THE POUTECAL DESTRIBUTION 0F PUBUC POLICY GOODS EN RURAL INDIA: RAEASTHAN, 1961-1971 fiissemtioa hr the Degree of Ph. D. MECHEGAR STATE UMVERSTFY BRIAN WILSON BUYER £975 This is to certify that the thesis entitled THE POLITICAL DISTRIBUTION OF PUBLIC POLICY GOODS IN RURAL INDIA: RAJASTHAN‘, 1961-1971 presented by Brian Wilson Coyer has been accepted towards fulfillment of the requirements for J3h.D. degreein Poh'tica] Science / / Major professor 7 (— ./ Date ' VS 7 f 0-7639 ABSTRACT THE POLITICAL DISTRIBUTION OF PUBLIC POLICY GOODS IN RURAL INDIA: RAJASTHAN, 1961-1971 By Brian Wilson Coyer Once levels of public policy goods allocation are set, policy decision-makers face the task of distribution. A number of environmental factors may influence these distributions, yet little is known about which factors are important and when they are most likely to be salient features of a distribution system. Understanding this process in poor, agrarian societies is particularly important now because of the dilemma of increasing population and static agricultural production systems. Before policy—makers attempt to change their own rural environments, analyses of how different institutional arrangements are associated with policy distribution and performance should be formulated. This study utilizes aggregate data for several time periods taken from one Indian State, Rajasthan, for the period of 1961-l97l, to make inferences about the associations between the policy environment and rural policy distribution. Two policy areas are chosen for examination. These are rural development funding disbursed through panchayati raj organizations and rural electrification administered and distributed through the Rajasthan State Electricity Board. These policy variables are associated with electoral factors -- electoral mobilization rates and party fragmentation; economic development factors, particularly irrigation potential; and, institutional factors -- socio-economic dominance and political party dominance. Analysis of covariance and regression techniques are used to test five multipart hypotheses. For policy resources controlled by the regime party in Rajasthan's parliamentary system, party fragmentation and party control (political party dominance) are important factors in distribution decisions. Constituencies with highly unequal concentrations of socio-economic resources (land and status) also receive greater amounts of policy resources controlled by the regime party. Policy resources controlled independently by administrators -- in this case, rural electrification ~- are less affected by policy environmental factors. It appears that administrators avoid political interference from elected officials and resist pressures generated from contexts in which there is inequality of resource distribution. THE POLITICAL DISTRIBUTION OF PUBLIC POLICY GOODS IN RURAL INDIA: RAJASTHAN, 1961-1971 By Brian Wilson Coyer A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1975 For Shahnaz and Jesse ii ACKNOWLEDGMENTS Professor Timothy M. Hennessey, the Chairman, has seen this work through and assisted in limiting it to manageable proportions. His scholarly corrlnitment and interest in policy analysis has been a suitable example. Professor Frank A. Pinner's critical appraisal of several preliminary drafts has kept attention on important issues. Professor Robert Jackman contributed to whatever methodological skills are applied here. My iritellectual debts are many and varied. Professor Richard N. Blue, University of Minnesota, became involved at an early point in this research and has been consistently enthusiastic about it. Parts of the work might reflect his insights. Others have contributed to my under- standing of important intellectual issues in studying agrarian societies and public policy. Professor Ralph W. Nicholas, Professor Lloyd 1. Rudolph, and Professor Susanne Hoeber Rudolph, University Of Chicago, are central in my understandings of Indian society and 500161 Structure and Rajasthan's political process. Professor A. Allan Schmid, Michigan State University, introduced me to the importance of institutional variables for policy decisions and outcomes. I wish to acknowledge the financial support of the Midwest Univer- sity Consortium for International Activities who funded this research for the year 1970. The funds were managed jointly by the Asian Studies Center, Michigan State University, William T. Ross, Director, and the Center for International Programs, Michigan State University. During various stages of this work the Center for the Comparative Study of Technological DevelOpment and Social Change, University of Minnesota, provided financial and clerical assistance for research preparation and data processing. I am deeply grateful for the support of Professor Robert Holt, Director of the Center. Professor Iqbal Narain, Depart- ment of Political Science, University of Rajasthan, faciliated the work during field work in 1970-1971 by smoothing the way and providing advice and assistance at several points. I wish also to thank Pratap Singh Verma and Neelam Gupta, Univer- sity of Rajasthan, who aided this research as field assistants. Sri Verma introduced me to his home district and constituency, Jhunjhunu and Khetri , which started my interest in Rajasthan's rural political systems and remains my basis departure. iv TABLE OF CONTENTS CHAPTER ONE THE POLITICAL DISTRIBUTIONS OF PUBLIC POLICY GOODS IN RURAL INDIA: RAJASTHAN, 1961-1971 ......... . . . . . . INTRODUCTION . . . ................. . ..... A PARTIAL MODEL OF PUBLIC POLICY DISTRIBUTION . . . . ..... Public Policy Goods .............. . ..... Transactions and the Policy Distribution Process ..... The Policy Environment: Economic Criteria . . . . . . . . The Policy Environment: Electoral Factors . . . ..... The Policy Environment: Institutional Factors ...... AGRICULTURE, POLICY AND POLITICS IN RAJASTHAN ......... Agriculture in Rajasthan . . ............... Agriculture Policy in Rajasthan .............. Rural Development Funding: Panchayati Raj Disbursements ..................... Autonomous Policy Resource Distribution: Rural Electrification ..................... EXPLAINING RURAL POLICY DISTRIBUTION IN RAJASTHAN ....... Economic Deve10pment and Rural Policy Distribution Electoral Mobilization Rates and Rural Policy Distribution ...................... Party Fragmentation and Rural Policy Distribution ..... Socio-economic Dominance and Rural Policy Distribution Political Party Dominance and Rural Policy Distribution ...................... CHAPTER TWO RURAL POLICY DISTRIBUTION IN RAJASTHAN: RESEARCH DESIGN, MEASUREMENT, AND THE LOGIC OF HYPOTHESIS TESTING ....... INTRODUCTION .......................... RESEARCH DESIGN .............. . . . . ...... 54 Most Similar Systems in Rajasthan ............. 54 The Analysis of Change ............ . ..... 59 Level of Analysis: The Legislative Constituency ..... 6l OPERATIONALIZATION OF THE MAJOR VARIABLES ..... . ..... 67 The Dependent Variables ........... . ..... 67 The Independent Variables ........... . ..... 7O Conceptualizfing Agricultural Development . . . . ..... 70 The Electoral Environment ........... . ..... 73 Institutional Factors ............. . . . . . . 77 THE LOGIC 0F HYPOTHESIS TESTING .......... . ..... 80 CHAPTER THREE ECONOMIC AND ELECTORAL FACTORS AND THE DISTRIBUTION OF RURAL PUBLIC POLICY GOODS . . ................ 88 INTRODUCTION .......... . ............... 88 HYPOTHESES . .......................... 92 Rural Public Policy Goods Distribution and Agricultural Development ....................... 93 Rural Public Policy Goods Distribution and Electoral Factors ......................... 95 TESTS OF HYPOTHESES ...................... 98 Rural Public Policy Goods and the Environment ....... 99 Summary of the Findings .................. 104 DISCUSSION ........................... 108 CHAPTER FOUR SOCIO-ECONOMIC DOMINANCE AND THE DISTRIBUTION OF RURAL PUBLIC POLICY GOODS ...................... 117 INTRODUCTION .......................... ll7 Socio~economic Dominance and Rural Public Policy Distribution ...................... ll7 Socio-economic Dominance and the Rajasthan Case ...... l23 vi HYPOTHESES . . . .............. . . . . . TESTS OF HYPOTHESES ............ . . . . . DISCUSSION . . ............... . . . . . CHAPTER FIVE POLITICAL PARTY DOMINANCE AND RURAL PUBLIC POLICY DISTRIBUTION . . .................. INTRODUCTION . . .................. . Public Policy Distribution and Political Party Dominance .................. Political Parties and Rural Policy Distribution ir1 Rajasthan ................. HYPOTHESES ..................... TESTS OF HYPOTHESES ...... . .......... DISCUSSION ..................... CHAPTE R S IX EXPLAINING RURAL POLICY DISTRIBUTIONS BY AN ANALYSIS OF THE POLICY ENVIRONMENT .............. INTRODUCTION ..................... Rural Public Policy Distribution, Socio-Economic Dominance, and Political Party Dominance in Rajasthan ................... Analyses and Findings .............. A SUW'lARY OF THE FINDINGS .............. CONCLUSIONS ..................... ACCOMPLISHMENTS, LIMITATIONS AND FURTHER RESEARCH . . APPENDIX A LIST OF RAJASTHAN SAMPLE CONSTITUENCIES WITH CENSUS UNIT AND PANCHAYAT SAMITI NAMES .................. APPENDIX B DESCRIPTION OF INDICATORS FOR RAJASTHAN'S ASSEMBLY CONSTITUENCIES ......................... 145 145 145 149 155 158 166 171 171 171 174 178 182 191 209 216 LIST OF TABLES Page Table 1.1 Sumnary of Hypotheses ..... . . . . . ..... 50 Table 2.1 Sample Districts and Constituencies . . ..... 65 Table 3.1 Public Policy Distribution and Agricultural Development (Irrigation Percentages) . ..... 100 Table 3.2 Public Policy Distribution and Electoral Mobilization Rates ......... . ..... 103 Table 3.3 Public Policy Distribution and Party Fragmentation . .......... . ..... 106 Table 3.4 Rural Public Policy Distribution and the Policy Environment ............... 109 Table 4.'| Sample Constituencies Grouped by Socio- economic Dominance Type ............. 129 Table 4.2 Socio-economic Dominance and Rural Public Policy Distribution . . . ............ 135 Table 4.3 Socio-economic Dominance, the Policy Environ- ment, and Public Policy Distribution ...... 137 Table 4.4 Rural Public Policy Distribution, The Policy Env1ronment, and Socio-economic Dominance . . . . 139 Table 5.1 Political Party Dominance and Rural Public Policy Distribution . . ............ 161 Table 5.2 Political Party Dominance, the Policy Environ- ment, and Rural Public Policy Distribution . . . 162 Table 5.3 Political Party Dominance, the Policy Environ- ment, and Rural Public Policy Distribution . . . 165 (summary) Table 6.1 Rural Public Policy Distribution, Socio-economic Dominance, and Political Party Dominance . . . 176 Table 6.2 Rural Public Policy Distribution, the Policy Environment, Socio—economic Dominance, and Political Party Dominance . . . . ........ 183 viii Figure Figure Figure Figure Figure Figure Figure 1.1 1.2 1.3 1.4 3.1 3.2 13.3 LIST OF FIGURES Simplified Political and Administrative Structure in Rajasthan . . . . . . . ..... Summary of Percentage of Voting Turncnrt in Five Elections to the Rajasthan Legislative Assembly . . . . . . . ..... Political Party Support in Rajasthan'ss Electorate and Legislative Assembly ..... . ..... Cabinet Membership in Rajasthan: 1961-1968 Rural Policy Distribution and Irrigation Percentages, Bivariate Slopes . . . ....... Rural Policy Distribution and Electoral Mobilization Rates, Bivariate Slopes ....... Rural Policy Distribution and Party Fragmentation, Bivariate SlOpes ......... ix Page 34 36 37 39 105 CHAPTER ONE THE POLITICAL DISTRIBUTION OF PUBLIC POLICY GOODS IN RURAL INDIA RAJASTHAN, 1961-1971 Introduction Many agricultural nations are investing heaviky to grow more food. Zhufia and its sub-national units, the state governments, have placed a special emphasis upon agricultural productivity. During the past decade the pressures on food supply brought about by an expanding population have placed additional strain’on India's policy-makers. India's 550 millRNIVWill expand to nearly one billion near the year 2000. While technologies are available for increasing agricultural production, the solutions 11) increasing food grain yields are bounded by political and institutional factors, also. Agrarian nations' ability to manage agricultural resources and technology has been questioned by numerous scholars. Those nations with decentralized decision-making structures have been most severely criticized. Francine Frankel and Karl von Vorys (1972) conclude for India that . the introduction of new agricultural technology commonly known as the green revolution, is accompanied by an accelerated disruption of traditional rural initiatives; forces already in motion will push tradi- tional socieites in rural areas to a total breakdown before an alternative system of mutual obligations can emerge and be established. . . . It is difficult to imagine just how democratic politics, or any political system, for that matter, can survive when it is based on a population so fragmented and radicalized. (1972: 37-38 This judgment is premature and preceeds a theoretical understanding of the relationships between the green revolution technologies and the management of resources.1 In fact, there is evidence to suggest the green revolution is not breaking down traditional rural structures but has solidified the power bases of rural elite and blocked the redistribution of policy and technological benefits. To discuss "total breakdown" of rural social structure is neither grounded in an accurate or compelling analysis of rural social structure nor is very useful in suggesting alternative technologies or policies. The current issue is not whether any given political system will survive, but is how political arrangements have an impact on which technological alternatives are chosen and how policies are made and policy goods distributed. Such analyses may lead to the discovery and manipulation of instrumental factors and evaluation of policy-makers. It is necessary to move to hypothesis testing and specifi- cation of feasible institutional alternatives. There is little analytic writing on how India's (or any) democratic structures effect the management of agricultural resources. Yet there is speculation about whether or not decentralized systems are appr0priate for economic development in agrarian societies.2 Much of the agricultural development literature tends to take constitutional arrangements and administrative structure for granted, however, particularly in empirical research. There is abundant literature explaining agricultural producti- vity as a function of specified factor inputs, e.g. seeds, water, labor, and so on.3 Yet organizational components bringing these imputs together are often ignored. This oversight might be due to the difficulty of analyzing how large-scale organizations may influence the agricultural policy process and, then, how to measure the specified variables. There are, however, partial models available and this work is a small step in the direction of ordering data for one democratic agrarian society.4 This work dimensions the rural policy distribution process in one Indian state, Rajasthan, and focusses upon one decade of policy distribution, 1961-1971. Before describing the empirical part of this monograph, attention is given to the theoretical factors which lead to the statement of hypotheses and measurement of important variables. A PARTIAL MODEL OF PUBLIC POLICY DISTRIBUTION One type of constitutional arrangement is analyzed here: a parlia- mentary democracy organized as a federal system. This means that regu- lar elections are held wherein universal adult suffrage and voter parti- cipation determine some of the officials who make policy. Candidates are chosen by political parties to run for elected office. The regime is formed by a political party or coalition of parties whose leader is the chief executive officer of the government. Opponents attempt to gain power while sitting in the legislative Assembly. An additional feature of this limiting case is that a non-elected, semi-autonomous administrative organization parallels a hierarchy of elected officials and is responsible for the implementation of policy. These administra- tors may exercise discretion in key areas. Administrators have greater control over policy resources when they are separated from political pressures brought to bear by a political party organization and when the policy goods they control are more easily dimensioned in economic (or rational) terms. When rational, planning criteria can be utilized for policy distributions, administrators exercise considerable authority. Organizations outside the governmental arena may attempt to exert influence on policy decisions and implementation. Such non-governmental organizations have been termed "interest groups" by some scholars and ”sectors" by others.5 Voters, as individuals, or as part of some political organization, may seek to receive larger portions of the policy good than others. The system is decentralized; pressures from political actors at the lowest level of representation and administration are applied to achieve higher levels of policy allocation. What happens at the bottom of this set of political institutions matters a great deal though there may be periods when the control of a legally superior authority at the center of the system may be exerted. The institutional feature of this policy environment may be conceptualized and measured. It makes sense to investigate the conditions under which the policy environment may affect policy distributions and what aspects of that environment are salient for the policy distribution process. Public Policy Goods Among the many characteristics of public policy goods, two are central to this analysis.6 First, public policy resources are, to some degree, divisible; and, second, public policy goods may be more or less specifically associated with a desired societal product. Governments can make decisions on the extent to which certain persons or groups can benefit or absorb loss from a given public policy resource. And, some public policies are more directly related to an economic production function that others. For example, we seem to know more about regulating the monetary system than we do about improving the "quality of life" in urban (or rural) areas. With regard to policy goods' divisibility, as Curry and Wade (1968) point out One may have more or less than whatever it is that current policy does in the way of allocating things people want. . . . Public highways are clearly divisible . . . as are military forces, taxes, maritime subsidies, welfare programs, government ' controls over business, public jobs, and all other public policies and programs. (1968: 3). Governments provide public policy goods which if consumed by one diminish the consumption of the same good by another and where the exclusion of potential customers is feasible.7 Policy goods may also vary on the cost of the first unit and the nature and extent of externalities, and have many of the same characteristics of priVate goods except they are produced and distributed by governments often by invoking authority or force. Any public policies which are produced by political processes may benefit some more than others or some may pay higher costs than others. Costs and benefits of public policies are more or less divisible. Public policy goods are distributed with social, political and/or economic goals in mind. The relationship of a specific public good to the desired outcome may be well-known or may be less well understood. In the case of producing food grains, there are well known functions combin- ing many material inputs. For example, policy—makers, given a commitment to certain technologies, can predict accurately the amounts of water necessary, in combination with other inputs, to produce a given amount of wheat. Policy-makers can choose alternatives of water supply to conform to these constraints. If the food production paradigm is known, fewer exogenous factors can influence the distribution of public policy goods intended to create a steady water supply. However, when the paradigm for a given produce is not familiar, the number of exogenous forces which may influence the allocation of public policy goods expands. Take, for example, the goal of producing "quality education," a goal prominent in discussions of racial desegregation, educational financial structure, and control of the educational process.8 Very few of the important hypotheses concerning "quality education" have been tested. What is quality education for some, say better college preparation, may not be quality education to others who might prefer more Specialized vocational education. Further, even if everyone agreed on what the outcome of the educational process might be, there are numerous strategies available. When governments provide public policy goods for a loosely defined goal, where the "production function" is less clearly formulated, political factors are more likely to explain policy goods distribution. A later section discusses this political process and policy goods distribution in a democracy. The divisibility of public policy goods and the degree of specificity of knowledge relating these goods to production functions are two important characteristics included here. Ilchman and Uphoff (1969) in a relevant discussion, comment on the "tangibility" and "intangibility" of some policy goods and resources. Their general list includes tangibles such as economic goods and services, and intangibles including information, status, force, legitimacy, authority and coercion. The emphasis upon "tangibility" or "intangibility," however, obscures the continuous process by which policy goods are given value. When we ask what is the value of something and make a new discovery about the relationship of a good to a human preference, the discovery becomes useful and somehow the policy good becomes more tangible when before it was listed as an "intangible." It is p0 - urhw a given produce is not familiar, the number of exogenous forces which may influence the allocation of public policy goods expands. Take, for example, the goal of producing "quality education," a goal prominent in discussions of racial desegregation, educational financial structure, and control of the educational process.8 Very few of the important hypotheses concerning "quality education" have been tested. What is quality education for some, say better college preparation, may not be quality education to others who might prefer more Specialized vocational education. Further, even if everyone agreed on what the outcome of the educational process might be, there are numerous strategies available. When governments provide public policy goods for a loosely defined goal, where the "production function" is less clearly formulated, political factors are more likely to explain policy goods distribution. A later section discusses this political process and policy goods distribution in a democracy. The divisibility of public policy goods and the degree of specificity of knowledge relating these goods to production functions are two important characteristics included here. Ilchman and Uphoff (1969) in a relevant discussion, comment on the "tangibility" and "intangibility" of some policy goods and resources. Their general list includes tangibles such as economic goods and services, and intangibles including information, status, force, legitimacy, authority and coercion. The emphasis upon "tangibility" or "intangibility," however, obscures the continuous process by which policy goods are given value. When we ask what is the value of something and make a new discovery about the relationship of a good to a human preference, the discovery becomes useful and somehow the policy good becomes more tangible when before it was listed as an "intangible.“ It is important to remember that even if our analyses are incapable of placing a policy good in one category or another, policy distribution processes proceed so that even if we do not know the “value" of a policy good, those involved in the process directly do make some estimation of its worth. Policy goods divisibility and production specificity are two categories for explaining the distributions of public policy goods in one political system. Transactions and the Policy Distribution Process Public policies and goods are produced by transactions between decision-makers in institutional arrangements. Decision-makers may consider a number of factors including those in the policy environment in determining levels of policy goods allocation. The total amount of public policy goods produced by one government can be understood by an analysis of intra-governmental bargaining and exchange, the supply of resources available, and the nature of the production and distribution systems.9 The level of public policy goods, produced for the system is assumed here. What is of interest is once a level of policy goods has been determined, what factors impinge upon its distribution. Why do some persons, organizations or regions get more of what is available than others? As such, this analysis is interested in making inferences about transactions between state-level organizations and policy actors at lower levels of the political process. Further, it can be suggested what conditions encourage and discourage transactions between state-level decision-makers and local-level political actors. Important transactions, or bargained exchanges, occur between the regime and local-level political actors. It is useful, in these terms, to distinguish between "nonpolitical" and "political" transactions: Most goods and services, status, information are exchanged directly between members of sectors without regime interference or involvement. These exchanges are not political. However, many exchanges involve authority or another resource held by the regime. Whenever sectors use resources to affect public policy or the statesman combines his resources into policies affecting resource allocation and aiming at compliance, these exchanges are political. (Ilchman and Uphoff, 1969: 94). One clarification might be added here. Bargained exchanges between regimes and local-level policy actors are transactions when compliance and authority are essential features of the activity.10 The government may be able to establish price unilaterally at some point in the bargain. Varied decision-makers control different resources to be used in trans- actions. Political parties may gain access to policy goods which are use— ful in the maintenance of their organization and winning elections; admin- istrators may control other policy goods. Party organizations may have a special affinity for divisible policy goods for which there is consider- able ambiguity in the application of the policy good to a production function. Such policy goods might be directed toward political and/or electoral outcomes. Administrators may have access to less divisible goods or those which are more easily related to an economic production function. Transactions occur when two or more sets of actors hold resources which are mutually desired. Parties may exchange policy goods fer votes; administrators may exchange policy goods for support or promises not to interfere in administrative affairs. Either set of decision-makers may trade a policy good for control of a process or to block rivals. Administrators or political parties may also withhold goods for local- level actors to deprive them of resources which might be used to undermine the decision-makers' positions. Transactions are not made without limit and are conducted in rule- defined arenas. Strategies for achieving an advantage in bargained exchange are developed by interacting with the formal and informal rules of the system. The rules of the system separate the policy-making activities from policy implementation. Constitutional arrangements provide for elected officials to "make" policy while non-elected officials, administrators, are formally responsible for seeing that policies are carried out. The informal rules of the system determine the degree to which policy-makers can exercise their will and the degree to which administrators can impose their own solutions upon elected officials and their clientele. Further, the rules of the system may allow and protect the participation of citizens in the political process by direct election of policy-makers. Each voter is given the right to choose a candidate or political party through a direct electoral process. Variance in voter participation and preferences of political parties is relevant to the strategies developed and pursued by the political party which forms a government. There may be only one regime party (or coalition) at a time and other parties and coalitions may seek to become the regime or govern- ment party. The strategies developed by the regime party are likely to be based upon transactions between the government and political supporters at the local-level which strengthen the regime party's electoral position. Administrators may be protected from intrusion by political party organizations and by influential local-level political actors by rules 10 against bribery and corruption. These formalities and other social and attitudinal factors, give reason for administrators to resist pressures from political party interferents. Transactions between state-level decision-makers and local-level political actors can and are more likely to be sought under some conditions than others. Planning criteria apply to a policy resource are a first limiting factor on these transactions. If the policy model calls for a policy good to those who match explicit criteria, others who do not match are excluded. But when criteria are not stated or are unenforcable, other actors, not originally intended to be beneficiaries of the policy good, may be included. A second consideration is when the regime party does not need additional Members to control a legisla- tive Assembly and the prospects for current Members retaining their seats are good. Under these circumstances, there is less reason to heed the demands and needs of constituents at the local-level. However, when the electoral environment becomes more competitive and when the margin between the government (or regime) party and opposition groups (or coalitions) is slim, more attention is given to local political processes. This redirection of attention by political parties from governing to seeking election may have significance for both the internal organization of the regime party and its actions on policy distribution.]] State-level administrators might perceive the electoral environment to be important when greater pressures are brought upon them by political party organizations for increased levels of resources. Clearly, administrators' views of the electoral or policy environment may follow those of elected decision-makers because elected officials make the ll policies administrators are to implement. At the same time, administra- tors may use policy resources to increase their control over policy processes and utilize information available from the policy environment for that objective. If it is true that administrators' perceptions of the policy environ- ment depend upon the saliency of that environment for elected officials, it may be reasoned many of the same factors important as policy distribu- tion criteria for elected officials will be critical for non-elected, administrators also. Administrators may react differently to the environ- ment than elected officials, however, even if the environmental factors are identical. The same policy environmental pressures which may cause the regime party to yield policy resources are those which can invoke negative sanctions from administrators. This is so when administrative agencies are defined as autonomous organizations and when these agencies select their own personnel without intrusion from external sources. Administrators are likely to utilize policy resources to control more resources, achieve specific performance criteria, and to resist inter- ference by other agencies, political parties and influential local-level actors. This analysis assumes that administrators and elected regime party (and opposition) party Members are in conflict and respond to similar policy environmental factors in a manner which reflects their own interests.12 Three aspects of the policy environment are related here to trans- actions between the regime party and local-level political interests. First, rational, planning criteria may be applied to the distribution of public policy goods, where appropriate. Even when administrative and 12 political systems are decentralized, planning occurs. It may be perceived that "a planned economy can be implemented rationally in a contemporary society . . . by a process of multi-level planning and decision-making." (Sau, 1971: 1759). Second, voter turnout and voter party preferences matter to elected officials and administrators; and, three, there are local-level "institutional" factors important to both sets of policy decision-makers. Voter turnout and preference are significant factors because electoral support allows the retention of a regime party or its removal. Institutional factors refer directly to control of a constituency. Political party organizations may mobilize support for candidates at the constituency-level and/or informal organizations based upon wealth and status may be important for both the election of Members to the Assembly and for the distribution of policy resources.13 Voting behavior is linked, no doubt, to the institutional characteristics of a constituency. For the sake of a clear understanding of each set of factors, informal organizational and party control are included separately in the analysis. The Policy Environment: Economic Criteria Policy resources are intended for many purposes and goals. Policy goods are allocated for regulation or for the distribution and redistribu- tion of benefits. Policies included here are those which distribute a society's scarce resources for the production of benefits for a rural population. These policies can be associated with a social or economic product in more or less clear terms. The policy goods either are intended for the socio-economic uplift of persons living in constituencies or the 13 improvement of agricultural productivity. A more careful consideration of economic criteria in policy goods distribution is given below in the section on agricultural policy in Rajasthan. The Policy Environment: Electoral Factors Voter turnout and party preferences are vital factors for elected officials and administrators who distribute public policy goods. Shifts in the percentage of the electorate who vote may have an independent impact on who wins in a constituency. If the number of voters who participate in an election increases by a large percentage over previous elections, the chances of a shift in representation are greater. There is a greater likelihood that a new party or group may be able to mobilize their supporters to defeat an opponent. The number of parties or candidates who compete in a constituency also has an impact on the changes of any given party for victory. -Where there are more parties competing, the chances are that winning may be accomplished by a smaller percentage of the votes in a plurality system. The character of the cleavages between political parties is a critical factor in predicting who may win an election and how policy resources may be invested to affect an electoral outcome. Voter turnout at one point in time is not an adequate indicator of voter participation considered by regime parties in the distribution of public policy resources. Chaggg_in voter turnout between two or more elections may be more carefully assessed by regime party leadership. If a party leader or administrator can guage the changes between an immediately preceding time and the present, it may be possible to direct policy resources for electoral outcomes. A measure of change in voter turnout 13 improvement of agricultural productivity. A more careful consideration of economic criteria in policy goods distribution is given below in the section on agricultural policy in Rajasthan. The Policy Environment: Electoral Factors Voter turnout and party preferences are vital factors for elected officials and administrators who distribute public policy goods. Shifts in the percentage of the electorate who vote may have an independent impact on who wins in a constituency. If the number of voters who participate in an election increases by a large percentage over previous elections, the chances of a shift in representation are greater. There is a greater likelihood that a new party or group may be able to mobilize