Annex X
APDC/01/6

PASTORAL RISK MANAGEMENT FOR DISASTER PREVENTION AND PREPAREDNESS IN CENTRAL ASIA WITH SPECIAL REFERENCE TO THE CASE OF MONGOLIA ^*

1. Introduction

1. Extensive pastoralism plays a key role in the rural economies of most Central Asian countries. It is a key economic adaptation of using under harsh climatic conditions (long and very cold winters and dry summers), the vast grassland areas, mountains and desert steppes, involving large numbers of people, occupying extensive land areas and making a significant contribution to the national economy. In addition, contrary to the trend in other parts of the world, there has been an increase in the number of people engaged in extensive pastoralism since decollectivization in several countries in Central Asia, as economic adjustment takes place, subsidized industrial and agricultural enterprises fail, and the artificial urban economies promoted by central planning falter.

2. In Mongolia, pastoralism employs between 1/3 and 1/2 of the national population and its output constitutes 1/3 of the GDP and 1/4 of export earnings. In Kyrgyzstan, mobile livestock keeping employs 1/4 of the agricultural population. Pastoralism does not play a large part in the Chinese economy as a whole, but in several provinces and autonomous regions - especially Inner Mongolia, Xinjiang, Tibet and Qinghai - extensive pastoralism is the main economic activity of millions of people and a substantial proportion of China's livestock (in 1990, 22 per cent of all large livestock, and 37 per cent of all sheep and goats) are raised under pastoral conditions. Therefore, the development of pastoral economies in these areas is a key to poverty alleviation and improved food security, as well as to the wider goal of creating sustainable livelihoods.

3. In spite of their importance pastoral production systems in Central Asia are, however, also subject to high levels of variability and risk mainly from environmental, but also from socio-economic and institutional causes. The most obvious risk for pastoral livestock keepers comes from periodic snow disasters, which cause high animal and human mortality. The most recent examples of snow disasters causing major damage in Central Asia occurred during Mongolia in the winters of 1999/2000 and 2000/2001 associated with continued summer droughts. The harsh winters with extended snowstorms and long spells of cold came at a time when risk preparedness and management strategies were almost non-existent. An estimated 2.4 million animals died during the 1999/2000 winter. More than 2,000 herding households lost all their livestock and were thus pushed into poverty. The damages of the 2000/01 winter had not yet been fully assessed at the moment of writing this paper, but in March 2001 over 2 million animals had already died and many more were expected to die later in spring ¹.

4. In addition to the direct impact on animal and human mortality, snow and other risks undermine pastoral development strategies, and are major causes of rural and urban poverty. Pastoral development will have to start anew every few years if risks are not better managed, thus reducing the effects of disasters.

5. The vulnerability of pastoral households to risks has substantially increased since decollectivization in Mongolia, China and the Central Asian countries of the former USSR. In the past, pastoral societies in the region had a range of coping mechanisms, based on customary institutions before and on state institutions ² during the socialist period, to limit the damages suffered by individual households as a result of shocks, and to reinsert impoverished households into a viable production process. However, transition governments can no longer afford the wide range of previously subsidized services. In Mongolia and Kyrgyzstan, with the liberalization of the economy starting in the early 1990s, the following changes have occurred: the employment guarantee has been removed, price control of many basic foodstuffs has been removed, the automatic resupply of collectively owned animals to households with insufficient herds has been discontinued, subsidized emergency fodder supplies no longer exist, cost-recovery for most services has been introduced, and social security guarantees have been reduced. Changes in China - marked by the introduction of livestock management contracts in 1983 and pasture management contracts in 1994 - have been less drastic, but have had a similar effect. This has meant a de facto increase in risk exposure of herding households ³. However, none of the countries of Central Asia has yet developed adequate strategies, policies and institutions dealing explicitly with risk preparedness and management. The recent example of Mongolia dramatically proves that gap: no pastoral risk management policy or strategy enacted by government existed, and the disaster situation - as will be shown later in this paper - was tackled only in an ad hoc manner.

6. This paper synthetizes the results of field work conducted since 1995 in Mongolia, China and Kyrgyzstan on institutional approaches to pastoral risk management in Central Asia, within the wider context of poverty alleviation and food insecurity ⁴. Taking the recent experience of Mongolia as an example, it proposes components and institutional mechanisms for mitigating risks and better preparedness in disaster situations, which are regularly faced by pastoral households and economies.

2. Approaching pastoral risk management in Central Asia

7. The following chapters outline the most important aspects, facets and dimensions for approaching systematically the issue of pastoral risk and risk management. They emphasize the different perspectives required to gain a comprehensive understanding of the complexity of risk management.

2.1 Defining pastoral risk

8. Risk is the subject of a large and rapidly growing literature. Risk in general terms refers to the occurrence of uncertain events or processes which jeopardize people's well being and can be expected to happen at any time. The classic conception distinguishes between risk - defined as the probability that a particular adverse event will occur during a certain period of time, or as the result of a particular challenge - and damage - defined as a numerical measure of the expected harm or loss associated with an adverse event. The literature further distinguishes between objective risk and damage -  past and likely future occurrence of risk and damage measured by experts - and perceived risk and damage - the way in which people anticipate future events or view past ones. Once heavy damage is caused as a result of a sudden unforeseen or foreseeable event (often referred to as `shock'), a disaster situation arises.

2.2 A typology of major risks faced by pastoralists

9. Risks threatening pastoral production systems have different origins. Risk can be categorized in various ways. A functional typology ⁵ gives four main categories whereas the first category is by far the most threatening.

10. The different types of risks are inter-related and often occur at the same time or as consequence of each other: for example, snow disaster may weaken surviving animals, making them particularly susceptible to heavy spring rains, associated with heavy storms when the animals are sheared or without wool protection. Different risks can also occur in combinations: the risk of economic losses caused by drought, for example, is likely to be higher in a remote area far away from big livestock markets, where spontaneous selling of animals (as an economic coping mechanism) is more difficult. Different ecological or geographic zones may have characteristic risk features: for example, areas close to markets are much less likely to suffer from market risk than remote hinterlands, but they may be more susceptible to animal theft.

