Previous Page Table of Contents Next Page




This chapter discusses natural hazards and the preparation of investment projects within the context of the agricultural sector in Latin America and the Caribbean. It includes a summarized review of key concepts and policy issues and of selected project formulation and appraisal methods which can be used to incorporate natural hazard information into investment project preparation.

A review of existing investment projects in Latin America and the Caribbean indicates that those in the agricultural sector are generally undertaken with little or no consideration of natural hazards. Hazards affect agricultural projects more than any other sector. Considering the estimated US$670 billion in investments that will be necessary in this sector between 1980 and the year 2000 (FAO, 1981), there is a great need for an improved understanding of natural hazards, their assessment, and their management.

A combination of geographic location, climatic conditions, and limited capabilities for natural hazard assessment and disaster mitigation makes Third World nations more susceptible to the disasters natural hazard events pose than post-industrialized nations. Furthermore, the agricultural sector in these countries is often the most vulnerable and least able to cope with natural hazards in terms of infrastructure and institutional support.

In the following discussion, emphasis is placed on the need to apply the methods described in the formulation stage of new investment projects, rather than in the review of already prepared projects.


1. Historical Disasters and Agricultural Losses
2. Economy-wide Effects of Disasters
3. Natural Hazards and Development Issues

1. Historical Disasters and Agricultural Losses

Data from a variety of sources indicate that approximately 90 percent of all natural disasters worldwide occur in developing countries (Long, 1978). Recent Latin American and Caribbean examples illustrate the magnitude of the problem. When Hurricanes David and Frederick struck the Dominican Republic in 1979, they caused an estimated US$342 million in damage to the agricultural sector (UNDRO, 1980), destroying 80 percent of all crops and 100 percent of the banana crop. As a result, agricultural production fell 26 percent in 1979 and continued to be down 16 percent in 1980. Agriculture accounts for 37 percent of the country's gross domestic product and employs 40 percent of the labor force (USAID/OFDA, 1982). In 1984, the worst floods in Colombia in a decade caused an estimated US$400 million in damage to crops and livestock, while floods in Ecuador in 1982 and 1983 shrank the value of the banana crop by US$4.3 million (UN/ECLA, 1983).

In short, from 1960 to 1989 natural disasters caused over US$54 billion in physical damage in Latin America and the Caribbean. While the information available on the amount of national and international funds committed to reconstruction in response to each disaster is limited, the need to redirect funds to post-disaster work curtailed the availability of funds otherwise targeted for new investment.

2. Economy-wide Effects of Disasters

Besides the indirect social and economic impacts on a given region or sector, disasters can affect employment, the balance of trade, foreign indebtedness, and competition for scarce development investment funds. It has even been said that "the effect of natural disasters in disaster prone developing countries tends to cancel out real growth in the countries" (Long, 1978).

Figure 2-1 shows, in simplified fashion, the impact natural disasters in the agricultural sector can have on the entire economy. Internally, farm products provide food for the urban population and primary inputs to industry. Externally, they are exported and earn foreign exchange. Earnings from internal and external markets provide capital for new investment in the economy. Furthermore, the sector's operation generates an important demand for products from other sectors (e.g., fertilizers, equipment, and machinery). Finally, agricultural employment generates an increased demand for consumption goods and services from urban sectors. Urban growth and rural exodus are important considerations in the management of natural hazards, since they result in overcrowding of peripheral urban areas and increase the probability of disasters in these areas as a result of floods, landslides, earthquakes, and other hazards.

Figure 2-1 - Potential economy-Wide impacts of natural hazards in the agricultural sector in Latin America and the Caribbean

3. Natural Hazards and Development Issues

Notwithstanding the term "natural," a natural hazard has an element of human involvement. A physical event, such as a volcanic eruption, that does not affect human being is a natural phenomenon but not a natural hazard. A natural phenomenon that occurs in a populated are is a hazardous event. A hazardous event that causes unacceptably large numbers of fatalities and/or overwhelming property damage is a natural disaster. In areas where there are no human interests, natural phenomena do not constitute hazards nor do they result in disasters. This definition is thus at odds with the perception of natural hazards as unavoidable havoc wreaked by the unrestrained forces of nature. It shifts the burden of cause from purely natural processes to the concurrent presence of human activities and natural events.

Figure 2-2 illustrates this approach incorporating another argument into the discussion: the relationship of human and economic losses to the severity of an event and the degree of vulnerability (or survival capability) of human and economic interests.

The survival capability of projects depends on many factors. Losses from a severe event may be no worse or even less than those from a milder event if the former occurs in an area where both the population is adequately prepared to respond and the physical structures are designed and built to withstand its impact. One of the main differences between losses suffered by industrialized and less developed countries is the extent to which natural hazards and mitigation measures have been considered in the development planning process.

Figure 2-2 - Natural Hazard events in the context of human and economic interests

Planning systems and planners in developing countries cannot always be held fully responsible for the inadequacy of the natural hazard assessment and mitigation measures implemented (see Chapter 1). There are several reasons for this. First, much development is based on already existing hazard-prone scenarios. Second, planners depend on the availability of hazard information. And last, the planning process takes place within the prevailing economic, political, social, technological, and cultural parameters of a society. Mexico City's vulnerability to earthquakes is a good illustration. The sprawling city rests on precarious and deteriorating geological foundations. In spite of a well documented history of seismic activity, economic and technological constraints and complex political, social, cultural, and demographic elements impede the introduction of non-structural mitigation measures.

On the other hand, planning systems and planners are responsible for some serious shortcomings of investment projects in hazard-prone areas. Irrigation systems, roads, reservoirs, dams, and other infrastructure facilities are prime examples. In these cases, where the system of constraints and parameters is less complex than in urban planning, planners should be able to incorporate more information and have greater control over decision-making. But even where sufficient hazard risk information was available, projects have been undertaken without minimum mitigation measures. It is not uncommon for an area periodically devastated by hurricanes or earthquakes to be rebuilt again and again in the same way. Other disasters occur routinely as a direct consequence of improper human intervention in areas with previously stable ecosystems. The following box lists the key elements for incorporating natural hazards into agricultural investment projects.

