0769-A1

Valuing Environmental Benefits of Agroforestry Practices in the Lake Okeechobee Watershed, Florida

Ram K. Shrestha and Janaki R.R. Alavalapati^[1]

Abstract

Silvopasture practice in the Lake Okeechobee watershed is described as an environmentally benign land-use that generates non-market benefits by limiting phosphorus runoff, sequestering atmospheric carbon and protecting wildlife habitat. We estimated household willingness to pay (WTP) for these public goods using stated preference method. The multinomial logit model results reveal that household WTP for improvements in the above environmental attributes is up to US$120.24 per year. With appropriate policy incentives, ranchers could adopt these practices and contribute to the improvement of environmental quality in the watershed. Our estimates provide key information to the government agencies and policy-makers in designing incentive policies to encourage agroforestry practices.

1. Introduction

Lake Okeechobee is a large and shallow fresh water lake with a surface area of 730 sq. miles. It is the largest freshwater lake in Florida and second largest in the U.S. after the Great Lakes (EPA 2000). As a source of drinking water, Lake Okeechobee is designated as a Class I water body. In addition, it is also a source of irrigation and ground water, habitat for fishes and waterfowls, as well as an avenue for flood control, navigation, and recreational uses.

The northern watershed of Lake Okeechobee extends nearly 5,000 sq. miles at the heart of the inter-connected ecosystems including the Kissimmee River to the north and the Everglades to the south (EPA-SFWMD 1999). Current land-use in the Lake Okeechobee watershed is primarily beef cattle ranches (65%) and dairy operations (3%) (Boggess et al., 1995). Improved and semi-improved pasture management are the key components in these livestock operations.^[2] Out of 7 million acres of total cattle pasture and ranchlands in Florida, over 5 million acres are improved pasture (FCA 1999). Florida currently has 1.8 million cows and calves and ranks 10^th in all states (3^rd in states east of the Mississippi River) for beef cattle herd size. Approximately, 66% of Florida's beef cattle are raised in 25 counties in central and south Florida (FAS 2000).

The freshwater ecosystem of the Lake Okeechobee in south-central Florida is degraded due to excessive phosphorus concentration originated primarily from cattle ranches in its northern watershed. The growth of excessive aquatic weeds was first recorded in the Lake Okeechobee in late 1960s. Since this time, many empirical studies have been conducted on the quality of lake waters (Boggess et al. 1995; Reddy et al. 1995; EPA-SFWMD 1999; SFWMD 2002; Zhang and Essex 1997). These studies consistently found that high input of phosphorus was the main cause of eutrophication in this lake, providing appropriate conditions for surface bloom of blue-green algie (Boggess et al. 1995). Furthermore, the special issue of Ecological Engineering, 5(2-3), 1995 presented ample evidence that the excess phosphorus in the lake originated primarily in the northern watershed of Lake Okeechobee.

Literature suggests that silvopasture, an agroforestry technology which combines trees, forages, and shrubs with livestock operation, provides environmental benefits such as water quality improvement, soil conservation, carbon sequestration, wildlife habitat protection, and aesthetics (Alavalapati and Nair 2001; Shrestha et al. 2002; Clason and Sharrow 2000; Garrett et al. 2000; Gold et al. 2000; Kurtz 2000). These benefits are considered public goods and private ranches will have little motivation to supply them at optimal levels unless market incentives are established.

In this study, we propose to estimate the value of public goods supplied by private ranchers through silvopasture land-use. We elicit public willingness to pay for the improvement in water quality, carbon sequestration, and wildlife habitat in the Lake Okeechobee watershed. Economists routinely use non-market valuation techniques to estimate the willingness to pay for public goods (Mitchell and Carson 1989). In particular, contingent valuation method (CVM) has been extensively used to directly estimate the public willingness to pay for public goods. The application of CVM and other stated preference methods in estimating non-market value of public goods are widely reported in economics literature (Adamowicz et al. 1998; Bateman and Willis 1999). However, the public goods supplied through environmentally beneficial agroforestry land-use are rarely valued. We use stated preference method of contingent valuation to estimate private provision of public goods. In particular, we estimate the environmental benefits of silvopasture land-use in the northern watershed of Lake Okeechobee, Florida. The public willingness to pay for improvements in water quality, carbon sequestration, and wildlife habitat in the watershed are elicited using the choice experiment design of stated preference method.

The remaining of this paper is organized as follows. Section 2 that follows describes the economic models and estimation methods. The survey design and data collection process are elaborated upon Section 3. Then the results and discussion are presented in Section 4, and finally conclusions are drawn in Section 5.

