Linking action situations: Coordination, conflicts, and evolution in electricity provision for irrigation in Andhra Pradesh, India

Ecological Economics 90 (2013) 150–158

Contents lists available at SciVerse ScienceDirect

Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon

Analysis

Linking action situations: Coordination, conflicts, and evolution in electricity provision for irrigation in Andhra Pradesh, India Christian Kimmich ⁎ Humboldt University Berlin, Division of Resource Economics, Philippstraße 13, D-10115 Berlin, Germany

a r t i c l e

i n f o

Article history: Received 5 November 2012 Received in revised form 1 March 2013 Accepted 22 March 2013 Available online 13 April 2013 Keywords: Technology adoption Irrigation Energy efficiency India Economic network analysis Game theory

a b s t r a c t Actor-centred institutional analysis can gain through an expanded focus from a focal action situation to the adjacent situations that make up its structure. Equilibrium outcomes in game models of a focal action situation may not be explainable without considering linked games. The concepts of an ‘ecology of games’, ‘nested games’ or economic network analysis indicate the relevance of this move, but a structured approach to heterogeneous networks of adjacent action situations encountered in resource and infrastructure governance has only recently been developed. This paper draws on the adjacency concept and proposes four types of links, a potential boundary for adjacency networks, and a condition for bidirectional causation between linked action situations. The relevance of the theoretical propositions laid out is empirically supported for the analysis of electricity governance of irrigation in Andhra Pradesh. The actual and empirically observed outcomes, as well as the potential capacity of an adjacent action situation to influence focal outcomes, are analysed through a set of stylised game theory models and their links. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Actor-centred institutional analysis frequently focuses on selected key situations within a game theory framework, enabling identification of the critical variables that structure interdependent choice, actions and related outcomes, and facilitating the use of multiple methods (Poteete et al., 2010). One of the most prominent research areas is resource and infrastructure governance, where analytical frameworks of the elementary variables structuring action situations have been developed and refined (see e.g. Hagedorn, 2008; Ostrom, 2005). Common-pool resource analysis is frequently associated with a game-model class known as the Prisoner's Dilemma, covering an extensive set of variations in the game structure. The origin of a situation's structure itself and structural changes are increasingly being considered (see e.g. Aoki, 2007; Bardhan and Ray, 2008). Particularly institutional structure is usually the outcome of other action situations, an example being the contact between resource users and a market for their products (Tarui, 2007). Institutions to govern common-pool resources are considered public goods with their own second-order provisioning challenges (Ostrom et al., 1994). Linked action situations are often decisive in shaping more complex settings (McGinnis, 2011). This paper contributes to the endeavour to analyse networks of action situations. The empirical evidence presented here makes the relevance of adjacent situations become visible. In electricity utilisation for irrigation ⁎ Tel.: +49 30 2093 6430; fax: +49 30 2093 6497. E-mail address: [email protected]. 0921-8009/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecolecon.2013.03.017

in Andhra Pradesh, India, farmers face a coordination problem regarding technological measures to improve power quality and prevent equipment damage. Coordination failures are common in rural infrastructure provision and agricultural technologies (Drèze and Sharma, 1998; Janssen, 2007). In the case analysed here, coordination is obstructed unless an underlying dilemma of grid-capacity overuse is tackled. Underprovision and unauthorised connections create overload and prevent farmers from using energy-efficient pumpsets and measures to effectively improve power quality. As Anderies et al. (2004) point out, the “link between resource users and public infrastructure providers is a key variable affecting the robustness of SESs that has frequently been ignored in the past”. Collusion between some farmers and electricity staff enables illegal tapping and capacity overexploitation. Decreasing groundwater tables through excessive exploitation requires additional pumping capacity, further increasing load on the electricity grid. The political economy of electricity subsidisation nurtures this vicious circle (Tongia, 2007). Apparently, the inhibitor of coordination is rooted in multiple linked action situations of infrastructure utilisation and provision. Recent conceptual developments and empirical studies of adjacent action situations (McGinnis, 2011) and the ecology of games (Lubell et al., 2010) enable a structured approach to capture the interdependencies of linked situations. The paper derives and tests theoretical propositions for the types of links, a potential boundary to delimit relevant situations, and to identify capacities of links that co-determine outcomes of adjacent situations. Six relevant situations and respective links for the case of electricity-driven irrigation in Andhra Pradesh are analysed and expected equilibria are derived from a basic game model. In a final stage, the

C. Kimmich / Ecological Economics 90 (2013) 150–158

network is synthesised to analyse interdependencies of co-determining situations and to identify situations that impede coordination to improve power quality and energy efficiency in the core action situation. The findings provide guidance to further structure the analysis of linked action situations. In practice, where electric infrastructure capacity and groundwater are overexploited, only an integrative approach to surmount these dilemmas will render effective any efforts to increase energy efficiency. Disregarding these linked situations will very likely even defeat measures to overcome the persisting political economy of electricity subsidisation. The remainder is structured as follows: Section 2 reviews the literature on action situation links, networks, and the evolutionary perspective on institutions. Section 3 deduces a set of theoretical propositions concerning links, boundaries, and interdependencies. Section 4 briefly describes data and methods used. Section 5 presents the background to the case, and in Section 6 the most crucial action situations in electricity provision for irrigation and their links are analysed. The paper draws conclusions for the case and derives theoretical implications for further research.

2. Theories for Linked Action Situations A game model requires the specification of a variety of parameters, before any equilibrium outcome can be derived. The number of actors involved, the choices available to them and the functional interdependence of choices together yield the dominant or most probable emerging equilibrium. Outcomes determine whether actors find themselves participating in situations of conflict or coordination. A wide variety of models can result, as even the taxonomy and topology of simple ordinal preference two-player, two-choice games indicate (Rapoport, 1966; Robinson and Goforth, 2005). Outcomes depend on the content of the actions, that is the properties of the transactions and their functional interdependence (Hagedorn, 2008), such as subtraction in common-pool resources, addition in public goods, or critical transactions in electricity provision (Künneke et al., 2010). The number of actors involved is crucial (Olson, 1965), as are the number and type of choices and outside options (Hirschman, 1970), communication, repetition, framing, and the rationales of actors (Bromley, 2006; Ostrom, 2010; Vatn, 2005). All parameters are shaped by both physical and institutional conditions. The parameters that structure one situation, be they physical or institutional, may stem from other situations. As Ostrom et al. (1994, p. 45) emphasise, even though “a ‘single’ arena may include large numbers of participants and complex chains of action, most of social reality is composed of multiple arenas linked sequentially or simultaneously”. Some individual parameters can be decisive for an outcome. It becomes necessary, then, to take into consideration a parameter's origin in a linked action, where the underlying structure can be quite different from that of the core action situation. The levels of operational, collective, and constitutional choice are a case in point (Kiser and Ostrom, 2000). Two-level games of domestic and international politics (Putnam, 1988), as well as the variety of games and time horizons within a game (Shubik, 1986) have been considered, but only recently have links become a more explicit matter of concern in the analysis of resource and polycentric governance, although the study of policy networks has long shown the importance of ‘nested games’ and the Ecology of Games (EG) approach (McGinnis, 2011). Important is the crucial analytical distinction between network connections and actual choices available in situations, where agency comes into play. Economic network analysis develops theories for a larger number of homogeneous games with varying network structures, relevant for example in studying physical and social infrastructures (Goyal, 2007). However, many empirical phenomena exhibit several co-existing types of links, and the linked situations may exhibit heterogeneous underlying structures, making such cases more complex for the researcher. Consequently, a more open concept is needed.

