Impact Evaluation of Forest Conservation Programs: Benefit-Cost Analysis, Without the Economics

Abstract

Economists are increasingly using impact evaluation methods to measure the effectiveness of forest conservation programs. Theoretical analysis of two complementary economic models demonstrates that the average treatment effect on the treated (ATT) typically reported by these studies can be related to an economic measure of program performance only under very restrictive conditions. This is because the ATT is usually expressed in purely physical terms (e.g., avoided deforestation) and ignores heterogeneity in the costs and benefits of conservation programs. For the same reasons, clinical trials are a misleading analogy for the evaluation of conservation programs. To be more useful for economic analyses of conservation programs, impact evaluations should work toward developing measures of program outcomes that are economically more relevant, data that would enable the evaluation of impacts on forest degradation (not just deforestation) and primary forests (not forests in general), better estimates of spatially disaggregated treatment effects (not program-wide averages), and better information on the accuracy of estimated treatment effects as predictors of future risks.

This is a preview of subscription content, log in to check access.

Notes

  1. 1.

    Experimental methods have also been used to evaluate conservation programs, but not as commonly. For an early example, see The and Ngoc (2006).

  2. 2.

    Several impact evaluations of conservation programs have used poverty alleviation as an outcome measure (Sims 2010; Andam et al. 2010; Ferraro et al. 2011; Ferraro and Hanauer 2014).

  3. 3.

    One could instead employ a continuous outcome measure, such as the proportion of a spatial unit that was deforested.

  4. 4.

    The assumption that the evaluator has information on site-specific values is a big one, of course, easier to make in a conceptual setting than in practice. Improvements in spatial datasets and ecosystem service models are making the estimation of conservation benefits at large spatial scales increasingly possible, however. For examples, see Naidoo and Ricketts (2006) and Bateman et al. (2013).

  5. 5.

    It is well-known that maximizing a benefit-cost ratio does not necessarily lead to the same ranking of alternative investment projects as maximizing NPV or NFV. Inspection of (7) indicates that the conservation program that maximized the benefit-cost ratio would be the same as the one that maximized NPV or NFV if the two programs were constrained to have the same aggregate cost. Otherwise, the programs could differ. In fact, in the simple context of (7), maximizing the benefit-cost ratio would imply protecting just the single site with the highest deforestation probability, which is hardly a reasonable conservation policy recommendation.

  6. 6.

    I note in passing that the existence of a feasible program with an average avoided deforestation rate higher than the agency’s program does not necessarily imply that the agency behaved irrationally by selecting the latter program. The agency’s program might have been the rational choice given the ex ante information on expected deforestation rates that was available to the agency when it selected sites for the program.

  7. 7.

    See Syrbe and Walz (2012) for a general discussion of spatial heterogeneity in the values of services provided by forests and other ecosystems. The most careful empirical analysis of the relationship between land cover, including forest cover, and the spatial distribution of multiple ecosystem services is probably a study by Eigenbrod et al. (2010) for England, which found that land cover (such as area forested) was a poor proxy for the value of services provided.

  8. 8.

    Costello and Polasky (2004, Table 1) provide a simple example that nicely illustrates this point. Their example includes three sites, which offer different benefits (numbers of species present) and face different conversion probabilities. Protection costs are uniform across the sites. The authors demonstrate that the dynamically optimal decision is to select the site that has the intermediate conversion probability, not the highest probability.

  9. 9.

    A more realistic model would include harvesting of roundwood (fuelwood or timber) in the country’s forests, as an economic activity that generates a valuable extractive resource but degrades the forests’ conservation value. The model in this section ignores degradation resulting from wood harvesting because nearly all impact evaluations of forest conservation programs have also ignored it. I return to this point in the final section of the paper.

  10. 10.

    The implicit assumption here is that a donor country randomly selects the area to be protected in a forested country, which is the only possible selection procedure in this aspatial model.

  11. 11.

    Because the donor country selects a single area for protection, unlike in Sect. 2 the deforestation rate is not averaged across multiple protected areas. Put another way, the deforestation rate given by (9) is the average across the landscape within which the donor country randomly selects a single area to protect.

  12. 12.

    The current-value Hamiltonian for land conversion in the country is \(\pi \left( {K(t)} \right) -C\left( {\dot{K}(t)} \right) +\lambda \left( t \right) \dot{K}(t)\), where \(\lambda (t)\) is the costate variable. The costate variable gives the present value of current and future agricultural profits from a marginal unit of agricultural land. The first-order condition with respect to the control variable \(\dot{K}(t)\) is \({\lambda (t)=\partial C\left( {\dot{K}(t)} \right) }/{\partial \dot{K}(t)}\). Hence, the marginal value of agricultural land equals the marginal cost of conversion: forest is converted to agriculture up to the point where the long-run marginal benefit of conversion equals the long-run marginal cost.