their supporters to defeat an opponent. The number of parties or candidates who compete in a constituency also has an impact on the changes of any given party for victory. oWhere there are more parties competing, the chances are that winning may be accomplished by a smaller percentage of the votes in a plurality system. The character of the cleavages between political parties is a critical factor in predicting who may win an election and how policy resources may be invested to affect an electoral outcome. Voter turnout at one point in time is not an adequate indicator of voter participation considered by regime parties in the distribution of public policy resources. ED229£.1" voter turnout between two or-more elections may be more carefully assessed by regime party leadership. If a party leader or administrator can guage the changes between an immediately preceding time and the present, it may be possible to direct policy resources for electoral outcomes. A measure of change in voter turnout 14 between two elections may be part of a decision-maker's model of policy resource distribution. Voter turnout change is here termed electoral mobilization rates and is defined as the positive or negative change occuring in voter turnout for one constituency between at least two 14 elections. Increases in electoral mobilization rates imply a fluidity in the electoral environment. The number of party preferences and the strength of these parties in a constituency may affect regime party and local-level policy transactions. There may be more than one or two parties competing in a constituency and if a sizable proportion of the voters chooses each party, only a small plurality of voters may elect a Member to an Assembly seat. Minor parties are crucial in these constituencies because small, but loyal, numbers may win a seat. Cleavages in the party system with a constituency can be equally or more salient for policy decisions than electoral mobilization rates. There may be a wide range of cleavage situations across constituencies which have a number of parties. It is not so much an absolute increase in voters as it is the percentage of the vote necessary to change an electoral outcome. As more parties claim a larger portion of the vote, margins between winning and losing become smaller. The character of party cleavages in the constituency for the Assembly seat can be summarized by the term party fragmentation.15 A constituency may be less fragmented as the number of parties declines to approach one and there is near unanimous support for one party; a constituency may be more fragmented when the number of parties increases and the proportion of those voting for each party increases to the point when each party has nearly the same proportion of those voting. 15 The Policy Environment: Institutional Factors Economic criteria and electoral factors -— electoral mobilization rates and party fragmentation -- are part of a partial model for policy goods distribution in a decentralized, democratic system. Yet, a voter's calculation of benefits acrued or costs incurred from his voting decision must include the temperance of group memberships. This is independent of whichever political party he might wish to choose or his innate desire to vote, say, as a function of political knowledge. If a voter is dependent upon another person or group for economic welfare or status in a community, there is a greater chance that some significant other will prevail in the voting decision. The voter will comply with a "dictator's" wishes and shed preferences he might otherwise have expressed. "Control" of a constituency and its voters becomes an important 16 Control of a factor to be added to the policy distribution model. constituency may be conceptualized in a number of ways, but two come to mind immediately. First, the bases of much political organization are the distribution of a society's valued resources including wealth and status. Inequalities in these distributions take many forms including political "machines" in urban areas to "patron-client“ relationships 17 A second form of control typically associated with agrarian societies. of a constituency important for policy distribution is which party sends a Member to the Assembly. While this institutional factor is less stable, it may explain a great deal about why certain constituencies receive more policy resources than others. 16 Socio-Economic Dominance. The strength of local-level political organization may rest on the degree of concentration of socio-economic resources. Where private property holdings are customary, inequalities in the holdings of status and economic goods may be present. These inequalities, whether induced through force or shrewd management, are regularized and affect the lives of those who depend upon the owners of economic production resources for livelihoods. High levels of inequality in economic goods and status covary with dominance by those who have much over those who have significantly less. The other pole of this discussion is seldom examined. When property rights and means of production are more equally distributed, there is less social, economic, and political dominance by those only slightly wealthier. Where socio-economic resources are concentrated in the hands of a few, one may speak of a high socio-economic dominance context. Where socio—economic resources are less concentrated, it is possible to speak of a low socio-economic dominance context. Because the distributions of social and economic resources in the constituency are slow to change, there is sufficient reason to believe that the relationships which depend upon social and economic resources are institutionalized or regularized with the passage of time. Political Party Dominance. A second local-level institutional factor is which political party sends a Member to the Assembly from a constituency. In democratic systems, regime parties govern and opposition parties await the opportunity. Regime parties may treat opposition party constituen- cies differently in the distribution of policy goods. The regime party controls some of the important policy-making agencies and may distribute .‘. n, he. '1 l 1 .1“ \ V. 1 ‘l A ~ Ali - . pus . uhv : lit a J .h . 17 policy goods to its own Members at the expense of Opposition Members. Regime parties do not control every constituency, however, and any party may compete in a constituency so long as it can send a Member to the Assembly from some region and can keep an organization. The regime party is under no obligation to supply parties with resources for use in the maintenance of opposition candidates or Members. Political party dominance reflects which party's candidate goes to the Assembly from the constituency. There are a number of logical relationships between socio-economic dominance and political party dominance. A constituency may be controlled by the regime party in a high socio—economic dominance context. A party and local elite may be intertwined. Similar to this is the close relationship between an opposition party and important local actors when an opposition party sends a Member to the Assembly from a high socio-economic dominance context. The regime political party may control a constituency in a low socio- economic dominance context; and, an opposition party may send a Member to the Assembly in a low socio-economic dominance context. A party may dominate a constituency in a low socio-economic dominance context, but parties must certainly deal with local actors in a high socio-economic dominance context. The Policy Environment and the Distribution Process Once the level of policy goods allocation is set for any given substantive policy area, political parties and administrators distribute policy resources. Policy resources are more or less divisible and may be more or less specifically applied to a societal outcome or goal. ‘.. Q p - Lt! (JV . u . C I “('1 '1'! .- ““3 . 18 Political parties may control goods apart from administrative inter- vention; administrative agencies may be able to distribute policy resources without recourse to authority from the regime party. It may be assumed that the regime party and state-level administrators are often in conflict with one another for control over scarce policy resources. Both sets of decision-makers distribute policy resources as a function of three sets of variables included in this analysis: economic factors, electoral factors, and institutional factors. These are only a few of the possible factors which may have an impact on policy distri- bution. It is assumed that each set of decision-makers attempts to maximize the utility of the resources controlled in transactions with local-level actors at the constituency-level. Resources held by local actors may become important under conditions described above. The question now becomes what is the logic of the distribution of policy goods in terms of each set of decision-makers? The regime party's leadership is comitted to maintaining a viable organization in the Assembly while winning enough seats to insure a majority in the Assembly. At least two electoral factors have been mentioned as important in assessing the policy environment for trans- actions between the regime party and local political actors. The party cleavages in the constituency would seem to be significant fer policy resource distribution. The relative differences between winner and loser are critical. When those differences are small, the regime party investments of policy goods may be perceived to bring greater returns than when those percentages are great. With smaller differences, the regime party's investments of policy goods might retain a seat for its ‘5!‘ ‘BC‘ ... .I‘ ‘- '~ In. 1'1 19 own candidate or unseat an opponent. The higher the level of party fragmentation, the higher the rate of policy distribution. With respect to party fragmentation, who controls a constituency might be important. When the regime party sends a Member to the Assembly, the Member's constituency might well expect to receive more policy resources than a Member of an Opposition party. As an Opposition party Member's constituency becomes more fragmented, however, the regime party is more likely to sense the possibility of a victory and distribute more policy resources to that constituency. Who controls a constituency may affect the association between party fragmentation and the distribution of policy resources controlled by the regime party. Because victory in the Assembly constituency is so important to both regime and opposition party groups, electoral mobilization rates could have an impact on regime policy distribution. However, the rate of change in voter turnout, or electoral mobilization rates, may have only an indirect influence on victory at the constituency level. While_ increasingynumbers of voters implies a greater uncertainty and fluidity, these increases (or perhaps decreases) are only important as a function of the party fragmentation of the constituency. With large differences between the winning and losing parties, very high rates of change might make no difference on the electoral outcome. And where the difference between winning and losing is small, only a slight change can modify the outcome. Electoral mobilization rates, for the regime political party, are likely to have little independent impact on policy goods distribution. It is implied above that electoral factors, alone, do not explain the distribution of policy resources by regime party leadership. Within 20 constituencies institutional factors intervene in the distribution of policy resources. Socio-economic dominance and political party dominance might have a determining impact. Because political parties are vulnerable to constituency-level pressures related to the threats to withhold electoral support, there is every likelihood that where socio-economic dominance is associated with electoral behavior, the regime party will reSpond. Higher levels of socio-economic dominance are likely to be associated with higher average rates of policy resource distribution. Lower levels of socio-economic dominance may receive lower levels of policy distribution. Further, which political party controls a constituency may have an impact on policy resource distribu- tion rates. The regime party may be more likely to help its own Members than those of opposition parties when the regime party control policy resources. Where opposition Members are returned to the Assembly, they are less likely to have high average rates of policy resource distribu- tion than regime party Members' constituencies. State-level administrators may control policy resources unavailable to the regime party. When this is the case administrators have wide discretion on how to distribute those goods. When economic criteria are stated for distribution and information is accurate pertaining to those criteria, administrators distribute resources on those bases. Yet administrators are not without political performance criteria which can be subsumed in economic goals. Both Tullock (1965) and Niskanen (1971) assert the acquisitiveness of bureaucrats; administrators may seek to Inaximize resources or expand their sphere of control over resources in other programs. The spectrum of control includes the deliberate 21 avoidance of "political interference" by organizations and individuals outside of the bureaucracy. When bureaus are established apart from political party control, encroachment either from a party leader or important local-level leader can be repulsed on the basis of "efficiency." If regime (or opposition) party Members request policy resources, they may be denied if relinquishment of administrative control is implied in the transaction. Where economic performance criteria may be stated specifically for a policy resource, administrators are committed to them. In addition, the electoral environment is, or may be, part of the administrators calculations. It has earlier been noted that regime party decision- makers may respond to party fragmentation and not consider, as closely, electoral mobilization rates. However, if electoral mobilization rates are associated with socio-economic dominance in constituencies and administrators wish to avoid interference from local-level political actors, it is likely that electoral mobilization rates will be negatively associated with the distribution of policy resources controlled by administrators. Because administrators are distinct from party organiza- tions when the rules of the system prescribe administrative autonomy, party fragmentation is unlikely to have any immediate importance for administrators. In the same vein, which political party sends a Member to the Assembly is unlikely to have any direct impact on the distribu- tion of administratively controlled policy resources. To be sure, all of the possible considerations of regime party leadership and state-level administrators in transactions with constituency- level actors are not covered. There may be variations of these general- izations which can be stated in subsequent chapters in this monograph. 22 Hypotheses to be tested are found at the end of this chapter and in appropriate chapters later on below. All of the above analysis has an empirical referent and the discussion turns now to a description of the substantive policy area and geographical location chosen for the empirical analysis. We turn to a discussion of agriculture, policy and politics in Rajasthan. AGRICULTURE, POLICY AND POLITICS IN RAJASTHAN In the Indian states, public policy in agriculture is formulated by an interlocking set of individuals and institutions. At the pinnacle of the apparatus is the Chief Minister and his Cabinet. At the base of the system politically are the various organs of self-government. In between are Members of the Assembly who, when not holding ministerial rank, often serve as brokers between local leaders and the regime party. While most policies are formulated by elected officials, they are implemented and supervised by administrative personnel. Administrative agencies are statutorily responsible for decisions on the distribution of some agricultural resources, in addition. Administrators function at the levels of the state, district, and sub-district and often have discretion on the enforcement and implementation of policy. Within India's decentralized policy-making and -implementation environment is a nearly universal concern for planning agricultural growth and the application of rational criteria to the policy process. While rational planning is important, political processes are at the center of policy allocation and policy resources distribution. 23 Agriculture in Rajasthan Rajasthan is an appropriate site for the investigation of hypotheses associating economic and political antecedents of policy distribution. Eighty percent (80%) of its population live in rural areas, seventy-nine percent (79%) of the work force are in agriculture, and over fifty percent (50%) of its wealth is taken from agriculture. Rajasthan is located in the northwestern sector of India and has a long border with Pakistan extending along the Thar Desert. Map 1 identifies Rajasthan's location in the Indian Union. It is the second largest state in India with 342,000 square kilometers and tenth in population with nearly 27,000,000 persons in 1971. Population density in Rajasthan is well below the Indian average. In 1971, there were seventy-five (75) persons per square kilometer; India's population density is 182. Until the early 1950's Rajasthan was a net importer of food grains. She has fluctuated since as an importer and exporter among Indian states. Crops are dependent upon the monsoon of late June and early July. If rains are insufficient, the kharif_crop (cultivated during the rainy season) is damaged and the following rabi_(the year's second crop cultivated during the cold season) is also affected. There are occasional rains in December and January in the eastern parts of the state which can abate the loss of rabi_but these are nearly always restricted to the plains and plateaus of the east. Irrigation is one way of managing water supplies in the absence of frequent or insufficient rainfall. Rajasthan ranks fourth among Indian states in the amount of irrigated land -- 8.3% of the Indian total -- but 24 MAP 1 Cease Fire Line. KASHMIR ‘ c \ 6‘ hundignrh‘J lmla \D \39 DELHI ' , Main Political sanagnré7mm & .5" DIVISIOHQ Pym“, 0 :‘(HIMACHAL PRADESH W. 1 Annular-’8 7 1X 47*001111 “\ u§§tos€2 a UTTAR p \ ,8 R G k s. 405 1. ‘32, ' m" 255 :N'thémuo ®le \. shillomtg lonitihaT. ,. - <' ',MEGH‘ALAYA 9" I Gandinagar \ B'HAR ® GUJARAT MANIPUR J\ ®Aizawl BENGAH TRIPURA MIZORAM MADHYA PRADESH 6530 1 4v a ORISSA 9% [38' 0‘ 0‘6““ @311 b NagarAvel'm \ MAHARASHTRA ‘1 “98"" Bombay0 Hyderabad 8“- MRI Pnnaji GOA} Madras 8 LAKSHADWEEP . g :1 POND'C"ERRY ANDAMAN -. pm an.» Kawarathy® "' E AND 0 *1 5' .3 NICOBAR ': 'l 7 . ISLANDS ' 1:1. Trlvnndrum '4 8:01.: ‘6 '1'“ 320 mllostotholneh . , 205 km. to tho cm. Q Source: Kingsbury (1974: 49). 