2.2.1 The levels of risks

11. Further classification of risks by levels is critical since there are no specific response systems to deal with single types of risks. Risks can occur at the micro, meso or macro level. They may affect individual households only (referred to as idiosyncratic or individual risks, at micro level), or all households in a given area (covariate risks, at meso level). Snow disasters or market collapse are covariate risks, affecting everyone in the area. Animal theft, predation or illness are generally risks faced by individuals.

12. The ability of herding households to help neighbours or kin to prepare against and/or recover from risk depends in part on the type of risk: in the case of individual risk, other households may be able to help, but in the case of covariate risk, everyone is in the same situation, and cannot. In terms of managing risk, the State (often supported externally) has in general a comparative advantage and is highly demanded to assist in and/or cover covariate risks, whereas most individual risks may be better handled by local communities, groups or individuals themselves. It is generally accepted in advanced economies that the coverage of major covariate risks is a public service that should be provided by government, while idiosyncratic risk is generally covered by individuals and the private sector.

2.2.2 Cultural, economic and ecological construction of risk

13. Risk is not just a series of random events striking unwary populations. Risk is culturally constructed. People in risk-prone zones, such as pastoral areas, experience risk and anticipate it in various ways depending on prior experiences, how they expect others to behave (especially their own kinship or residential community and public authorities), and on their resources and objectives. Different stakeholders in rural/pastoral development perceive risks in different ways. In addition, their responses to a variety of ecological conditions involving the same type of risk will lead to different outcomes.

14. Herders' attitudes to risk and their perception of the probability and magnitude of future adverse events, are partly shaped by their previous experience of risk and its outcomes, e.g. their expectations of the respective roles of government versus the local community in helping risk victims. Based on experience, these attitudes are continuously modified throughout their life. Different categories of herders perceive risks in different ways. Their perceptions are determined by a range of variables, including:

• wealth: rich herders perceive and can react to risks differently than poor herders;
• labour: herders with plentiful labour perceive and can react to risks differently that labour-scarce households; household life cycle stage is especially important in determining what any particular risk means for a particular household;
• species herded: herders of drought or snow-resistant species can afford to face the risk of drought or extreme snowfall in a different way than herders of more vulnerable species;
• ecological conditions and opportunities: herders with access to emergency reserves, if they exist, do not need to take the burden of long emergency moves;
• social security/networks/social capital: herders who have access to social services or insurance, belong to strong networks, large multi-generational families or cooperatives with good leadership, perceive and can react differently to risks than isolated households or households without such social capital;
• links to urban areas and to public administration: a particularly important category of social capital for Central Asian pastoralists in risk perception and management is constituted by kinship, patronage or economic links to the urban economy or to public administration, since this provides a strong safety net in the case of hardships.

15. The studies carried out in 1998/99 among herders and officials in Qinghai, China, provide a good example of how risk is perceived by different stakeholders. Though all stakeholders confirmed the overwhelming threat resulting from heavy snowstorms, the perceptions of the different stakeholder groups regarding the importance of different types of risks to pastoral livelihoods varied significantly, as shown in the table below. Snow storms were ranked as the greatest risk, with animal thefts next, followed by wolf predation; animal disease was ranked least important, as it was considered to be under control. County and township officials who had no animals of their own ranked animal disease as the second most important risk, and animal theft least important. However, officials who had animals of their own ranked the risks in the same order as the herders. These different perceptions have important implications for policy making.

16. Risk management strategies to tackle those risks perceived as the most threatening are most likely to be supported and adopted by local stakeholders, and thus provide good entry points for interventions and the mobilization of self-help initiatives.

3. Case Sample Mongolia - the risk of "zud"

17. "Zud" is ranked as the greatest risk by most herders in Central Asia. Definitions of zud vary. In Mongolia, the meteorological service defines zud as a snow cover of more than 25 cm, a sudden, prolonged snowstorm, 2-3 cm of frozen snow cover, or extreme prolonged spells of cold. As can be expected, herders have a rich vocabulary to describe and analyze zud. The main criteria are not so much the snow cover itself, as its consequences, especially in terms of animal mortality. A serious zud is one during which heavy snow freezes to a permanent ice cover which prevents the animals from grazing. Zuds are also defined by their geographic coverage (a single valley or a whole ecological or administrative zone), and by their duration (heavy zuds exceed 20 days). Mongolian herders also classify several types of zud: black zud, which does not involve snow at all, but is the freezing of surface water, which makes it inaccessible to the animals; white zud, the classic snow disaster; storm zud, when a severe snow storm lasts for several days, forming deep drifts and driving the animals down-wind where they go astray; and freezing zud, when the temperature is so low that people and animals cannot maintain their body temperature, and die of cold.

3.1 The nature and dimension of the 1999/2000 zud

18. From a meteorological point of view the 1999/2000 "zud" had its roots in the previous year when the country faced a severe drought. Exhausted animals had to go through a dry summer in 1999 without the possibility of developing the necessary fat to withstand the cold winter. The dry summer also prevented haymaking. First heavy snowfalls in some areas were recorded in the first days of September-October, which was much earlier than in normal years. The early, heavy snowfall was followed by exceptional warm days in November and December 1999. During these warm days, the snow cover melted and hardened creating an ice cover on pastures which hindered animal grazing. At the same time, the snow cover increased regularly. The extremely cold days in January-February 2000 worsened the situation created by the deep snow, which prevented animals from grazing almost completely.

