Ewers, H.-J., 1993: Marktversagen und Wirtschaftspolitik: mikroökonomische Grundlagen staatlichen Handelns. Vahlen, München.

Glück, P., 1976: Die Rolle der Verbände in der theoretischen Forst- und Holzwirtschaftspolitik. Habilitationsschrift an der Universität für Bodenkultur, Wien.

Glück, P. and Pleschberger W., 1982: Das Harmoniedenken in der Forstpolitik. Allgemeine Forstzeitschrift 37 (22): p. 650-655.

Glück, P. and Weber, M. (eds.), 1998: Mountain Forestry in Europe – Evaluation of Silvicultural and Policy Means. Publication Series of the Institute for Forest Sector Policy and Economics 35. Universität für Bodenkultur Wien.

Glück, P. and Weiss, G., 1997: Evaluierung der Instrumente zur Sicherung der Schutzleistung des Waldes – Umweltpolitische Prinzipien in der Implementierung. In BUWAL, Ökosystem Wald - Rohstoff Holz - Prinzip Nachhaltigkeit. Umweltmaterialien Nr. 67. Bundesamt für Umwelt, Wald und Landschaft, Bern, p. 45-56.

Heeg, B., 1973: Forstliche Funktionenlehre und forstpolitische Theorie. In: Allgemeine Forst- und Jagdzeitschrift 144 (12): p. 242-251.

Héritier, A. (ed.), 1993: Policy-Analyse. Kritik und Neuorientierung. Politische Vierteljahresschrift (PVS), Sonderband 24/1993. Westdeutscher Verlag, Opladen.

Hollenstein, K., 1995: Analyse und Bewertung von Risiko und Sicherheit bei Naturgefahren. In: Schweizer Zeitschrift für Forstwesen 146/9: p. 687-700.

Jäggi, M., 1996: Gedanken zu den Grenzen der Vorbeugung. Österreichische Wasser- und Abfallwirtschaft 48, H. 5/6: p. 179-186.

Jänicke, M., 1986: Staatsversagen: die Ohnmacht der Politik in der Industriegesellschaft. Piper, München, Zürich.

Johnson, B.B. and Covello, V.T., 1987: The Social and Cultural Construction of Risk. Reidel, Dordrecht.

Jungermann, H. and Slovic, P., 1993: Die Psychologie der Kognition und Evaluation von Risiko. In: Bechmann (ed.): p. 167-207.

Kaufmann, F.-X. (ed.), 1991: The Public Sector. Challenge for Coordination and Learning. deGruyter, Berlin, New York.

Kienholz, H., 1993: Naturgefahren – Naturrisiken im Gebirge. Forum für Wissen: p. 7-22.

Koch, N.E. and Rasmussen, J.N. (eds.), 1998: Forestry in the Context of Rural Development. Final Report of COST Action E3. Horsholm: Danish Forest and Landscape Research Institute.

Krohn, G. and Krücken, W., 1993: Risiko als Konstruktion und Wirklichkeit. Eine Einführung in die sozialwissenschaftliche Risikoforschung. In Krohn/Krücken (eds.): p. 9-44.

Krohn, G. and Krücken, W. (eds.), 1993: Riskante Technologien: Reflexion und Regulation. Einführung in die sozialwissenschaftliche Risikoforschung. Suhrkamp, Frankfurt/Main.

Krott, M., 1985: Zu den Waldfunktionen als Instrument der forstpolitischen Wissenschaft und Praxis. Centralblatt für das gesamte Forstwesen 102 (1): p. 1-28.

Krott, M., 1990: Öffentliche Verwaltung im Umweltschutz. Ergebnisse einer behördenorientierten Policy-Analyse am Beispiel Waldschutz. Braumüller, Wien.

Lowrance, W.W., 1980: The nature of risk. In Schwing/Albers (eds): p. 5-14.

Luhmann, N., 1966: Politische Planung. Jahrbuch für Sozialwissenschaft. Göttingen.

Luhmann, N., 1993: Risiko und Gefahr. In: Krohn/Krücken (eds.): p. 138-185.

Mayer, H., 1967a: Das Fichten-Naturwaldreservat Rauterriegel am Eisenhut bei Turrach (Österreich). Centralblatt für das gesamte Forstwesen 84: p. 2-6.

Mayer, H., 1967b: Zur Behandlung überalterter Gebirgswälder. In: Schweizer Zeitschrift für Forstwesen 118 (6): p. 335-372.

Mayer, H., 1976: Gebirgswaldbau-Schutzwaldpflege. Stuttgart.

Mayntz, Renate, 1985: Soziologie der öffentlichen Verwaltung. 3., überarb. Auflage. Müller, Heidelberg.

Mazur, A., 1985: Bias in Risk-Benefit Analysis. Technology in Society 7: p. 25-30.

Meister, G., 1985: Übergangsstrategie zur Sicherung der Schutzfunktionen des Gebirgswaldes. Mitteilungen des Deutschen Alpenvereins, H. 3: p. 149-155.

Müller, W., 1993: Risiko und Ungewißheit. Enzyklopädie der Betriebswirtschaft: p. 3813-3825.

