5. From theory to practice

Although the PES approach is intuitively appealing, putting it into practice is far from simple (Pagiola and Platais, 2003). The theoretical elegance of a blackboard concept needs to be translated into actual implementation arrangements on the ground. The remainder of this paper describes the approach adopted to do so by the RISEMP project.

The RISEMP, which began implementation in July 2002, is seeking to pilot the use of the payment for environmental services approach to encourage the adoption of silvopastoral practices in degraded pastures areas in Central and South America (World Bank, 2002). The project is being implemented in three microwatersheds: Quindío, in Colombia; Esparza, in Costa Rica; and Matiguás-Río Blanco, in Nicaragua. Participating land users enter into contracts under which they receive a payment for the environmental services that they generate. They receive annual payments over a two- or four-year period, based on the increment in environmental services provided relative to the baseline situation for that particular farm. Through this mechanism, the project aims to establish silvopastoral systems on 3,500 ha, thus enhancing the environmental benefits generated in watersheds covering about 12,000 ha.

The project was prepared with support of the multi-donor Livestock, Environment and Development Initiative (LEAD), hosted by the Food and Agriculture Organisation (FAO). It is financed by a US$4.5 million GEF grant, with the World Bank acting as implementing agency. In each country, field activities are being undertaken by local non-governmental organizations (NGOs): the Centre for Research on Sustainable Agricultural Production Systems (Centro para la Investigación en Sistemas Sostenibles de Producción Agropecuaria, CIPAV) in Colombia, the Center for Teaching and Research on Tropical Agronomy (Centro Agronómico Tropical de Investigación y Enseñanza, CATIE) in Costa Rica, and Nitlapan in Nicaragua, with CATIE coordinating the work. The American Bird Conservancy (ABC) is providing technical assistance for the development of a common and consistent methodology for the monitoring of biodiversity at the three project sites.

What is being paid for?

Contracting for land users to provide biodiversity benefits is all very well in theory, but in practice it is clearly unrealistic to ask them to deliver biodiversity. A way is needed to communicate what is desired to potential participants in ways that they can understand. The typical solution has been to offer to pay not for biodiversity itself, but for land uses that are hospitable to biodiversity (Pagiola and others, 2002).

But land use alone can be a relatively blunt instrument. In Costa Rica’s PSA program, for example, most contracts call for conservation of existing forest, and pay all participants the same amount (FONAFIFO, 2000; Pagiola, 2002). While this approach has the virtue of simplicity, it fails to recognize the very different levels of services that different land uses can provide. The biodiversity-friendliness of agricultural practices is not a binary, yes/no proposition. On the contrary, there is a spectrum of effects, ranging from relatively inhospitable systems such as monocultures with heavy agrochemical use to relatively hospitable systems such as organic coffee grown under a diverse shade canopy of native species. Location also matters: biodiversity-friendly practices in proximity to protected areas, for example, might be more valuable by helping to buffer and protect them. Failing to take these differences into account risks either under-paying for desirable land uses, or over-paying for relatively less desirable ones (Pagiola and Platais, forthcoming).

The solution adopted in the RISEMP was to prepare a list of land uses and associate each with a point system upon which payments are based. This approach is similar to that of the Environmental Benefits Index (EBI) used in US Conservation Reserve Program (CRP) (NCEE, 2001). Separate indices were developed for the biodiversity conservation and carbon sequestration benefits of each land use. These two indices were then aggregated to form an environmental service index to be employed as the basis for calculating payments to participants. A similar index for water benefits was not included, partly because of the lack of data needed to develop it, and partly because improved water flows would be national benefits, and thus are not eligible for GEF funding. The biodiversity conservation and carbon sequestration indices are presented in Table 4.

The biodiversity conservation index was scaled with the most biodiversity-poor land use (annual crops) set at 0.0 and the most biodiversity-rich land use (primary forest) set at 1.0. Within this spectrum, the points given to each specific land use were set by a panel of experts, taking into consideration factors such as the number of species (of plants, birds, small mammals, and insects), their spatial arrangement, stratification, plot size, and fruit production. Higher scores were given to land uses that have greater potential to maintain the original biodiversity of the region. Note that the index estimates the environmental benefits of all land uses, and not only silvopastoral practices.