24 MAP 1 Main Political -W ® JAMMU &J Sfimsar KASHMIR‘ - ° - -, \. DIVISIons \ PUNJAB 0 gillMACHAL PRADESH hundigath ""13 \' ‘6 DELHI ) M {MI}. ’9‘}? Delhi "\‘ s\\g\K\ ' @‘S, 0 5 9 UTTAR pRAD'\ cat I“. [S , Gangtok ~ J. ® 8” “m...” ' “58 NAGALANO RNASTHAN Lucknow Patna . 9 Kohim. g \. SIIIIIongO l-mpm ",MEISII‘ALAYA 0' MANIPUR Aizawl ‘00—- "' Gandinagar \ ® BIHAR W. I A‘“mh§0’l.® @ma' BENGALx \ MIZDRAM GUJARAT MADHYA PRAOESH @Ol "“90“ 0/ 6‘ «(a ORISSA “We Dodralu 0'6 ' ' ® IegarAveII/x'MAIIAII ASHTRA B“""‘”°“’" BombayO Panoji GO LAIIsIIADIyEEP , " ‘ PONDICHERRY ANDAMAN - . Port Blair Kawanthy ’0 "’3 E AND O 1.1. .9 a; NICOBAR 1. f '7 ..' ISLANDS . '1; 1' Trivnndmm ' 5““: 3'4 K.“ 320 «III» Io tho Inch , 205 km. to the cm. Q Source: Kingsbury (1974: 49). 25 does not produce a proportionate amount of India's food supply. In 1968, Rajasthan grew 7.7% of India's cereal food grains. The major portion of these crops were grown on Rajasthan's estimated 2,161,000 irrigated hectares (5.3 million acres). The most important cereal grains are wheat, barley, jgwar_(sorghum), bajrg_(millet) and maize. A small amount of rice is grown in some areas along with ground nuts and chilis. ggwar, bajra_and maize are grown during the kharif_season; wheat and barley are cultivated during the rabi_or winter season. The production of these crops has been unpredictable and highly dependent upon favorable weather notwithstanding the large total of irrigated land. Using 1952-53 to 1955- 56 as base years, there has been uneven expansion in the production of cereals. There was no improvement in 1962-63 and 1966—67, a significant decline in 1965-66 and good years in 1964-65 and 1967-68. More recently, 1970-71 provided bumper crops with drastic declines in 1972-73 and 1973- 74. Cereals are not the only agricultural commodities in Rajasthan, however; and it is necessary to distinguish zones or regions within the state where food grains are grown and where animal husbandry prevails. Rajasthan may be divided into two divisions on rainfall. The "wet" districts are those in the east. The "dry" districts are in the arid west. The geographical boundary for these divisions is the Aravalli Hills which rise in Rajasthan's southwest and fall diagonally to the northwest. The eastern "wet" region can be further subdivided into highlands and plains. The "dry" region, or those districts with lower rainfall, comprise the western half of the state. Animal husbandry is the main agricultural 26 sector in this division. There are several exceptions to this characterization, however. In the south, Sirohi district has an extensive system of tubewells which brings its effective level of water availability to higher levels than other districts in the "dry" region. In the northwest, Ganganagar district is interlaced with canal irrigation dependent upon headwaters originating in the snow- fields of the Himalayas. Finally, Sikar, Nagaur, and Jhunjhunu districts are cyclically "wet“ and "dry" with greater variation in rainfall. For purposes of this analysis of agricultural systems, they are viewed as closer in character to the systems of the eastern plains and plateau than to the desert of the western part of the state. Map 2 indicates the divisions mentioned in the preceeding paragraphs. Agricultural Policy in Rajasthan Growing food grains requires combining numerous inputs at critical periods of time. Water, soil quality, seeds, fertilizer, labor and markets are among the necessary ingredients. Rapid technological advances have provided seeds which allow significantly higher water and fertilizer applications. Water and means by which it is provided are crucial. Wells, tanks, and canals can each be made more efficient through the use of tubewells and pumpsets. Tubewells with sufficient ground water can provide steady and adequate supplies for crops; tanks and canals are enhanced by the use of tubewells and pumpsets. Though pumpsets are initially expensive, there are data to suggest investments can be paid off within one crop year and certainly within two. Pumpsets run by electricity are considerably cheaper than those powered by 27 O I ' .0. ~ -. ‘0 ( ‘. 5 2o 3. , MAP 2 11' 15' —Fo l U i I 1 I ,f ' l I RAJASTHAN I Q ‘36 6 ADMINISTRATIVE DIVISIONS ~ I a \ J 1 , c . ‘1 (mm; o “ .5"... 0 A a 2:1- I I I I '\ DIILA‘I’IEII .‘~ ".“O’N’WW ‘- «21' // .‘.’.’ o 0 a.“ a ,. I! 1 “in-"v1 u r v u / "t ‘3". noun 1' l i '~.p.“" " . -1 c'..“. !""‘.-\ " “50. ‘I 0 0+- -‘ '0 .DDNWI 1O ' \ , ‘26 . O ’ c..' I ‘ 'Ofi\\.-'o' .1 d \ I. ‘- f'."‘ ’ 4M"""~‘vv‘\.- ,9 s. , ””‘w" (.J’) o ' I N " I mu: 1". -.- . ._ 'UVI‘NIIWI“ m...“ 1‘ . -.-.- "a" mono-I'm Ml—b—J 90.... .‘O'"" .zi: ; .1. :om" l I 1 ' A J tort 71' n 71 1 _J Based upon Survey of India map with the permission of the Surveyor General of India. Government of India arid, but with exten- sive canal j irrigation Copyright, 1962. k [4, cyclically L wet and dry, (semi-arid) "wet" plains and plateau 28 petroleum based products. Because the supply of oil in India is largely imported, there is some possibility of a shortage or prices too high for the agriculturalist to pay. Of all the inputs necessary for increas- ing food grain production with the new high yield variety seeds, water is among the most critical, and as a means of supplying water, electricity is a most attractive power supply. But physical inputs and water neither produce food grains nor account for changes in agricultural productivity. As Hunter (1969), Nair (1969), and Elder (1968) point out, numerous nonphysical factors impinge upon increased food grain yields. Farmers must be made aware of new technologies and taught to use them. Further, the stimulation of the economy may rest on citizen awareness of the importance of their political activities and their ability to receive policy resources from the state government. In addition to supplying factor inputs into the agricultural production system, it may be important for governments to teach citizens how a new political system works while providing policy resources to make it work. Along with technical information, there is a major dosage of political information and values included in policy programs. Often citizen interest in the political system can be generated by the creation of local-level institutions which encourage active citizen participation. This has been the aim of the Government of India and some of its state governments, including Rajasthan. Policy resources intended for agricultural productivity and the development of citizen awareness through local political institutions are well suited to this analysis of policy resource distribution. In general, Rajasthan's elected Chief Minister, the head of the regime 29 party (the Indian National Congress) in the state's Assembly (Vighan Sabha) makes policy decisions. Most policies are formulated by a Cabinet of elected Members of the Assembly who are appointed by the Chief Minister. Members of the Cabinet hold portfolios according to substantive policy matters. Within the Cabinet are ranks of seniority and authority. Ministers outrank State and Deputy Ministers who outrank the recently created position of Parliamentary Secretary. The remaining Members of the Chief Minister's party are considered back- benchers and have less formal input into policy deliberation. Policy is implemented by a hierarchially arranged administrative network. Ministers with portfolios are served by members of the Indian Administrative Service (I.A.S.). At the district (5113) level, I.A.S. officers serve as Collectors who perform a multitude of functions from being the districts' chief judicial officers to being the District Development Officers. At the sub-district (tghsjl) level members of the Rajasthan State Administrative service (R.A.S.) perform supportive and some discretionary duties. The tghsjl_level functions include maintenance of revenue and land records and action in minor civil suits. Below the tghsjl, a government employed village level worker (VLW) serves as an extension agent and a patwari keeps village land records. The political and administrative hierarchies are closely related and Members and administrators are in regular contact over many issues. At the same time, there are autonomous agencies created for single purposes and separated from continuous Assembly scrutiny. These boards and agencies manage a single policy resource according to a qualitatively IM'I u.‘ 30 different set of restraints. They can more easily be involved in adversary relations with elected officials unless coordinating bodies are created to resolve conflicts over goals and control of resources. TWo policy resources are analyzed in this description of policy distribution in Rajasthan. The first are rural development funds distributed through local self-government institutions which are part of Rajasthan's panchayati raj system; the second is rural electrifica- tion administered by an autonomous agency called the Rajasthan State Electricity Board. Rural Development Funding: Panchayati Raj Disbursements Parallel to the elective and administrative hierarchies described above are panchayati raj structures. Panchayati raj introduced rule by local committee and was legislated in 1959 as a means of decentraliz- ing rural policy initiatives and activities. The purpose of the program was to increase citizen awareness and promote the uplift of the average rural village dweller. Groups of one to five villages are organized and authorized to formulate policy programs, submit budgets, and administer funds disbursed from the state government. The local, village- 1evel council is called a panchayat, is directly elected and is chaired by a sarpanch. Panchayati raj is three-tiered linking the village to the district level through an intermediary organization called the panchayat samiti which includes representatives from each panchayat at the tehsil- and Assembly constituency-levels. Each panchayat samiti is combined at the district level into the zila parishad (district council) which includes the pradhans (chairmen of the panchayat samiti), Members 31 of the Assembly and representatives to the national parliament from the district, and the president of the district central cooperative bank. The panchayat samiti is responsible for important expenditures and receives relatively large amounts of money from the state govern- ment through the Development Department. These funds are distributed in the forms of loans and grants and are supervised by a Block Development Officer and the Pradhan. The major category for grants is primary education; other major categories are agriculture and community development. Loans are made in several classes depending upon the type of activity a farmer undertakes. These are short-term, medium-term and long-term loans for seeds, fertilizers, bullocks and other equipment. In addition to the disbursement of funds the panchayat samiti encourages seed and fertilizer distribution. Autonomous Policy Resource Distribution: Rural Electrification Some policy resources are distributed by autonomous agencies which are less accountable to either an elected policy-making body or to the administrative infrastructure. One such agency is the Rajasthan State Electricity Board which determines how electrification should be distributed in the state. Electrification and the consumption of electricity are regulated and provided through the activity of the Electricity (Supply) Act of 1948 promulgated by the Indian Parliament. The Act created both a Central Electricity Authority and enabled the creation of state electricity boards. The Rajasthan State Electricity Board was created in 1957 after states' reorganization to regulate and supply electricity for Rajasthan. 32 The Rajasthan State Electricity Board consists of three permanent members -- a trained technician who is the Chief Engineer and two general administrators, one of whom is the Chairman. There are four temporary members chosen jointly by the Chief Minister and the Chairman. The Rajasthan State Electricity Board is funded from state money and receives capital from the Centre Government through the Ministry of Irrigation and Power. There are no formal restraints upon the distribution decisions of the Rajasthan State Electricity Board though state funding decisions can be used as a partial check. The Board's autonomy is largely due to the technical characteristics of its policy resource. Electrification's economic value is more easily determined and decisions on electrification, it is reasoned, are better left to technicians and administrators than to politicians. The Board attempts to establish programs in anticipation of Centre power and irrigation policy. In Rajasthan, with limited industrial capacity and great need for agricultural development, schemes have been devised to be capable of providing electrification to rural areas in anticipation of Centre and state funding. Most Board administrators are part of the all-India I.A.S. though many are engineers. The autonomy of the Board has generated dissent from Members of the Assembly who feel the Board should be amenable to "political" control. A step in this direction was imposed at the insistence of the Govern- ment of India's Ministry of Power and Irrigation in 1969. The Ministry directed the Board to decentralize its village selection process. Prior to 1969, while economic criteria were applied, the decisions were made without consultation or information supplied from sub—state authorities. 33 With the Ministry's directive, the Board was required to act upon recommendations made by the District Agricultural Production Committee (D.A.P.C.) which operates in each district of Rajasthan. The D.A.P.C. consists of elected leaders (Members of the Assembly and delegates to the national parliament) and administrators assigned to development posts. The Collector, a member of the I.A.S. chairs the Committee. This rule change introduced the possibilities of local political demand influencing electrification decisions and better information being made available to the Board for distribution decisions. Figure 1 below illustrates the relationships between elected and non-elected officials, panchayati raj institutions and the administration of electrification in Rajasthan. The distribution mechanisms for two policy resources are summarized in this figure. These are rural develop- ment funding through panchayati raj institutions and rural electrifica- tion through the Rajasthan State Electricity Board. Before stating a set of hypotheses explaining their distributions, it should be made clear when policy environmental factors impinge on those distributions and further what characteristics of the distributions are sensitive to the policy environment. The decade of the 1960's in Rajasthan is an appropriate time to examine these associations as the following discussion should clarify. EXPLAINING RURAL POLICY DISTRIBUTION IN RAJASTHAN Earlier the logic of policy distribution for elected officials and administrators is outlined in the context of Rajasthan. No comment has been made to this point on what aspects of policy resource distributions 4 3 AmomP .cmpczz Eoce umwwwuoev cmgpmmhmm aw mczuuchm m>wpmcwmwcwevm use quwpwpoazvmwwwpaswm 11 _.F mmeHm mmpp_esou cowpuauocm _aL=p_=uPLm< puwtpmwo camcwmcu wmwgu ucaom spwupepam_u 33m cg casc_a;u gmowwmo __u=:ou wmmPPp> gmxgoz m==m>mm :mEngcu.:ocmagmm Fw>m4 mmm__m> .vcmzpwa _ _ u4u>uk uw<44H> Lmuweyo ngsmmm< pcmnwmoga wuwsem Lmuwemo “cue mscm>mm m>wmemwam4 $0 mgmaemz pmausocma.cmsumga Inchm>mo xqum .cmupmmcm» ” _ _ ”su>ms goosm - . LmUw%%O “cmE . Iqopm>wo puwspmwo xuzamo . umcmwsma mFFN .zgmuwgumm . casewmgu Lmuwwmo pcmsaopm>mo ucmsmwficea eo mcmnsmz nocmwcaa «PAN .cxasmca “Became: can copumP—ou "4m>m4 hummhmgo acmEpcmqma “cuspcmamc mc_oem< newspcmqma pcmsucmamo m3=m>m copumgmno mac: peasaopm>ma asap aupcm xgmuocuom cucumw em acmpmsumm cm:o_mm_esou mgmumwcwz_eo umcpamu Lmumwclpz A—ww—B _ . A>smzumm< msHA<4mH¢wuv I I I I :m Susan me y «m _I . , RAJASTHAN o ADMINISTRATIVE DIVISIONS w UTTAI _._ __I-1un1uhowu Dow-out 0 .