Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Lessons from contract and auction theory Uwe Latacz-Lohmann Department of Agricultural Economics, University of Kiel and School of Agricultural and Resource Economics, The University of Western Australia Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Incentive theory Principal-agent theory Mechanims design theory design of proper institutions for successful economic exchange between a principal and an agent Contract theory Auction theory Delegation of a task Procurement of a good Maximize social welfare Maximize value-for-money Asymmetric information Conflict of interest between principal and agent Hierarchical relationship Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Payments for Environmental Services implemented through Conservation contracts Conservation auctions Management prescriptions Compensation payment Contract allocation mechanism Alternative contract allocation mechanisms: First-come, first-serve (the common model in EU AEP) Individual contract negotiation (SSSI management agreements UK) Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Outline 1. 2. 3. 4. 5. 6. The nature of agri-environmental contracting Contract design to address adverse selection Contract design to address moral hazard Conservation auction theory - bidding models Auction performance estimates Conclusions Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects The nature of agri-environmental contracting 1. Uncertainty about the traded good 2. 3. 4. 5. Measurement problems Contracting on effort rather than output Noise in relationship between effort and env. output Non-separabilities in benefit functions Uncertainty about the value of the traded good Information asymmetry Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Information asymmetry Agri-environmental contracting as a game with asymmetric information time N chooses the type of L (’bad’ or ‘good’) A designs the contract L accepts (or rejects) and receives payment L supplies effort (‘high’ or ‘low’) N chooses the state of nature Only observed by L (‘hidden information’) not perfectly observed by A (‘hidden action’) resulting in ‘adverse selection’ resulting in ‘moral hazard’ observed by A and L ex post, but not verifiable by a third party (‘non-verifiability’) implying non-contractability of environ. output L = landholder; A = environmental agency; N = nature Environmental output (public good) Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Contract theory Information asymmetry has fundamental implications for the design of (conservation) contracts Contracts must elicit information from agent(s) … at the expense of some information rent Contract theory concerned with design of second-best contracts under asymmetric information Hidden information: trade-off between efficiency and rent extraction Hidden action: trade-off btw. efficiency and insurance Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects The first-best, full-info contract t � *= V(q) - t 𝑊 �*= t- θ�q = 0 𝑈 A* * t 𝑡̅ * W* = V(q) - t First-best production level: equating principal‘s marginal value and agent‘s marginal cost. First-best payment level = agent‘s individual cost B* 𝑞�* U*= t- θq = 0 q* Source: Adapted from Laffont and Martimort (2002) q Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Contract design to address adverse selection: self-selection contracts Challenge: devise a menu of contacts{(t, q); (𝑡̅, 𝑞� )} such that (t, q) is preferred to (𝑡̅, 𝑞� ) by the low-cost agent and (𝑡̅, 𝑞� ) is preferred to (t, q) by the high-cost agent. � ) - 𝒕𝒕̅) (1) Max p(V(q) – t) + (1 – p)(V(𝒒𝒒 � )} { (t, q); (𝒕𝒕̅, 𝒒𝒒 subject to t – θq ≥ 0 Participation constrain low-cost agent 𝒕𝒕̅ – � 𝛉𝛉� 𝒒𝒒 ≥ 0 Participation constrain high-cost agent � θ Self-selection constraint low-cost agent t – θq ≥ 𝒕𝒕̅ – 𝒒𝒒 � Self-selection constraint low-cost agent � ≥ t – q𝛉𝛉 𝒕𝒕̅ – � 𝛉𝛉𝒒𝒒 Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Solution to the self-selection problem � – θ) 𝑞𝑞�SB,q*); (θ � 𝑞𝑞�SB, 𝑞𝑞�SB< 𝑞𝑞�*)} ̅ , 