151

2.1. Networks of Adjacent Action Situations According to the concept of Networks of Adjacent Action Situations (NAAS), “an action situation Xi is adjacent to Y if the outcome of Xi directly influences the value of one or more of the working components of Y” (McGinnis, 2011, p. 53). McGinnis proposes an extension of the Institutional Analysis and Development framework to identify related types of adjacent action situations. A systematic procedure to identify action-situation networks based on the generic tasks of polycentric governance (Ostrom et al., 1961) is proposed, including the tasks of production, provision, financing, consumption, coordination, dispute resolution, rule making, and monitoring, while the tasks of constructing collective entities and internalising norms are proposed as potentially relevant, though likely remote from the focal action situation. Each of these generic tasks can constitute multiple action situations, with a variety of actors involved in each of them. The theoretical relevance of EG and NAAS has been demonstrated with several empirical cases (Dutton et al., 2012; Lubell et al., 2010; McGinnis, 2011) that are instructive examples revealing the complexity of the adjacent situations that condition each other. With the extension of analysis from the focal situation to the adjacency network, the researcher is confronted with different types of outcome. While the outcome of the focal situation in the resource context is usually the level of exploitation or a product or service delivered, the outcome of an adjacent situation is a working component of the focal situation (McGinnis, 2011). This outcome can be physical in nature, but is often an institution. The underlying situation can be the singular legislation of a law or a highly repetitive and long-lasting situation of reproducing habits and norms of behaviour. Especially in the latter case, the outcome is part of an underlying and persistent institutional structure that shapes the focal situation. Here, evolutionary concepts may be required. 2.2. Evolutionary Theories and Technology Adoption Institutional economics and game theory have created some of the most fruitful insights into understanding the performance of institutions (see e.g. Ostrom et al., 1994). Evolutionary economics and evolutionary game theory focus on the emergence and persistence of institutions or stable strategies respectively. Both complement each other: institutions may, for example, contribute to coordination failure, inhibiting gains for the actors involved in an action situation, yet they may still survive. The institutions required for a change in performance may not easily emerge, let alone persist. If the research aims at explaining empirical phenomena emerging from linked action situations, both may be required. Accordingly, institutions have been defined as “rules and conventions of society that facilitate coordination among people regarding their behavior” (Bromley, 1989, p. 22), and as “durable systems of established and embedded social rules that structure social interactions” (Hodgson, 2004, p. 14). Law, conventions, custom, habits, routines and norms, and even strategies and heuristics may all be considered institutions, and both perspectives are helpful in explaining some of the linked action situations. Probably the best known example of evolutionary game theory is the evolutionary stable strategy of tit-for-tat in the indefinitely repeated Prisoner's Dilemma with a sufficient likely continuing interaction (Axelrod, 1981). Analogously, several imitation and best-response strategies have been tested in repeated interactions in a class of coordination problems (Skyrms, 2004), known as the Stag Hunt or assurance problem (Sen, 1967), with two Nash equilibria in pure strategies, one being payoff-dominant, the other risk-dominant (Harsanyi and Selten, 1988). A coordination problem is at the core of the focal action situation analysed in Section 5. In both models, a Pareto-superior outcome has been shown to be feasible, given the evolution of strategies. While classical game theory bases its models on (boundedly) rational agents with common knowledge and frequently complete or perfect information,

152

C. Kimmich / Ecological Economics 90 (2013) 150–158

evolutionary game theory is located at the opposite extreme, with agents often identical to one single strategy, without choice and individual rationality calculus or only simple learning mechanisms. Evolutionary game theory includes selection and mutation processes as well as inheritance mechanisms. These structures, processes and mechanisms are exogenous to the model, which limits their possibility space to the assumptions and reductions made (Hodgson and Huang, 2010), although model variations and mutations already cover large areas of research, especially in evolutionary biology (see e.g. Skyrms, 2004). Evolutionary models analyse for example the comparative advantages of cultural versus individual learning in a dynamic setting, where culture helps in coordination (Golman and Page, 2010), and the ability to switch learning strategies outperforms ‘acultural’ populations (Kameda and Nakanishi, 2003). Evolutionary economics and co-evolutionary frames provide guidance for a broader perspective (Norgaard, 1984; van den Bergh and Stagl, 2003). Evolutionary dynamics of technology adoption and diffusion play an important role in electricity utilisation for irrigation, as in agriculture more generally, and respective learning processes have been extensively studied (Foster and Rosenzweig, 1995; Goyal, 2007; Jara-Rojas et al., 2012). The literature shows, for example, how type of production technology and related observables have significant effects on the existence of social learning strategies with regard to variation, experimentation and adoption (Munshi, 2004). As will be shown, the stochastic process of equipment damage and unobservable functioning of the electricity grid make learning to coordinate difficult. 3. Theoretical Propositions An adjacent action situation Xi “directly influences the value of one or more of the working components of [the core action situation] Y” (McGinnis, 2011, p. 53). The EG approach offers some general statements on the links between action situations: “Plays (i.e., moves or actions) made in one game can affect the play of others. Also, the outcome of one game might affect the rules or play of another”, and “the players' moves in one game might be constrained by their moves within other games” (Dutton, 1995, p. 382). Focusing on the links, Lubell et al. (2010, p. 289) state “that issues may be interconnected through biophysical, economic, or social processes, so decisions made in the context of one issue may directly affect payoffs in other issues”. From the structure of action situations and the concepts of adjacency and EG, three theoretical propositions are derived. The working components wj of an action situation Y can be influenced by (a) biophysical transactions, (b) information, (c) institutions, and (d) actors involved. Thus a proposition can be derived concerning links: Link proposition (1). Every adjacent action situation Xi must be linked through one or more of the types a-d influencing one or more working components wj(Y), so that Xi directly influences Y. By definition, the working components wj of an action situation Y must explain the empirically observed equilibrium outcome. However, one or more of its working components may fluctuate or even remain unobserved, if an adjacent action situation Xi is not analysed. This may easily be the case for pure information links, where the mere presence of an adjacent action situation may change Y. The range of potentially linked action situations may be unlimited, which makes a condition for delimitation and criteria for relevance necessary: Boundary proposition (2). If the equilibrium outcome of Y is co-determined by the equilibrium outcome of an adjacent action situation Xi, and at least one of the core actors is involved in Y and Xi, then Xi is in the range of influence of Y.