Survival capability depends on many factors, and mitigation can make a substantial difference in minimizing the effects of disasters. While planners and planning systems are not responsible for some problems associated with natural hazards, they can exert influence in correcting some of the shortcomings. The following section discusses the process of integrating natural hazard information into the preparation of investment projects.


1. Probability
2. Risk
3. Risk Aversion
4. Risk Assessment
5. Risk Management
6. Investment Project

To facilitate the understanding of the subsequent sections, several key concepts are defined and explained below.

1. Probability

Probability is the likelihood of occurrence of a particular event. This is often based on historical frequency. For example, the probability of a hurricane in any given year could be 0.1, or 10 percent, if hurricanes have struck in two of the past 20 years. For the purpose of decision-making, however, probabilities are rarely based strictly on historical information but are usually adjusted to take account of currently available information may be then referred to as subjective probabilities. For example, the observation that tropical storms have recently occurred in other parts of the world can result in the assignment of a higher subjective probability to a local storm than would be indicated by the historical frequency.

2. Risk

Risk is generally defined as the probability of loss. In economic terms, this refers to a decline in income due to losses resulting from a natural hazard. Here risk will be used more generally to refer to uncertainty in the variables used in economic planning. For instance, in assessing the benefits and costs of a planned irrigation project, prices and yields of agricultural crops may fluctuate during the life of the project. These fluctuations can be caused by natural hazard events, but can also be caused by changing market conditions and weather cycles.

3. Risk Aversion

Risk aversion refers to an individual's attitude toward risk. Most people are risk-averse; that is, they are willing to incur some cost to avoid risk. But there is a wide range in degrees of risk aversion (Binswanger, 1980, and Young, 1979). In other words, to avoid a given level of risk, some people will pay more than others.

4. Risk Assessment

Risk assessment refers to the quantification of a risk. It requires a determination of both the consequences of an event and the likelihood of its occurrence. For example, a risk assessment of the potential economic effects of an earthquake on an agricultural project would require an estimate of its impact on farming activities and structural components, and of the probability of earthquakes in the region during the life of the project.

5. Risk Management

Risk management refers to actions taken to reduce the consequences or probability of unfavorable events. Similarly, natural hazard management refers to activities undertaken to reduce the negative effects of natural hazards. For example, a farmer may choose to plant a windbreak along a field to reduce the chances that wind will damage his sugar crops. While this may reduce his average income if he has to remove land from production, he may still do it to mitigate against an uncertain but potentially damaging storm.


- Natural hazard risk information is rarely considered in the preparation of investment projects in the agricultural sector. Agricultural projects can sustain severe losses from natural hazards. Losses can have serious sectoral and intersectoral economic implications.

- An adequate consideration of natural hazard issues in the preparation of projects, particularly during the formulation stage, can greatly reduce the vulnerability of investments and minimize human and economic losses.

- While basic natural hazard information is generally available, more specific and sophisticated risk and vulnerability information can be generated economically and used in the different project planning stages.

- Several analytical methods are available for incorporating natural hazard risk information into the evaluation of agricultural investment projects.

6. Investment Project

An investment project is the use of capital to create assets capable of generating a stream of benefits over time. Agricultural investment projects include land settlement, agricultural extension, irrigation, and soil conservation. Projects can be independent or part of an integrated regional development package.


1. Preliminary Mission
2. Phase I - Development Diagnosis
3. Phase II - Project Formulation and Definition of Action Plan
4. Project Implementation

Minimizing the effects of natural hazards on the agricultural sector, and on an entire economy, can reduce the vulnerabilities and increase the ability to survive natural disasters. This can be achieved by incorporating natural hazard information into the preparation of agricultural investment projects. How it is done, and its relationship to an integrated development study, are discussed in this section.

Integrated development planning is a multisectoral and multidisciplinary approach to generating plans and proposals for economic and social development. It brings together issues concerning various sectors and analyzes them in an integrated fashion vis-a-vis the needs of the population and the characteristics of the natural resource base. Appropriate natural resource use along sound environmental management guidelines seeks to maximize development opportunities while minimizing environmental conflicts (see Chapter 3). The creation of an integrated development planning study is a complex process, within which the preparation of investment projects is only one step. The preparation of planning studies and investment projects is very similar. That similarity is often a source of confusion.

An integrated development planning study is composed of four basic stages: the Preliminary Mission, Phase I or the Development Diagnosis, Phase II or Project Formulation and Action Plan, and Implementation. (See Chapter I for a detailed discussion of the four stages of integrated development planning.) The preparation of investment projects within the development planning study also entails four steps: Project Profile, Prefeasibility Analysis, Feasibility Analysis, and Implementation. The information needs of the four development planning study stages are described in the box below.

Although most institutions do not require risk information in project preparation guidelines except at the engineering design stage, both integrated development planning studies and investment project preparation are improved when analysts incorporate natural hazard information into all stages of development planning. Guidelines for the use of natural hazard information in project preparation are listed in Figure 2-3 and discussed below.


- Natural hazards should be considered throughout the entire development study.

- Preliminary Mission: Natural hazard information should be collected and analyzed at this time. This information should be used to define the study area, the objectives and characteristics of the study, and the preparation of the intended work program.

- Phase I: Hazardous event information and risk maps should be used in the development diagnosis for the identification of hazard-prone areas, land-use zoning, and preliminary identification of mitigation measures.

- Phase II: Vulnerability information can be used to refine project costs and benefits at the prefeasibility level. Risk considerations can be incorporated into the different project formulation studies (market, size and location, engineering, etc.); and structural and non-structural mitigation measures can be selected. At the feasibility analysis level, available information can be complemented by specific hazard assessments and used to further refine cost and benefit calculations. More sophisticated probabilistic evaluation methods can be used by generating probabilistic risk distributions.

- Implementation: The implementation of an integrated development strategy and of individual projects should include monitoring of construction procedures to ensure compliance with engineering standards of buildings and structural mitigation measures. Long-term monitoring for ensuring compliance with designed non-structural mitigation measures should also be programmed

1. Preliminary Mission

Risk information should be introduced at the earliest stages of project planning. (See Appendix A for more details on the types of available natural hazard information.) When this information is included at the Preliminary Mission stage, the design of the integrated study and the investment projects can accommodate risk factors; if the risks are too great, alternative overall development strategies can be considered. When risk information is not included until the feasibility analysis stage, it is usually too late for anything but remedial actions.