19. Official statistics suggest that the natural disasters that occurred in Mongolia in 1999/2000 covered 13 provinces out of 21 and 157 districts of 310, which is about 70 per cent of Mongolia's land area. In general, 452,000 people from 80,900 herding households herding 7 million head of livestock were affected to some extent by the 1999/2000 winter-spring zud. One of the specific patterns of these disasters was that the zud affected the most vulnerable areas: the Gobi, the steppe and the high, treeless mountains, where the availability of pasture and forage is limited by such ecological factors as low rainfall, low grass growth and high fluctuations in weather conditions.

20. At the national level, the total loss in terms of monetary value amounted to 91.7 billion Tugriks (the official exchange rate, May 2001: 1 US$ = 1,091 Tugrik). As of 1 June 2000, total livestock loss amounted to 2.4 million heads, i.e. almost 10 per cent of the total national herd. Over 2,370 herding households lost all their livestock and the number of households with less than 10 animals increased by over 10,000.

21. The disasters clearly demonstrated that the zud is the worst natural disaster in Mongolia, and is almost of equal importance to herding societies as droughts in other pastoral areas.

3.2 Lessons learned from the 1999/2000 "zud"

22. Retrospective analysis of 1999/2000 provides the following picture of several main causes that in combination caused high vulnerability and hardship for Mongolian herders.

A. Constraints and shortcomings regarding risk preparedness

23. Unclear responsibilities about risk management: After transition, the Mongolian authorities assumed that the privatization of livestock and their distribution into private hands would lead to better management and increased risk-mitigation practices at decentralized levels. In conjunction with the privatization of livestock the State decided to shift major risk management responsibilities previously held by the State and cooperatives to individual livestock keepers. However, many herders still believe that the government should continue to provide risk management related services and act in the same way as during the centrally planned economy. In that sense there has been a misjudgement of the responsibilities of two major actors in the management of natural and related risks.

24. Unclear information strategy: Herders report that during the summer 1999, the government had announced its readiness and willingness to support the herding population in the case of emergencies. Herders assumed therefore that the government would provide help as the negdels (state co-operatives) did in the past.

25. Lack of decentralized fodder reserves: No permanent fodder reserves were available at decentralized level to provide supplementary fodder during emergency situations ⁶. High input prices and maintenance costs are the main reasons why the previously existing reserves were abandoned.

26. Exceedingly centralized emergency reserves: Some unofficial sources suggest that in 1999/2000 over 60 per cent of all assistance and aid came from reserves in the central part of the country. This type of localization of important resources and coordination of their distribution added unnecessary costs to the initial prices which were unaffordable to many zud-affected herders. A quick survey ⁷ in Uvs province identified the price of a 25 kg hay bale transported from the central part (Selenge and Bulgan provinces) as 3,800-4,200 Tugriks (the official exchange rate, May 2001: 1 US$ = 1,091 Tugrik), while the same bale arriving from Bayanulgii province cost only 2,400 Tugriks.

27. Lack of local preparedness: Herders lacked basic items during the "zud", such as hay, minerals and sufficient amounts of manufactured and hand-made feed. This led to huge demands for external assistance over short periods of time, which it was impossible to provide.

28. Lack of a clear government plans to respond to emergencies: There were no clear and consistent government regulations indicating under what conditions the State would provide assistance and/or coordinate assistance received from donors and NGOs.

29. Lack of common definitions of natural disasters: A common tendency was observed that once snowfall had occurred, local administrations were urging to declare a "zud" in their territories and immediately requested higher authorities to give them access to the reserves and/or aid packages before the others. Consequently, local governments applied their own judgement - often varying from each other - in declaring the severity of the risk.

30. Unclear rules and procedures to declare emergency situations: The above shortcoming made it very difficult to differentiate between the severity of disaster situations and led to another set of problems. The only means to verify a disaster applied by the government was to field their staff to see and understand what had happened, and to decide what to do. Local representatives were not allowed to declare an emergency until the situation was verified by province, or sometimes district administration. This procedure tied the hands of the local authorities to act quickly enough to prevent desperate situations. People were urged to inform the authorities before taking action to mobilize their own reserves and resources.

31. Lack of early warning systems: No unified and integrated early warning systems to predict natural disasters were in place. Local meteorological stations played only a minor role in information dissemination. Most of the information flowed from district to province administration and from there to the centre. No further verifications or assessments were carried out by central government institutions.

32. Insufficient use of long-term weather forecasts: The reliability of long-term weather forecasting is still weak and was not properly used for disaster planning and risk assessment purposes.

33. Wrong timing of surveys and field assessment: Some international agencies conducted surveys in early winter to assess if there were food shortages or starvation among the rural population, including those in zud-affected areas. It is known that early winter is the only time when all Mongolian households, if not extremely poor or socially disadvantaged, enjoy peak consumption of winter food.

B. Constraints and shortcomings observed during emergency relief operations

34. Lack of coordinated system for nation-wide disaster relief planning: The information received from local administrations was often inaccurate and not up-to-date. This caused serious difficulties in operational planning and the delivery of aid to the zud-affected areas.

35. Delay in the provision of aid: This was one of the main issues of the extensive discussions that emerged among all the people involved: from herders to government officials. For example, the main bulk of assistance arrived in February 2000 when the disasters had already caused all the damage: high animal mortality, abortions, and death from freezing, as well as human starvation, including shortages of medicines, clothes and food.

36. Poor local coordination of aid distribution: Many herders and people who were in need of external assistance criticise the distribution of all kinds of goods as exceptionally poor. This is again connected with the lack of local preparedness and the high demands placed on such poor, local capacity. In some cases, aid shipments were either delayed or insufficient to meet and satisfy the requirements. Local governments had neither clearly identified nor planned the requirements. Instead there was a rush for everything without any differentiation between the various goods: food, fodder, medicines and other. On the other hand, the coordinating agencies distributed what they had at hand almost without planning for the next steps to be taken.

37. Very few measures were directed to long-term needs and sustainable post-shock recovery of herding households badly affected by the zud.