Table 4. Environmental service indices used in the RISEMP
(Points per hectare, unless otherwise specified)

This approach can take into consideration the different impact that different land uses are likely to have on biodiversity. There are, of course, limitations. The biodiversity impact depends not only on the characteristics of the land use, but also on its location, its extent, and its relationship to other land uses. At the pilot scale of the RISEMP, issues of location are not significant, as all three pilot areas were specifically chosen for their proximity to protected areas or to corridors between them. All three project areas were selected in part because of their location in ecologically-sensitive areas. The Quindío project site is in one of the most severely degraded regions of Colombia, with few, mostly unconnected remnants of natural habitats. Restoring a degree of habitat heterogeneity and connectivity would increase the chances of survival of species requiring large home ranges in an area considered as a priority for bird conservation. The Esparza area in Costa Rica is in the vicinity of conservation areas such as La Fortuna, the Monteverde Reserve complex, and the Alberto Brenes Biological Reserve. More biodiversity-friendly land use practices would help the chances of survival of several species occurring in these protected areas. The Matiguás-Río Blanco watershed in Nicaragua is part of the buffer zone of the Cerro Musún Natural Reserve, and is very close to one of the priority areas for bird conservation in the country. If this approach were to be scaled up and applied on a broader scale, location effects could be incorporated either by varying the points for activities in different locations or by varying the payment per incremental point. Issues of scale and contiguity are harder to address. Some biodiversity benefits may be obtained only after appropriate land uses cover a minimum area, or if the areas covered are contiguous rather than scattered. To an extent, these effects might be addressed by adding bonus points if the area covered by a given land use passes a threshold. Such an approach could quickly result in an excessively complex point system, however. Another approach would be to set a minimum participation threshold for the PES program to take effect; this approach was followed by New York City, for example (A. Appleton, pers. comm.)

A similar procedure was used to establish the carbon sequestration index, with different land uses given points according to their capacity to sequester stable carbon in the soil and in hard wood through the years. Recent studies indicate that secondary forest can fix an average of 10 tonnes of carbon per year in wood and in the soil. As secondary forest has a value of 1.0 in the index, 0.1 points correspond to an estimated sequestration of 1 tonne of carbon. Data from studies conducted by CATIE were used to calibrate the carbon sequestration index.

As data were insufficient to derive country-specific indices, the same index is being used in all three countries. Data from the monitoring efforts will be used to improve the indices, and it is expected that these will differ from country to country.

Should downstream water users be willing to pay for hydrological services, the approach could also be extended by adding an index denoting the contribution of each land use type to the desired water services, though developing such an index would certainly prove difficult.

Note that under RISEMP, biodiversity and carbon sequestration benefits are given equal weight in calculating payments. The two indices could easily be de-coupled, however, with separate payment levels for each kind of environmental service. Alternatively, different weighting schemes could be used to give proportionally more weight to one or the other, depending on the interest of those making the payments.

This index approach was tested with potential participants, and is proving quite intelligible to them in practice. Dissemination materials such as posters and handbooks have been prepared showing precisely what the payments would be for specific land uses.

How should payments be made?

The second challenge in developing an appropriate contract is the need to understand the economics of the farming system, so that the appropriate amount and form of payment can be determined. Payments for environmental services will have the desired effect only if they reach the land users in ways that influence their decisions on how to use the land.

Analysis of the time path of benefits generated by silvopastoral systems showed that they are unattractive to land users primarily because of their substantial initial investment, and because of the time lag between investment and returns, as shown in Figure 1 above. This leads to the hypothesis that a relatively small payment provided in the early period of adoption would be sufficient to ‘tip the balance’ between current and silvopastoral systems. This effect works by increasing the net present value of investments in silvopastoral practices, but also by reducing the initial period in which adoption of these systems imposes net costs on land users. By the time payments end, the silvopastoral practices themselves are ready to begin generating income for land users. The payments also alleviate the liquidity problems faced by many land users and help them finance the required investments.

Based on this analysis, it was decided to provide a relatively small, up-front payment to participating land users. This payment is of US$75 per incremental point, per year over a four-year period, up to a maximum of US$4,500 per farm (US$6,000 in Colombia, where input prices are higher). Both of these aspects deserve further discussion.

In principle, the amount should be no less than the land users’ opportunity cost (or they will not participate), and no more than the value of the benefit provided (or it would not be worthwhile to provide the service). In practice, the actual value of the benefit provided is extremely difficult to estimate, and particularly so for benefits such as biodiversity conservation. In contrast, the farmers’ opportunity cost can usually be estimated relatively easily. For this reason, as well as to limit the budgetary requirements of the payment, payment levels are usually set at slightly more than the opportunity cost of the main alternative land uses. All the existing systems of payments for environmental services implicitly or explicitly use this approach. Costa Rica’s PSA program, for example, currently pays US$45/ha/year for forest conservation. This payment has proven to be quite attractive, with far more applications for this contract than the program has been able to finance. (In contrast, a payment of US$538/ha over 5 years for reforestation has proven to be less popular, as many landowners consider the payment offered insufficient to justify the investment.) In Mexico, a specific study was commissioned of the opportunity cost of land (Jaramillo, 2003) to provide a basis for payments levels under the PSAH; no study was made of the magnitude of benefits. Zelek and Shively (2003) propose a scheme to pay the opportunity costs of Philipino farmers who adopt practices that sequester carbon. Paying the opportunity cost of adopting the desired practice also accords well with GEF’s policy of paying for the incremental costs of generating global environmental benefits.

