This perspective explores in greater depth part of the cost-benefit analysis undertaken by Markandya, particularly for Sustainable Development Goal 15.2: By 2020, promote the implementation of sustainable management of all type of forests, halt deforestation, restore degraded forests, and increase afforestation and reforestation by x percent globally.

Although the implementation costs of measures to achieve the Aichi targets have been estimated during a specific program of work, the same is not the case for the benefits. Looking at ecosystem services for both tropical and temperate forest biomes, we find that the largest category for tropical forest is provisioning services (43 percent of value), while for temperate forest 49 percent relates to cultural services (largely recreation).

The majority of tropical forest valuations use market value type estimates. By contrast, nearly 60 percent of the temperate forest values are from stated preference techniques (contingent valuation, choice experiments). This raises a number of important issues; arguably the key ones with respect to the use of these extant values are passive use values and nonmarginal values.

The estimates collected for valuation databases typically reflect marginal changes in the provision of ecosystem services. They are not appropriate for estimating the total value of an ecosystem. Put simply, can we defensibly use a $/ha value originally elicited for changing forest extent at one site by 10 ha to value changes totalling thousands of hectares on a global scale? Nevertheless, existing value estimates are likely to be the only reasonable source for global cost-benefit analysis.

We can estimate benefits on a regional basis by using spatial variables to account for differences. Value functions can be used to allow for the impact of different variables. For example, smaller sites have higher values per hectare, consistent with diminishing marginal utility, and the extent of urban areas within 50 km also increases per ha values of forest. Values across regions also reflect variables such as income levels.

The total losses from deforestation from 2000 to 2050 reported by Markandya range from US$334 billion to US$1,118 billion, compared to a total undiscounted value of US$1,322 billion implied by the value functions in this perspective.