Adjacency per se does not require any of the actors of Y to be involved in Xi. The focus of the presented analysis, however, is only on those adjacent situations that are in the range of influence of Y through actors involved in both, although other adjacent action situations may as well exert strong influence on the outcomes of Y. Drawing an analytical boundary around the reach of the core actors has been determined by the research objective of identifying the agency potential of core actors to improve conditions for irrigation and energy efficiency. Interdependency proposition (3). If Xi is in the range of influence of Y through at least one actor, the causal mechanisms between Y and Xi are unidirectional or bidirectional: Xi (Y) co-determines Y (Xi), or Y co-determines Xi, and Xi co-determines Y. Note that the empirically observed equilibrium outcomes of Y may reveal only one of the potential capacities of the working components of Y and the adjacent action situation Xi. An adjacent action situation Xi may exist that has the capacity to change Y, if one or more of the working components of Y or Xi change. Empirical observation can yield only manifest outcomes, while an analysis of the underlying game model structure can expose potential equilibrium outcomes of the generative mechanisms. 4. Data and Methods Identification and analysis of action situation networks and links require in-depth study of the patterns of each situation. No established, systematic method exists yet to empirically identify adjacency or types of links and their effects on a core action situation. This renders a mixed-method research design effective, as the focus lies in deliberatively investigating the context (of linked action situations) because its relevance is not clearly understood (Beckmann and Padmanabhan, 2009; Yin, 2003). The selected case is instrumental (Stake, 2000) and its study based on an iterative research process (Eisenhardt, 1989) of qualitative and quantitative observations (Gerring, 2007), involving semi-structured interviews, a cross-section survey, econometric analysis, and game theory model building in a between-method triangulation process (Downward and Mearman, 2007), to make theoretical propositions testable. The interviews, conducted between 2009 and 2011, included farmers, repair workshop owners, pumpset manufacturers and local retailers, utility managers and ground staff at sub-stations, as well as village revenue officers, elected heads of the village level government, and members of the Electricity Regulatory Commission. The results framed the cross-section survey design, which covered N = 305 observations through a stratified random sample. Table 1 summarises selected survey variables. The qualitative data informed the model building process to develop more robust formal models (Walker and Sinclair, 1998). The results of the econometric analysis (Kimmich, in press) and game models inspired, in turn, structured interviews. The game theory models allow for consideration of the ‘counterfactuals’ within the reasoning processes of actors and their plausible alternatives in a specific situation. They do not require strong assumptions about rationality, but rather paying attention to the perceptive dimension, for which “it is absolutely essential that the beliefs, ideas, and experiences of the actors themselves are moved onto center stage” (Swedberg, 2001, p. 325). 5. The Electricity–Irrigation Nexus in India In his seminal work on groundwater irrigation, Tushar Shah (2009) analysed electricity policies for irrigation in South Asia. Most Indian states conduct a low flat-rate policy that has persisted for three decades, predominantly due to the emergence of a political economic constellation of party competition and ‘vote bank’ politics, and the characteristics of electricity governance (Birner et al., 2007; Chakrabarty, 2008;

C. Kimmich / Ecological Economics 90 (2013) 150–158 Table 1 Summary statistics for selected survey variables. Variable

Mean

SD Median Min

Branded pump-set (1 = yes) 0,67 ISI-marked pump-set (1 = yes) 0,37 BEE-rated pump-set (1 = yes) 0,06 Capacitor successfully installed (1 = yes) 0,10 Motor burn-outs per year 1,86 1,64 1 Costs for motor repair (INR) 2693,15 1513,11 Age of the pump-set (years) 7,21 5,94 Transformer burn-outs per year 1,02 1,04 620,58 869,65 Costs for transformer repair (INR)1

2 2500 5 0,70 400

Max

0 1 0 1 0 1 0 1 0 12 200 8500 0 30 0 7 0 8000

1

: 62 INR = 1 €2010.

Jenkins, 1999; Kale, 2004; Kimmich, 2013). The eastern states increasingly struggle with maintaining their electricity infrastructure for agriculture, forcing ‘dieselization’, thermodynamically and environmentally the least preferable option. A World Bank (2001) study, conducted at the outset of electricity regulation reform, indicated that improvements in service quality might be a precondition for legitimising increasing tariffs. The implications, however, have so far hardly been considered: How can service quality be improved? What are the prevailing actions and related institutions that keep both agriculture and utilities in a low equilibrium? The following analysis focuses on Andhra Pradesh, where 36% of all final electric energy available was used in agriculture in 2007, compared to a national average of 22% (CMIE, 2008). With more than 2.5 million pumpsets connected to the grid in 2008 (APTRANSCO, 2008), one of the major challenges there is increasing energy efficiency while maintaining the viability of the agricultural sector. With irregular monsoon precipitation, access to groundwater has contributed to increasingly secured irrigation and food supply (Fig. 1). Given the flat-rate power supply regime, financial incentives to implement demand side measures (DSM) to improve energy efficiency are mostly absent. Inefficient pumpsets contribute toward deteriorating power quality, increasing pumpset and transformer damages. Farmers and distribution utilities are incurring high repair costs, discouraging any investment in better equipment (Tongia, 2007). Farmers even pay partly for repairing transformers, despite transformers being part of the distribution companies' property. Some DSM – such as the use of standard-approved pumpsets with energy-efficient motors (‘ISI-marked’ by the Bureau of Indian Standards) and the installation of capacitors – could simultaneously reduce equipment damage and energy consumption. A capacitor or condenser is an electric circuit element, which can correct the power factor in an electricity grid. It

Fig. 1. Agricultural area irrigated per source in Andhra Pradesh, India in 100,000 ha (Government of Andhra Pradesh, 2008). Tube wells provide a secured access to groundwater that resists fluctuations from irregular monsoon precipitation.