38. Ignoring of refugee herding households: Herders normally move from their area to another (communities, districts and provinces) to escape disasters. Traditionally, high-standing bodies of province and district administration negotiate the sharing of grazing resources with (disaster induced) refugee migrants. However, in the zud situation 1999/2000, such negotiations, if materialized at all, hardly covered any other needs which migrant refugees have - such as ensuring equal access to aid assistance and other inputs and the provision of social, commercial, health and education services.

39. Delivery of inappropriate goods: The delivery of items in large quantities as pursued by local distribution agents made them often close their eyes to whether the items were actually needed. In addition, distribution was not always correct and some people benefited more than others. For some households the aid obtained was, reportedly, barely enough to support them for less than a week.

40. Lacking continuity of assistance: The small quantities of fodder assistance failed to provide any real support to herders. Many herders and people had to wait for the next instalment for weeks or longer, while starvation was becoming serious and worse.

41. Different eligibility and inconsistent aid eligibility criteria: Eligibility to benefit from assistance and aid was one of the basic issues that was continuously and controversially discussed during the emergency. This was the case of national and international NGOs and other voluntary groups that used a variety of social status-based approaches and different schemes for the selection of beneficiary households: poverty, input development, human support, improved health and temporary zud-relief. These sometimes led to overlaps and ignorance, as well.

C. Government responses initiated after the 1999/2000 zud

Lessons learnt related to poor preparedness:

42. The central and provincial governments issued orders to local governments at lower level to implement measures to enhance hay-making and feed preparation for the winter, animal fattening and the repair of winter and spring shelters to improve the disaster preparedness of individual herding households. However, in 2000 only over 700,000 tons of natural hay and other types of animal feed were prepared, i.e. at same level of 1999. No notable increase in haymaking was achieved, due to the 2000 summer drought. There was some increase though in commercial haymaking.

43. Measures were also taken to supply more vaccines, and dipping and deworming drugs for disease prevention purposes.

44. The State provided a budget to cover the costs for herders' children in district and village level boarding schools. This exempted many herders from compulsory payments previously imposed and allowed them to pay more attention to herding.

45. Several activities to evaluate the 1999/2000 zud were undertaken with the support of UNDP to refine relevant lessons for improving the anti-disaster capacities of central and rural structures.

46. Further, the government issued a decree (# 190 of 2000) on the assessment and declaration of zud. The decree sets out the basic rules and criteria for evaluating the severity of disasters. In addition, the participation of central and local bodies was also proposed by the decree.

47. Nation-wide appeals were launched to mobilize local organizations and people to assist zud-affected societies by providing additional inputs.

48. The government has enacted a national programme for the provision of support to improve the protection of livestock from drought and zud (# 31, 2001). This programme aims at developing better nation-wide networks of risk preparedness through improved feed and water supply, manpower development, encouraging herders and the private sector to increase their efforts for fodder production and improve livestock productivity and processing, as well as effective cooperation.

49. A new draft land law has been submitted to the parliament which places more emphasis on the rational use of natural grazing resources, including natural pastures and crop land to increase the forage supply for the livestock sector. The law contains some new elements, such as leases or long-term titling of winter and spring pastures to individual and groups of herders and auctioning crop land to different entrepreneurs to improve land conservation and use.

Lessons related to delayed government actions:

50. In 2000/2001, the government attempted, as local people confirm, to react quicker than in the previous year to the upcoming emergencies, but the 2000/2001 zud was even more extensive (19 provinces) than the previous year (13-15 provinces) thus again hampering government interventions.

51. Two different coordination channels were set up in order to improve the coordination and distribution of external assistance: (i) the State Emergency Standing Committee (SESC) and (ii) the Mongolian Red Cross Society.

52. In order to improve administrative procedures some progress has been made in reviewing the responsibilities of central ministries and in promoting positions of relevant ministries in SESC. This enabled decision-making and relief implementation procedures to be more efficient and integrated.

53. Finally, the government has asked some big international agencies, such as IFAD, WB, ADB and IMF, to increase the amount of loans and grants to assist Mongolia in making its livestock production more sustainable and productive, as well as less vulnerable to natural disasters.

54. However, a comprehensive risk management framework and concrete plans of action at provincial level are still missing.

4. A framework for improved risk management planning

55. The framework presented below is the intermediate result of the field analysis conducted by SDAR between 1995 and 2001, reflecting as well the lessons learned from the recent disaster in Mongolia. ⁸ It serves as an entry point for the Technical Cooperation Project "Pastoral Risk Management Strategy", to be launched in May 2001 in Mongolia, which will be further fine-tuned and developed into a concrete plan of action in selected provinces.

4.1 Phases of risk management

56. Pastoral risk management involves three phases, each with different characteristics, actors and activities:

Phase A: Risk preparedness

57. The preparations for coping with risks in the pastoral context can be distinguished as (a) long-term strategies for risk reduction and avoidance, which prepare the herding economy and authorities on an on-going basis for stress periods, such as winter, and for unexpected shocks; and (b) annual (short-term) planning for risk, which includes activities that prepare the herding households for regular (foreseeable) stress periods such as winter.

Phase B: Responding to risk (when disaster occurs)

58. The need to respond to risk arises once a risk turns into an acute threat and hardship. This phase requires immediate action. It is more likely that successful action can be taken if good planning and preparatory steps have been taken by all institutions involved and immediate and integrated actions are taken.

Phase C: Recovering from risk (when damage has occurred)

59. Recovery from risk is a key process, since until the household economy has recovered it remains especially vulnerable to new risks. The level of damage/loss and the support needed at this stage are determined, on the one hand, by the gravity of the calamity and, on the other hand, by the quality, quantity and co-ordination of actions during the previous stages.

4.2 Key aspects of improved risk management

60. In the following section, the main components and activities of a comprehensive risk management framework for remote pastoral areas are organized according to the three above mentioned phases of risk management. For each phase a set of activities is presented against which the current status quo in a particular study area can be assessed. Each aspect is mentioned only once: at the stage where it is considered to be of utmost importance. This, however, does not mean that some of the aspects might not be important also in other stages. Highest emphasis and attention is given to risk preparedness.