Nelkin, D. and Pollak M., 1980: Problems and Procedures in the Regulation of Technological Risk. In: Schwing/Albers (eds.): p. 233-248.

Olson, M., 1965: The Logic of Collective Action. Cambrigde (Mass).

Ottitsch, A. and Weiss, G., 1998: Mountain Forest Policies in European Countries – a Comparison Using RAGIN’s Qualitative Comparative Analysis Method. In Glück/Weber (eds.): p. 263-284.

Ottitsch, A., Pregernig, M. and Rappold, G., 1998: Bürgerbeteiligung in der forsttechnischen Planung. Research report prepared for the Bundesministerium für Land- und Forstwirtschaft, Wien.

Perrow, C., 1984: Normal Accidents. Living with High-Risk Technologies. Basic Books, New York. Dt.: Normale Katastrophen. Die unvermeidbaren Risiken der Großtechnik. Campus, Frankfurt/Main and New York.

Peters, B.G., 1984: The Politics of Bureaucracy. New York and London.

Pleschberger, W., 1989: Staat und Wirtschaft. Am Beispiel der österreichischen Forstgesetzgebung von 1950 bis 1987. Studien zu Politik und Verwaltung Bd. 28. Böhlau, Wien and Köln.

Pregernig, M. and Weiss G., 1998: Forest Policy in Austria: Policy Making by the Sector for the Sector. Paper presented at the IFF (Intergovernmental Forum on Forests) European Regional Meeting on the Underlying Causes of Deforestation and Forest Degradation in Europe, 28-29 October, Bonn.

Prittwitz, V., 1990: Das Katastrophenparadox. Elemente einer Theorie der Umweltpolitik. Leske+Budrich, Opladen.

Rosa, E.A., 1997: Modern Theories of Society and the Environment: The Risk Society. Paper presented at the Conference "Sociological Theory and the Environment", 20-23 March 1997, Woudschoten, The Netherlands.

Sabatier, P.A., 1986: The Top-Down and Bottom-Up Approaches to Implementation Research: a Critical Analysis and Suggested Synthesis. Journal of Public Policy 6(1) 21-48.

Sabatier, P.A., 1993: Advocacy-Koalitionen, Policy-Wandel und Policy-Lernen: Eine Alternative zur Phasenheuristik. In: Héritier (ed.): p. 116-148.

SBN, 1994: Naturschutz im Bergwald. Grundlagenpapier in Zusammenarbeit mit der Schweizerischen Gebirgswaldpflegegruppe. Schweizerischer Bund für Naturschutz, Basel.

Scharpf, Fritz W., 1997: Games Real Actors Play. Actor-Centered Institutionalism in Policy Research. Westview Press, Boulder, Oxford.

Schmiderer, H. and Weiss, G., 1999: Die Bannwaldpolitik in Österreich. Publication Series of the Institute for Forest Sector Policy and Economics 38. Universität für Bodenkultur, Wien.

Schwing, R.C. and Albers, W.A. (eds.), 1980: Societal Risk Assessment. How Safe is Safe Enough? Plenum Press, New York.

Shrader-Frechette, K.S., 1995: Evaluating the expertise of experts. Risk 6(1).

Shrader-Frechette, K.S., 1997: How some risk frameworks disenfranchise the public. Risk 8(4).

Sieferle, R.P., 1997: Rückblick auf die Natur: eine Geschichte des Menschen und seiner Umwelt. Luchterhand, München.

Starr, C., 1969: Societal Benefit versus Technological Risk. What is Our Society Willing to Pay for Safety? Science 165: p. 1232-1238.

Starr, C., 1980: Introduction to the session. In Schwing/Albers (eds): p. 1-4.

Thompson, P.B. and Dean, W., 1996: Competing conceptions of risk. Risk 7(3).

Weinmeister, H.W., 1988: Ökonomie und Ökologie am Beispiel des Schutzwasserbaues (landschaftsökologische und volkswirtschaftliche Grenzen des technischen Schutzwasserbaus). Proceedings of the conference Interpraevent 1988, vol. 4: p. 339-376.

Weiss, G., 1998a: Bewertung der Maßnahmen zur Schutzwaldverbesserung. Research Report prepared for the Bundesministerium für Land- und Forstwirtschaft, Wien, 4 volumes.

Weiss, G., 1998b: Evaluation of Mountain Forest Policy in Austria. In Glück/Weber (eds.): p. 107-136.

Weiss, G., 1999: Die Schutzwaldpolitik in Österreich. Einsatz forstpolitischer Instrumente zum Schutz vor Naturgefahren. Proceedings of the Institute of Forest Sector Policy and Economics 39, Wien. Also: dissertation at the Universität für Bodenkultur, Wien.

Weiss, G. (in press) Perceptions of risks in mountain forests. In M. Price (ed.), State-of-knowledge-report on forests in mountain development. IUFRO/FAO.

Wildavsky, A., 1989: Searching for safety. Oxford (USA).

Wolf, R., 1986: Der Stand der Technik. Geschichte, Strukturelemente und Funktion der Verrechtlichung technischer Risiken am Beispiel des Immissionsschutzes. Westdeutscher Verlag, Opladen.