2. Method and Model Specification

The theoretical construct of stated preference choice experiment (SP-CE) approach is similar to dichotomous choice contingent valuation (DC-CV), but CE provides a more rigorous framework to WTP elicitation (Adamowicz et al. 1998; Hanley 2001; DeShazo and Fermo 2002). Beyond discrete choice elicitation, CE uses series of experiments in the valuation process. In CE, consumers make repeated tradeoffs between consumption of public goods and private consumption bundles subject to a budget constraint.

Random utility theory (RUT) provides the basis for CE valuation method. Under this model, each of the environmental attributes of the improved land-use in the watershed forms an alternative, j, in a choice set, c. Alternative j would be one specific type of consumption bundle representing an improvement in the environmental quality of the watershed with its conditional indirect utility level u_j for an individual household i expressed as

u_ij = v_ij + e_ij (1)

Where v_ij is the deterministic component of the model and e_ij is the random component (Louviere et al. 2000). Thus, selection of alternative j over alternative h implies that the utility of u_ij is greater than that of u_ih (Equation 2). This suggests that one can only analyze the probability of choice of one alternative over another. The probability of an individual i choosing alternative j, p(·), may be expressed as

p(ij|c) = p[u_ij > u_ih] = p[(v_ij + e_ij) > (v_ih + e_ih)] (2)

Assuming the error terms of the utility function are independently and identically distributed (IID) and follow an extreme-value (Weibull) distribution, the choice probabilities have a closed-form solution, and are appropriately estimated using a multinomial logit (MNL) specification (Adamowicz et al. 1998; Louviere et al. 2000). The MNL model structure represents the probability of choosing an alternative j such that the utility of the alternative is greater than the utility of all other alternatives.

Respondent's WTP representing the compensating surplus (CS) can be estimated using MNL model (Louviere et al. 2000; Adamowicz et al. 1998). A simplified method of estimating WTP for a level of change in environmental attributes is to take the ratio of the estimated coefficient of the attribute â_j and the coefficient of the cost attribute â_c. This ratio is commonly known as part-worth or utility representing the marginal value of a change in the attribute, i.e., the marginal rate of substitution between income change and the change in the attribute under consideration.

3. Survey design and data collection

We used choice experiment design approach of stated preference method to elicit pubic willingness to pay for environmental benefits. The experimental design was used to characterize public goods with various attributes and levels. The design was then presented in the survey with several valuation scenarios. Finally, respondents were asked to evaluate alternative bundles of goods (options) with different configuration of attribute levels (profile) including price attribute in each scenario and record their choices. Typically the respondents evaluate two options and a 'statue quo' option in one occasion (scenario), and continue the choice task for several scenarios as required by a particular design.

The state utility tax was used as the cost attribute in our experimental design. We used this as the payment vehicle because Florida prohibits state income tax that is commonly used in valuation surveys (Milon et al., 1999). The utility tax is a form of utility bill which is familiar to respondents and closely linked to the goods valued in the survey (Mitchell and Carson, 1989). Each environmental attribute was described using three levels, and tax attribute was detailed to six levels. We followed an orthogonal main effect experimental design representing 3³ x 6 x 2 factorials. Following the smallest orthogonal main effect plan we identified 12 pairs of profiles in two blocks using autocall macros of SAS version 8 (Kuhfeld et al., 1994).

In the survey, a map of the Lake Okeechobee watershed was presented. The rationale for selecting the study area shown in the map was also given in the survey questionnaire. First, a brief description of the traditional ranching, water pollution issues in the watershed, as well as potential environmental benefits and costs of silvopasture practices were outlined in the survey. Color photos and drawings were used to illustrate the silvopasture land-use changes. Second, the directions and an example of how to answer the survey questions were provided. Third, the respondents were given six choice tasks each with two options (A, B) and a status quo option (C) representing current condition of the Lake Okeechobee watershed, and asked to choose one of the options. The valuation scenarios were designed in a referendum format (Figure 1). The choice question was followed by a rating question that asked respondents to rate their confidence level in each choice task. Finally, the respondents were asked to provide their socioeconomic and demographic information.

Figure 1: Example of a choice set presented in choice experiment tasks.

Choice Scenario 1

Please vote for the plan that you prefer:

If the above plans are in a referendum, which one would you vote for?