𝑞𝑞�SB)} = {(θq* + (θ {(tSB,qSB); (𝑡𝑡SB In English: Second-best contract menu under hidden information: for the low-cost agent: the first-best, full-information production level and overcompensation of costs; for the high-cost agent: reduced output level with a payment equal to costs Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Self-selection contracts Wu and Babcock (1996) • Model setup: Maximise net welfare gain from the contract subject to self-selection constraint and participation constraint Contract variables = mgt. prescriptions and payment rate Continuum of agent types • Key findings: Offer a uniform payment to all farmers equal to the highest compliance costs Less restrictive management prescriptions than under perfect information Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Self-selection contracts Moxey, White, Ozanne (1999) A model of two farmer ‘types’ (high and low productivity) Model setup: Maximise ‘net social welfare of a contract’ subject to two self-selection constraints and two participation constraints (one for each farmer type) Key findings: Overcompensate the low-productivity farmer Demand less input reduction from the high-productivity farmer than under perfect information Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Contract design to address moral hazard: overview Asymmetric information about agents' effort level Standard contract theory: Observe output as a proxy of effort and link payments to output (high/low). Trade-off between incentive provision and insurance Conservation contracting: Output not observable/verifiable/contractable Principal must monitor effort directly and penalize low effort Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Contract design to address moral hazard: models Principal’s objective: cost minimisation or social welfare maximisation Contract variables: Monitoring rate, level of fine, payment level, management prescriptions (effort level) Risk-neutral and risk-averse agents Honest or dishonest farmers Single agent or multiple agents One, two or continuum of farmer types (costs) Static (one-shot) or dynamic enforcement situation Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Contract design to address moral hazard: some insights and policy conclusions Set fine as high as possible Overcompensate agents to encourage compliance Harness risk aversion of agents: mean-penality preserving shift in compliance instruments Target compliance monitoring on high-cost agents Target compliance monitoring on those who have cheated in the past Reduce stringency of management prescriptions Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction theory: why auctions? Alternative mechanism to address adverse selection Rely on competition rather than self-selection Price discovery: harness information held privately by bidders in determining prices for public goods Cost revelation: agents reveal their costs with the bids Cost-effectiveness: Auctions reduce information rents and enhance cost-effectiveness Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Characteristics of conservation auctions Sealed-bid, multiple-unit procurement auctions Trading a heterogeneous good Repeated auctions: multiple bidding rounds Budget-constrained versus target-constrained auctions Discriminatory-price versus uniform-price auctions Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction theory for budget-constrained auctions (Latacz-Lohmann and Van der Hamsvoort, 1997) bidding strategies predicated on the belief of a maximum acceptable bid, or bid cap, β. β = implicit reserve price Bidders form expectations about β Probability that a bid b is accepted Optimal bid: Max Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction theory for budget-constrained auctions (Latacz-Lohmann and Van der Hamsvoort, 1997) Optimal bid formula for risk-neutral agent Overbidding info rent Imperfect cost revelation Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction theory for target-constrained auctions (Hailu, Schilizzi and Thoyer, 2005) model of the