153

balances the phase between current and voltage (Dugan, 2003; Meier, 2006) and can thus improve power quality and energy efficiency. If implemented, farmers and utilities could save on repairs, and fiscal expenditures on subsidies could be reduced, contributing to the viability of agriculture and benefitting distribution utilities as well as the overall economy through reduced fiscal burdens. Table 1 provides an overview of the share of adopted DSM by farmers, including ISI-marked and BEE-rated pumpsets (efficiency-rated by the Bureau of Energy Efficiency), and capacitors. Soon after its constitution, the Andhra Pradesh Electricity Regulatory Commission (APERC) realised the importance of DSM. In its tariff order (APERC, 2001), the transmission corporation and the distribution licensees had committed themselves to distributing transformers and erecting capacitors for agricultural pumpsets: “To improve the power factor, it must be made compulsory for the farmers to use capacitors with the pumpsets” (p. 79). APERC also ordered a sanction of discontinuation of services if capacitors were not installed (p. 117), and provided specifications for choosing an adequate capacitor (p. 306). Four years later, APERC stated that the “DSM measures, especially the capacitor compensation for the inductive load of the agricultural sector has been the biggest techno operational problem encountered in the power sector especially in Andhra Pradesh. Past experience appears to indicate that the initiatives taken earlier [ . . .] did not achieve the results as the consumers have not been made a party to the scheme” (APERC, 2005, p. 74). In a public hearing conducted by APERC (2010, p. 40), a farmers' association representative stated that “though Capacitors are purchased by the farmers, the Licensees are reluctant to fix them, citing shortages in staff to do this work”. Interviewed farmers reported that they had been forced to buy capacitors, supported by several large campaigns to distribute capacitors. Still, only 10% of the surveyed farmers were using a capacitor during the research period. Given that these DSM are supposed to be beneficial for power quality and adoption could be self-enforcing, this seems surprising, but becomes intelligible through an analysis of the action situation network. 6. Analysis: Adjacency Networks in the Electricity–Irrigation Nexus in Andhra Pradesh 6.1. The Core Action Situation: A Coordination Problem (AS1) Power quality is a shared resource within the nested hierarchical structure of the electric power distribution grid. Sub-stations transform power to an 11 kV level, covering several villages and distribution transformers. Depending on a transformer's capacity, on average 17 pumpsets are connected. Each pumpset can subtract from power quality, especially if of low standards, while exclusion is difficult. The choice of one farmer affects all other farmers connected to the same transformer. If all farmers choose to install a low-quality pumpset, the utilisation of a standard-approved pumpset by only one farmer increases equipment damages. If all farmers install a standard-approved pumpset, repair costs are drastically reduced, and all farmers are better off. The use of a capacitor to balance out voltage fluctuations is subject to a similar coordination problem, as “the equipment installed to increase the productivity is also often the equipment that suffers the most from common power disruptions. And the equipment is sometimes the source of additional power quality problems” (Dugan, 2003, p. 3). Unlike in a dilemma situation, no farmer has an incentive to deviate from the Pareto-superior equilibrium, once reached, as a standardapproved pumpset and capacitor reduce equipment damages and improve pumping efficiency. A simplified bi-matrix model of the coordination problem highlights the two Nash equilibria in pure strategies in bold print (Table 2). The equal payoff for the strategy not to invest ~I, and the loss incurred by the one not coordinating, makes this model type an assurance problem. An econometric analysis of AS1 revealed that, under the given conditions, the rational strategy is not to adopt any DSM (Kimmich,

154

C. Kimmich / Ecological Economics 90 (2013) 150–158

in press). This is the low Nash equilibrium of the underlying coordination problem, predicting no adoption at all. Yet, despite their negative impact on the frequency of equipment damages, a significant share of the surveyed farmers has adopted DSM, due to the legal order to make the use of DSM compulsory and related campaigns and partial enforcement when capacitors were distributed. Interviews with farmers indicated that the functioning of the electricity system is only rarely understood. The underlying assurance problem cannot be conceived, let alone the outcome of a payoff-dominant strategy. Some farmers reported that the capacitor prevented their motor from starting, due to low voltage. In other cases, either the motor or the capacitor had been burnt out shortly after installation. Almost no farmer was aware that a simultaneous installation of capacitors is required. The farmers thus applied another evolutionary strategy of technology adoption.

Fig. 2. The adjacent action situation of social learning in technology adoption (AS2) is depicted as a sequential game tree with two farmers (F1; F2) and a separate move by technology (T).

6.2. Factual Versus Actual Games: Social Learning (AS2) Farmers resorted to a common practice in technology adoption: Only selected farmers experimented with the use of a capacitor, and their neighbours and peer network adapted to their experiences. As a result, no experience with coordinated use by all farmers has been possible, impeding any beneficial experience. This actual action situation exemplifies a social learning strategy. From an evolutionary perspective, this is a dominant adoption strategy for many production technologies, but when faced with a coordination problem, it is unlikely to work. In fact, sequential adoption even undermines simultaneous experimentation and the emergence of successful variation. Only when the coordination problem becomes resolved in one instance can social learning potentially catalyse adoption, if coordination requirements are learnt and transmitted. This evolutionary stable strategy can then outperform sequential experimentation. Fig. 2 depicts a basic sequential game tree of learning and adoption, where the number of farmers using DSM determines the move of technology (T). In equilibrium only one farmer invests, the technology will fail, and no farmer observing the outcome invests in the same technology. This finding demonstrates the necessity of not only analysing the core factual action situation as generated by the physical properties, but also how the situation is actually dealt with by actors, which, in the given case, involves evolved shared routines of technology adoption. AS1 and AS2 are not completely distinct, as both have the same actors and underlying physical structure, but it appears to be crucial to analytically separate them in order to understand the outcomes of the core action situation. The crucial link between AS1 and AS2 is thus purely informational. 6.3. Infrastructure Capacity: A Dilemma of Provision (AS3) As some farmers have reported from their experience, not only low power quality, but also low voltage levels prevent capacitors and ISI-marked motors from working. Only locally assembled, nonstandardised pumpsets can stand low-voltage conditions, partly caused by the total load connected (Dugan, 2003, p. 19) exceeding the capacity of the transformer. In addition, the simultaneous start of pumps requires high starting currents and creates short-duration voltage variations (Dugan, 2003, p. 20). A conflict results from infrastructure Table 2 The two farmers (F1; F2) have the choices to invest (I) or not to invest (~I) into measures to improve power quality. Outcomes are ordinal ranks. Investing (I) carries costs, but reduces equipment damage. If both F1 and F2 invest, the payoff is Pareto-optimal, if only one farmer invests, he carries the costs without improvements in power quality. AS1: F1

I ~I

F2 I 2, 2 1, 0

~I 0, 1 1, 1

under-provision or, respectively, overuse of existing capacity. The provision of sufficient capacity for every additional connection becomes necessary. This constellation results in a social dilemma. If some farmers have already provided for infrastructure capacity through their connection payments made to the utility, an additional farmer can evade contributing to that capacity. Infrastructure provision is managed by a distribution utility, which requires a onetime connection charge to be paid by every farmer for the provision of additional transformer capacity. According to farmers' reports, the authorization process is influenced by many informal arrangements at the sub-station level, as a substantial amount of ‘informal payments’ in addition to the official connection charge reveals. While all surveyed farmers stated to have paid connection charges, there is also an unknown share of unauthorised connections. Survey statistics indicate that, for 11% of the transformers, the calculated capacity is below the average pumpset power, and only 37% of the farmers saw the present capacity to be sufficient for an additional pumpset. Regulators, farmers, and electrical engineers have mentioned estimates of between 20 and 30% unauthorised connections. Here, informal arrangements with sub-station personnel enable such unauthorised connections. While AS1 is concerned with the utilisation of power, this adjacent action situation AS3 shapes the provisioning of electricity infrastructure. Both are linked through the same actors, but especially through the physical properties of the infrastructure and the transactions of the involved actors. The underlying model structure is a Prisoner's Dilemma with asymmetric payoffs, resulting in an asymmetric equilibrium (Table 3). Most farmers have authorised connections (A) and paid for the provision of infrastructure capacity, while some have been able to set up an informal arrangement to receive an unauthorised connection (~A). In many cases, the number of informal connections is higher than the capacity that the transformer would allow. Yet, the voltage level is enough to still operate all connected pumpsets, albeit with negative consequences on their efficiency. Apparently, institutions have evolved that make the informal arrangements possible.

6.4. The Implicit Conflict: Collusion and Solidarity (AS4) The established relations between farmers and sub-station personnel, and also solidarity among farmers with those who cannot afford the connection charge, play a role in shaping capacity provision in AS3. Although most of the sub-station personnel do not reside in the villages, kinship or evolved personal relations can strengthen mutual exchange. Obtaining access without authorization is an ‘open secret’ in many cases. In a discussion with an electrical engineer from a sub-station and several villagers, one farmer openly admitted not to have authorised his connection.