61. It is important to note here, that the various aspects presented for each phase are intended only to provide a skeleton, for the analysis and planning of any location-specific pastoral risk management strategy, in the context of which further location- and situation-specific adaptations and/or modifications may be needed.

Phase A: Risk preparedness

A.1 Long-term risk reduction and risk avoidance strategies

• Institutional development and herder organizations: The degree to which formal and informal herder organizations operate in Central Asian countries varies. In many CA countries, an effort was made during the socialist period to strengthen authority relations in a different form than during the pre-existing feudal regimes, at the expense of relations of trust and freely-established norms. A deliberate attempt was made to reduce or destroy kinship relations and other types of customary social relationships as the basis of economic activity. Today, an organizational vacuum can often be observed in the countryside hampering co-ordinated risk and environmental management by herders. Often herders cooperate in daily activities, including range and risk management, through informal camp and neighbourhood groupings, however, such informal groupings are usually not recognized as economic entities and as a result have none of the legal protection or tax advantages which more formal groups may enjoy. There is a need to encourage further evolution of formal as well as informal herder organizational structures (for example through grass-roots herder associations) which facilitate local-level collective action in pastoral economic development, including risk management and environmental conservation. Livestock marketing is one area where small-scale collective action may provide a partial solution to some market failure risks.
  An important part of the reconstruction of liberalized herding economies in the region will be the way new types of authority relations, trusts and socially accepted norms are recreated. It is possible that this will happen partly through the creation of hybrid institutions - part customary, part formal - such as extended camps operating as micro-enterprises for commercial hay production or milk processing. It is likely that the success or failure of this reconstruction of social capital will largely determine the success of poverty-alleviation strategies and perhaps of the pastoral economy itself. Social capital is a key determinant of the ability of pastoral groupings to survive extreme risk, by facilitating collective local action, in particular since it has been recognized that social capital is a key component of any group to pursue effective economic development.
• Appropriate financial institutions. There is an urgent need for savings, credit and insurance institutions appropriate for the remote countryside. Mongolia is an example of a country that used to have a successful economic system of universal insurance against animal loss as a result of disaster, although since economic liberalization only around 10 per cent of animals have been insured. Animal insurance needs to be established in a form suitable for market conditions. Key tasks include the development of appropriate savings mechanisms for herders, especially in the context of the need to market more livestock each autumn so that fewer need to be carried through the winter-spring pasture shortage period, and appropriate credit mechanisms to provide working capital to herders (e.g. for marketing) and facilitate restocking of herds that have fallen below a viable size or for livestock fattening or other livelihood diversification strategies.
• Risk-avoiding herd management techniques: Pastoral herding technology comprise several risk-avoiding practices, including breed selection for survival, mixed species herding, and individual household hay-making and fodder reserves. These are often specific to particular eco-systems. Many of these techniques involve trade-offs with individual short-term animal productivity. There is a need to understand better these techniques, perhaps through the experience of successful herders, and to provide training in such techniques for less experienced herders.
• Incentive systems to reduce animal stocking rates in areas of over-grazing: Over-grazing of pastures causing severe range deterioration and consequently a reduced fodder base is a severe problem for instance in NW China. Over-grazing and desertification have been reported on the increase in Mongolia over the last five years. Policies and approaches to promote a reduction in the number of animals in large herds and rotational grazing patterns can be powerful tools for risk mitigation in the long term.
• Participatory technology development: Existing local knowledge on risk management and pastoral livestock production can be ideally complemented and matched with farm research and extension approaches. Innovative technologies for environmental conservation and management, range monitoring and production increase, can significantly contribute to a long-term reduction of risk exposure.
• Better co-ordination of key risk actors at provincial and lower administrative levels: Key risk actors include the herders themselves, local officials at provincial and lower administrative levels, the technical services and civil defence. These actors need to be linked by a specific plan for large-scale emergencies, with an appropriate institutional backing and enough resources for an initial rapid reaction that can be implemented at short notice. There is a need for an operational definition of natural disasters, agreed by herders and officials, to trigger action, and a standing committee with specialized working groups on different risks, with herder representation.
• Long-term weather forecasting and improved broadcasting: This is already undertaken by the different Institutes of Meteorology and Hydrology in most Central Asian countries, and transmitted to the provinces and districts. Further work needs to be done, however, to ensure that the information is made available to herders and their organizations in a timely, accessible and understandable manner.
• Market development: The development of a more efficient marketing system for livestock and livestock products, for example by providing assistance to marketing co-operatives or the organization of livestock auctions at province level, would reduce market risks to herders, and would assist in responding to other types of risks. There is a role for teaching herders about the functioning of the market economy, and for improved market information.
• Creation of emergency grazing reserves at local/district level: Herders make their own decisions about pasture use, but there is an important role for local authorities in pasture allocation, settling disputes over pasture use, and in creating grazing reserves for exceptional situations. Such reserves could be situated on the border between different administrative areas, and could thus be used by different groups. Some such reserves already exist, or have survived since the collective period. However, comprehensive plans need to be elaborated, in collaboration with local authorities and user groups.
• Strategic fodder storage reserves at provincial or district level: Strategic fodder emergency stores (please refer also to footnote 6) at decentralized levels are central to any pastoral risk management strategy in high risk environments. However, they need to be maintained, regularly replenished and appropriately equipped to be ready to distribute fodder to herders in emergency situations. Herders themselves could contribute in kind (or cash) on an annual basis to the production of emergency fodder for storage. It must be ensured that the stores remain genuine reserves to be used in exceptional situations and do not become a source of subsidized fodder every year.
• Pastoral mobility: Any pastoral system in the Central Asian Region requires seasonal (e.g. transhumance) and/or periodic moves of animals between different grazing areas. Policies which do not leave enough space for herd mobility reduce the chances of successful risk management and coping. There is a particular need for herd mobility during snow storms in order to move animals away from the threatened areas. Arrangements for temporary use of emergency target areas (e.g. through leases), which may require coordination at district or provincial level, need to be prepared and up-dated on a regular basis. Any long-term risk management strategy should incorporate such options.
• Herders' risk planning capacities: Training should be provided to herders in risk planning and management, including the preparation of animals for the winter. Economic liberalization in Mongolia and Kyrgyzstan has created a situation where there are many inexperienced herders, ignorant of basic herd management, and no framework such as that provided by the former collectives, through which herders could learn and receive supervision. In Mongolia, the situation is made worse by the substantial urban-rural migration which took place after 1990, when the urban economy contracted and inexperienced people sought a new livelihood in herding.
• Physical Infrastructure: In the few areas where they do not already exist, there is an urgent need to construct winter livestock shelters (e.g. through a cost-sharing arrangement between herders and local government, as already practised through the four counter measures in China. It should be ensured however, that such shelters are affordable to all herders, even the poorest). More generally, there is a need to review available infrastructure (including transport and communications) taking into account the requirements in an emergency situation.