Nash equilibrium risk-neutral bid function in a multi-unit procurement auction (Harris and Raviv, 1981) Optimal bid formula for risk-neutral agent v = bidder’s private value (cost) uniformly distributed on [0, 1]; n = number of bidders m = number of units demanded by auctioneer (i.e. target) u = integrand for values between v and 1 Overbidding info rent Imperfect cost revelation Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction performance estimates: overview Data source: From the field or from the lab Counterfactual: fixed-rate payment first-best contract self-selection contract Alternative auction designs: Payment format: discriminatory-price versus uniform-price Auction format: budget-constrained versus targetconstrained Information policy: information revealed or concealed Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction performance estimates: auction versus fixed-rate payments BushTender pilot auction (2001): Same amount of biodiversity benefits would have cost seven times as much!!? (Stoneham et al., 2003) Auction for Landscape Recovery: efficiency gains between 200 and 315% (White and Burton, 2005) Auction experiment Kiel/Perth: Cost savings in the range of 30 to 60%, quickly eroding with repetition (Schilizzi and LataczLohmann, 2007) Simulated cost savings between 16 and 29% (Latacz-Lohmann and Van der Hamsvoort, 1997) Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Auction performance estimates: auction versus self-selection contracts Cost-effectiveness of auction versus self-selection contract as a function of bidders‘ expectation about the maximum acceptable bid Source: Glebe (2008, ERAE) Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Conclusions Contract models have clarified the principles, but Assumptions too restrictive (e.g. two farmer types, knowing their costs, rational behaviour) Often (incomprehensible) corner solutions Recommendations often out of step with the intuition of policy administrators Models not adapted to the complex regulatory process of conservation contracting Not really useful to inform practical contract design, but papers publish well in high-impact journals Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Conclusions Auction theory has had a significant impact on policy, but Bidding models too simple to cater for the complexity of agri-environmental contracting (uncertainty) Little scope for extending auction theory (except combinatorial auctions, … ) Economic experiments more promising than theory development, but Evidence of performance advantages inconclusive, preference for cooperative approaches, equal treatment mentality Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Alternatives? Contract attributes (z) CONTRACT 1 CONTRACT 2 CONTRACT 3 No contract Permitted fertilizers Organic fertilizer No fertilizer Organic and mineral fertilizers No mowing before 22 June 1 June 22 June Max. stocking rate 4 LU/ha 2 LU/ha 4 LU/ha Contract duration 1 year 10 years 5 years Annual payment €450/ha €350/ha €250/ha I would choose none of the contracts I would choose … O O O O Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Cost estimates for individual farms 𝑾𝑾𝑾𝒊 = 𝑪𝑪𝑪𝑪𝒊 = 𝟏𝟏𝟏, 𝟎𝟎 ∗ 𝒛𝟏 + 𝟐𝟐𝟐, 𝟖𝟖 ∗ 𝒛𝟐 − 𝟒, 𝟏𝟏 ∗ 𝒛𝟑 + 𝟏𝟏𝟏, 𝟔𝟔 ∗ 𝒛𝟒 − 𝟏𝟏. 𝟑𝟑 ∗ 𝒛𝟓 + 𝟏𝟏𝟏, 𝟏𝟏 ∗ 𝒔𝒊𝒊 + 𝟔𝟔, 𝟒𝟒 ∗ 𝒔𝒊𝒊 − 𝟑𝟑, 𝟓𝟓 ∗ 𝒔𝒊𝒊𝒊 + 𝟐, 𝟎𝟎 ∗ 𝒔𝒊𝒊𝒊 − 𝟏, 𝟔𝟔 ∗ 𝒔𝒊𝒊𝒊 + 𝟏, 𝟒𝟒 ∗ 𝒔𝒊𝒊 − 𝟎, 𝟑𝟑 ∗ 𝒔𝒊𝒊 − 𝟐, 𝟒𝟒 ∗ 𝒔𝒊𝒊 − 𝟎, 𝟗𝟗 ∗ 𝒔𝒊𝒊𝒊 + 𝟎, 𝟕𝟕 ∗ 𝒔𝒊𝒊𝒊 − 𝟖𝟖, 𝟐𝟐 ∗ 𝒔𝒊𝒊𝒊 + 𝟑, 𝟕𝟕 ∗ 𝒔𝒊𝒊𝒊 − 𝟑, 𝟐𝟐 ∗ 𝒔𝒊𝒊𝒊 ≤ 𝒂 Z variables = contract attributes S variables = farm / farmer characteristics Estimate of average participation cost for contract 1 = €323/ha Lowest cost in sample = €0/ha Highest cost in sample = €795/ha Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Supply curve for area under contract Conservation Tenders in Developed and Developing Countries Status Quo, Challenges and Prospects Lessons from contract and auction theory Uwe Latacz-Lohmann Department of Agricultural Economics, University of Kiel and School of Agricultural and Resource Economics, The University of Western Australia