2. Phase I - Development Diagnosis

Natural hazard issues should be considered further in the development diagnosis stage. Risk maps and hazard event frequencies should be consulted in order to identify the area's problems and opportunities. For example, a floodplain map produced by remote sensing techniques would depict areas that are prone to severe flooding. From the start of the project planning process, planners might want to avoid designating these areas for agricultural activities requiring extensive capital investment and propose instead an alternative land use less sensitive to flooding. Or planners might want to consider hazard mitigation practices to reduce the risk to acceptable levels. (See Chapter 8 for a discussion of flood hazard assessments and remote sensing techniques.)

The design of investment projects begins at this stage with the development of alternative project profiles. A project profile should include project objectives and principal characteristics, rough estimates of costs and benefits, and a preliminary identification of alternatives for design and implementation. These activities should reflect the natural hazard information collected between the Preliminary Mission and the Development Diagnosis stages of the integrated development planning study.



- What type of natural hazard information is available for the study area, and what additional information is needed?

- What are the benefits of incorporating natural hazard information at this stage?


- What values will be assigned to natural hazard issues as they relate to defining characteristics of a potential project(s), e.g., its area, objectives, target population?

- These values having been specified, what are the definitions of the defining characteristics



- What type of information needs to be used?

- How can planners complement non-probabilistic hazard information in order to evaluate project uncertainty?

- Are natural hazards a significant variable in the identification of investment projects?

- Can mitigation measures be included as part of the development strategy? Which ones?


- What is an acceptable level of risk for the affected population and the investment project?

- What are the relevant natural hazard issues in the project area? What priority should be given to each?

- Are modifications of the project profile needed?

- Will additional hazard assessment studies be needed for investment project formulation

3. Phase II - Project Formulation and Definition of Action Plan

In Phase II, investment projects are generated and selected. This phase includes prefeasibility and feasibility analyses and is based on a standardized project formulation methodology. The prefeasibility analysis involves a preliminary evaluation of the technical and economic viability of a proposed project: alternative approaches to various elements of it are compared, the best are recommended for further analysis, and investment and operating costs are estimated. The feasibility analysis constitutes the final determination of the viability of the project, reexamining every aspect of it and refining the estimate of its benefits (income stream, increases in production, generation of employment, etc.), costs (construction, operation and maintenance, depletion of resources, pollution effects, etc.), and valuative criteria (net present value, internal rate of return, benefit-cost ratio, and repayment probabilities).

Figure 2-3 - The use of Natural Hazard information in investment project preparation within the context of integrated development planning studies

The design of individual investment projects should, but in current practice ordinarily does not, incorporate the following types of natural hazard information:

- Incidence of hazard risks in the project area

- Incidence of hazard risks in the project's market areas and commercialization routes

- Vulnerability of the supply and/or cost of production inputs (e.g., raw materials, equipment, energy resources) to natural hazard events

- Vulnerability of the project's output prices to natural hazard events

- Vulnerability of physical structures and production processes to natural hazard events

- Existence of current and/or proposed legislation that establishes guidelines for natural hazard risk mitigation in project design

- Effectiveness and cost of alternative natural hazard mitigation measures



- Which components of a project formulation study should include natural hazard considerations?

- Is the available hazard information adequate to formulate investment projects?

- What type of mitigation measures should be considered during the project formulation stage?


- How will the identification and technical analysis of natural hazard mitigation measures take place?

- How will the best project options and most suitable mitigation measures be determined?

The critical factor for the successful incorporation of natural hazard considerations into the project formulation phase is the ability of project planners to use hazard information in the design. The identification of cost-effective mitigation measures that will significantly reduce risks is of crucial importance. Not every mitigation measure should be implemented-only those whose benefits exceed their costs.

Mitigation measures may be structural or non-structural. Structural mitigation includes physical measures or standards such as building codes, materials specifications, and performance standards for new buildings; the retrofitting of existing structures to make them more hazard-resistant; and protective devices such as dikes. Non-structural measures typically concentrate on identifying hazard-prone areas and limiting their use. Examples include land-use zoning, the selection of building sites, tax incentives, insurance programs, relocation of residents to remove them from the path of a hazard, and the establishment of forecasting and warning systems. Figure 2-4 presents some examples of structural and non-structural mitigation measures relevant to the agricultural sector. For a more detailed discussion of mitigation measures related to specific hazards, see Chapters 8 through 12.

A strong case can be made for emphasizing non-structural measures in developing countries. Essentially, all structural mitigation measures have a direct cost that must be added to the project under consideration. Given the prevailing lack of awareness of risks from natural hazards, additional costs will appear unjustified vis-a-vis expected costs and benefits. This does not mean that non-structural mitigation measures will add no cost to projects or society, but that in an area subject to flooding, for example, the economic and social costs of measures such as zoning policies and crop insurance are likely to be much lower than those of large-scale flood control systems in terms of initial cost, operation, and maintenance. Furthermore, the agricultural activities that have been the most affected by natural hazards are large-scale agricultural development projects.

When project characteristics impede the adoption of non-structural mitigation measures, more costly structural mitigation systems should be explored as a way to reduce risks to a socially acceptable and economically feasible level.

4. Project Implementation

The Implementation stage begins once the investment projects and the action plan of a development planning study have been determined. Depending on the nature and scope of the overall study and of the individual projects selected, implementation can be simultaneous with or preceded by the implementation of sectoral and regional support programs and the development of legal and institutional frameworks.

Figure 2-4 - Mitigation measures for the agricultural sector


- Do natural hazard considerations enter into the project implementation phase? Which ones?

- What actions can be taken to ensure that the implementation of projects complies with the specified engineering standards and that mitigation measures are adequately put in place?


- Does the final project design include the necessary mitigation measures?

- Are monitoring and maintenance procedures established for the implementation of mitigation measures within a project's implementation phase?