A.2 Annual planning for risk preparedness

• Winter preparation by herders: Winter preparation is an important activity undertaken by herders, and includes proper animal fattening during the short summer season, accustoming the animals progressively to the cold, making decisions on the number of animals that can be carried through the winter and early spring, and which animals should be sold or slaughtered in the autumn, preparation and stocking of hay and other animal feed, and reinforcement of winter shelters and other enclosures. Herders need assistance with this, and many new or young herders need training by more experienced herders.
• Annual winter/drought fodder storage by households: Improved hay-making techniques and local winter fodder storage realized by individual households or groups of households need to be established/increased. Hay storage at the household level is critical for pastoralists to feed their animals through periods of frost and heavy snow. The distribution of fodder from central stores to households is often delayed or made impossible during periods of heavy snow or frost, even when such stores do exist at the community level. The four counter measures implemented in China, which establish special hayfields for individual households, is an improvement in the right direction. However, the amounts of hay currently available at household level are often insufficient to withstand extended periods of snow - the most critical periods. Transportation from the fields at the time of harvest and the hay storage facilities at the household level are other constraining factors. Inter-household collaborative mechanisms to support each other in hay-making and transport, and/or the strategic demarcation of additional communal hay-making areas at community level are options that may contribute to further increasing the availability of hay for herding households during winter time.
• Seasonal pasture allocation planning: This is already undertaken by county authorities, but needs to be further co-ordinated between the herders themselves, especially through further agreements between herders' camp and neighbourhood groups regarding seasonal movements between pastures and seasonal pasture reservation.
• Early warning: There needs to be an institutional base, ideally at provincial level, to monitor potential emergencies and provide early warning to herders and provincial authorities. More accessible weather forecasting in the five to seven day range is important, but the early warning system should also include other sources of information, especially information from experienced herders. Getting warnings to herders in time for them to react is critical (use of the existing radio system is a possibility, but often radios in the countryside are not working).

Phase B: Responding to risk

• Co-ordinated emergency management: The central point of administrative emergency management is the provincial government, in collaboration with authorities at county level. In major emergencies, such as serious snow disasters, the army can assist in co-ordinating responses by all government actors at provincial level. But in the administrative structures emerging from liberalization, a better articulation of State and private actors is needed (including non-governmental organizations and formal and informal herders' groupings), especially to share information, decide on priorities and make decisions about specific responses. These institutional arrangements must be created and tested in advance.
• Facilitate herder mobility and access to emergency fodder reserves/grazing: During snow disasters and major droughts, the ability of herders to move their animals away from the affected area is even more important than at other times, but is constrained. During snow disasters, heavy machinery may be needed (usually from the army) to clear tracks. New, more easily accessible pasture areas must be selected, transportation may have to be arranged, and water and fodder provided at waypoints on the route. Special emergency services (veterinary, human health) may be needed in places where they are not normally available. Emergency fodder reserves need to be opened and delivery of emergency fodder organized.
• Labour mobilization: Some crises are more labour-intensive than other. For example, snow disasters create additional labour at household and communal level for clearing the snow, rescuing isolated animals, and moving animals. Often there is an under-employed population at county level and in county centres, which could be mobilized by local authorities to assist with these tasks during crises.
• Emergency food distribution: There is some experience in food distribution during crises in China, but such situations could become more frequent in the future and need to be adapted to the pastoral environment.

Phase C: Recovering from risk

• Appropriate credit for restocking. Some interesting experience has already been acquired in pastoral areas in China. Such experiences provide a basis for developing appropriate schemes targeted to the particular conditions of other ecological zones.
• Alternative livelihood strategies for impoverished herders: Not all poor herders will be able to return to full-time pastoralism immediately after a crisis. Interesting work is being done to develop vegetable production by former herders either as a new branch of the economy in remote pastoral areas or as a supplementary temporary source of income. This and other possible ways of diversifying rural livelihoods need to be further explored.
• Assistance where necessary with short-term consumption needs: Restocking and the development of alternative livelihood strategies do not solve the immediate problem of households impoverished by a crisis and the need to feed their member's daily. Such a situation may last for months after a crisis. There may be a role for specific public work programme, such as food-for-work or food-for-cash, as part of a wider risk management strategy.