--- - . O'S'IICI :0. i.” 1‘]. ;‘ - - :6. l.E Based upon Survey of India map with the permission of the Surveyor General of India. @ Government of India Capyright, 1962. fourteen sample districts 59 a Mann-Whitney U test, it caII be asserted that the fourteen sample districts have significantly! more agricultural development potential than the remaining twelve districts in Rajasthan.2 This leads to the conclusion that the noise-creating effect of agricultural productivity potential might be reduced. The fourteen districts which have been chosen for this research are similar with respect to the nature of their agricultural productivity system outputs and they all produce small grains, and predominantly. wheat. They also have more productivity potential than those districts which have been excluded from the sample. lhe Analysis of Change In addition to controlling for exogenous variables, this research design accounts for time. The design incorporates, to the extent possible, a means by which hypotheses stated in terms of change might be tested. Herbert Jacob and Kenneth Vines' epilogue to Politics in the American States (1971) accurately appraises much of the comparative mnflic policy literature with respect to data and measurement: Data used are cross-sectional, representing one point of time rather than longitudinal, representing observa- tions over many points of time. Present findings make the need for the analysis of time series obvious. (1971: 560). lhfile the models and hypotheses attempt to explain change, data and nethods tend to reflect synchronic processes. J. S. Coleman (1969) notes that the use of cross-sectional data at one point in time assumes: either implicityly or explicitly, that the causal processes have resulted in a equilibrium state . . . (T)he implicit assumption in regression analysis is that this is a stable relationship, which would give ”g‘rs "1“? ‘1 I ll'. wi 60 the same values for the regression coefficients in a later cross-section unless an exogenous factor disturbed the situation. (1969: 444). Since this study deals with a country (India) and a State (Rajasthan) for which accurate time series data for sub-district units are largely unavailable, care has been taken to examine the underlying time assumptions in each variable. A minimal requirement for this study is that one dependent variable and one independent variable be amenable to measurement at more than one point in time. Other factors for which there is less change variation over time may be measured at one time point. Unless there are drastic institutional changes, e.g. revolution or civil war, the relationships of farmers to landlords and the distribution of land change quite slowly. In addition, the basic constitutional rules of a system change less frequently. In short, the admonition is to measure that which changes as a function of more than one mm in time while it is permissible to treat relatively static variables as measureable at one point in time. It is asserted for this study in Chapter One that change in one electoral variable -- electoral mobilization -- is associated with rural public policy distributions. Electoral variables then should be measured for at 193513 two time points and, if the dependent variable is defined in terms of change. rural public policy distribution might be measured at two time points, as well. Because it is hypothesized that electoral factors effect rural public policy distribution change, the first electoral time point should precede the first public Policy distribution time point and the second electoral time point Shonld precede the second public policy distribution time point. If the data are 50 arranged and collected, changes in electoral variables may be said to be antecedent to changes in rural public policy distribution. '1' mt .'. pd. IPA '. H.‘ 61 In Rajasthan the Third General Election (1962) is the first electoral time point and the combined budget years of 1963-64 and 1964-65 are the first rural public policy funding time period. The second electoral time point is the Fourth General Election (1967) and the corresponding rural public policy funding distribution period is the combined budget years 1968-69 and 1969-70. Rural electrification of villages creates special measurement problems which are dealt with below. It may be said that rural electrification is measured to allow the interpretation of the impact of electoral variables upon its distribution. Level of Analysis: The Legislative Constituency This analysis is concerned with how policy decision-makers utilize knowledge of the political and economic environments in distributing scarce rural public policy goods. Inferences are made about state policy outputs on the basis of information obtained from the legislative constituency and other approximately congruent administrative units. The assertion that data obtained from legislative constituency-level units can be associated "111“ Policy goods distribution is new in the study of Indian public policy- Macro-studies of Indian political behavior based upon aggregate data use the district as a data base because of the ease of collection and the Potential of conducting all-India, cross-state research. Government documents, "ePOI‘ting policy, electoral and census information, are accessible readily for the district. A second level of analysis, the village, has been utilized f" Some aggregate data studies, as well.3 Because of the lack of data and/or interest, few scholars have attempted to collect policy relevant data for units equivalent to the legislative constituency. 62 The legislative constituency, however, is the most appropriate unit of study for a state-level explanation of rural policy distribution. The legislative constituency is the base for which a Member of the Assembly is chosen. Because the Member may transmit demands from constituents or important local organizations, a seat in the Assembly is a coveted position. Nominations by political parties are sensitive to pressure from important local-level political actors, as well. If the Member is from the dominant political party in Rajasthan, the Indian National Congress (INC), he may formulate or effect policies in state government which have an impact on his constituency and the constituencies of rivals and Members of their factions. From the perspective of representation, policy-making, and Presizige, the legislative constituency is a critical electoral unit. The choice of the legislative constituency raises data gathering constraints, however. Because the relevant local -level political and adminis- trative units in question -- the panchayat samiti, tehsil, and constituency -- are not entirely congruent, there are fittings to be performed before testing hYpotheses. The selection of the data starts by choosing the legislative constituency and averages the values of the variables for tehsils and Maya samitis across the constituency.4 This method requires that sample constituencies have the same boundaries over time. Legislative constituencies in Rajasthan vary in size and boundaries across units and over time. Reapportionment creates new constituencies, and, occI‘isionally, eliminates them. The Indian electoral system has also a system of reserved constituencies. Until 1962, there were double-member constituencies from which one member of a scheduled caste or scheduled tribe was elected maples u Z‘Ei’fid In :znjatipr 2‘: schedu' ‘rniiun i "'Egll ar c 31 Sta histgpg Ease o if: muse v cl) m 63 along with another member from the remainder of the communities or castes in the constituency.5 The participating electorate voted for two candidates, one on each list. For the Third General Elections (1962) double-member constituencies were eliminated. In their place, reserved constituencies were created in 1962 with large percentages of scheduled caste or schedule tribe population. From a reserved constituency only members of the scheduled caste or scheduled tribe may be nominated for the Assembly position. Voting Participation in these constituencies has been consistently lower than in other regular constituencies.6 In addition, nominations have tended to be dictated by state-level party leaders rather than emerge from candidates relative support positions in the constituency.7 Because only those constituencies with the same boundaries over time could be chosen, some of the possible legislative constituencies within the Sample districts had to be eliminated. From the potential 124 constituency units delimited in the 1962 and 1967 electoral statutes for the fourteen district groups; sixty-four (64) are chosen as the sample for the tests of hypotheses. The boundaries of these constituencies remained the same for the Period of 1962 through 1971.8 Data for the Third and Fourth General Elections (1962 and 1967) can be utilized for these constituencies. There is no reason to believe that a bias has been introduced into the analySiS by se199111119 constituencies on this basis. The samP1E3 approximates the pOPUEa' tion on important variables such as percent voting, percent voting for political parties, distribution of regular and scheduled seats, and so on. The method of averaging depends also on the consistent application of teChm'(lues for combining non-constituency units which are not exactly congruent ”1th the constituency. Legislative constituencies are defined with reference I“"Ir --l EL LII 64 to t_eh_s_i_l_ and village boundaries. In most cases tehsils and legislative constituencies are congruent. In a number of other cases tehsils overlap the constituencies. Of the total number of sample constituencies (64) there are several cases in which there are multiple tehsils for one constituency. Of these cases, there are numerous for which one of the tehsils is wholly contained in the constituency and one or two tehsils overlap. A similar situation exists in the relationship between the constituency and the panchayat samiti. A number of constituencies contain panchayat samitis within them and have one or two mnchayat samitis which overlap. The assignment of a variable value taken from t_e_h_s_i__l_or panchayat $31111“ data is accomplished by first determining whether or not there are overlapping MOP panchayat samitis on the constituency. For cases in which there are no overlapping tehsils or panchayat samitis, the value for the variable is recorded without further calculation. In those cases where two or more M or panchayat samitis either are contained within or overlap the constituency, the variable is calculated as an average for the tehsils or mchayat samitis. Previous research using district level data for India Show there is no significant difference for results which utilize a similar technique as compared to a method which disaggregates and regroups areas to make all data reporting units congruent.9 While a regrouping technique is feasible at India's district level, it is prohibitive at the sub-district level. Table 2.1 names all of the sample constituencies for this study and indicates the districts within which they are contained. Appendix I at the end of this monograph gives the names of all tehsils and panchayat samitis which have been grouped to obtain values for the important variables in the s . . . tUdY- The discussion now turns to the measurement of the major variables. 65 'TABLE 2.1 SAMPLE DISTRICTS AND CONSTITUENCIES DISTRICT/Constituency DISTRICT/Constituency JHUNJHUNU BHARATPUR Pilani Kaman Khetri Deeg Gudha Bharatpur Nawalgarh Weir Jhunjhunu Bayana Mandawa Rajakhera Surajgarh Dholpur Nadbai Bari SIKAR Fatehpur SAWAI MADHOPUR Singrawat Danta-Ramgarh Karauli Sri Madhopur Mahuwa Neem-ka-Thana Gangapur Lachhmangarh Sawai Madhopur Hindaun Malarna Chour/Bamanwas JAIPUR Khandar Nadoti/Todabhim Amber Phulera Dudu TONK Bandikui Bairath Uniara Kotputli Malpura Phagi Newai Lalsot Sikrai Chomu AJMER Bassi Dausa Kishangarh Jamwa-Ramgarh Nasirabad Pubhkar Beawar Masuda Bhinai 66 TABLE 2.1 (Continued) SAMPLE DISTRICTSIAND CONSTITUENCIES DISTRICT/Constituency DISTRICTYConstituency ALWAR BUNDI Ramgarh Bundi Thanagazi Hindoli Rajgarh Kathuman Tijara BHILWARA KOTA Mandal Bhilwara Digod Mandalgarh Chabbra Banera Ramganj Mandi Asind Pipalda Sahada Jahazpur UDAIPUR GANGANAGAR Mavli Nathdwara Karanpur Kumbhalgarh Ganganagar Bhim Suratgarh Gogunda Hanumangarh Phalasia Nohar Sarada Lasadia Salumber NAGAUR Rajsamand Nagaur Jayal Ladnu Deedwana Nawan Degana Merta Parbatsar 67 OPERATIONALIZATION OF THE MAJOR VARIABLES Rural public policy goods distribution is attributed to political system characteristics, socio-pol itical institutional factors and the level of agricultural development. These variables have been described in Chapter One. Here the discussion turns to measurement. The Dcpendent Variables The distributions of two rural public policy goods, important to agricultural productivity in Rajasthan, have been chosen as dependent variables. They are l) the rate of change in the disbursement of rural development funds to the panchayat samiti; and, 2) the electrification of villages reported at the _t_e_h§_il level in Rajasthan. The first dependent variable includes change by definition. Because only ten years of data are avaiable, time series analysis is impossible. Change, therefore, has been included in the measurement definition of the dependent variable dealing With rural funding. This is theoretically appropriate because change in funding is more likely to be related to political and electoral character- istics than might the total amount of funding in any given period. Change in electrification of villages in any given tcflgfl is more difficult to measure because Rajasthan's electrification programs do not provide electri- Clty connections to every t_eh_sil_ for the time period under consideration. This issue is discussed below in the section on rural electrification as a dePendent variable. Ml Development Funding Rates. State-level development organizations dleurse funds to panchayat samitis for programs administered at the local- levei. These funds are for education, agricultural development, and community 68 welfare. The funds are disbursed in the forms of grants and loans. There is. little accountability for the monies which are disbursed and, in practice, some money for loans may not be repaid. ‘ There is no clear picture, however, of"the magnitude of unrepaid loans. In this analysis, all development funding, whether for agricultural development or education, is combined intx) one anmunt for the purposes of measuring the rates of change in develop- ment funding at the Enchayat samiti level. The first dependent variable is defined as a first difference equation. It is the per capita amount of funds disbursed for the period 1969 plus 1970 (TOT70) minus the per capita amount 'for~ 1964 plus 1965 (TOT65) divided by TOT65. This translates into a simple: equation: _ TOT7O - TOT65 Y1 ‘ TOT65 ‘ "here ’ Y1 = the rate of rural development funding; TOT7O = the total per capita amount disbursed to a panchayat samiti for the combined budget years of 1968-69 and 1969-70; and TOT65 = the total per capita amount disbursed to a panchayat samiti for the combined budget’years of 1963-64 and 1964-65. Accurate rural development funding data for each panchayat samiti for each year are difficullt to obtain. All years from 1961-62 through 1970-71 were collected for the sample units and a second collection, to insure accuracy, for 1963-64, 1964~65, and 1968-69 and 1969-70 was conducted with the assistance of the Development Department of Rajasthan's Ministry of Agriculture.10 69 Rural Electrification. The second dependent variable is the distribu- tion of rural electrification. Only that part of the rural electrification program which brings electrical connections into tehsils and villages is included in this variable. The discussion turns on the distinction between electrification access which can be provided by a public agency, the Rajasthan State Electricity Board (RSEB), and electricity itself which is provided largely when an individual farmer has sufficient resources to obtain a connection. Other governmental agencies, i.e., credit, may be in a position to assist the individual farmer to obtain a connection. But our concern is not with credit institutions; rather we are concerned with the distribution of electrification to villages which are aggregated at the 3% unit. As stated before, the _t_e_h_§_i_l_ is equivalent to the legislative constituency and in cases where there is overlap the averaging technique is applied. The distribution of rural electrification is measured by first determining the total rural population given access to electrification from 1961 through April 1971. The proportion of the total ruraljopulation gaining_access to electrification facilities from March 1967 through April 1971 is the second dependent variablfie. Hereafter, this variable is referred to as rural electrification and is represented symbolically as Y2. This time period is chosen so that an association between electrification and electoral variables defined for 1967 and as a rate of change for 1962 and 1967 can be tested. It is reasoned that if electoral factors have an impact on electrification, the relationship is more likely to appear 111g; a critical election when decision-makers may pursue an "electoral strategy." It might be preferable to measure electrification as a rate or as a first difference equation, as has 70 been defined for rural development funding. However, there are many tehsils for which there are no villages with access to electrification for the period from 1962 through 1966, or the period following the Third General Election. Values for a first difference equation with zero as a denominator are difficult to interpret. Rural electrification data are obtained from the records of the Rajasthan State Electricity Board which maintains records of villages with access to electrification. These villages are grouped by M; the villages' popula- tions are determined from the District Census Handbook of the Census of India, 1961, and the percentages of the rural population with access to electrification is computed. The Indejendent Variables There are three sets of endogenous, independent variables. These variables comprise a partial model of factors which influence the distribution of rural public policy goods. These are sets of variables which 1) describe the potential for agricultural growth based upon an understanding of the agricultural production process, 2) indicate short-term electoral factors important to elected and non-elected policy-makers, and 3) include socio-political institu- tional factors which may set the boundaries for public policy goods alloca- tion. Conceptual 1 Zing Agricultural Devel ppment The section above which outlined a I'most similar systems" research design incl uded two exogenous variables: the type of agricultural productivity system and the overall level of agricultural development potential for the er". I E. e. ; . 