C. Kimmich / Ecological Economics 90 (2013) 150–158

Several interviewees mentioned the large investments for the pumpset and tube well, financed through credits, as inhibiting additional payments for authorization. Many farmers were already highly indebted, partly due to the risky investment into a large share of failed tube wells. Groundwater access decides on the economic success of such farms, where subsistence is not an option, as payment obligations have to be served. Under these circumstances, it is conceivable that norms of solidarity have developed, and neighbouring farmers support access to the electricity grid, although the transformer may already be overloaded. Even some sub-station personnels were aware of the circumstances and tolerated open access. The underlying model has a simple but powerful structure, with several equilibria of mutual exchange, and may be omnipresent within some social network ties. In the case of authorization, this relation can create collusion and lead to informal payments. Table 4 depicts a bi-matrix model with simple mutual exchange between a farmer (F) and sub-station personnel (U) involving the decision to exchange favours (E; ~E), with two equilibria in pure strategies. A network link is established if the Pareto-superior equilibrium of mutual exchange is reached. Ties within social networks can be subject to fluctuations on their own and can be strengthened or reshaped. Exchange ties are most likely established and enforced between those actors who can offer equal levels of services or commodities, or a kinship relation exists. This action situation of collusion, linked to AS3, is highly contingent upon the social networks and the established norms within the village. Furthermore, it not only improves the conditions of the involved actors, but also makes the dilemma of providing capacity asymmetric and worsens the conditions for the uninvolved. This exemplifies one of the differences between dyads and triads and the role of coalitions in nested games (Tsebelis, 1990), while also indicating how norms can emerge through asymmetries of power (Knight, 1992). Electricity theft entails relational links and aspects of trust (Winther, 2012). AS4 is strongly linked to AS3 through some farmers being involved in both, institutions governing the relation with the utility, and the resulting changes in payoffs in AS3. AS1 and AS3 can be attributed to the operational-choice level. The ubiquitous adjacent situations AS4 explained here constitute part of the institution-shaping collectivechoice level, although there is apparently no singular collective decision. However, actions and outcomes can only be explained if additional relations between farmers and their electricity utility are taken into account. Why do some farmers collude against the utility, and others collude with sub-station personnel, supporting unauthorised access?

155

Table 4 The matrix represents an adjacent action situation of mutual exchange and collusion (AS4) between a farmer (F) and utility's sub-station personnel (U). The available choices are to exchange (E) and not to exchange (~E) a service. AS4: F

E ~E

U E 1, 1 1, 0

~E 0, 1 0, 0

Local electricity infrastructure, covering several districts of a State, is provided and managed by a distribution company. This utility is in most cases fully state-owned and controlled by a commission which regulates according to aggregate revenue requirements (Pani et al., 2007). After some fundamental power sector reforms, including the organisational separation (unbundling) of generation, transmission and distribution, state electricity regulatory commissions have been set up to develop many of the collective rules in the sector. Thus unlike in many common-pool resource conditions, the formal institutional setting for provision and appropriation is not developed by the

farmers. Even the utility itself is only partly involved in designing the formal institutions for its governance. The action network is larger, involving the regulatory agency and its governance (Thatcher, 2002), with some discretionary power to shape the institutional environment for the utilities, but being itself a product of contextual factors (Dubash and Rao, 2008). The focus of this work lies only on the institutions for providing infrastructure capacity within the boundary proposed in Section 3. The formal procedure of receiving an electricity connection is to apply and then pay a fixed connection fee to the distribution company. The connection is registered and the distribution company provides sufficient load and transformer capacity to allow for the additional connection. However, the sub-station personnel can exert discretionary agency power over these institutions. Even with authorization the provision of capacity might still be delayed for several months, because of reduced ground staff. In addition to formal payments, informal payments are often expected, depending on the farm size, income, and the distance to the sub-station. The farmers' experiences and expectations related with this provision problem can have important effects on their strategies. Table 5 shows a bi-matrix model of the interaction between a farmer (F) who can decide for authorization (A; ~A) and sub-station personnel (U) who can decide to install additional capacity (I; ~I). This setting is only feasible if the farmer has the choice to get a connection without authorization, which depends on established connections as in AS4, the crucial link being the same actors and institutions of collusion involved. The only equilibrium is then (~A; ~I). The involvement of F2 and the low equilibrium create the asymmetry of the payoff structure in AS3, by reducing the infrastructure capacity available to the farmers. The delay in service adds another problem of provision: When a distribution transformer gets burnt out, repair by sub-station personnel can take several days. Especially during the phase when crops need regular irrigation, any delay can result in sharp drops in yields, and farmers are forced to take the initiative of repairing transformers on their own. A large share of the farmers incurs significant costs for such repairing of infrastructure equipment (see Table 5), which can also be extremely dangerous. In five of the 18 villages studied between one and three maintenance-related deaths of farmers have been reported within one year, findings that are also confirmed by public hearings (Andhra Pradesh Electricity Regulatory Commission, 2010). The analysis of AS1 has shown that the farmers can achieve improvements, even without the involvement of the distribution company, when the conditions of infrastructure capacity permit this. However, the provision of sufficient transformer capacity that would reduce the load on each of them lies to a large extent in the hands of the utility. Regular transformer maintenance and measures at the sub-station level can also increase power quality. The interviews suggest that farmers cannot clearly distinguish the effects of their own contributions to

Table 3 The game matrix depicts an asymmetric dilemma of infrastructure capacity provision (AS3) between two farmers (F1; F2) and the choices of authorising (A) and not authorising (~A) the connection of electricity supply.

Table 5 The matrix depicts the process of authorization and capacity provision (AS5) between the farmer (F2) from AS3 and utility's sub-station personnel (U) with the choices to authorise (A; ~A) and to invest in infrastructure capacity (I; ~I).