The implementation of investment projects is a critical phase in the successful incorporation of natural hazard considerations into the development planning process. All the efforts made in the previous stages will be lost unless the projects are carefully monitored throughout the implementation process to ensure that structural mitigation measures are adhered to and non-structural mitigation measures have been selected and adopted.


1. Attitudes Toward the Risks from Natural Hazards
2. Establishing Evaluation Criteria and Priorities

1. Attitudes Toward the Risks from Natural Hazards

While risk aversion at the individual level is well documented, the question of whether or not government institutions should be risk-neutral has been the subject of controversy. Should risk be considered in the analysis of public sector projects?

It has been argued that although individuals are risk-averse, governments should take a risk-neutral stance because, given that project benefits and costs are spread over a large number of individuals in the society, the risk faced by each one is negligible. This implies that governments should be indifferent between a high-risk and a low-risk project provided that the two have the same expected net present value (NPV) (Arrow and Lind, 1970).

This argument is valid only up to a point. The reality of developing countries suggests otherwise. Governmental decisions should be based on the opportunity cost to society of the resources invested in the project and on the loss of economic assets, functions, and products. In view of the responsibility vested in the public sector for the administration of scarce resources, and considering issues such as fiscal debt, trade balances, income distribution, and a wide range of other economic, social, and political concerns, governments should must not be risk-neutral.

Suppose there are two projects under consideration in a coastal area of a developing country. The NPV of Project A is US$2 million, and of Project B US$1.5 million. Because Project A has the higher NPV, it would be selected if risks were ignored. However, Project A is vulnerable to floods and its actual NPV, depending on their frequency and severity, could be between US$0.5 and US$2.5 million. Project B is less susceptible to flood damage, and therefore has an NPV range of US$1.3 to US$1.7 million. Since the returns on Project B are more stable, the participants directly involved might prefer the project with the lower NPV. Furthermore, they would probably be unimpressed by arguments about the merit of societal risk sharing, since the risk (the variation in NPV) that their community directly bears from these projects is rather large.

In practice, most Latin America and Caribbean governments and their planning agencies lack awareness of the need to reduce the vulnerability of investment projects to natural hazards, and tend to disregard it in their evaluations. Some of the reasons for this lack of awareness are listed in the following box.


- The magnitude of the risks and potential savings of mitigation are perceived as low.

- Political and financial pressures make it unappealing to take careful steps now to avoid losses in the future.

- If losses occur, international agencies will provide assistance.

- Natural hazards are accepted as inevitable and little is known about non-structural mitigation.

- The burden of analysis, institution-building, and implementation discourages the effort of assessment. Political, financial, economic, and social costs of natural hazard assessments and mitigation may not always be less than their benefits.

- The costs of undertaking natural hazard assessments and mitigation fall on government institutions that cannot recapture directly the benefits of preventing losses in the future.

- The property rights surrounding the Investments do not necessarily provide a strong incentive for preventing losses from hazards

National and international banking institutions also tend toward neutrality in the treatment of risks from natural hazards. They are generally more concerned with how macroeconomic and political factors may affect a government's overall repayment ability than with the effect of risk factors on cost recovery. As a result, loans are routinely made with little or no risk assessment. While this attitude makes sense for the bank because it grants loans against overall government credit worthiness and does not share the risk of any individual project, it does not necessarily make sense for borrowing nations.

2. Establishing Evaluation Criteria and Priorities

In dealing with governmental and societal attitudes toward natural hazards, planners can benefit from multicriteria analysis or, as it is sometimes called, multiple conflicting objectives analysis. This method has been used in environmental assessments and is gaining increasing acceptance for the incorporation of societal goals and priorities into the selection of investment projects.

Multicriteria analysis entails the establishment of a set of objectives and a subset of attributes representing alternative social, economic, political, and environmental goals which are to be fulfilled by specific projects. The relevant social groups (government, interest groups, community leaders, etc.) participate in establishing the objectives and attributes and placing discriminatory weights on them. Projects can then be evaluated in terms of their capacity to fulfill the stated goal. If the establishment of the objectives and attributes is properly oriented, natural hazard vulnerability criteria can be introduced into the analysis along with the other goals (Vira and Haimes, 1983; Haimes et al., 1978; Keeney and Raiffa, 1976).

It is important to remember that regardless of the methods used in project evaluation, it is not planners but decision-makers who will ultimately rule on public investment options. Multicriteria analysis forces decision-makers to state their evaluation criteria explicitly. While most decision-makers will give low vulnerability a high priority in project selection for economic or political reasons, natural hazards will not always be considered in the final decision.

Multicriteria analysis can be applied throughout the project cycle, from the profile stage to the feasibility study, but since it is effective in the early identification of more desirable projects and project components, its use at the beginning stages of project planning maximizes its benefits.


1. Measuring Costs
2. Measuring Benefits
3. Discounting Net Project Flows
4. Project Evaluation

Economic or cost-benefit analysis is a method that evaluates the efficiency of public sector activities, permitting a comparison of the merits of different government projects over time. A number of techniques are available, and analysts should choose the one best suited to each case.

When private individuals consider whether or not to make an investment, they consider only the benefits that have a direct personal impact on them; this is financial analysis. In economic analysis the societal perspective is taken, incorporating all benefits and costs affecting society.

Another important aspect of economic analysis is the "with-and-without" criterion: what the state of affairs would be with versus without the project in place. The "with-and-without" analysis helps to sort out the benefits and costs of a project. Suppose an irrigation project is being considered for an area where crop yields are increasing. The project will raise them even more. The assessment of potential benefits would be erroneous if it attributed all the increase to the project, since some of it would have occurred anyway (Howe, 1971). In areas that are growing rapidly, it is particularly important to ensure that benefits and costs are properly accounted for and do not include changes that would have taken place without the project.

The economic appraisal of projects can be organized into four main steps:

- Identification and computation of all the costs of the proposed projects;
- Identification and computation of all the benefits of the proposed project;
- Discounting future net benefits and expressing them in current dollar terms; and
- Evaluation of the net project flow of proposed projects.

