1. Evaluating Agri-Environmental Schemes – The Marginal Cost of Ecosystem Services Johannes Sauer and Ada Wossink Economics, School of Social Sciences, University of Manchester

2. 2 What is the issue? Growing literature on the societal relevance and valuation of ecosystems services This knowledge is important but understanding and modelling the underlying processes leading to service provision is essential for predicting and managing change in ecosystem services (Nicholson et al., 2009). We address this issue in the context of agri- environmental agreements

3. 3 Research Question How to determine the cost of marginal ecosystem changes and the effectiveness of green payments based on a theoretical and empirical analysis of the bio-economic production relationships at the micro (farm) level.  supply curves should be estimated at a low level of aggregation accounting for biophysical and socio- economic variability  relationships between marketed output and non- marketed ES, and assessment of direct and opportunity costs at the margin

4. 4 Approach/Contribution New theoretical approach based on generalized joint production model which allows complementary, substitutive and competitive relationships. We implement this theoretical framework empirically as a transformation function. Third, we include farm/farmer specific impacts and use panel data analysis. We apply our approach to UK data on the Environmental Stewardship Scheme (ESS) and the Hill Farm Allowance (HFA).

5. 5 Change in the supply of marketed output Y simultaneously affects the supply of non-marketed ecosystem service denoted by Z Inputs contribute to both outputs and Z affects Y Economies of scale and scope ES-Dimension of Agricultural Production Yield ES positive Z1Z1 ES negative Z2Z2 PPF 1

6. 6 Theoretical Model The two outputs are produced simultaneously but since these are multiple outputs a separate production function is used for each output. This leads to a generalised joint production model. This model allows for joint inputs and the possibility of varying the proportion of agricultural output and ES.

7. 7 Theoretical Model where Y = agricultural output Z = non-marketed ecosystem service X = input contributing to Y and Z D = site specific biotic and a-biotic environment. constraint on level of ecosystem service to establish marginal cost of trading-off Y for Z

8. 8 F.O.C.: Envelope theorem and FOC: Theoretical Model Envelope theorem and FOC : for an optimal solution where 1 and 2 as Lagrange multiplier for technology and ecosystem service constraints 2 as shadow price for Z; techn inter- dependencie s and non- allocatable inputs for production of Z and Y

9. 9 Three Cases Case 1: Complementary. The sum of the direct yield effect and the indirect yield effect of input use is positive (but decreasing) and the farmer can produce more Z while also increasing his commodity output Y. λ* 2 = 0 Case 2: Substitutive. Either the direct yield effect or the indirect yield effect, is non-positive but the net yield effect of the rearrangement of input X is positive. λ* 2 = 0 Case 3: Competitive. The direct yield effect of reallocating X is nil and there are yield losses caused by the required increase in Z needed to satisfy the constraint on the ecosystem services.

10. 10 Implementation Two schemes: the Environmental Stewardship Scheme (ESS) and the Hill Farm Allowance (HFA)

11. Implementation 11 Proportion of land under ESSAgro-env. payment per ha (ESS and other)

12. Implementation 12 Panel data collected in England and Wales, 2005-2007 Theoretical model implemented as a transformation function Y 1 = G(Y -1,X,T), where Y 1 is agricultural output of the farms and Y-1 the vector of other outputs (Z payments and non-agricultural output YNAO). Generalized linear functional form, random effects specification.

13. 13 Estimated Model

14. 14 Results Estimated Direct and Indirect Effects Effect evaluatedMeanStd. Dev.MinMax dYAO/dX173.978259.197-440.0661591.110 dYAO/dZESS.3722.887-8.23312.288 dYAO/dZHFA-2.5296.310-39.07123.947 (dYAO/dZESS)(dZESS/dX)0.0650.0320.0060.192 (dYAO/dZHFA)(dZHFA/dX)0.0710.0580.0070.438 (dYAO/dZESS)(dZESS/dZHFA)-6.61e-045.61e-04-0.004-7.01e-05 (dYAO/dYNAO)(dYNAO/dZHFA)9.03e-057.74E-051.21e-055.83e-04 (dYAO/dYNAO)(dYNAO/dZESS)-5.03e-053.09E-05-2.24e-045.24e-06 (dYAO/dYNAO)(dYNAO/dX)-0.0080.005-0.043-7.11e-04

15. 15 Results # Observations per case for Product-Product Relationships RelationAgric. output ESS Agric. output HFA ESS HFA Direct effect dYAO/dX dYAO/dZHFAdYAO/dZESS Indirect effect (dYAO/dZESS)* (dESS/dX) (dYAO/dZHFA)* (dZHFA/dX) (dYAO/dZESS)* (dZESS/dZHFA) (dYAO/dZHFA)* (dZHFA/dZESS) Case I314 00 Case II00121202 Case III79 272191 Total Obs.393

16. 6/30/2015vTI Options for Efficient Reorganisation of Production Y – Z relationship Case 1 – complementary Case 2 – substitutive Case 3 – competitive AO - ESS + agricultural output + environm output (314 farms) (79) AO - HFA+ agricultural output + environm output (314 ) (79) HFA - ESS (121)+ environm output 2 (272) ESS - HFA + environm output 1 or + environm output 2 (202) (191) HFA - NAO (121) + non-agricultural output (272) ESS - NAO + environm output 1 or + non-agricultural output (202)(191) AO - NAO + agricultural output oder + non-agricultural output (314)(79)

17. 17 Conclusions - Empirics The majority of farms produce agricultural output and ecosystem services in a complementary relationship (λ* 2 = 0 ) Generation of multiple ecosystem services on the same farm showed either a substitutive or competitive relationship. Changing the composition of the ecosystem services output (HFA, ESS) would have very different implications for individual farms.

18. 18 Current / Further Work. modelling issues: monetary payments vs. “service quantities” correction for potential selection bias look at spatial patterns investigate significant characteristics of the farms being part of the classes I-III as estimated in our paper more sophisticated modelling approaches (dynamic panel data approach, mixed-effects logistic approach)