2'" " fl" 71 sample districts. It was concluded that the sample districts and constitu- encies were all sufficiently similar to assume the effects of the variables were controlled. Within that sample, however, there may be variation on other important factors considered by policy-makers who distribute rural public policy goods. These factors become part of the policy-maker's decision model and are considered here. It is one thing to distinguish between animal-husbandry and small-grain farming and conclude that those small-grain districts chosen have more potential than the animal-husbandry districts. It is quite another to say that there if variation m small-grain districts on their productivity potential.H Socio-economic development at the sub-district or constituency level in India is a complex phenomenon and has rarely been quantified. There are measures for district-level socio-economic development, however. Adams and Bumb's (l973) factor analysis of Rajasthan's twenty-six (26) districts defines four dimensions of rural development. For one of those dimensions, agricultural develOpment, agricultural productivity, fertilizer use, irrigation develOpment, and scheduled caste population are the strongest variables. Benjamin and Blue (l969) and Morris-Jones (l969) also conceptu- alize district-level modernization and measure it largely as a function of urbanization and literacy. AQricultural production depends upon seeds, fertilizers, water, equip- ment, and Ski"- These factors are interrelated and must be coordinated to grow Small-grains . Ideally, agricultural development might be conceived as having at least two dimensions. The first reflects the availability of sufficient quantities of the material resources and skills necessary to grow grains. The second concerns the distribution of these resources among .vdv H) 72 farmers who grow grains. Agricultural develOpment potential, as it is conceived here, means the former. This variable isolates only the material goods available from growing small-grains. The distribution of resources is considered as an institutional factor and is discussed below in another section of this Chapter. The availability of material resources and skill and the distribution of these resources are related, but for the purposes of analysis the two variables are separated. If a decision is made by a government and a set of important farmers to develop small-grain farming based upon water availability, irrigation becomes a basic and critical factor in agricultural productivity. Because the Government of Rajasthan pursues a strategy associated with the Green Revolution, i.e., to distribute high yield variety seeds and chemical fertilizers, water is an essential and critical ingredient for success. In planning terms, where there is less water availability, there is less compelling reason for the investment of money for rural development projects or for rural electrification. The Adams and Bumb (1973) factor analysis reflects this pattern. Agricultural productivity is higher with higher levels 0f irrigation facilities. Irrigation facilities are highly correlated With fertilizer use, as well. In the absence of reliable data on agricultural PrOdUCtiVlty or fertilizer usage at the constituency (sub-district) level, one can turn to data maintained by the Rajasthan Land Revenue Board for irrigation at the constituency level. when this analysis discusses agricultural develgment levels among the sample constituencies, the empirical reference is to the netiercent of the cropped land with irrigation facilities. Hereafter, this variable is referred to as irrigation and is represented symbolically as X]. 73 The net percent of cropped land includes those acres which are double- cropped. If an acreage is double-cropped, it is counted twice in the value. The Electoral Environment Electoral variables are hypothesized to be associated with the distribu- tion of rural public policy goods. There is an assumption that policy makers in Rajasthan perceive the electoral environment as a source of information about the political consequences of agricultural policy outputs. It_i§_ observed by Kothari 0970) and Brass (1965) for India, generally, that voters are capable of assessinLthe actions of government in a self-interested way. Rajasthan's voters appear to be no different than their compatriots in other Indian states. Verma, M. (l967), Papichristou (1968) and my own inter- views with farmers and politicians attest to the rationality of the "average" voter.12 There may be a change among Rajasthan's voters toward a more critical appraisal of governmental policy in agriculture as the advantages of scientific farming are more visible. Whether decision-makers perceive votes to be demands or supports, at the very least, election results serve as a barometer on policy performance. Electoral results may be modified by institutional factors, however, and this possibility is discussed theoretically in Chapter One. Electoral mobilization rates (turnout) and party fragmentation are operationalized as measures of the electoral environment associated with rurai public Policy allocation. What follows is a description of how each of these variables is measured for the Rajasthan electoral data. Electoral Mobilization Rates. The first electoral variable is based on the percentage who vote (turnout) in the Third and Fourth (l962 and l967) 74 General Elections. As mentioned in Chapter One, there is a general increase in the percentage of voting participation. A higher percentage of persons is voting in each of the elections since 1952.13 Electoral mobilization rates are measured for the Third and Fourth General Elections and are calculated as a first difference equation. Rates of change in voting are hypothesized to be associated with rural public policy goods allocation. These change rates are defined as the difference between the voting percent- ages in the Third and Fourth General Elections divided by the voting percent- age in the Third General Election. This may be expressed in the following algebraic form: _ VTP67 - VTP62 X2 ' VTP62, "here X2 = electoral mobilization rates; VTP67 = the percent of the electorate voting in the Fourth General Election (1967); and, VTP62 = the percent of the electorate voting in the Third General Election (1962). The higher the numerical value, the higher the rate of change. Where voting percentages declined in 1967 from l962, the rate would be expressed as a negative number; positive rates indicate increases in the percentages of those voting for l967. Party Fragmentation. Policy decision-makers may be interested in the number 0i Political parties competing in a constituency and the percentages of voters who support those parties. Rajasthan's political party system includes several party organizations including the Indian National Congress, 1‘ a. ... l ... 4"} 75 the Swatantra Party, the Jana Sangh, and others. If the system had only two parties, it might be sufficient to measure the percentage difference between the two parties -- perhaps as a function of the rate of increase or decrease in electoral mobi'lization. However, with more than two parties competing in many constituencies a summary measure of the cleavages between parties is necessary. Party cleavages in an election preceeding policy decision may have an impact on those allocation decisions. The character of political party cleavages in Rajasthan's constituencies for the Fourth General Election should be associated with rural public policy distribution decisions for l967 and l97l. And, so, party cleavages for the Fourth General election are those measured. There has been considerable experimentation toward finding an appropriate measure of party cleavages and/or interparty competition at the constituency 14 An adequate measure for this study would level in multiparty systems. reflect the proportion of voters in a constituency who are divided by political party. If the conmunity of voters is conceived as a set of individuals and political parties bring individuals together in paris, then some measure which relates the total number of pairs of individuals supporting different parties as a fraction of the total number of pairs of voters in the constitu- ency would be appropriate as a measure of party cleavages. Rae and Tay‘l or's (1970) measure of fragmentation is such a measure. Rae and T3340" assume nominal groups (or for this case, political parties -- voters casting votes for party candidates) and the fragmentation index is constructed as fol lows: 76 X = F = number of mixed pairs of individuals, where 3 total number of—pairs of individuals’ ’ the total number of pairs =(2) = 1/2 (N(N-l)); the number of mixed pairs is n U = 2 (f. f.); and i,j=l ‘ 3 n _ 2 2 F“ NN-l i,j=l (fi fj)’ (j‘i) Where, X = party fragmentation; N = the number of persons voting; f = the proportion of the voting electorate choosing the 'i'th political party; fj = the proportion of the voting electorate choosing the 'j'th political party; and, n = the number of political parties. It lmas been suggested that the measure of fragmentation developed by Rae and Taylor distrots by overweighting constituencies with splinter parties, i.e., higher fragmentation scores are reported when some very small party may not actually have an impact on the electoral results. The range of: values for party fragmentation (F) is 0.00 to l.00. The higher the value of the index, the greater the fragmentation. In other terms, the higher the numerical value of F, the higher the pr0portion of the voting electorate which is divided by political party cleavages. a". .v" 77 Institutional Factors Institutional arrangements, particularly those at the constituency level, are hypothesized to have an independent effect on the distribution of rural public policy goods. These institutional factors are dominance by a landed elite and which political party wins, and is able to control, the seat to the Assembly. Socio-economic Dominance. Land, labor, and the cohesiveness of the socio-political organizations are intrinsically part of the political process in agrarian societies. The archetypal agrarian political system -- the patron-client association -- is based upon inequalities in resource distribu- tion and face to face interaction between landed, aristocratic elite and their vassals. The Indian variation of this system is based upon much the same set of factors. Two factors emerge from a description of socio-political patterns in rural India (see Chapter One and Four for a more extensive discussion): the inequality of land distribution and the ready supply of low status persons for agricultural labor and political support. Where one finds a high level of resource distribution inequality along with large quantities of economically deprived labor force, the potential for political dominance is great. The converse is true as well. The first measurement concern is specifying variables to rank each constituency on the potential for dominance by a landed elite. To do this, an index of socio-economic dominance is created by combining a coefficient of land distribution inequality (a Gini coefficient) with the percentage of 78 scheduled caste populations. There have been several approaches to measure- ment of social and economic dominance using aggregate data indicators in the Indian context. Zagoria (l97l) and Adams and Bumb (1973) use concepts based on the percentage of households owned or rented as an indicator. Elkins (l972) has developed a "hardy peasant" index to measure the concentration of social and economic power at the regional level in southern Indian districts. Land distribution used to calculate Gini is reported in the District Census Handbook for the Census of India, l96l for a twenty percent (20%) sample of households. Percentages of scheduled castes and scheduled tribes are also reported in the District Census Handbook and each tehsil, The concentration of land and the control of landed peasants over scheduled caste labor are conditions that may be conducive to rural investment and technical change and are clearly important as independent variables. A Gini coefficient is calculated multiplying the mid-point of each interval by the number of cases and summing the products to determine the total land in the twenty percent (20%) sample. Next, the cumulative per- cent of land held by the cumulative percent of households in graphed to produce a Lorenz curve. The Gini coefficient is the area between the curve and a line of perfect inequality multiplied by two. Alker (l965) describes Gini and Benson (l969) provides a convenient calculational tool which is easily programmed for the computer. A Gini coefficient is computed for each political unit. The logic of the combination of Gini and the percent of scheduled castes is straightforward. The percent of scheduled caste population in a 79 political unit approximates the percent of the population available to be dominated by a political, economic and/or social elite. The degree to which they are manipulable may be a function of the inequality of land distribu- tion. Gini, as described here, has a relatively low variance, but a rather considerable range suggesting many cases in the middle and several extreme cases. This may be a reason to reject Gini as an explanatory variable. The compelling nature of the variable rather than its statistical qualities urge, however, that the variable be retained in the analysis. In order to create a more normal variation, a standard score is computed for Gini which is combined with the percent of scheduled castes.15 The index of socio— economic dominance is equal to the percentage of scheduled castes plus the standard score for land distribution inequality divided by two. The index has the properties of near normal distribution and possible range from 0.00 to 100.00. It should be interpreted that scores closer to 100.00 are high on socio-economic dominance and those scores closer to 0.00 are low on socio- economic dominance.16 The purpose of the index of socio-economic dominance is to discriminate between constituencies which have great potential fOr elite domination on one end of the scale and those which are less likely to be dominated by a landed elite on the other. The nondominance end could be egalitarian as it is characterized by relative equality of land distribution (in the Rajasthan context) and low percentages of scheduled castes. Egrty dominance. One simple hypothesis about the rates of policy diStribution is that the regime party tends to help its own members. That is, if a constituency has a member of the regime party, all considerations equal, 80 it will award policy goods to that constituency rather than one of the opposition party because parties which govern seek first to maintain or gain power. In order to add this factor to the analysis a nominal scale for party dominance is included. A constituency is considered to be part of the regime party apparatus if the constituency's Member is consistently of the regime party. Any constituency where a legislator has either defected to or from the dominant party_is excluded from this regimeparty list. Party dominance is determined for the period from the Fourth General Election through April, 1971." A LOGIC OF HYPOTHESIS TESTING One important consideration in research design is the use of appropriate statistical tools in hypothesis testing. Often multivariate public policy studies have utilized partial correlational techniques to link policy outputs with political and economic factors. A focus upon standardized coefficients rather than unstandardized coefficients eliminates the possibility of decision- makers proposing meaningful changes in independent variables to produce a policy outcome. A sample beta does little for the policy analyst whose goal is a policy outcome change in a proscribed direction; it tells him only that an independent variable makes an independent contribution to an explanation of variance in the dependent variable. A focus upon metrig_coefficients might aid the policy maker in changing the direction or level of performance in a way which fulfills his own goals. To date, policy analysis has few variables appropriate for direct applicability. The policy analyst can say little about‘which political and institutional variables produce concomitant change in Policy outcomes. This study takes only a small step in the application of metric and unstandardized coefficients to policy analysis.18 :m i l :térzzs n ‘\ jgj‘wnp‘. Q {3&5 lS ' W"! ‘ l\ . ' wed. 81 Chapter One presents several hypotheses which may be tested within thee contexts of this analysis. Hypotheses One, Two, and Three concern the relationships between economic and political environmental factors and rural policy goods distribution. Hypothesis Four concerns average levels of rural public policy goods allocation and the functional relationships between policy variables and independent environmental variables controlling for socio-economic dominance. Hypothesis Five concerns average levels of rural policy goods distribution and the functional relationships between policy variables and independent environmental variables controlling for a second institutional factor, political party dominance. Each institutional variable -- political party dominance and socio-economic dominance -- is dichotomized. A method suited to this problem of analysis combining nominal scales and interval scales while testing for the significance of intercepts and slopes is the analysis of covariance. Rao and Miller (l97l), Blalock (1972), Johnston (l972), Kmenta (l97l) and others discuss covariance analysis. An important feature of covariance analysis is that nominal scales may be introduced into regression analyses theough the use of dummy variables. Daniel Suits (1957) clarifies the use of dummy variables in regression analysis. Analysis of covariance techniques allow fbr the tests of significance of both intercepts and slopes. Most statistical packages for computers which include regression analyses may be adopted for the use of dummy variables and covariance techniques. The tests of hypotheses proceeds by first considering the bivariate equations relating one independent variable to one dependent variable. The equation takes the following form: Y1 = a + ka; where, i is a subscript for a dependent variable; and, 82i k is a subscript for an independent variable. Hypotheses One, Two and Three are tested using this equation. However, because there may be multicollinearity in the relationships between independergt variables, the full multiple regression equation is estimated in the follow- ing form: vi = a + blx1 + bZXZ + b3x3; where, i is a subscript for a dependent variable; X ll 1 irrigation percentages; electoral mobilization rates; and, X N ll >< ll 3 party fragmentation. Hypothesis Four is tested in a straightforward application of covariance techniques. Socio-economic dominance is introduced into the regression equation as a dummy variable along with an interaction term for each independent variable and each dependent variable: Y, = a + blxk + bzx5 + b3(Xk*X5); where i is a subscript for a dependent variable; k is a subscript for an independent variable; X5 = l, when socio-economic dominance is low in a constituency, 0 otherwise; and (Xk*X5) are interaction terms. The interested reader may consult Suits (1957) to determine why only one of the levels of the nominal scale is included in the equation. If both had been 83 included, the solution would have been indeterminate. For high socio- economic dominance constituencies, the intercept is the set of (a); for low socio-economic dominance constituencies, the intercept is the set of (a + b2). The sl0pes for high