While these steps may appear simple, a thorough analysis requires considerable effort. The economist or planner carrying out the analysis should work with other specialists such as agronomists, engineers, and hydrologists to ensure that all relevant factors are taken into account and that technical and institutional relationships are property reflected. This integrated, interdisciplinary approach to planning has been advocated by the OAS (OAS, 1984).

1. Measuring Costs

In measuring the costs of a project, it is important that all of them be accurately reflected, including those that may not be immediately apparent. There are, of course, the direct costs. Materials and administration are among these, as is the use of natural resources. The costs of natural hazard vulnerability reduction, both structural-canal systems, dams, dikes, windbreaks-and in some cases non-structural are direct costs. Additionally, there are indirect costs. For example, if a new project will draw water resources from nearby farmland, any decline in agricultural production in that area should be counted as a project cost. And then there are the "opportunity costs"-the loss of the benefits that would accrue from some alternative use of the resources that are being devoted to the project.

The analyst must also be aware that, owing to market distortions, the prices of inputs may not reflect their true valuation by society. In such cases, prices should be adjusted to correct for these distortions. If a government subsidy lowers the cost of the fertilizer used in the project, the economic analysis must add the amount of the subsidy to the market price of the fertilizer to reflect its true cost to society. Adjusted prices are referred to as "shadow prices."

2. Measuring Benefits

Direct benefits of an agricultural project can result from an increase in the value or quantity of farm output and from a lowering of production costs. The benefits from natural hazard mitigation can be measured in terms of income losses avoided. Projects generate indirect benefits as well. For example, an irrigation project might have the "spillover" benefit of increasing the productivity of land adjacent to the land actually being irrigated by the project.

An evaluation of the benefits of a project should include only real increases in output. A flood control project may raise the value of farmland in the protected area, but since this higher value reflects the increased output potential of the land, counting it as a benefit would result in counting the benefits of the project twice.

The consideration of natural hazard risks requires differentiating between the concepts of income stream and benefit stream of a project. While the income generated by a project is a major component of the benefits, it does not reflect certain essential variables. For instance, income and job stability from the project and associated enterprises might be severely affected by a hazardous event, but merely adjusting the income stream to the uncertainty associated with natural hazard events will not reflect the economic and social losses that would accrue from income and job disruption. The benefit stream reflects these losses. In the case of a project that includes mitigation measures, the economic analysis should include the added benefit of avoiding losses. A proper identification of the benefit stream of a project allows analysts to evaluate the net effect of introducing mitigation measures into the project design, since both the direct cost of these measures and their expected; benefit will be included in the evaluation process.

3. Discounting Net Project Flows

The third step in project analysis is to discount the future benefits and costs. This is done by using a discount rate to convert future values into present values. The need to discount future costs and benefits arises because a given amount of money is worth more today than in the future: money today can earn interest between now and then. An investment of US$100 at an annual interest rate of 10 percent will be worth US$121 at the end of two years. Future benefits and costs must be discounted in order to express them with a common denominator-today's dollars or present value.

The project analyst must choose the discount rate, and often more than one rate is used in a project. For financial analysis, the discount rate is usually the rate at which the firm for which the analysis is being done is able to borrow money. In economic analysis, three alternatives for the discount rate are suggested: the opportunity cost of capital, the borrowing rate, and the social time preference rate (Gittinger, 1982). Probably the best is the opportunity cost of capital, which is the rate that will result in the utilization of all the capital in the economy if all possible investments that yield as much or more in return are undertaken. The opportunity cost of capital cannot be known with certainty, but in most developing countries is considered to be between 8 and 15 percent in real terms.

The borrowing rate is most commonly proposed when the country expects to borrow from abroad for investment projects. Financial rates of interest, however, are generally too low to justify their use in economic analysis, and may even be negative in real terms when the rate of inflation is high. The social time preference rate differs from the opportunity cost of capital in that it assigns a different (usually lower) discount rate for public projects than for private ones, given that society has a longer time horizon.

4. Project Evaluation

The discounted or net present value (NPV) of a project is represented mathematically as:

where B = benefits, C = costs, r = discount rate, t = time period, n = life of the project in years, and S = summation operator. After benefits and costs are evaluated and a discount rate is selected, this equation will indicate the NPV of the project under consideration. The economic criteria used to determine the value of a project are (a) whether the NPV is positive and (b) whether the NPV is higher than that of alternative projects. Another way to compare benefits and costs is to set the equation equal to zero and solve for the value of r. This value is referred to as the "internal rate of return" (IRR).

The equation is often rearranged as a benefit-cost ratio in order to facilitate comparison of projects:

The higher the NPV of the project, the higher the ratio will be. A benefit-cost ratio greater than one indicates that the discounted benefits exceed the discounted costs.


1. Decision Criteria with Limited Information
2. Decision Criteria with Probabilistic Information

Several methods are available for evaluating the natural hazard components in the economic analysis of a project. Some can be applied when little hazard information is available, others are appropriate when information on probability distributions can be obtained. All can be used to compare different projects or alternatives within a project. The methods used when limited information is available can be applied at the project profile and prefeasibility levels of analysis. Those using probabilistic information are usually used in feasibility studies, but may also be used at the prefeasibility stage. In all cases the methods should be applied as early as possible in the project cycle.

1. Decision Criteria with Limited Information

a. Cut-Off Period
b. Discount Rate Adjustments
c. Game Theory Approaches
d. Sensitivity Analysis

Four methods of risk evaluation compensate for a lack of information: cut-off period, discount rate adjustment, game theory, and sensitivity analysis.

a. Cut-Off Period

The crudest procedure for incorporating risk into economic analyses is the use of a cut-off period (Mishan, 1982). It is primarily used by private investment agencies interested in capital return rather than in long-term development. Under this method, economically feasible projects must accrue enough benefits to surpass project costs in relatively few years.

For very risky projects, the cut-off period might be set as low as two or three years, whereas for low-risk projects it would be much longer, say 30 years. The underlying logic is that the benefits and costs are so uncertain beyond the cut-off date that they can be ignored in determining project feasibility. The cut-off period should be determined at the prefeasibility stage of project preparation.