  13. 13.

    The marginal benefit described here is a current value, received only in the period when the donor country makes its protection decision. In contrast, the marginal cost of protection described in the previous paragraph is a long-run value that is equivalent to the present value of current and future marginal agricultural profits. An apples-to-apples comparison requires expressing the marginal benefit to the donor as a present value, too. All that needs to be shown here, however, is that the ranking of long-run marginal benefits is the same as the ranking of current marginal benefits, and doing this does not require formal analysis. The present value of future benefits from a marginal forest area is higher in the forest-poor country than in the forest-rich country because economic conditions in the forest-poor country at time \(t\) are equivalent to those in the forest-rich country at a later time \(t+x\). So, period-by-period into the future, the current marginal benefit of forest protection cannot be lower in the forest-poor country than in the forest-rich country, and so neither can the long-run marginal benefit calculated at a common point (i.e., the same time period) in the two countries.

  14. 14.

    Similarly, I might have obtained different results if I had developed a more complicated model than the simple one in this section, but that would just reinforce this point: the relationship between the ATT and economically rational conservation decisions is context-dependent. The ATT does not necessarily provide a useful proxy for the latter.

  15. 15.

    A study on community-based forest management by Somanathan et al. (2009) provides a rare example of combining impact evaluation results with information on conservation costs.

  16. 16.

    Similarly, spatial data on the ranges of threatened species from IUCN (http://www.iucn.org/) or NatureServe (http://www.natureserve.org/) could be used to develop outcome measures related to reduced species loss. Species loss correlates less closely with biodiversity-related benefits than reduced carbon emissions correlate with climate-related benefits, however, because different species or assemblages of species do not necessarily have identical values.

  17. 17.

    I am grateful to an anonymous reviewer for very helpful suggestions regarding the issues discussed in this paragraph.

  18. 18.

    Ferraro, Hanauer, Miteva et al. (2013, p. 2) state, however, that “most [impact evaluations] treat ‘protection’ as if it were homogeneous.”

  19. 19.

    It follows that the ATTs estimated by these studies will necessarily be small in most cases: annual deforestation rates are only a few percentage points even in countries with “rapid” deforestation, and so nearly all of the land that was forested at the start of the interval will still be forested at the end, regardless of protection status. It also follows that a conservation program with an ATT \(=\) 0 during such a short interval could nevertheless reflect smart economic decisions: protection could generate a positive net return in the long run, after accounting for benefits and costs that occur beyond the interval evaluated.

  20. 20.

    For example, see the University of Texas at Austin’s Perry-Castañeda Library Map Collection (http://www.lib.utexas.edu/maps/ams/).

References

  1. Alix-Garcia JM, Shapiro EN, Sims KRE (2012) Forest conservation and slippage: evidence from Mexico’s national payments for ecosystem services program. Land Econ 88:613–638

    Article  Google Scholar 

  2. Andam KS, Ferraro PJ, Sims KRE et al (2010) Protected areas reduced poverty in Costa Rica and Thailand. Proc Natl Acad Sci USA 107:9996–10001

    Article  Google Scholar 

  3. Ando A, Camm J, Polasky S et al (1998) Species distributions, land values, and efficient conservation. Science 279:2126–2128

    Article  Google Scholar 

  4. Angrist JD, Pischke JS (2009) Mostly harmless econometrics: an empiricist’s companion. Princeton University Press, Princeton

    Google Scholar 

  5. Asner GP, Powell GVN, Mascaroet J et al (2010) High-resolution forest carbon stocks and emissions in the Amazon. Proc Natl Acad Sci USA 107:16738–16742

    Article  Google Scholar 

  6. Bateman IJ, Harwood AR, Mace GM et al (2013) Bringing ecosystem services into economic decision-making: land use in the United Kingdom. Science 341(6141):45–50

    Article  Google Scholar 

  7. Blackman A (2013) Evaluating forest conservation policies in developing countries using remote sensing data: an introduction and practical guide. For Policy Econ. doi:10.1016/j.forpol.2013.04.006

  8. Brauman KA, Daily GC, Duarte TK et al (2007) The nature and value of ecosystem services: an overview highlighting hydrologic services. Annu Rev Env Resour 32:67–98

    Article  Google Scholar 

  9. Chiang AC (1992) Elements of dynamic optimization. Waveland Press, Long Grove

    Google Scholar 

  10. Costello C, Polasky S (2004) Dynamic reserve site selection. Resour Energy Econ 26:157–174

    Article  Google Scholar 

  11. Deaton A (2010) Instruments, randomization, and learning about development. J Econ Lit 48:424–455

    Article  Google Scholar 

  12. Eigenbrod F et al (2010) The impact of proxy-based methods on mapping the distribution of ecosystem services. J Appl Ecol 47:377–385