6.5. Taking Into Account the Utility's Presence: Service, Uncertainty, and Fault (AS5)

AS3: F1

A ~A

F2 A 2, 2 0, 1

AS5: ~A 1, 3 1, 1

F2

A ~A

U I 2, 1 3, 0

~I 0, 2 1, 1

156

C. Kimmich / Ecological Economics 90 (2013) 150–158

power quality from the actions of the utility. Consequently, many farmers ascribe all of the reasons for voltage variations and fluctuations to the utility, further complicating the situation. The incentives for the utility are more difficult to capture than AS5 may indicate. It provides distribution transformers and incurs repair costs of transformer burnouts. Still, the utility has so far hardly tried to reduce voltage fluctuations and to increase power quality for irrigation. The reasons for this inertia require a detailed analysis of utility governance, which goes beyond the boundary range of influence of the core actors. 6.6. Capacities of the Electricity–Groundwater Nexus (AS6) Certainly the most visible link is located between power utilisation and groundwater exploitation. Groundwater exploitation could well be considered another core action situation, because of the drastic decrease in groundwater availability in the case study area since the last two decades (Fig. 3). Electricity pricing may improve allocation of energy and water (Kumar, 2005), and an integrative perspective on heterogeneous common-pool aquifers and energy policies is crucial (Shah, 2009). The groundwater dimension itself has received considerable attention (World Bank, 2010), which has long been studied from an institutional and game theory perspective (Blomquist and Ostrom, 1985). Decreasing groundwater tables require deepened wells, which steadily increases energy use and the respective loads on the electricity grid. This makes successful coordination more difficult, as AS3 revealed. While on a global scale many actors dependent on aquifers have not surmounted this dilemma, there have also been few successful cases (van Steenbergen, 2006), one of them being a village located in Andhra Pradesh, where the council introduced a ban on tube wells and the community voluntarily imposed limitations on crop choices to reduce water intensity. This unique case shows how the core action situation of electricity utilisation can practically vanish. But there was no similar situation encountered in the case study area, although the small hard-rock aquifers of the Deccan plateau could potentially ease cooperation to restrict groundwater exploitation (Shah, 2009). On the other side of the nexus, improvements in pumpset efficiency through successful coordination and infrastructure capacity provision could aggravate groundwater exploitation. A bi-matrix model of the dilemma faced by the farmers extracting groundwater is represented in Table 6. While the consequences of the other action situations can immediately be felt, groundwater exploitation exhibits a more dynamic character. The extraction equilibrium (E; E) initially creates the highest outcome, but with a decreasing groundwater level, the costs of pumping (γ) increase and the equilibrium falls below the payoffs for restraining pumping. Although the actors in AS6 and AS3 are identical, both action situations are

decisively linked through the physical electric load requirements of decreasing groundwater levels. 6.7. Extending the Network of Adjacent Action Situations The farmers are involved in several other situations that exert influence within the adjacency network. Directly linked to AS6 is the development and maintenance of tanks, the traditional South Indian water reservoirs. Surface irrigation and percolation can reduce the pressure on groundwater exploitation. Alternative income sources extend the available choices, reducing the pressure on irrigation use, and were found to have a statistically significant influence on equipment damages (Kimmich, in press). Education and training play another significant role in shaping the core action situation. Subsistence farming, as well as the presence of markets for different crops and labour markets all exert influence on the conditions for the farmers. These action situations are considered exogenous to the analysis as they are either only indirectly linked to AS1 or their influence is estimated to be lower than the impact of the selected situations. The sub-station personnel are part of several situations within the utility itself. Effective monitoring of unauthorised connections by a vigilance unit might influence the payoff structure of some of the sub-station personnel and farmers. Although not directly involved, farmers' associations provide for influence in political and regulatory actions. A broader and more complex network is involved in creating an exchange of services for votes among the local constituency, which is itself embedded in the State level political system, where party competition has led to ‘vote bank’ policies of electricity subsidisation. The political economy dimension is treated as exogenous to the adjacency network analysed here. 6.8. Linking the Action Situations The action situations analysed above exert influence on the working components of AS1 (Fig. 4). If the infrastructure capacity provided through AS3 is not sufficient, the measures to improve power quality cannot work. Thus the reduced capacity also reduces the payoffs in AS1, symbolised by a parameter τ. Similarly, a reduced groundwater level requires additional energy and thus negatively affects the capacity of the electricity grid. In the case of AS4 linked to AS5, it is shared norms and respective delta parameters δ that change the payoff structure of AS3. The case of AS2 is more intricate. Here, cognitive institutions with regard to the information received from AS1, and a respective parameter ι prevent the simultaneous experimentation with power quality measures. These institutions transform AS1 into a sequential game, impeding coordination on the high equilibrium. The adjacency network is also linked through one or more of the actors involved. In AS1-3 and AS6 all actors are farmers. In AS4 and AS5 at least one is a farmer, and the other sub-station personnel. This reveals that the range of influence extends over several situations that can potentially improve the outcome of AS1. The action situations can also co-determine each other, supporting the interdependency proposition made in Section 3. Sufficient capacity (AS3) can intensify groundwater exploitation (AS6). Only through collusion (AS4) can the option of unauthorised connections (AS5) actually emerge. Successful coordination (AS1) requires sufficient capacity as a precondition, but can afterwards even itself improve Table 6 Groundwater exploitation (AS6) is represented as a dilemma between two farmers (F1; F2) with the choices to extract (E; ~E) and a normalised parameter (γ) for pumping costs. AS6: F1

Fig. 3. Depth versus age of the bore wells in the case study area.

E ~E

F2 E 2-γ, 2-γ 0, 3

~E 3, 0 1, 1

C. Kimmich / Ecological Economics 90 (2013) 150–158

Fig. 4. The network of adjacent action situations is linked to the focal action situation AS1 with direct and indirect causal links and effect parameters.

the payoffs in AS3, because DSM reduce electric load on the grid. This makes AS3 crucial in those cases, where capacity is insufficient. Infrastructure capacity is only a problem in a share of the empirical cases, which makes a focus on AS1 and AS2 a practical guidance in the other cases. Best practice demonstration projects can thus focus on regions where electric utilities are cooperatively governed and successful watershed management schemes are established to ease coordination success.

7. Conclusions and Theoretical Implications This paper analysed adjacent action situations, which exert influence on the working components and outcomes of a core action situation. While the outcome of the focal situation is a physically manifest phenomenon, the adjacent action situations can shape one or more of its working components. Institutional components can be a result of singular action situations, but often emerge from repeated interactions, evolutionary in nature. A set of propositions of (1) potential types of links, (2) the boundary range of influence of the core actors, and (3) the dynamic capacities of linked action situations have been deduced from the adjacency and ‘ecology of games’ concept and tested empirically. The empirical evidence, given by the unexplained outcomes of the core action situation in the exemplary case of electricity supply for irrigation in South India, led to the identification of a highly influential network of situations of appropriation, provision, as well as collective and constitutional choice, linked through transactions, information, institutions, and actors. This adjacency network is heterogeneous with regard to the set of actors involved, the structure of the game model, and the time span and frequency of repeated interaction. A singular governmental order and a related campaign, as well as the evolved networks of exchange between farmers and the ground staff of the utility, were found to change the core action situation. The theoretical implications are threefold: First, as the empirical evidence suggests, there are a potentially unlimited number of adjacent situations, but not all may be relevant for the core action situation. The theoretical propositions and empirical findings indicate, how relevant action situations may be extracted, beyond magnitude of influence: An adjacent action situation may only exert its influence in the context of another situation, or the influence may be absent in the empirically observed cases, but activated under different conditions. Consequently, the extent of potential influence of an adjacent situation may be decisive with regard to its relevance. Ultimately, however, the network boundary, like the epistemological premise itself, remains empirically underdetermined and is influenced by the research purpose. Delimiting the range of influence of the core actors is only one option to draw a boundary for analysis, but has been sufficient to explain the outcome in the empirical case analysed.