Some information is necessary to determine the relative risk of the project. The most useful data are a list of historical natural disasters or episodic information, meteorological records, land-use maps, agricultural crop maps, and previous damage assessments. This information provides economists with a rough idea of the inherent risks. In addition, satellite photography of the impacts of natural hazards can be useful in deciding on a cut-off period. In many cases it is not too difficult to obtain this type of information for short periods of time.

A cut-off period should only be considered when few records are available and the nature and magnitude of the hazards can potentially pose a great risk to development, e.g., severe storms and floods. It is more difficult to establish a cut-off period in the case of slow-onset hazards such as droughts or desertification.

As an example, the cut-off period method could be applied to a ten-year, large-scale vegetable and livestock farming project. This project may have a high risk if the area is subject to periodic flooding, which would damage crops and destroy livestock. In this case, a four- or six-year cut-off period might be chosen. Figure 2-5 illustrates this example.

Figure 2-5 Cut-off-period method

While this approach considers the effects of risks, it does have some limitations. Too short a cut-off date can ignore economic information associated with much of the project's life, since it discards all information beyond the cut-off period. This may be particularly important when considering the sustainability of economic returns from a project as resources, renewable or non-renewable, are depleted after the cut-off period. If benefits and costs are highly variable beyond the cut-off date, there are more appropriate methods which can address the risk of benefit-cost variability.

b. Discount Rate Adjustments

Another ad hoc way to reflect uncertainty in project analysis is to add a risk premium to the discount rate. The effect of increasing the discount rate is to give less weight to the increasingly uncertain costs and benefits in future time periods (Anderson et al, 1977). This is consistent with what has been observed in the private sector: managers generally require higher internal rates of return for riskier investments. A variation of this is to add a premium to the discount rate for the benefits and subtract a premium for the costs, a procedure consistent with the fact that hazards decrease benefits and increase costs.

This technique is based on a subjective decision as to the risk premium to be added to and/or subtracted from the discount rate. The same type of information that is useful for a cut-off period can be used to determine the discount rate. This information should be available by the prefeasibility stage of project planning.

A subjective decision on the discount rate can incorporate the information available on the possibility of a slow-onset hazard in addition to short-term, immediate impact hazards such as severe storms and flash floods. Once again, this method should be employed when the information is limited.

In the previous farming example, any indication of flooding increases the risk of the project. If normally a discount rate of 10 percent for benefits is used, the discount rate might be increased to 12 or 15 percent, as shown in Figure 2-6.

Figure 2-6 Discount rate adjustment method

This approach is preferable to the cut-off-period method because it includes information about the future benefits and costs. However, the risk adjustment of the discount rate is arbitrary, and the approach does not recognize risk differences across project components. More rigorous and defensible approaches which are capable of quantitatively assessing the uncertainty of benefits and costs over time are discussed below.

c. Game Theory Approaches

When there is no reliable information on probability distributions of hazards, two strategies from game theory can be useful: the maximin-gain strategy and minimax-regret strategy. Both can be applied in the early stages of project formulation as the necessary minimum of information-records of historical events, climatological and meteorological data, and previous natural hazard damage records-becomes available. From this information it is possible to estimate the comparative benefits of equivalent alternatives under varying degrees of natural hazard severity. Game theory approaches are better suited for short-term, immediate-impact hazards which can be easily divided into least/most-damage scenarios.

Maximin-Gain Strategy

To illustrate the maximin-gain approach, which derives its name from maximizing the minimum, suppose that a decision has been made to augment the previously discussed farming project with a structural mitigation measure aimed at reducing the effects of potential flooding. Three alternative flood control projects, Projects A, B, and C, equal in cost, are under consideration (Anderson and Settle, 1977). For convenience, it is assumed that there are two possible scenarios-heavy rainfall and normal rainfall. If heavy rainfall occurs, the NPV of benefits from the three projects are: Project A = $100 million. Project B = $120 million, and Project C = $150 million. If the rainfall is normal, the projects will provide irrigation and other discounted benefits of $30 million, $60 million, and $20 million, respectively. The benefits will be greater in the case of heavy rainfall, because the primary benefit is the prevention of flood damage. The different outcomes are summarized below and shown in Figure 2-7.

Figure 2-7 Maximin-Gain strategy


Heavy rainfall

Normal rainfall

Project A

$100 million

$30 million

Project B

$120 million

$60 million

Project C

$150 million

$20 million

The maximin-gain strategy would result in choosing Project B, since its minimum benefit is $60 million, as compared to $30 million for Project A and $20 million for Project C. The maximin-gain strategy is based entirely on security and has the drawback of being very conservative: even if the benefits of A and C were 10 times larger than those of B under heavy rainfall conditions, Project B would still be selected. Thus, it can lead to the selection of projects which most people would agree are inferior.

Minimax-Regret Strategy

An alternative approach is the minimax-regret strategy. This consists in minimizing the maximum regret or loss that could be realized. Using the same example as above, if heavy rainfall does occur Project C would result in the greatest benefit, $150 million. If Project A was selected, the regret or forgone benefits from not selecting C would be $50 million ($150 million minus $100 million) and from not selecting B would be $30 million ($150 million minus $120 million). If the rainfall is normal instead of heavy, Project B would produce the most benefits, $60 million. In that case the forgone benefits would be $40 million for Project C and $30 million for Project A. Now considering both possible weather conditions, heavy and normal rainfall, the maximum regret would be $50 million, $30 million, and $40 million respectively for Projects A, B, and C. Therefore, the minimax-regret strategy would lead to a choice of B since it has the smallest maximum regret, as is shown in Figure 2-8.

d. Sensitivity Analysis

In a sensitivity analysis, the analyst changes the value of key parameters that are subject to risk to determine the effects on the NPV of a project. Usually, the values are changed one at a time, but sometimes they are changed in combination with one another. This can be useful when the available information indicates how much each parameter should be changed (Irwin, 1978). Typically, values are changed by an arbitrary amount, say five percent.

Sensitivity analyses can help to identify project elements that need further consideration and thus can be used at the project profile stage before a more sophisticated risk analysis is completed. They can also be used to test the effect of mitigation measures. They are suited to all types of hazards, even when the information available is minimal.