socio-economic dominance constituencies are the set of (b]) and for low socio-economic dominance constitutencies, the slopes are the set of (b1 + b3).]9 Hypothesis Five is tested in a manner identical to that described above. The dummy variables entered into the equation are for regime and opposition party dominance. These variables are: X6 = l, where the regime (INC) party is in power, 0 otherwise; and, X I 7 - l, where the opposition is in power; 0 otherwise. Hypotheses Four and Five contribute variables to the overall solution of the goodness-of-fit for the model. Where interaction terms are significant, they may be included in the model which attempts to explain the variance in policy allocations. As for the other parameters in the model, a test of significance can be applied to each increment in the coefficient of multiple determination (R2). Finally, in an effort to weight the overall validity of the model presented in Chapter One and expanded in subsequent chapters, all terms are included in one linear regression model. The equation takes the following general form: 2X + bm(xk*x5) + l l 5 Y. = a + ijk + b X + b 7 bn(Xk*X7); where, 84 i = a subscript for a dependent variable; k is a subscript for independent variables; j is a set of metric coefficients for k variables with i variables; X I 5 - l, where there is low socio-economic dominance in a constituency; 0 otherwise; X I 7 - l, where the opposition party is in control of a constituency; 0 otherwise; (Xk*X5) are interaction terms for socio-economic dominance; (Xk*X7) are interaction terms for political party dominace; m is a set of metric coefficients for socio- economic dominance interaction terms; and, n is a set of metric coefficients for political party dominance. Each chapter briefly discusses the appropriate equations being estimated and presents the intercepts and sl0pes of interest. In the next chapter, we turn to the examination of rural policy allocation and economic and electoral environmental variables. €35 FOOTNOTES 1Ganganagar, Jhunjhunu, Nagaur, and Sikar districts are included in the original sample of districts for various reasons. Misra (l967) places Ganganagar in the "canal region" of Rajasthan. Ganganagar was the scene of intensive colonization and irrigation during the pre-independence administra- tion of Maharajah Ganga Singh of Bikaner state. During his administrative control, the Gang Canal was constructed bringing water from northern rivers. Since then, the Bhakra Canal and The Rajasthan Canal have allowed the cultivation of much of the district and promoted rapid inward migration. Jhunjhunu, Nagaur, and Sikar Districts are part of a zone considered "semi- arid" by Misra. These districts are alternatively "wet" and "dry" over a period of a few years. These three districts are included because their cropping patterns are more similar to those of the eastern plains and plateau than they are to the desert districts of the western region. 22 = 2.0l, p. = .022, one-tailed, see Blalock (1972) for one discussion of the Mann-Whitney "U" test and other nonparametric tests of significance. 3Morris-Jones and Dasgupta (197l), Benjamin and Blue (l97l), Zagoria (197l), Elkins (l972), and Adams and Bumb (l973) utilize district-level data; Adelman (1971) has used village-level data for a factor analysis of rural development. 4The panchayat samiti forms the middle rung of the three-tier structure of local seTf-governemtniin Rajasthan. The panchayat samiti is congruent with the tehsil which is a sub-district administrative unit for which revenue data are kept, minor judicial and police functions are performed, and census data collected. In the great majority of cases, the panchayat samiti, and tehsil, and the Assembly constituency are congruent. 5Scheduled castes and scheduled tribes are those castes and tribes designated by the Government of India to be low in economic resources and educational Opportunities. These groups are enumerated in the Census (since l93l these are the only castes and tribes listed, by name, in the Census of India) and separate tables are prepared. The purpose of the enumeration is to assist the Government in planning for the scheduled caste and scheduled tribe up-lift programs. 6See L. Dushkin (l972) for a careful description of this phenomenon on an all-India footing. 7My own interviews with political party workers confirm this observa- tion. Rarely does a scheduled caste or scheduled tribe candidate or Member receive a nomination more than once. 86 80f the original 124 constituencies in the fourteen (14) districts chosen for the study, ninety-six (96) have the same constituency boundaries from l962 (Third General Election) throu h 197l's Parliamentary Elections. Of these ninety-six (96), sixty-four (64 are chosen as the sample for hypothesis testing. This second group of sixty-four (64) was created to insure variance on one of the independent variables, socio-economic dominance (see below in this Chapter). The strategy was to find constituencies within the sample of ninety-six (96) which were unambiguously high and low on socio- economic dominance. Care was taken to check for sample bias at this stage. The sixty-four (64) constituencies do not differ significantly on other variables from the original 124 or ninety-six (96). 9Benjamin and Blue (l969) utilized two sources of data for each variable in their study of political modernization at the district-level in India. When districts and parliamentary constituencies were not congruent, they first chose the overlapping district with the hi hest value for a variable and ran that value against other important variables in the study. Then, they chose the overlapping district with the lowest value and used that variable in the analysis. They found no significant difference in the results when comparing a method which used the highest value for a parliament- ary constituency with the lowest value for the parliamentary constituency. 10Rajasthan's budget year runs from April 1 through March 3l. 11Earlier in this Chapter a control was introduced for levels of agri- cultural productivity. Sharma's (1964) index of agricultural productivity potential placed all of Rajasthan's (and India's) districts into one category regardless of agricultural productivity system types. At that point, the purpose was to distinguish between the districts chosen and those excluded. This is a precaution against picking districts where the Governments take no interest at all in rapid agricultural growth! Where agricultural productivity potential is relatively high and where the government pursues a strategy to combine high yield variety seeds with fertilizers and irrigation, variance in one important economic component could be part of the decision- makers' models for rural public policy goods allocation. 12Voter rationality is nearly an undefinable charactertistic and may mean simply acting in one's own interest assuming an ability to order alternatives. The voter's model for electoral choice appears to have expanded to include government action as a factor independent of patron dominance or caste or factional loyalities. While there is no direct proof of this assertion, there is general agreement among Indian political scientists about this point. 13See a discussion of voter mobilization and participation rates in Chapter One. 14Party fragmentation may be analogous to some measures of electoral competition in that the more highly fragmented the constituency, the smaller the differences between winning and losing parties. A number of measures were used for this study, but none were as useful as Rae and Taylor's (1970) index discussed below. For an excellent summary of measures of interparty competition assuming several types of data see Wilcox (l973). 87 15A standard score for Gini is calculated here according to the following formula: 5 = X(r-O); where, S = the standard score for Gini; X = lOO/the range of the Gini coefficient; r = the value for the individual constituency unit; and, O = the lower limit of the Gini coefficient range. l6 See Chapter Four for a more detailed discussion of the characteristics of this variable. It is to be referred to as X4 when constituencies are high on socio-economic dominance and X5 when constituencies are low on socio- economic dominance. 17See Chapter Five for a discussion of this factor. It will be presented symbolically as X5 when the regime political party has been in power from March l967 through April 197] without interruption and X7 when an opposition political party Member, independent Member, or a defector has represented the constituency in the Assembly. 18There is a growing emphasis upon develo ing public policy models with instrumental variables. Cain and Watts (l97l criticize the "Coleman Report" for this shortcoming. Holt and Turner (1974) state a case fbr artisanship among public policy analysts. While this work seeks to move in such a direction, the materials presented here do not fit all of the important qualifications stated by either Cain and Watts or Holt and Turner. 19A test of significance is applied for each intercept and partial slope. For the intercepts a test of the significance of differences between two intercepts is given; for the partial slopes the test is whether the partial slope differs significantly from zero in a predicted direction HO: b=O; Ha:b>0, or Ha: b< ll 2 electoral mobilization rates; >< II 3 party fragmentation, and, E = and error term. Sn. 99 The testing proceeds by examining each slope of the regression equation for both the bivariate and multivariate cases. Only linear functions are tested here; however, tests for non-linear functions are reported in appropriate footnotes and comments are offered about possible interrelationships between independent variables. An F-test is reported for each parameter of the regression equation. A graphical representation of the findings on each hypothesis is offered also to illustrate findings. Rural Public Policy Goods and the Environment Hypothesis l.l and Hypothesis l.2. Does irrigation relate to rates of rural development funding and electrification? The data and regression analysis indicate this is the case. Table 3.] shows the simple correlation between irrigation and rates of rural development funding is inverse (r = -.26) and the correlation between irrigation and electrification is postive (r = .l7). While these are not high correlations, the parameters of the regression equation reflect that Hypothesis l.l and Hypothesis l.2 can be supported. The multivariate regression equation yields a negative and significant slope for irrigation and rural development funding rates (b = -.783, p. = .05) and a positive and significant slope for irrigation and rural electrification (b = .364, p. = .05). These sl0pes suggest when controlling for electoral mobilization and.party fragmentation, irrigation is significantly associated with the Figure 3.l illustrates the unit changes in irrigation and rates od development funding and electrification reSpectively. The slopes of the lines indicate support for Hypothesis One.3 TOO aucmzpwpmcou :_ cowummwecw “smegma _x mmcesu mm m_m:cm mmmmo mumwgm>wupze to» mmmmucmucma Cowpeuwmwcuqum Fees; n N> Aoo .m u .e.vv mxmn + Nxmn + —x—n + m u w>ee mmumc mc_u::m ucmanFw>mu Foes; u _> Ame .P u .m.vv xxn + < u P>e Amo.nav ¢PN.N omo.o up.o me.o aom.o oom.o e«_x mw¢._ omo.o NF.o oo~.o cup.o mm~.o «Fx N> Amc.uav mee.~ meo.o em.o- Nee.o mme.o- Nee.o- ee_x Amo.uav mmm.e meo.o eN.o- eme.o mmo._- mmm.o sex _> Any PANLV Agv Anmv Any Any mpamwgm> mpaewcm> pcmucmqmvcH “cavemamo mug<~zmu¢mm onHQH40a uHemza P.m m4mPupas ”LoeF mommucmucma coepmowewepumpm _ecsc u ~> Aoc .m u .e.uv mxmn + Nxmn + Pxpn + a u w>ee meow; meeucze ucmsqo_m>mc Fees; u _> Ame ._ u .m.uv xxa + a u _>n A_o.uav mme.ne om,.o mm.o- eep.o _mm.o- oo~.o «tux Apo.uav N__.N oop.o mm.o- Pop.o mm¢.o- «me.o «Nx N> Nmm.o omo.o om.o- Nmm.o mom.o- mmo.o- temx Ao_.uav oom.m oao.o cm.o- mom.o owm.o- mmo.o emx P> Aav PANLV Acv Anmv ADV Amv anweLm> mpnm_cm> pcmucmamncH ucmucmamo mmhuHaoa oHamzm m.m m4mwupzs Loep mmmmpcmoema Cowpeowmwcuumwm Pegs; u m» Aoo .m u .e.cv mxmn + Nxmn + PXFQ + w u _>«e mmumc m:_uc:e paman~m>mo Fees; u _> “mm .P u .m.nv xxa + e u _>« Ao_.nav moo.m omo.o mo.o- aoo.o moo.o- oo~.o temx au¢.o moo.o mo.o- aoo.o moo.o- mmm.o tmx AFo.uav Pao.m ¢P_.o oe.o woo.o mmo.o Nmo.ou *«mx Apo.uav mm_.mp eep.o oe.o moo.o emo.o eoz._- emx _> Aav _ANLV Acv Anmv Anv Amy m—amwcm> mfinm_cm> pcmncmqmucH “consumes onHhxumqom omumzm m.m mgm Anmv ADV Aev mpamwga> mpnm_em> ucmucmawvcH ucmucmqma A>mzm >uHsoa are oz< zoeeamHeech >ofisoa oasmna s<¢3a ¢.m m4m cmucoaoucv cm toe unwgumnam a me x "apnewee> acmvcmamu a so» ua_gomn:m m me up mean: .A xexxvma+ mx 3 + xxpn + m n F» me umumevumm cowumaam copmmmgmms «gem .m>onm _ moo: cw mmugsom mmm .o n mg to .o u Afimn+pav . Fay co Ama+an u Fa to» we ammo mucmowmecmwm anew .Amme "Nu F .xoo—mFmv ucm .A¢0NumoN " nap ..pm pm .Fngooaemzw .Amnp ”N~m_ .copmccoa mwmv mo n me go .o u AA n+mv . av on ucwpm>waco me A n+ov toe ammo wucmuewwc _m weep “mm.o mpo.c ooo.o e~m.o ~oo.c- mme.o mx Amo.uav mee.m mm_.o 9mm.o- epwm.o tome.o- eeum.o Nx Ao_.uav mmo._ meo.o ewo.o eemmm.o eame.o «eem~.o ex N» Amo.uav mem.e emp.o e_o.o mem.o- emmo.o teem.o- mx Amo.uav eon.~ mo_.o me~.o- emee.c Nee.o- «_mm.o Nx Amo.uav ~om.m mep.o Nee.P- mm“.c omm.o- e_~m.o _x _> MAGS Away Nama+_av FA~a+ev apav Aev o_aaeta> o_aeece> 3o; saw: acmucmaeuca acmucmamo m.¢ m4m x523u we“ mo ecommpmu «go cho >53 P; we mucm:_sou oweocoum1owoom meme; ._ n ax use mxpm>Puomammc .cowpmacmEmecm xuemn new moon; cowum~w_wnos Pocopompm .cowgmmwecw mew mx .Nx .Fx mmpnmwcm> ucmucmamw a com pa_eumnzm m me P mews: .ean + Fxpn + a flop." A_o.uav Ame." Ace." A_o.u Ame. Aeo.uav mmm.m u a FN¢.N mpm.o o¢¢.NP mNm.N Npo.o N—o.o mmp.o wNo.o AoF.uav .wm.e u 2 www.— CNN.m mmo.o mom.N a: euo.o Fmo.o moo.o moo.o Ammcmzu Nay mON.o u mp.o mo.on Nm.on NF.o mNN.o n om.ou ov.o ON.ou 0N.on Rev N N m m Nuo.o voo.o oop.o mFN.o emp.o moo.o mmm.o o~¢.o as Npp.o coo.o: mwm.on mem.o mmP.ou 0No.on mop.o- emu.os Anv muzzm >umdom e.e mam we umpmswumm corpuscm wee Q X N") ><><>< apnewee> “coucmamucH _ mpnewee> ucmucoaoo mzh .onhamHmhmHo >uH4om uagmaa 4 pcmucmamc a com pawcumnzm a me P mews: .Anxsxxvmn + Nan + xx_n + a u w> we cmumewumm cowumzcm :o_mmmsmme mgpm .m>one F moo: cw mmucsom mom .0 n mg to .o u AAma+_nv 1 an so Amn+pnv u Fa toe we owe» mocmuwewcmwm mcpN .Amme ”Nump .xuofim_mv ace .AaON-MON "mmmp ..~e um .Pgwuooasezv .AmNF "Nnm_ .COpmccowv mam mo u Na to .o u AANn+mv . av on pampe>pacm we ANn+ev toe ammo mucmuwmwcmwm weep Nee.o mmo.o eoo.o “a“.o eoo.o mem.o mx Ace.uav mme.~ mo_.o eom.o- ome.o sspmm.o- Fme.o Nx Fme.o omo.o mem.o mem.o om~.o m_m.o ex N> Amo.uav mmm.m mo~.o seemo.o ssmom._- ese_o.o teem.o- mx eme._ meo.o ome.o- sepmm.o eom.o- eemmm.o Nx Ao_.uav mmm.~ eoF.o com..- .om.o esmem.o- mem.o _x P» MAGS Ammv Nama+_av _A~e+ev Apav Aev oeaeete> opeaete> ocmucmamucm “consumes cowuemoaao mewmmm muzHmuHaoa uHamaa 4zm >uH40a mrh .mozhm<¢ 4one F muoc mmm fluxes + mxma + Nxma + —x_a + m u N> we cmpeewumm cowuezcm mspm usmcpo .0 .gmzoa cw m_ mecca cowuwmoaao 10mg» xpcma Low Esau cowouecmpcw cm can mew mx .mx .Nx .Px mews: .Amxamxvmn Amo.uav me_.m "a eeo.o AoF.uav meo.~ Apoo.uav mom.o_ Amo.ua0 oae.m N _A,o.ua0.~em.e "a Amo.uav mme.m A_o.ua0 emm.m mmm.o emm.o A_o.ua0 mme.m A00 NNP.O "Na F00.0 0N0.0 0Np.0 0N0.0 00N.0 000.0 N00.0 0FF.0 0N0.0 «No.0 Ammcmzu va «0.01 00.0: Nm.0u m—.0 00.0- m_.0u 0v.0 0N.0- 0N.0u Atv 000.0 «00.0 00_.0 mNN.0 mF0.0 m00.0 0—0.0 mmm.0 N0¢.0 Anmv .mm .m u .w.u ”mmpnmwsm> x5530 eo cowmmzumwu a sow Asmmp .muesmv mum ”one: so can: .F u nx mews: .mucmcweou apemq Pmuwuepoa 0cm cowueucwe .cowpmpcmemmsw a can .mmpms :oneNPFPnoE Pecopompm .cowummwece + fixes + mxma + an + Fxpn + a u _> we umueewumm co_pm=cm weep NP0.01 000.01 Nmm.0a 000.0 0N0.0 NF0.