157

Second, co-determination of outcomes through interdependence has been found to be crucial. Adjacent situations can impede improvements in the outcomes of a core action situation, as the example of technology adoption strategies indicates, but could well catalyse a higher equilibrium if factual coordination requirements were to be included in the learning process. A project-based intervention to demonstrate the outcome of coordinated investment into power quality measures may likely induce a shift in the social learning strategy, and thus have an impact far beyond the demonstration project itself (Mohan and Sreekumar, 2010). In addition, groundwater exploitation aggravates the conditions for providing infrastructure capacity, but could well ease the conditions for energy provision and coordination, if constraining institutions for groundwater extraction was successfully implemented. Third, a soft-link of model parameters helped to explain the empirical outcome. A formal analytical concept, available for homogeneous models in economic network analysis (Goyal, 2007), remains to be developed for linking heterogeneous adjacent action situations and their underlying game models. As Dutton et al. (2012, p. 59) point out: “Formal modelling of nested and interconnected games is primarily limited to sequentially connected games. Mathematical models of multiple games with multiple, and heterogeneous, players such as suggested by evolutionary game theory, is at very early stages of development”. The gap between classical and evolutionary game theory needs to be reduced in order to link many empirically relevant adjacent action situations. While the bounds to rationality are empirically often stronger than modelled in classical game theory, many of the mechanisms, such as inheritance and learning, in evolutionary game theory appear to be too weak (Hodgson and Huang, 2010; Scharpf, 1997), especially when contrasted with empirical evidence in the social spheres of infrastructure and resource governance, as demonstrated here.

Acknowledgements I would like to thank my colleagues from the Division of Resource Economics at Humboldt University Berlin, and the Workshop in Political Theory and Policy Analysis at Indiana University Bloomington. I would especially like to acknowledge the contributions of Elinor Ostrom, Michael McGinnis, Christine Werthmann, and Jens Rommel, who gave me valuable comments on earlier drafts of this work. I thank Philip N. Kumar and his team, Rama Mohan from the Centre for World Solidarity and Dr. Sreekumar from Prayas Energy Group Pune for the great collaboration in the field. Financial support from the German Ministry of Education and Research is gratefully acknowledged. All errors are my own.

References Anderies, J.M., Janssen, M.A., Ostrom, E., 2004. A framework to analyze the robustness of social–ecological systems from an institutional perspective. Ecology and Society 9. Andhra Pradesh Electricity Regulatory Commission, 2001. In: APERC (Ed.), Tariff Order: Review of Tariff Filings FY2001-02 (Hyderabad). Andhra Pradesh Electricity Regulatory Commission, 2005. In: Andhra Pradesh Electricity Regulatory Commission (Ed.), Tariff Order 2005-06 (Hyderabad). Andhra Pradesh Electricity Regulatory Commission, 2010. In: Andhra Pradesh Electricity Regulatory Commission (Ed.), Tariff Order: Retail Supply Tariffs for FY2010-11 (Hyderabad). Aoki, M., 2007. Endogenizing institutions and institutional changes. Journal of Institutional Economics 3, 1–31. APTRANSCO, 2008. Administration Report 2007–2008. Transmission Corporation of Andhra Pradesh Limited and AP Distribution Companies, Tenth Administration Report, Hyderabad. Axelrod, R., 1981. The emergence of cooperation among egoists. The American Political Science Review 75, 306–318. Bardhan, P., Ray, I., 2008. The Contested Commons: Conversations between Economists and Anthropologists. Blackwell, Malden, MA, USA; Oxford, UK. Beckmann, V., Padmanabhan, M.A., 2009. Analysing Institutions: What Method to Apply? In: Beckmann, V., Padmanabhan, M.A. (Eds.), Institutions and Sustainability: Political Economy of Agriculture and the Environment — Essays in Honour of Konrad Hagedorn. Springer, Dordrecht, pp. 341–372.

158

C. Kimmich / Ecological Economics 90 (2013) 150–158

Birner, R., Gupta, S., Sharma, N., Palaniswamy, N., 2007. The Political Economy of Agricultural Policy Reform in India. The Case of Fertilizer Supply and Electricity Supply for Groundwater Irrigation.IFPRI, New Delhi and Washington 207. Blomquist, W., Ostrom, E., 1985. Institutional capacity and the resolution of a commons dilemma. Review of Policy Research 5, 383–394. Bromley, D.W., 1989. Economic interests and institutions: the conceptual foundations of public policy. Basil Blackwell, New York, NY, USA. Bromley, D.W., 2006. Sufficient Reason: Volitional Pragmatism and the Meaning of Economic Institutions. Princeton University Press, Princeton. Chakrabarty, B., 2008. Indian Politics and Society since Independence: Events, Processes and Ideology. Routledge, London; New York. CMIE, 2008. In: Centre for Monitoring Indian Economy (Ed.), Economic Intelligence Service - Energy, November 2008 (Mumbai). Downward, P., Mearman, A., 2007. Retroduction as mixed-methods triangulation in economic research: reorienting economics into social science. Cambridge Journal of Economics 31, 77–99. Drèze, J., Sharma, N., 1998. Palanpur: population, society, economy. In: Lanjouw, P., Stern, N.H. (Eds.), Economic Development in Palanpur over Five Decades. Clarendon Press, Oxford University Press, Oxford, New York, p. xxviii (640 pp.). Dubash, N.K., Rao, D.N., 2008. Regulatory practice and politics: lessons from independent regulation in Indian electricity. Utilities Policy 16, 321–331. Dugan, R.C., 2003. Electrical power systems quality, 2nd ed. McGraw-Hill, New York. Dutton, W.H., 1995. The ecology of games. Communication Theory 5, 379–392. Dutton, W.H., Schneider, V., Vedel, T., 2012. Ecologies of games shaping large technical systems: cases from telecommunications to the Internet. In: Bauer, J., Lang, A., Schneider, V. (Eds.), Innovation Policy and Governance in High-Tech Industries The Complexity of Coordination. Springer, Berlin, Heidelberg, p. 1 (online resource). Eisenhardt, K.M., 1989. Building theories from case study research. The Academy of Management Review 14, 532–550. Foster, A.D., Rosenzweig, M.R., 1995. Learning by doing and learning from others: human capital and technical change in agriculture. Journal of Political Economy 103, 1176–1209. Gerring, J., 2007. Case Study Research: Principles and Practices. Cambridge University Press, New York. Golman, R., Page, S.E., 2010. Individual and cultural learning in stag hunt games with multiple actions. Journal of Economic Behavior & Organization 73, 359–376. Government of Andhra Pradesh, 2008. Socio-Economic Survey 2007–08. State Planning Department, Hyderabad. Goyal, S., 2007. Connections: An Introduction to the Economics of Networks. Princeton University Press, Princeton. Hagedorn, K., 2008. Particular requirements for institutional analysis in nature-related sectors. European Review of Agricultural Economics 35, 357–384. Harsanyi, J.C., Selten, R., 1988. A General Theory of Equilibrium Selection in Games. MIT Press, Cambridge, Mass. Hirschman, A.O., 1970. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Harvard University Press, Cambridge, MA. Hodgson, G.M., 2004. The Evolution of Institutional Economics: Agency, Structure, and Darwinism in American Institutionalism. Routledge, New York. Hodgson, G., Huang, K., 2010. Evolutionary game theory and evolutionary economics: are they different species? Journal of Evolutionary Economics 1–22. Janssen, M.A., 2007. Coordination in irrigation systems: an analysis of the Lansing– Kremer model of Bali. Agricultural Systems 93, 170–190. Jara-Rojas, R., Bravo-Ureta, B.E., Días, J., 2012. Adoption of water conservation practices: a socioeconomic analysis of small-scale farmers in Central Chile. Agricultural Systems 110, 54–62. Jenkins, R., 1999. Democratic Politics and Economic Reform in India. Cambridge University Press, New York ix (250 pp.). Kale, S.S., 2004. Current reforms: the politics of policy change in India's electricity sector. Pacific Affairs 77, 467–491. Kameda, T., Nakanishi, D., 2003. Does social/cultural learning increase human adaptability? Rogers's question revisited. Evolution and Human Behavior 24, 242–260. Kimmich, C., 2013. Networks of Coordination and Conflict: Governing Electricity Transactions for Irrigation in Andhra Pradesh, India. Shaker, Aachen. Kimmich, C., 2013. Incentives for energy-efficient irrigation: Empirical evidence of technology adoption in Andhra Pradesh. India. Energy for Sustainable Development. http://dx.doi.org/10.1016/j.esd.2013.02.004 (in press). Kiser, L.L., Ostrom, E., 2000. The three worlds of action: a metatheoretical synthesis of institutional approaches. In: McGinnis, M.D. (Ed.), Polycentric Games and Institutions: Readings from the Workshop in Political Theory and Policy Analysis. University of Michigan Press, Ann Arbor, p. xvi (541 pp.). Knight, J., 1992. Institutions and Social Conflict. Cambridge University Press, Cambridge. Kumar, M.D., 2005. Impact of electricity prices and volumetric water allocation on energy and groundwater demand management: analysis from Western India. Energy Policy 33, 39–51. Künneke, R., Groenewegen, J., Ménard, C., 2010. Aligning modes of organization with technology: critical transactions in the reform of infrastructures. Journal of Economic Behavior & Organization 75, 494–505.