Figure 2-8 Minimax-Regret strategy

The types of information that are useful for this analysis are event histories, climatological and meteorological data, and previous damage reports. These data assist economists in estimating percentage variations in parameters from previous hazard information.

The example of the farming project can be used here to demonstrate this method. With the aid of a personal computer or even a hand calculator, a sensitivity analysis can be performed on each cost and benefit to determine their effects on the rest of the project. For example, a sensitivity analysis performed on crop yields may demonstrate that if production falls by 40 percent in the first year as the result of an intermediate-level flood, the overall project benefits may be greatly decreased, or it would take much longer to recover the costs.

The best way to report the results of sensitivity analysis is by means of "switching values" (Baum, 1980). These are the values of the key variables at which the NPV of the project becomes zero or the benefit-cost ratio falls below one. Switching values can be presented as shown below and in Figure 2-9.


Switching Value

Corn price


Corn yield


Construction costs


Fertilizer price


Labor cost


In this example, corn yields would only have to decline from their expected value by 20 percent to make the project NPV equal zero. On the other hand, labor costs could increase up to 60 percent before the NPV falls to zero.

2. Decision Criteria with Probabilistic Information

a. Mean-Variance Analysis
b. Safety-First Analysis

If probability distributions for key economic variables are available, a more rigorous evaluation of risk can be carried out. The probability distributions may be based on the subjective assessments of experts or on historical information such as episodic, climatologic, meteorologic, and agronomic data. For example, if adequate data are available, the probability distribution for crop yields can be estimated from historical farm or experiment station records. Where these data are not available, as is often the case, subjective probabilities can be elicited from farmers, extension agents, or agronomists.

Figure 2-9 Sensitivity Analysis

One relatively simple way to obtain subjective probabilities is the triangular distribution method. Analysts can estimate the most likely, the best, and the worst possible yields. The mean and variance of the probability distribution can then be estimated (Anderson et al, 1977). Subjective distributions of yields can be provided for projects with or without natural hazard mitigation measures.

Since natural hazards can affect both the benefits of a project (for example, by destroying crops) and the costs (for example, by damaging irrigation systems), in some cases it will be desirable to obtain probability distributions of natural hazard events. Probabilistic information can be obtained for any type of natural hazard with measurable magnitude and frequency, but of course the quality of the information can vary widely.

In estimating the probability distribution of economic feasibility measures, such as NPV, only a limited number of variables are considered random or subject to fluctuations; others are considered fixed for the purposes of the analysis. The variables that are allowed to fluctuate can be determined either by making a sensitivity analysis to identify those that are important or by observing those that fluctuate widely. Various probability distributions can be combined mathematically or with computer simulation methods to form a probability distribution of NPV. The distribution conveniently conveys information about the risks of a project.

After the probability distributions have been calculated, the mean or average values of each distribution can be compared to make a selection between projects, or between alternatives within a project. But using averages alone ignores the relative risks of the projects, even though this information is available from the already prepared probability distributions. Two methods are suggested to compensate for this: mean-variance analysis and safety-first analysis.

a. Mean-Variance Analysis

With mean-variance analysis, which can be applied in the prefeasibility stage of project development, projects can be compared by graphing the NPV probability functions. In Figure 2-10, Project A and Project B have similar probability distributions-that is, they have the same risk-but the distribution for Project B is further to the right, indicating that the average NPV is greater. Project B, then, is preferable to A.

Figure 2-10 - Mean-Variance analysis: projects with equal risk and different NPVs

In Figure 2-11 Projects C and D have the same mean, but Project D has a greater dispersion around the mean, and thus is riskier. If only the mean values of the projects' NPV are considered, society will be indifferent between Projects C and D. However, if society considers this a critical project and cannot afford to have it give low yields, Project C will be preferred, since there is less chance that the NPV will fall below the mean. The comparison of Project C with Project E is less clear-cut: Project E has a much higher mean than Project C, but its variance is also greater. Clearly, there is a trade-off between a higher expected NPV and the acceptance of greater risk. The decision-maker, not the analyst, will have to decide what weights to apply to higher mean NPV versus greater risk.

A mean-variance analysis can be easily applied to the example of the flood control projects presented earlier. The information needed includes historical data on past flood events-magnitudes and frequency of occurrence-from which statistical means and variances can be calculated to provide sufficient data for determining the probability of flooding. This information can be used by planners in making a decision. It can also be used to calculate the probability distribution of the NPV of alternative flood control projects and, in turn, the means and the variances of the projects' NPV. This analysis enables the project planner to view the variance, or the risk, of the NPV resulting from flood events.

b. Safety-First Analysis

Since risk management is concerned primarily with reducing losses, the left-hand side of a probability distribution is of more interest to an analyst than the right-hand side. If the distribution is symmetrical, as is normal, decisions based on the variance will be suitable for risk management because negative and positive fluctuations around the mean are equally likely. However, some real-world phenomena of interest to risk analysts appear to follow distributions that are skewed in one direction or the other. For example, corn yields may average 100 bushels per acre, and a drought that occurs every five years could cause yields to fall to zero, but there will probably never be yields fluctuating as far above the mean as 200 bushels. Thus, analysts may want to choose a decision criterion that focuses on the lower tail of a distribution. An additional advantage of such an approach is that it lends itself more easily to discussions of minimizing losses, which can be useful when considering hazard mitigation measures. Safety-first criteria can be applied to relatively frequent natural hazards, such as floods and severe storms, but they are not as useful for low-frequency catastrophic events such as volcanic eruptions and tsunamis.

Figure 2-11 - Mean-Variance analysis: trade-off between higher NPV and greater risk


Maximize NPV subject to P(NPV < C) < a

The box above shows the following values: NPV = net present value, P = probability, C = critical threshold value, and a = small probability value. The decision criterion is to maximize expected NPV subject to a constraint that there is only a small probability that it will fall below some constant value. For example, the decision-maker might choose the project with the highest expected NPV as long as the probability of its falling below zero is less than 5 percent (Pandey, 1983).