5. Final remarks and conclusions

• Most Central Asian countries ⁹ pay too little attention to risk reduction and prevention and rely too much on interventions after disaster strikes. While coping mechanisms will remain vital for dealing with unforeseen shocks, the balance however needs to shift more from coping policies to better preparedness and mitigation of risks.
• A worrying feature observed in Central Asia since transition and increased privatization is that governments (mainly due to resource constraints) tend to decrease their assistance and services to herders in preparing for and managing risks. This tendency has to be reviewed. In spite of the important role herders themselves can play in risk management under free market conditions, governments need to continue their key role of coordination and setting the policy and legal frameworks required for herders and herding communities to get better prepared for risks. As the case of Mongolia shows, it is of utmost importance that governments are constantly prepared for possible upcoming disaster situations, as well.
• Research in Central Asia stresses the overwhelming importance of the institutional and organizational dimension to improve risk management. Capacity building, transparent and partnership-based co-ordination among actors, as well as a favourable legal and policy environment, were identified as the basic determinants of sustainable pastoral development and improved risk management. The development, promotion and introduction of new technologies are also important, but with lower priority.
• Pastoral risk has to be monitored, managed and counteracted at several levels, with a well co-ordinated and integrated approach. New modalities for sharing planning and implementation responsibilities in joint risk management need to be defined at national, provincial and district levels through extensive stakeholder negotiations, including representatives from the local to the highest institutional levels.
• After transition to a market economy and privatization of assets, herders and herding communities are able to prepare for and handle idiosyncratic (individual) risks. Coping with covariate risks, however, continues to require government coordination and support.
• The main actors in day-to-day pastoral risk management in all the countries concerned are currently the herders themselves. However, this is rarely recognized and rarely acted upon. Customary herder institutions - such as mechanisms for immediate assistance from fellow camp members or neighbours in assuring food security - constitute the main protection for herders who suffer individual livelihood losses, but as long as such institutions are not officially recognized and adequately supported, they are easily overwhelmed during major disasters covering large areas and many households.
• Risk and poverty are closely inter-related ¹⁰. Perhaps the most common route to extreme poverty in Central Asian pastoral economies is the case of labour-deficient households that are not very skilled in herd management, own small herds close to viability, and are hit by a major climatic disaster. Once present, poverty increases scale of risks and damage that may occur as result. Poor households are more vulnerable to risk. They have less capacities and means to prepare for and counteract risk. Due to lack of opportunities, poor households are often forced to use the most marginal and therefore most fragile and high-risk eco-niches, where they possibly cause severe ecological damage. Poverty alleviation strategies for pastoral areas in Central Asia need to consider risk management as an integral part.
• Risk also jeopardizes major investments by government and donors. For example, IFAD and the Mongolian Government are investing US$ 5 million in Arkhangay and Huvsgul provinces for poverty alleviation, especially by restocking the herds of poor herders; other donors have similar restocking programmes in other provinces. A single major snow disaster could not only threaten the success of these initiatives, but could wipe out investments already made.
• Traditionally, mobility was the main strategy to avoid pastoral risks on the Central Asian steppes, at the same time successfully preventing over-grazing and environmental degradation. The mobility of pastoralists has in many places decreased substantially during the last decades for various economic as well as political reasons. This bears major environmental risks. The mobility of their livestock, but not necessarily of the whole herding households, should remain critical in the risk management strategies of Central Asian herders.
• The increasing numbers of herds and animals are often considered a recipe for mitigating production risks. However, this is a misconception. Rather than leading to long-term production increases, large herd increases often lead to over-grazing and trampling, and ultimately to a decreasing natural resource/fodder base and reduced resilience of the natural environment against the harsh climatic conditions. A reduced fodder base, however, increases long-term production risks by causing animal losses during long winters.
• The traditional analysis of risk assumes that all risk is negative, and that risk produces only losers. However, risks offer opportunities as well as threats and there are substantial rewards to certain forms of risk-taking. Thus, a herder in a cool mountain area who herds cattle instead of yaks or yak-cattle cross-breeds is likely to fare much better in producing for the market than a herder who keeps only yaks, but at the cost of substantially increased risk of losing animals in cases of exceptional or severe snowfall. Herders, like other people, are continuously trading off risk and potential loss against potential gain. They base their calculations on their perception of risk, the economic and social frameworks which surround risk-taking, the potential rewards and damages of risk-taking, and the safety nets available in case the worst happens. These calculations change as the key elements in the calculation alter the potential rewards or the potential damage ¹¹. Better risk preparedness will increase the perception among herders to see risks not only as threads.

62. The Kyrgyzstan situation (where items are scored out of two) shows a substantial shift of risk responsibility from the government to herders following economic liberalisation. This is especially true for snow disasters, market risk, illness and accident. In China (where the scores are out of five), the changes are less dramatic, but substantial. There have been major shifts from government to individuals in the case of economic risks (animal theft and market risk), illness and accident, but the shifts in the case of most environmental risks (snow disaster and animal disease), although they have taken place, are less dramatic; the Chinese state has retained substantial responsibilities in these areas.

Annex 1: Changing risk responsibility in Kyrgyzstan and China following liberalisation

References

Adams, John, 1995, Risk. London: University College London Press.

Beck, Ulrich, 1992, Risk Society. London: Sage (first published in 1986 as Risikogesellschaft).

Behnke, R.H., I. Scoones, C. Kerven, eds, 1993, Range Ecology at Disequilibrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas. London: ODI, IIED, Commonwealth Secretariat; I. Scoones, ed., 1996, Living with Uncertainty.

Bruno, M., Nurbek Omuraliev, 2001; Strengthening pastoral institutions in Kyrgyzstan for risk management and poverty alleviation; SDAR, Case study report (unpublished).

Ellis, J., Chuluun, T., 1993, Cross-country survey of climate, ecology and land use among Mongolian pastoralists. Unpublished note, IDS: PALD project.

FAO, 1997, Rural development in Pastoral Areas, Arkhangay Province, Mongolia. TCP/MON/4553. Terminal Statement Prepared for the Government of Mongolia. Rome: FAO.

FAO, 1996, Trends in Pastoral Development in Central Asia (Buryat, Inner Mongolia, Kalmyk, Kazakstan, Kyrgyzstan, Mongolia). Rome: FAO.

Hazell, Peter 1998, "Public policy and drought management in agro-pastoral systems", paper presented a the international symposium on Property Rights, Risk and Livestock Development, Feldafing, Germany, September 1998.