In terms of payments for carbon emissions reductions, the US$75/point/year payment level is equivalent to paying US$7.5 per tonne of carbon sequestered, or US$2 per tonne of CO₂ equivalent. This compares favorably to current world prices for carbon emissions reduction of US$3-5 per tonne of CO₂ equivalent (World Bank, 2003b)—although these payments typically require a higher degree of assurance of the permanence of the emissions reduction and a more intensive monitoring regime than the RISEMP offers. No similar comparison is possible for payments for biodiversity conservation.

In general, emerging guidelines for payments for environmental services indicate that payments should be on-going rather than finite (Pagiola and Platais, forthcoming). In Costa Rica’s PSA program, for example, payments for forest conservation contracts are for 5 years, but they are renewable indefinitely by mutual agreement. The logic for this is simple: if environmental services are to be generated over a long period of time (presumably, indefinitely), then payments for these services should also be made over a similarly long period. Ending payments sooner creates the risk that land users will revert to their previous land use practices. This is a risk that has been observed time and time again in projects that attempted to change land use practices, such as soil conservation or reforestation projects (Lutz and others, 1994). This risk was thought to be relatively low in this instance, as the silvopastoral practices, once established, are privately more profitable (see Figure 1). Moreover, the payments represent only a small portion of the necessary investment costs, thus making it unlikely that land users would adopt practices they intend to abandon solely to receive the payments. In an effort to determine the long-term sustainability of the mechanism, a sub-group of participants is being given a slightly modified contract, in which the payments are frontloaded: rather than receiving them over a four-year period, farmers with this alternative contract will receive a similar amount over a two-year period. Farmers were assigned randomly to one or the other contract.

Table 5 illustrates the application of this contract for the 20 ha farm in Matiguás, Nicaragua, used in the previous example. In the baseline year, the farm has 2.5 ha under annual crops, 14.5 ha under natural pasture without trees, and 3 ha under brush (tacotal). Motivated by the project, it converts 3 ha of its pasture to higher tree densities: 1 ha a year for the first three years. It also plants an 0.5 ha fodder bank, and fences off the scrub areas so that secondary forest can regenerate. Finally, it plants trees along 1.5 km of its fence lines. Using the environmental service index in Table 4, the resulting scores can be calculated for the baseline and for each subsequent year. These scores are then used to compute the payments due to the farmer, including the initial baseline payment for existing services (see below) and the main payment for incremental services provided under the project. Figure 2 shows the impact of these payments on the time profile of benefits to adopting silvopastoral practices, and the resulting impact on the profitability of the investment. What had been a marginally viable investment now becomes more attractive.

Table 5. Example of payment computation

Figure 2. Effects of PES on the profitability of silvopastoral systems

Note: 20ha farm in Matiguás, Nicaragua

Avoiding perverse incentives

The initial plan involved paying land users only for incremental improvements in land use practices. The extent to which land users had already adopted practices that conserved biodiversity or sequestered carbon prior to the project was to have been reflected in their baseline environmental service index, and only increments to this index were to be compensated. It soon became clear that this approach entailed a substantial risk of creating perverse incentives. “Bueno, corto todo,” was a common reaction by land users when told they would not be compensated for pre-existing trees: “fine, I’ll cut them all.” It might have been possible to avoid this risk among project participants by imposing contractual restrictions on such actions, though this would certainly have required an increased monitoring effort, and thus increased costs. But there was also a broader risk that non-participants in surrounding areas would postpone adopting silvopastoral practices that they might have been tempted to adopt, so that they might wait for a project to come and compensate them for doing so. As a result, the initial plan was modified to allow for a payment to be made for pre-existing environmental services. A one-time payment of US$10/point will be made for the baseline points, up to a maximum of US$500 per farm. This payment has the further benefit of helping to alleviate financing constraints to implementing silvopastoral practices.

As part of the effort to avoid perverse incentives, the contract also specifies that burning in pastures is banned (except in areas devoted to food security, where burning is allowed in the first two years), and that the contract will be terminated if the participants cut down primary or secondary forest in their farms.

A related problem is that of minimizing ‘leakage’—that is, avoiding environmentally-damaging activities being simply displaced, so that there is little net benefit. The RISEMP minimizes this problem by computing the points on which payments are made over the entire farm and basing payments on the net points over the entire farm. If land users cut down trees in one plot even as they plant them in another, the negative points earned from the adverse change will offset those gained from the positive change.