    Article  Google Scholar 

  13. Eisner R, Strotz RH (1963) Determinants of business investment. In: Suits DB et al (eds) Impacts of monetary policy. Prentice-Hall, Englewood Cliffs

    Google Scholar 

  14. FAO (2010) Global forest resource assessment 2010: Main report. Rome

  15. Ferraro PJ, Hanauer MM (2011) Protecting ecosystems and alleviating poverty with parks and reserves: ‘win-win’ or tradeoffs? Environ Resour Econ 48:269–286

    Article  Google Scholar 

  16. Ferraro PJ, Hanauer MM (2014) Quantifying causal mechanisms to determine how protected areas affect poverty through changes in ecosystem services and infrastructure. Proc Natl Acad Sci USA. doi:10.1073/pnas.1307712111

  17. Ferraro PJ, Hanauer MM, Miteva DA et al (2013) More strictly protected areas are not necessarily more protective: evidence from Bolivia, Costa Rica, Indonesia, and Thailand. Environ Res Lett. doi:10.1088/1748-9326/8/2/025011

  18. Ferraro PJ, Hanauer MM, Sims KRE (2011) Conditions associated with protected area success in conservation and poverty reduction. Proc Natl Acad Sci USA 108:13913–13918

    Article  Google Scholar 

  19. Ferraro PJ, Lawlor K, Mullan KL et al (2012) Forest figures: ecosystems services valuation and policy evaluation in developing countries. Rev Environ Econ Policy 6:20–44

    Article  Google Scholar 

  20. Ferraro PJ, Pattanayak SK (2006) Money for nothing? A call for empirical evaluation of biodiversity conservation investments. PLoS Biol 4:482–488

    Article  Google Scholar 

  21. Geist HJ, Lambin EF (2001) What drives tropical deforestation? A meta-analysis of proximate and underlying causes of deforestation based on subnational case study evidence. LUCC Report Series No. 4, Department of Geography, University of Louvain, Louvain-la-Neuve, Belgium

  22. Gibson L, Lee TM, Koh LP et al (2011) Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478:378–381

    Article  Google Scholar 

  23. Imbens GW, Wooldridge JM (2009) Recent developments in the econometrics of program evaluation. J Econ Lit 47:5–86

    Article  Google Scholar 

  24. Joppa L, Pfaff A (2010) Global protected area impacts. Proc Roy Soc B Biol Sci. doi:10.1098/rspb.2010.1713

  25. Kennedy P (2008) A Guide to econometrics, 6th edn. Blackwell, Malden

    Google Scholar 

  26. Le Saout S, Hoffmann M, Shi Y et al (2013) Protected areas and effective biodiversity conservation. Science 342:803–805

    Article  Google Scholar 

  27. Mackey B, DellaSala DA, Kormos C et al (2014) Policy options for the world’s primary forests in multilateral environmental agreements. Conserv Lett. doi:10.1111/conl.12120

  28. Miteva DA, Pattanayak SK, Ferraro PJ (2012) Evaluation of biodiversity policy instruments: what works and what doesn’t? Oxford Rev Econ Policy 28:69–92

    Article  Google Scholar 

  29. Miteva DA, Murray B, Pattanayak SK (2014) Do protected areas reduce blue carbon emissions? A quasi-experimental evaluation of mangroves in Indonesia. Manuscript, Duke University

  30. Myers N (1988) Threatened biotas: “hot spots” in tropical forests. Environmentalist 8:187–208

    Article  Google Scholar 

  31. Myers N, Mittermeier R, Mittermeier C et al (2000) Biodiversity hotspots for conservation priorities. Nature 403:853–858

    Article  Google Scholar 

  32. Naidoo R, Balmford A, Ferraro PJ et al (2006) Integrating economic costs into conservation planning. Trends Ecol Evol 21:681–687

    Article  Google Scholar 

  33. Naidoo R, Ricketts T (2006) Mapping the economic costs and benefits of conservation. PLoS Biol 4:2153–2164

    Article  Google Scholar 

  34. Pattanayak SK, Kramer R (2001) Worth of watersheds: a producer surplus approach for valuing drought mitigation in Eastern Indonesia. Environ Dev Econ 6:123–146

    Article  Google Scholar 

  35. Pattanayak SK, Wunder S, Ferraro PJ (2010) Show me the money: do payments supply environmental services in developing countries? Rev Environ Econ Policy 4:254–274

    Article  Google Scholar 

  36. Pearce D, Barbier E, Markandya A (1990) Sustainable development: economics and environment in the third World. Earthscan, London

    Google Scholar 

  37. Pfaff A, Robalino J, Sanchez-Azofeifa GA et al. (2009) Park location affects forest protection: land characteristics cause differences in park impacts across Costa Rica. BE J Econ Anal Policy 9: article 5