Lubell, M., Henry, A.D., McCoy, M., 2010. Collaborative institutions in an ecology of games. American Journal of Political Science 54, 287–300. McGinnis, M.D., 2011. Networks of adjacent action situations in polycentric governance. Policy Studies Journal 39, 51–78. Meier, A.v., 2006. Electric Power Systems: A Conceptual Introduction. IEEE Press: Wiley-Interscience, Hoboken, N.J. Mohan, R., Sreekumar, N., 2010. Improving Efficiency of Groundwater Pumping for Agriculture: Thinking Through Together, Discussion Paper. Centre for World Solidarity, Prayas Energy Group, Hyderabad, Pune. Munshi, K., 2004. Social learning in a heterogeneous population: technology diffusion in the Indian Green Revolution. Journal of Development Economics 73, 185–213. Norgaard, R.B., 1984. Coevolutionary agricultural development. Economic Development and Cultural Change 32, 525–546. Olson, M., 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Harvard University Press, Cambridge, Mass. Ostrom, E., 2005. Understanding Institutional Diversity. Princeton University Press, Princeton, Oxford. Ostrom, E., 2010. Analyzing collective action. Agricultural Economics 41, 155–166. Ostrom, V., Tiebout, C.M., Warren, R., 1961. The organization of government in metropolitan areas: a theoretical inquiry. The American Political Science Review 55, 831–842. Ostrom, E., Gardner, R., Walker, J., 1994. Rules, Games and Common-Pool Resources. The University of Michigan Press, Michigan. Pani, B.S., Sreekumar, N., Reddy, M.T., 2007. Power sector reforms in Andhra Pradesh: their impact and policy gaps. In: Centre for Economic and Social Studies (Ed.), Governance and Policy Spaces. Centre for Economic and Social Studies, Hyderabad, p. 62. Poteete, A.R., Janssen, M., Ostrom, E., 2010. Working Together: Collective Action, the Commons, and Multiple Methods in Practice. Princeton University Press, Princeton, NJ. Putnam, R.D., 1988. Diplomacy and domestic politics: the logic of two-level games. International Organization 42, 427–460. Rapoport, A., 1966. Two-Person Game Theory; The Essential Ideas. University of Michigan Press, Ann Arbor. Robinson, D., Goforth, D., 2005. The Topology of the 2x2 Games: A New Periodic Table. Routledge, London; New York. Scharpf, F.W., 1997. Games Real Actors Play: Actor-Centered Institutionalism in Policy Research. Westview Press, Boulder, Colo. Sen, A.K., 1967. Isolation, assurance and the social rate of discount. Quarterly Journal of Economics 81, 112–124. Shah, T., 2009. Taming the anarchy: groundwater governance in South Asia. Resources for the Future.International Water Management Institute, Washington, D.C., Colombo, Sri Lanka. Shubik, M., 1986. The Games within the Game: Modeling Politico-economic Structures. W. de Gruyter, Berlin; New York. Skyrms, B., 2004. The Stag Hunt and the Evolution of Social Structure. Cambridge University Press, Cambridge, UK; New York. Stake, R.E., 2000. Case studies. In: Denzin, N.K., Lincoln, Y.S. (Eds.), Handbook of Qualitative Research. Sage Publications, Thousand Oaks, Calif., pp. 435–454. Swedberg, R., 2001. Sociology and game theory: contemporary and historical perspectives. Theory and Society 30, 301–335. Tarui, N., 2007. Inequality and outside options in common-property resource use. Journal of Development Economics 83, 214–239. Thatcher, M., 2002. Delegation to independent regulatory agencies: pressures, functions and contextual mediation. West European Politics 25, 125–147. Tongia, R., 2007. The political economy of Indian power sector reforms. In: Victor, D.G., Heller, T.C. (Eds.), The Political Economy of Power Sector Reform: The Experiences of Five Major Developing Countries. Cambridge University Press, Cambridge; New York, p. xviii (330 pp.). Tsebelis, G., 1990. Nested Games: Rational Choice in Comparative Politics. University of California Press, Berkeley. van den Bergh, J.C.J.M., Stagl, S., 2003. Coevolution of economic behaviour and institutions: towards a theory of institutional change. Journal of Evolutionary Economics 13, 289–317. van Steenbergen, F., 2006. Promoting local management in groundwater. Hydrogeology Journal 14, 380–391. Vatn, A., 2005. Institutions and the Environment. Edward Elgar, Cheltenham, UK. Walker, D.H., Sinclair, F.L., 1998. Acquiring qualitative knowledge about complex agroecosystems. Part 2: formal representation. Agricultural Systems 56, 365–386. Winther, T., 2012. Electricity theft as a relational issue: a comparative look at Zanzibar, Tanzania, and the Sunderban Islands, India. Energy for Sustainable Development 16, 111–119. World Bank, 2001. India: Power Supply to Agriculture. Summary Report, volume 1. World Bank. World Bank, 2010. Deep Wells and Prudence: Towards Pragmatic Action for Addressing Groundwater Overexploitation in India. World Bank, Washington D.C. Yin, R.K., 2003. Case Study Research: Design and Methods, 3rd ed. Sage Publications, Thousand Oaks, CA.