~1 0P0.0 mmm.01 —u0.0u A30 Axgm553mv onhamHmhmHo >0H400 0H0000 4<¢0¢ 02< .hzmzzomm>zu m.m mgm u:mu=mamuc~ w_newcm> ucwncmamo >0H400 NIP .mozhm< -< Ii II 60 70 80 9O regime party's constituencies rural development funding rates party fragmentation for the Fourth (1967) General Election CHAPTER SIX EXPLAINING RURAL POLICY DISTRIBUTIONS BY AN ANALYSIS OF THE POLICY ENVIRONMENT INTRODUCTION Does the rural policy environment have an impact on the distribution of public policy goods in Rajasthan? Is there enough evidence to infer that policy environmental factors matter in policy decisions? The evidence points to positive answers for both questions for some factors. Before evaluating the findings there is one additional question remaining. It is a logical extension of the literature on patron-client associations which finds resource distribution inequality to be a fundamental under-pinning for political party organization. We may wish to determine whether and how a combination of sociO-economic dominance and political party dominance affects rural public policy distributions. After treating this question, if briefly, a summary of the model, with all factors, is given. A statement of the utility and validity of the model, a brief critique Of what has been done and suggestions of additional steps to refine and improve the work are discussed. RURAL PUBLIC POLICY DISTRIBUTION, SOCIO-ECONOMIC DOMINANCE, AND POLITICAL PARTY DOMINANCE IN RAJASTHAN Scott (1972), Elkins (1972) and Powell (1970) and others find patron- client relationships and political party organizations in democratic societies strongly associated. There is an implicit notion that political party organiza- tions, to be viable, rest upon patron-client patterns. When patrons have power 171 172 based upon land, they mobilize voters who are dependent upon them economically and socially. Yet, variance has been shown in the degree to which patron-client associations exist in agrarian societies; and, in the absence of patron-client ties, political parties grg_maintained, and candidates chosen and elected. Qualititatively different kinds of organizations may appear which are not based upon the support and dominance of a local elite. These organizations may be more ideologically based for example, or they may function more like movements than political parties. Whatever the form the organization takes, local party organiza- tions not directly linked to rural, landed elites are likely to be different than those organizations based upon landed elites. In societies where access to government for certain policy resources is provided through party organi- zation, and an independent source of power, status or authority to exchange for those resources is necessary, it is unlikely that party organizations not based upon such an "elitist" social structure will be successful in obtaining policy resources. In more general terms, constituencies controlled by a regime party and with strong patron-client associations should receive more policy resources distributed through party organizations which respond to electoral pressures than those constituencies with other characteristics. Even when an opposition party Member is sent to the Assembly, if there are strong patron- client structures in a constituency, that constituency may well receive more policy resources, on the average, than one which sends a Member from the regime party. Where neither the regime party controls a constituency nor there are strong patron-client structures, there is every possibility that 173 such a constituency would receive the lowest average amount of policy resources which are sensitive to political party influence. These remarks do not apply to policy resources administered by autonomous bureaus; no combination of patron-client structures or political party control should affect policy resource distribution. Chapters Four and Five investigate hypotheses associating socio-economic dominance and political party dominance, respectively, with rural public policy distribution in Rajasthan. Rural development funding through panchayati raj institutions responded both to sociO-economic dominance and political dominance. High sociO-economic dominance constituencies received higher average rates of rural development funding; regime party (Indian National Congress) constituencies received higher average rates of rural development funding than Opposition Members' constituencies. Rural electrification seemed to be only sensitive to socio-economic dominance contexts. Low socio-economic dominance contexts received more electrifi- cation than high socio-economic dominance constituencies for the period being considered; the results Of the test, however, are marginal (p = .10). In Rajasthan, constituencies controlled by the regime party in high sociO-economic dominance contexts are likely to receive the highest average rates of rural develOpment funding through panchayati raj institutions. Constituencies sending Opposition (and independent) Members to the Assembly in a high sociO-economic dominance context should receive higher average funding rates than low socio-economic dominance constituencies sending regime party Members to the Assembly. The lowest average rates of rural development funding will be found in low sociO-economic dominance constituencies sending opposition Members to the Assembly. As suggested above, rural 174 electrification percentages are not likely to be affected by the combination of political party dominance and sociO-economic dominance. It has earlier been observed that rural electrification decisions are made largely outside the context of direct political party or even patron influences. Analyses and Findings The sixty-four (64) constituencies in our sample Of Rajasthan's legislative Assembly seats are grouped by both sociO-economic dominance and political party dominance. SociO-economic dominance is described in Chapter One and measurement criteria are given in Chapter TWO. High levels of socio-economic dominance in Rajasthan have high land distribution inequal- ity and higher percentages of scheduled caste populations than low socio- economic dominance characterized by land distribution equality and lower percentages of scheduled caste populations. Political party dominance refers directly to which party sends a Member to the Assembly. The regime party may send a Member or the constituency may send a Member from an opposition party (or an independent). Sample constituencies may be grouped both by sociO-economic dominance and political party dominance. Of the sixty-four (64) constituencies, the numbers in each category do not deviate from what might be expected by chance. In the sample, it appears there is no association between political party dominance and sociO-economic dominance. A regime party Members is as likely to come from a high sociO-economic dominance context as from a low sociO-economic dominance context, and so on. Having grouped Rajasthan's sample constituencies by both sociO-economic dominance and political party dominance, we may investigate the combined 175 influence Of the two factors upon rural public policy distribution. In fact, it is possible to utilize statistical techniques which make it unnecessary to have explicit controls for each variable. Nonetheless, the reader interested in knowing which constituencies fall into each category may examine them. Table 6.1 shows there are significant differences between how the two factors are associated with each of the two rural public policy distributions. For rural develOpment funding (Y1), the average distribution rates are highest for regime party constituencies with high sociO-economic dominance and rates are lowest fro Opposition parties' constituencies with low socio-economic dominance. It was predicted that Opposition party Members' constituencies with high socio-economic dominance would have higher rates Of rural development funding than either regime party constituencies with low socio-economic dominance or Opposition party constituencies with low socio-economic dominance. This prediction is supported by the evidence. The equation estimated for rural development funding rates and the coefficients are significant. Rural electrification percentages are unrelated apparently to any combination of sociO-economic dominance or political party dominance. The results from Table 6.1 do not allow any other interpretation because neither equation or coefficients are significant. The percentage of rural electri- fication distributed for 1967-1971 as a percent Of the total decade (1961- 1971) appears not to be associated with sociO-economic dominance and which party sends a Member to the Assembly. An institutional context, socio-economic dominance, and which politi- cal party sends a Member to the Assembly are associated strongly with the distribution of rural public policy resources through Rajasthan's 176 _e .N u .c.e mexma + mx_a + e u N>sa eep.o u a “Neo.o 1 am memo.o- 1 Na Neo.F u a m~35 u an meeo.o 1 Fa mwm.o u a .mem.o u a Aepuev ep-=0 Nam.o- m_m.o A_uexv 1 mega; ANn+_a+e0 "ANn+aV cowpwmoaao Am_ue0 Ampuev . Aouexv ufi_n+m0 m¢m.0um mswmmm Aeumxv Acumxv sod ewe: mucmcwsoo u_eocoum-owoom eeAN>0 muoeezmumma 20HH<0HOHOH0040 4<¢0¢ .e .N - .e.e mexma + mxea + a u P>a NN~._ u a som_.o u an smeP.o- 1 Na eee.m n a mom; u an m$3. u _a Nae.m u c .eem.o u a Ae_-e0 Aec-e0 A»uex0 . - . beta Rem o- Noe o ea_o Ama+_a+av "Ama+e0 -_moaao Am_uev Ae_uev . “ouexv "Apn+mv nn0.0um mewmmm A_umx0 Aoumxv so; see: mucmcesoo oweocoomuowuom afi_>v mmeam quozau ezmzaoau>ua samzm muzHm0H400 0H4000 4<000 —.0 mam<~ 177 panchayati raj structure. Where there is a highly unequal distribution of land and status, there are highest average rates of rural development fund- ing. The concentration of land and status appears to be a stronger factor than political party dominance in determining rates of rural development funding. Even Rajasthan's Opposition parties' Members from constituencies with high levels of sociO-economic dominance receive higher average rates than do Members from the regime party elected from low sociO-economic dominance constituencies. Political party Membership and socio-economic dominance are critical factors for rural funding through panchayati raj institutions. The reasons why rural development funding and rural electrification are differently associated with sociO-economic dominance and political party dominance deserves brief mention. These reasons are likely to be found in the differences between the types of policy resources distributed and who controls them. Rajasthan's panchayati rgj_institutions control funding available in moderate amounts. Accounting procedures and formal administra- tive oversight functions are defined, but administrators responsible for these functions are beseiged by the realities of local public life. It is Often difficult to COpe with a strong pradhan and his coterie Of loyal supporters. The administrator's allies may include a schoolmaster (usually not from the same village or region) and a village level worker, who is not part of the local social and political network. Even when complaints can be made about personnel indiscretions and alleged misuse of funds, the complaints can be easily shortcircuited by sympathetic and watchful Cabinet Ministers who attempt to protect their own clients and brokers at the constituency- 1evel. Some panchayati raj administrators (block development officers) 178 are more flexible in handling the possible conflict between themselves and local political leaders; Mathur (1970) observes that administrators with "generalist" backgrounds have been more successful in coping with political influence than have specialists assigned to administer panchayat samiti_(or block) affairs. It is not uncommon for block development officers to ask for an elected official's intercession and assistance with their administrative superiors for transfers to better postings. In brief, panchayati raj institutions bring together administrators and elected officials (and local brokers) who seek to control the same resources. Conflict is often the result. The data presented here indicate that panchayati raj elected officials were able, for the period of this research, able to utilize policy resources toward their own performance goals. A SUMMARY OF THE FINDINGS Chapters Three, Four and Five, in addition to the first section of this chapter, investigate associations between the major independent variables and policy distribution. The first policy variable is the rate of rural development funding (Y1) measured as a first different equation between one time period before the Fourth General Election (1963-64, 1964- 65) and another time period following (l968-69, 1969-70). Rural electrifi- cation (Y2) is measured as a constituency's rural population whose villages have access to electrification from 1967-1971 as a percentage of the total rural population receiving access to electrification from 1961-1971. This variable is so defined to measure the affect of electoral variables upon the distribution of rural electrification after a critical election. The 179 independent variables are the percentage of irrigated land in a constituency (X1), electoral mobilization rates (X2), and party fragmentation (X3). These first three independent variables are measured as metric scales. Irrigated land is a percentage of the total arable land counting irrigated, double- cropped land twice. Electoral mobilization rates is a first difference equation for the percent of the electorate voting in the Third (1962) and Fourth (1967) General Elections. Party fragmentation is measured for the Fourth General Election (1967). Each of these independent variables is chosen so that their impact on post-1967 policy distributions might be examined. Two additional independent factors are added: sociO-economic dominance (high and low) and political party dominance (regime and Opposition party). These factors are measured as qualitative, nominal variables. Electoral factors are seen as being salient for the policy distribution process because the regime party's percentages of support in the Assembly and the electroate declined. Opposition parties came close in 1962 and nearly defeated the regime party in l967. Post-1967 policy distributions should reflect the regime party's occupation with the electoral environment. The set of hypotheses tested has associated an independent variable with each dependent variable. Rural develOpment funding rates are associated with irrigation percentages, electoral mobilization rates, and party fragmenta- tion; the same independent variables are associated with rural electrification percentages. The associations are estimated through bivariate and multi- variate regression techniques. Each nominal scale (or institutional factor) is also associated with rural policy distributions. Since each independent institutional factor is qualitative and the dependent variables are metric, analysis of variance techniques are used to test the significance of differ- ences between adjusted means for the groupings of the independent variables. 180 Each nominal scale is included, then, in a multiple regression and inter- action effects are included. Hypotheses predicting the coefficients for interaction terms are presented and tested. Finally, all independent variables, including significant interaction effects, are included in a summary multiple regression to show the variance explained by the factors and the controlled associations between independent and dependent variables. The brief discussion which introduces this section is an outline of the manner in which the hypotheses were tested. The remaining part of this section reviews the findings of the important chapters. It is apparent that irrigation percentages are significantly associated with both the distribution of rural development funds and rural electrification. For rural development funding the sign of the coefficient is negative, as predicted; for rural electrification, the signs are positive, again as predicted. The signs are significant for both bivariate and multivariate equations. Among the environmental factors remaining electoral mobilization rates are related to policy distribution. However, where the negative coefficient for rural development funding rates and electoral mobilization rates is significant in the bivariate case, it is nonsignificant in the multivariate case. For rural electrification, electoral mobilization rates is an important explanatory variable. In both the bivariate and multi- variate cases the signs are negative and the factor explains a sizable portion of the variance for rural electrification. Rural development funding is associated positively with party fragmentation in both bivariate and multi- variate equations. The coefficients are significant at a high level. Rural electrification and party fragmentation are not associated significantly. 181 Institutional factors are related to rural policy distribution. Rural development funding rate averages are higher for high socio- economic dominance constituencies than for low socio-economic dominance constituencies; the average rates are higher, also, for regime party constituencies than for opposition party constituencies in Rajasthan. For socio-economic dominance the difference between the adjusted means is highly significant; for political party dominance, the different is marginally significant. The impact of institutional factors upon functional associations are predicted in parts of Hypotheses 4 and 5. These hypotheses are tested in Chapters Four and Five. There appears to be no significant differences in the partial slopes for rural development funding and irrigation controlling for socio-economic dominance (though the slight change observed occurs in the predicted direction). Also, there is no interaction effect for socio-economic dominance with the association between party fragmentation and rural development funding; in addition, a prediction that sociO-economic dominance might change the association between irrigation and rural electrification is not supported. There is no interaction effect for political party dominance and the relationship between irrigation and rural development funding rates. However, there is a significant interaction effect for political party dominance and the association between party fragmentation and rural development funding rates. This finding is discussed below in the section on the significance of the findings. 182 TO this point, all of the conceptualized factors have been included in the model. Table 6.2 presents a summation of the associations for all of the independent factors and each dependent variable. For rural development funding rates, it is clear that irrigation percentages are a significant factor. Earlier results show that party fragmentation is associated with rural development funding, also, but with the full equation, while party fragmentation explains a large portion of the variance, the interaction term (X3*X7) for political party dominance appears to be significant. Both institutional factors, socio-economic dominance and political party dominance, have significant coefficients. Irrigation percentages, party fragmentation, socio-economic dominance and political party dominance are the critical factors, in our model, for the explana- tion of rural development funding rates. Rural electrification percentages for the period following 1967's Fourth General Election in Rajasthan appear to be associated most strongly with three factors: irrigation percentages, electoral mobilization rates, and socio-economic dominance. Party fragmentation and political party dominance do not matter in this statistical analysis. The amount of variance in rural electrification percentages explained for the period under consideration is R2 = 0.21. CONCLUSIONS This work began with an analysis Of public policy resources by their degree of divisibility and specificity of association with known policy "production functions." Rural development funding controlled by Rajasthan's panchayati raj institutions were considered to be highly divisible and ambiguously associated with some economic or social product. .mm .m u .e.v ”scrumuwewcuompm Pegs; so» magma cowpumcmpcP ucmueewcmem on age mean“ ”meopumm xuwpoa “cantons? Ppm mcwmucou coepmacm mceN .mm .0 n .w.u magma :owuuecmpcp pceu_ewcmwm a van meouuem xuwpoq “cabbage? Pym mcwmpcou cowumscm weep 183 Nflmo.uav _mo.m "a FN.o "Na mem.o mm_.o Noo.o eo.o eeo.o mmo.o- ex Ame." ome.N «mo.o mF.o Neo.o m_P.o mx emm.o mmo.o mo.o- eoo.o eoo.o- mx Aeo.ua0 oem.~F om_.o mm.o- wep.o omm.o- Nx “mo.ua0 _e~.~ omo.o e_.o m-.o mmm.o _x N» _A_o.uav mae.m "a em~.o "we m.m.o Amo.uav eom.~ eoo.o mo.o- m_o.o emo.o Aexsmxv Apo.ua0 e~_.m amo.o m_.o- maa.o eme._- ex flop.ua0 m_m._ m_o.o om.o- ~m_.o e-.o- mx mem.o mPF.o oe.o Peo.o moo.o mx m_m.o mmo.o om.o- Nem.o _eN.o- Nx A_o.nav NmN.m eeo.o e~.o- eee.o Fem.o- ex F» “my Ammcmgu va Acv Anmv Anv Amv «Famwcm> mpnmwcm> pcmucwamnca pcmccmama Repossemv muzemzu >uHaoa NIH .oneamHmemHo >uHaoa ufismaa s<¢=m N.0 mgm