Please circle one number to indicate how certain you are with the choice you have just made:

Survey sample was drawn randomly from the households of ten counties in south-central Florida, namely, Glades, Hendry, Highlands, Martin, Okeechobee, Orange, Osceola, Palm Beach, Polk, and St. Lucie. The Bureau of Economic and Business Research (BEBR), University of Florida assisted us with the sampling design. We mailed the survey packet to 504 households. The survey contained a questionnaire, cover letter, and a magnate sticker with School of Forest Resources and Conservation and University of Florida logos. Reminder post cards were sent to individuals who had not responded within two weeks. With second mailing and post card follow-ups we received total of 152 survey responses. There were 34 surveys undelivered and 7 incomplete leading to the response rate in our survey was just above 32%.

5. Results and Discussion

The responses were analyzed using multinomial logit models. Two levels of improvements in water quality, carbon sequestration, and wildlife habitat are estimated in the model. Qualitative attributes in the CE model are analyzed using effect codes, which allows us to estimate slope coefficients for the status-quo or excluded category of each qualitative attribute (Adamowicz, 1998). The variables included in the analysis are defined in Table 1.

Table 1: Variable definition

The signs and significance of the coefficients in the model are found to be consistent as expected. All slope coefficients on attribute variables, except the coefficient on higher level wildlife habitat attribute (WILDLIFE2), are significant at p<0.10. All attribute coefficients are positive suggesting a positive utility of the environmental quality attributes. The cost variable (TAX) representing state utility tax has negative and highly significant coefficient as expected, implying a higher tax reduces the utility of the environmental quality improvement (Table 2).

Table 2: The MNL model with economic and behavioral factors.

** is coefficient significant at p<0.05, * is coefficient significant at p<0.10.

The implicit prices or part-worth utilities of the improvement in environmental attributes are derived using MNL model results. We estimate WTP value for water quality, carbon sequestration, and wildlife habitat improvements as reported in Table 3.

Table 3: Welfare estimate for environmental improvement (in US $ per household)

Using the MNL model we estimate an annual aggregate value of $120.24 and $116.44 per household for the moderate and higher level improvements, respectively. A slightly lower WTP value for higher level improvement may be due to non-linearity in the utility function associated with carbon sequestration and wildlife habitat attributes.

6. Conclusions

The manner in which cattle ranches are managed in the Lake Okeechobee watershed has a serious impact on downstream environmental quality. Water quality in the Lake Okeechobee has deteriorated due to excess phosphorus loading from the private pasture and ranchlands. South Florida Water Management Districts (SFWMD) and the U.S. Environmental Protection Agency (EPA) have developed action plans to mitigate the lake water eutrophication through reducing the total maximum daily load (TMDL) of phosphorus from 110 to 40 parts per billion (ppb) and limiting ranchland phosphorus runoff within 0.35 - 1.2 mg/L depending on site conditions (EPA 2000). Because of the non-point sources of the pollution runoff, there is an increasing focus on voluntary best management practices (FCA 1999; Bottcher et al. 1995), where the success in meeting the water quality standards is quite uncertain.

One of the potential solutions for this issue is to encourage environmentally sound land-use practices in the watershed. Silvopasture land-use can be considered as a practical approach whereby ranchers could sustain their current operations with incremental change, and also meet the environmental standards. Furthermore, silvopasture provides added benefits of carbon sequestration and habitat protection for wildlife.

Our estimates of environmental benefits of the silvopasture land-use in the Lake Okeechobee watershed show that the households from south-central Florida are willing to pay up to $120.24 per year for the improvement in water quality, carbon sequestration, and wildlife habitat. These environmental values could be generated from the sound management of private ranchlands, however, it is the responsibility of the government agencies to design appropriate policy incentives that facilitate transactions in the provision of these public goods.

Current literature in agroforestry strongly supports the premise that environmental benefits of silvopasture operations are substantial, but no systematic effort has been made to internalize these benefits. As agroforestry is gaining popularity in the countries around the world as an attractive farming activity and environmentally benign land-use system our estimation of the environmental benefits can be a basis for the policies to encourage agroforestry practices.

References

Adamowicz, W., P. Boxall, M. Williams, and J. Louviere, 1998. Stated preference approach for measuring passive use values: Choice experiments and contingent valuation, Amer. J. Agr. Econ. 80: 64-75.

Alavalapati, J.R.R. and P.K. Nair, 2001. Socioeconomic and institutional perspectives of agroforestry. In: Palo, M. and J. Uusivuori (eds). World forests, society, and environment- Markets and policies, 52-62. Kluwer Academic Publishers, Dordrecht.