Suppose the safety-first criterion is established as follows: maximize NPV subject to no more than a 20 percent chance that NPV will fall below $20,000. The cumulative probability of the NPV for two different projects is shown in Figure 2-12. As the graph indicates, the probability is 40 percent for Project A and 15 percent for Project B. The safety-first criterion would eliminate A from further consideration. If there were other projects with less than a 20 percent chance of having an NPV smaller than $20,000, then the one with the highest NPV would be recommended for implementation.

A safety-first approach can be applied to the flood control example. The project planner can decide what level of NPV is the absolute minimum for the project to continue. If the minimum acceptable NPV is $1 million and the probability of falling below that is 40 percent, 20 percent, and 70 percent, respectively, for the different flood control projects, the one with the smallest probability might be preferred.

With the methods described in this section, projects can reflect the additional costs that natural hazards pose and the additional benefits resulting from mitigation measures. Figure 2-13 summarizes the relationships between these methods and the investment preparation process. Some of the key considerations for incorporating natural hazards into the evaluation of investment projects are listed in the following box.

Figure 2-12 - Safety-First approach

Figure 2-13 - Applicability of economic appraisal methods for incorporating Natural Hazard considerations into the evaluation of investment projects


- Natural hazards should be considered in the evaluation of public and private sector projects alike.

- Multicriteria analysis can be used to incorporate natural hazard considerations into general development and project investment planning.

- When little risk information is available, cut-off period, discount rate adjustments, maximin-gain, minimax-regret, and sensitivity analysis methods can be used to consider natural hazards in the economic evaluation of investment projects.

- When probabilistic information is available, mean-variance analysis and safety-first analysis, among other methods, can be used to consider natural hazards in the economic evaluation of investment projects.


Natural hazards can have considerable human and economic impacts on the agricultural sector in developing countries. Since these and other forms of risk can make the outcome of development projects uncertain, they need to be considered early in the development process. For this to happen, a large effort will be required to modify current project formulation and evaluation practices. But the changes should not be limited to project planning. If natural disasters are to be reduced significantly and consistently, not just in isolated projects, changes will also have to come about in government agencies, development assistance agencies, banking institutions, scientific communities, and attitudes toward natural hazards. Without a doubt, the availability of timely and adequate information will be a key factor in making these groups aware of the human and economic significance of disasters and of the necessity to support hazard mitigation at different levels. As intermediaries, development assistance agencies should take advantage of their inherent capabilities and assume a leading role in this process.

Because resources are scarce and costly, hazard mitigation actions should be focused and well articulated. Natural hazard mitigation actions should reflect legitimate social, economic, and political priorities, and new investment projects in key economic sectors, such as agriculture, should be given preference over retrofitting mitigation measures into already existing projects.


Anderson, J.R., Dillon, J.L, and Hardaker, J.B. Agricultural Decision Analysis (Ames, Iowa: Iowa State University Press, 1977).

Anderson, LG., and Settle, R.F. Benefit-Cost Analysis: A Practical Guide (Lexington, Massachusetts: 1977).

Arrow, K.J., and Lind, R.C. "Uncertainty and the Evaluation of Public Investment Decisions" in American Economic Review, vol. 60 (1970).

Baum, W.C. Risk and Sensitivity Analysis in the Economic Analysis of Projects. World Bank Central Projects, Note 2.02 (July 1980).

Binswanger, H.P. "Attitudes Toward Risk: Experimental Measures in Rural India" in American Journal of Agricultural Economics, vol. 62 (1980), pp. 395-407.

Food and Agriculture Organization (FAO). Agriculture: Toward 2000 (Rome: United Nations, 1981).

Gittinger, J.P. Economic Analysis of Agricultural Projects, 2nd ed. (Baltimore, Maryland: Johns Hopkins University Press, 1982).

Haimes, Y.Y., et al Multi-objective Optimization in Water Resources Systems (New York: E.S.P. Corp., 1978).

Howe, C.W. Benefit-Cost Analysis for Water Systems Planning, Water Resources Monograph 2 (Washington, D.C.: American Geophysical Union, 1971).

Hyman, D.N. The Economics of Governmental Activity (New York: Holt, Rinehart, and Winston, 1973).

Irwin, G. Modern Cost-Benefit Methods (London: Macmillan, 1978).

Keeney, R.C., and Raiffa, H. Decision Analysis with Multiple Conflicting Objectives: Preferences and Value Trade-Offs (New York: John Wiley and Sons, 1976).

Long, F. "The Impact of Natural Disasters on Third World Agriculture" in American Journal of Economics and Sociology, vol. 37, no. 2 (April 1978).

Mishan, E.J. Cost-Benefit Analysis: An Informal Introduction, 3rd ed. (Boston: George Alien and Unwin, 1982).

Organization of American States. Integrated Regional Development Planning: Guidelines and Case Studies from OAS Experience (Washington, D.C.: Organization of American States, 1984)

Pandey, S. Incorporating Risk in Project Appraisal: A Case Study of a Nepalese Irrigation Project, A/D/C - APROSC, Research Paper Series #18 (Kathmandu, Nepal: March 1983).

United Nations Economic Commission for Latin America (UN/ECLA). Ecuador: Evaluation of the Effect of the 1982/83 Floods on Economic and Social Development (May 1983).

United Nations Disaster Relief Organization. Case Report on Hurricanes David and Frederick in the Dominican Republic (Geneva: UNDRO, 1980).

U.S. Agency for International Development, Office of Foreign Disaster Assistance. Countries of the Caribbean Community (Washington, D.C.: USAID/OFDA, 1982).

- Disaster History. Significant Data on Major Disasters Worldwide, 1990-Present. (Washington, DC: USAID/OFDA, 1989).

Vira, C., and Haimes, Y.Y. Multi-objectives Decision Making: Theory and Methodology (New York: North Holland, 1983).

World Bank. Memorandum on Recent Economic Development and Prospects of Honduras (Washington, D.C.: World Bank, 1979).

Young, D.L "Risk Preferences of Agricultural Producers: Their Use in Extension and Research" In American Journal of Agricultural Economics, vol. 61 (1979), pp. 1063-1070.

Previous Page Top of Page Next Page