  38. Polasky S (2008) Why conservation planning needs socioeconomic data. Proc Natl Acad Sci USA 105:6505–6506

    Article  Google Scholar 

  39. Polasky S, Camm JD, Garber-Yonts B (2001) Selecting biological reserves cost-effectively: an application to terrestrial vertebrates in Oregon. Land Econ 77:68–78

    Article  Google Scholar 

  40. Polasky S, Costello C, Solow A (2005) The economics of biodiversity. In: Maler KG, Vincent JR (eds) Handbook of environmental economics, vol 3. North-Holland, Amsterdam

    Google Scholar 

  41. Polasky S, Nelson E, Camm J et al (2008) Where to put things? Spatial land management to sustain biodiversity and economic returns. Biol Conserv 141:1505–1524

    Article  Google Scholar 

  42. Ravallion M (2008) Evaluating anti-poverty programs. In: Schultz T, Strauss J (eds) Handbook of development economics, vol 4. North-Holland, Amsterdam

    Google Scholar 

  43. Robalino JA, Pfaff A (2012) Contagious development: neighbor interactions in deforestation. J Dev Econ 97:427–436

    Article  Google Scholar 

  44. Robalino JA, Pfaff A (2013) Ecopayments and deforestation in Costa Rica: a nationwide analysis of PSA’s initial years. Land Econ 89:432–448

    Article  Google Scholar 

  45. Saatchi SS, Harris NL, Brown S et al (2011) Benchmark map of forest carbon stocks in tropical regions across three continents. Proc Natl Acad Sci USA 108:9899–9904

    Article  Google Scholar 

  46. Sims KRE (2010) Conservation and development: evidence from Thai protected areas. J Environ Econ Manage 60:94–114

    Article  Google Scholar 

  47. Somanathan E, Prabhakar R, Singh Mehta BS (2009) Decentralization for cost-effective conservation. Proc Natl Acad Sci USA 106:4143–4147

    Article  Google Scholar 

  48. Syrbe RU, Walz U (2012) Spatial indicators for the assessment of ecosystem services: providing, benefiting and connecting areas and landscape metrics. Ecol Indic 21:80–88

    Article  Google Scholar 

  49. The BD, Ngoc HB (2006) Payments for environmental services in Vietnam: assessing an economic approach to sustainable forest management. EEPSEA Research Report 2006-RR3, Economy and Environment Program for Southeast Asia, Singapore

  50. Vincent JR, Rozali MA (2005) Managing natural wealth: environment and development in Malaysia. Resources for the Future Press, Washington

    Google Scholar 

  51. Wong IFT (1971) The present land use of West Malaysia (1966). Ministry of Agriculture and Lands, Kuala Lumpur

    Google Scholar 

Download references

Author information

Affiliations

Authors

Corresponding author

Correspondence to Jeffrey R. Vincent.

Appendix

Appendix

Differentiating (9) with respect to time yields:

$$\begin{aligned} \frac{\ddot{K}\left( t \right) \left( {L-K(t)} \right) +\dot{K}(t)^{2}}{\left( {L-K(t)} \right) ^{2}} \end{aligned}$$
(10)

Chiang (1992) demonstrates that the following relationship holds along the optimal path,

$$\begin{aligned} K\left( t \right) =\overline{{K}}-\left( {\overline{{K}}-K(0)} \right) e^{r_2 t}, \end{aligned}$$

where \(r_{2}\) is a negative constant, \(\frac{1}{2}\left( {\rho -\sqrt{\rho ^{2}+\frac{4\beta }{a}}} \right) \). This simplifies to \(K(t)=\overline{{K}}\left( {1-e^{r_2 t}} \right) \) if \(K\)(0) \(=\) 0 (all land is initially in forest). It follows that \(\dot{K}(t)=-r_2 e^{r_2 t}\overline{{K}}\) and \(\ddot{K}\left( t \right) =-r_2 ^{2}e^{r_2 t}\overline{{K}}\).

Substituting these expressions into (10) and simplifying, we obtain

$$\begin{aligned} \frac{r_2 \dot{K}(t)\left( {L-\overline{{K}}} \right) }{\left( {L-K(t)} \right) ^{2}}. \end{aligned}$$

This expression is negative because \(r_{2}\) is negative and all the other terms are positive. Hence, the deforestation rate declines over time.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Vincent, J.R. Impact Evaluation of Forest Conservation Programs: Benefit-Cost Analysis, Without the Economics. Environ Resource Econ 63, 395–408 (2016). https://doi.org/10.1007/s10640-015-9896-y

Download citation

Keywords

  • Conservation economics
  • Deforestation
  • Degradation
  • Impact evaluation
  • Primary forest
  • Protected area