Bateman, I.J. and K.G. Willis (eds.), 1999. Valuing Environmental Preferences: Theory and Practice of the Contingent Valuation Method in the US, EU, and Developing Countries, Oxford University Press, New York.

Boggess, C.F., E.G. Flaig, and R.C. Fluck, 1995. Phosphorus budget-basin relationships for Lake Okeechobee tributary basins, Ecological Engineering 5: 143-162.

Bottcher, A.B., T.K. Tremwel, and K.L Campbell, 1995. Best management practices for water quality improvement in the Lake Okeechobee watershed, Ecological Engineering 5: 341-356.

Clason, T. R. and S. H. Sharrow, 2000. Silvopasture practices. In: Garrett, H.E., W.J. Rietveld, and R.F. Fisher (eds.). North American Agroforestry: An Integrated Science and Practice, 119-148. American Society of Agronomy, Inc. Madison, Wisconsin, USA.

DeShazo, J.R. and G. Fermo, 2002. Designing choice sets for stated preference methods: The effects of complexity on choice consistency, Journal of Environmental Economics and Management 44: 123-143.

EPA, 2000. Total maximum daily load (TMDL) development for total phosphorus, Lake Okeechobee, Florida, US Environmental Protection Agency.

EPA-SFWMD, 1999. Lake Okeechobee Action Plan, The Lake Okeechobee Issue Team, US Environmental Protection Agency and South Florida Water Management District, West Palm Beach, Florida.

FCA, 1999. Water quality best management practices for cow/calf operations in Florida, Florida Cattlemen's Association, Kissimmee, Florida.

FAS, 2000. Beef cattle and calf inventory by county, in Livestock, dairy, and poultry summary, Florida Agricultural Statistics Service.

Garrett, H.E., W.J. Rietveld, and R.F. Fisher (eds.), 2000. North American Agroforestry: An Integrated Science and Practice. American Society of Agronomy, Inc. Madison, Wisconsin, USA.

Gold, M.A., W.J. Rietveld, H.E. Garrett, and R.F. Fisher, 2000. Agroforestry nomenclature, concepts, and practices for the USA In: Garrett, H.E., W.J. Rietveld, and R.F. Fisher (eds.). North American Agroforestry: An Integrated Science and Practice, 119-148. American Society of Agronomy, Inc. Madison, Wisconsin, USA.

Hanley, N., S. Mourato, and R.E. Wright, 2001. Choice modeling approaches: A superior alternative for environmental valuation? Journal of Economic Surveys 15: 435-462.

Kuhfeld, W.F., R.D. Tobias, and M. Garratt, 1994. Efficient experimental design with marketing research applications, Journal of Marketing Research 31: 545-557.

Kurtz, W. B., 2000. Economics and policy of agroforestry. In: Garrett, H.E., W.J. Rietveld, and R.F. Fisher (eds.). North American Agroforestry: An Integrated Science and Practice, 119-148. American Society of Agronomy, Inc. Madison, Wisconsin, USA.

Louviere, J.J., D.A. Hensher, and J.D. Swait, 2000. State Choice Methods: Analysis and Applications, Cambridge University Press, Cambridge, U.K.

Milon, J.W., A.W. Hodges, A. Rimal, C.F. Kiker, and F. Casey, 1999. Public preferences and economic values for restoration of the Everglades/South Florida ecosystems, Economics Report 99-1, University of Florida, Gainesville.

Mitchell, R.C. and R.T. Carson. 1989. Using Surveys to Value Public Goods: The Contingent Valuation Method, Resources for the Future, Washington, D.C.

Reddy, K.R., O.A. Diaz, L.J. Scinto, and M. Agami, 1995. Phosphorus dynamics in selected wetlands and streams of the Lake Okeechobee Basin, Ecological Engineering, 5(1-2): 183-207.

SFWMD, 2002. Lake Okeechobee surface water improvement and management (SWIM) plan: planning document, July 2002 draft. South Florida Water Management District, West Palm Beach, Florida.

Shrestha, R.K., J.R.R. Alavalapati, and R.S. Kalmbacher, 2002. Factors influencing the adoption of silvopasture in central and south Florida: A SWOT-AHP analysis, Agricultural Systems, under review.

Zhang, J. and A. Essex, 1997. Phosphorus load reductions from out-of-compliance sites in the Lake Okeechobee Watershed, Florida, Applied Engineering in Agriculture 13(2): 193-198.