Templer, G., J. Swift, P. Payne, 1993, "The changing significance of risk in the Mongolian pastoral economy." Nomadic Peoples 33:105-122.

Liu, Y., Wang G, H. 1998: Risk Management and Poverty Alleviation in Pastoral Areas of Qinghai Province China, FAO/CIAD field study report.

Liu, Y., Baas, S., et al. 2000 Strengthening Pastoral Institutions in North-West China Pastoral Area to access improved Extension Services for Risk Management and Poverty Alleviation, FAO/CIAD field study report 1999.

Potkanski, T., S. Szynkiewicz, 1993, The Social Context of Liberalisation in the Mongolian Pastoral Economy. Brighton, Institute of development Studies, Policy Alternatives for Livestock Development Research Report 4. and Cooper, L., 1995, Wealth and Poverty in the Mongolian Pastoral Economy. Brighton, Institute of Development Studies. Policy Alternatives for Livestock Development, Research Report.

Swift , J., and Baas, S., 1999, Pastoral institutions and approaches to risk management and poverty alleviation in Central Asian countries in transition; FAO-SD website.

Swallow, B., 1994 The role of mobility within the risk management strategies of pastoralists and agro-pastoralists, IIED, Gatekeeper Series no. 47.

^* Prepared by Stephan Baas, Rural Institutions and Participation Service, FAO; Erdenebatar Batjargal, Agricultural State University, Ulanbator, Mongolia; and Jeremy Swift, Institute of Development Studies (IDS), Sussex, UK.

¹ Other disasters on record in Mongolia were the winter of 1944/45, when a zud killed 8.1 million animals; in 1993, 1.6 million animals were lost to zud. The same phenomenon of snow disasters is common also in other Central Asian countries. In Northwest China the high risks for the environment and production are well documented by the particularly cold winter of 1997/98 which witnessed, for instance, in Dari county, Quinghai, temperatures below -35oC for more than a month. Severe dislocations and heavy animal losses were recorded: 1,400 people, and 330,000 animals, had to make emergency moves away from the worst affected areas; over 900 animals froze to death, and nearly 12,000 died of starvation.

² During collectivization of the herding economies of the CA region from the 1920s and 1930s in the republics of the former Soviet Union, and from the 1950s in Mongolia and China, substantial safety nets were institutionalized through herding collectives which provided: employment guarantee, provision of basic non-pastoral foodstuffs at controlled prices; restocking with collectively-owned animals in case of loss; collective livestock insurance against accidental loss; collective provision of free services, such as human and veterinary health and education, including emergency fodder provision; a comprehensive system of social security, including pensions and other social security payments. During that period local community driven risk coping mechanisms lost importance and weakened.

³ The tables in Annex 1 provide the results on Changing risk responsibility in China and Kyrgyzstan following liberalization as assessed during field studies in 1998.

⁴ This work was conducted by the Rural Institutions and Participation Service (SDAR) of FAO in close collaboration with the Institute of Development Studies, Sussex UK, the Research and Training Institute for Animal Husbandry (RIAH), Ulanbator, Mongolia, the Center for Integrated Agricultural Development (CIAD), China Agricultural University Beijing, China, the Centre for Social Research, (CSR), Academy of Science of the Kyrgyz Republic, and the Social and Economic Research Associates, (SERA), London.

⁵ Alternative ways of categorizing risks become useful once solutions and mechanisms for coordinated risk planning and management are sought. An important criterion at that stage may be the geographical scope of measures, operational entities, collaborative networks, etc., and responsibilities in risk management.

⁶ In Mongolia, the State Emergency Fodder Fund (SEFF) had been operational before liberalizations in all 22 provinces, with a total of 69 emergency fodder stores. By the late 1990s, responsibility, but no funds, were shifted from the central government to provincial authorities; in 2000 only two SEFFs were reported as being operational. As a result, there are insufficient emergency fodder reserves, making snow disasters much more dangerous.

⁷ Done by one of the authors.

⁸ A first draft of the framework was designed by the regional workshop organized by SDAR, FAO in 1998 in Mongolia on "The Role of customary pastoral institutions in risk management, poverty alleviation and food security in Central Asian countries in transition", and has since then been further developed.

⁹ As well as the disaster-related support provided by developed countries.

¹⁰ However, it is important to note that risk is certainly not the only determinant of poverty. In Mongolia, for instance, other major factors shaping todays poverty are that:

• some ecological areas used by pastoralists are so resource-poor that it is difficult for even the most skilled herder to be successful;
• households with insufficient labour risk poverty; this is common during certain stages of the household life cycle, especially households with large numbers of children below working age, or households where one or two old people are left after the departure of married children;
• the management skills of individual herders are an important reason for success in herding; after the decollectivization in Central Asia, many households that were allocated animals and machinery as part of the division of the assets of the former collectives, were not skilled enough in herding to build up a herd and make a living; this was particularly the case of many urban-rural migrants who left the cities in the hope of benefiting from the distribution of animals by the former collectives, but who lacked basic herd management skills;
• the initial allocation of animals to herding households at the time of decollectivization was not done by calculating the minimum size of a viable household herd, and households were often left with too small herds for subsistence; such households were trapped in poverty from the start, which no amount of herding skill could have remedied.

¹¹ A good example of risk-related behaviour among pastoralists in CA has been observed during the last decade in Mongolia. The existence of a State Emergency Fodder Fund (SEFF) during the period of central planning led herders to reduce their own production of winter fodder, and to abandon some previous risk-avoiding livestock management practices, since there was an assurance that the State would make fodder available in most circumstances and that even if they didn't and animals were lost, the collectives would restock the unlucky herder. When the SEFF was disbanded at the start of liberalization, there was an immediate modification in livestock management by herders towards more risk-coping practices, such as mixed-species herding, and herders started to cut hay for themselves again.