American Political Science Review (2018)112.4.918-938 doi:10.1017/S000305541800031X American Political Science Association 2018 Technological Change and Political Turnover:The Democratizing Effects of the Green Revolution in India ADITYA DASGUPTA University of California,Merced Yan technological change contribute to political turnover?Influential theories suggest that tech- nological change represents a form of creative destruction that can weaken incumbents and strengthen outsiders,leading to political turnover.This paper investigates a large-scale historical natural experiment:the impact of the green revolution on single-party dominance in India.Drawing on a theoretical framework based on models of contests,this paper argues that high-yielding variety (HYV) crops strengthened the incentives and capacity of a politically excluded group,in this case agricultural producers,to seek greater political representation.Exploiting the timing of the introduction of HYV crops, together with district-level variation in suitability for the new crop technology,instrumental variables analyses show that the green revolution played a pivotal role in the rise of agrarian opposition parties and decline of single-party dominance.The findings support theories linking technological change to political turnover,with important implications for the political economy of democratization. INTRODUCTION the sense that increases in the wealth of a politically excluded group are more likely to result in improve- conomic historians agree that technological ments in its political representation if also accompa- change is among the most important deter- nied by strengthened incentives to invest that wealth minants of long-run economic growth (Mokyr in competing for political power or greater efficiency 1992),but also a form of "creative destruction"that of collective action. can have economically and societally disruptive effects To test the theoretical argument,this paper lever- (Schumpeter 1942).Does technological change also ages a large-scale historical natural experiment:the disrupt the distribution of political power in society? impact of the green revolution on single-party dom- Different schools of thought,including modernization theory (e.g.,Deutsch 1961),the new institutionalism inance in India.In terms of its impact on agricul- ture,the green revolution was comparable in mag- (e.g.,Acemoglu and Robinson 2006),as well as Marxist nitude to the impact of the industrial revolution on approaches (e.g.,Boix 2015),suggest that technologi- manufacturing (Evenson and Gollin 2003),making it cal change can weaken entrenched political elites and a strengthen political outsiders,representing a poten- valuable large-scale historical natural experiment through which to investigate the political consequences tially powerful force for political turnover and democ- of transformative technological change.Integrating the ratization.However,empirical evidence for a linkage theoretical argument with historical evidence,this pa- between technological change and political turnover is per contends that the spread of high-yielding variety surprisingly scarce. (HYV)crops contributed to the emergence of a ris- Drawing on a theoretical framework based on mod- ing commercial class of agricultural producers with els of contests,this paper highlights three channels the wealth,incentives,and collective action capac- through which technological change can contribute to political turnover:(i)by increasing the financial capac- ity needed to seek a greater share of political rep- resentation.In a context of rural exclusion from the ity of politically excluded groups to contest political urban-and elite-dominated leadership of the dominant power,(ii)by strengthening their economic incentives Congress party (Lipton 1977;Rudolph and Rudolph to seek political representation,and (iii)by facilitating collective action.The channels are complementary,in 1987),this contributed to the rise of agrarian opposi- tion parties and played a pivotal role in India's his- torical transition from single-party dominance to mul- tiparty competition,an important historical episode Aditya Dasgupta is an Assistant Professor of Political Science,Uni- versity of California,Merced,School of Social Sciences,Human- of democratic deepening or democratization.Cumu- ities,and Arts,5200 North Lake Road,Merced,CA 95343,USA latively.the political changes wrought by the green (adasgupta3@ucmerced.edu). revolution represent a quintessential case of "polit- I thank Volha Charnysh,Dipak Dasgupta,Daniel de Kadt, ical creative destruction,"with important theoretical Devesh Kapur,Tesalia Rizzo,James Robinson,Prerna Singh,Sandip implications for the political economy of growth and Sukhtankar,Pavithra Suryanarayan,Steven Wilkinson,George Yin. democratization. Daniel Ziblatt,Adam Ziegfeld;seminar participants at the Histori- cal Political Economy Working Group,UC-Berkeley Contemporary To test the argument,this paper exploits district- India Politics Seminar,Harvard Economic History Workshop,Yale level panel data on the share of agricultural land Conference on History and Politics,University of Pennsylvania Com- planted with HYV crops over time,geocoded to over parative Politics Workshop,and Carnegie Endowment for Interna- tional Peace;and three anonymous reviewers for valuable feedback. Replication files are available on the American Political Science Re- view Dataverse:https://doi.org/10.7910/DVN/PSGCJH Many scholars hold de facto alternation of power to be an impor- tant criterion for effective democracy (e.g.,Przeworski 2000).From Received:May 13.2017:revised:December 16.2017:accepted:May this perspective,India's transition from single-party dominance to 18,2018.First published online:July 11,2018. multiparty competition represents a case of democratization. 918
American Political Science Review (2018) 112, 4, 918–938 doi:10.1017/S000305541800031X © American Political Science Association 2018 Technological Change and Political Turnover: The Democratizing Effects of the Green Revolution in India ADITYA DASGUPTA University of California, Merced Can technological change contribute to political turnover? Influential theories suggest that technological change represents a form of creative destruction that can weaken incumbents and strengthen outsiders, leading to political turnover. This paper investigates a large-scale historical natural experiment: the impact of the green revolution on single-party dominance in India. Drawing on a theoretical framework based on models of contests, this paper argues that high-yielding variety (HYV) crops strengthened the incentives and capacity of a politically excluded group, in this case agricultural producers,to seek greater political representation. Exploiting the timing of the introduction of HYV crops, together with district-level variation in suitability for the new crop technology, instrumental variables analyses show that the green revolution played a pivotal role in the rise of agrarian opposition parties and decline of single-party dominance. The findings support theories linking technological change to political turnover, with important implications for the political economy of democratization. INTRODUCTION Economic historians agree that technological change is among the most important determinants of long-run economic growth (Mokyr 1992), but also a form of “creative destruction” that can have economically and societally disruptive effects (Schumpeter 1942). Does technological change also disrupt the distribution of political power in society? Different schools of thought, including modernization theory (e.g., Deutsch 1961), the new institutionalism (e.g., Acemoglu and Robinson 2006), as well as Marxist approaches (e.g., Boix 2015), suggest that technological change can weaken entrenched political elites and strengthen political outsiders, representing a potentially powerful force for political turnover and democratization. However, empirical evidence for a linkage between technological change and political turnover is surprisingly scarce. Drawing on a theoretical framework based on models of contests, this paper highlights three channels through which technological change can contribute to political turnover: (i) by increasing the financial capacity of politically excluded groups to contest political power, (ii) by strengthening their economic incentives to seek political representation, and (iii) by facilitating collective action. The channels are complementary, in Aditya Dasgupta is an Assistant Professor of Political Science, University of California, Merced, School of Social Sciences, Humanities, and Arts, 5200 North Lake Road, Merced, CA 95343, USA (adasgupta3@ucmerced.edu). I thank Volha Charnysh, Dipak Dasgupta, Daniel de Kadt, Devesh Kapur, Tesalia Rizzo, James Robinson, Prerna Singh, Sandip Sukhtankar, Pavithra Suryanarayan, Steven Wilkinson, George Yin, Daniel Ziblatt, Adam Ziegfeld; seminar participants at the Historical Political Economy Working Group, UC-Berkeley Contemporary India Politics Seminar, Harvard Economic History Workshop, Yale Conference on History and Politics,University of Pennsylvania Comparative Politics Workshop, and Carnegie Endowment for International Peace; and three anonymous reviewers for valuable feedback. Replication files are available on the American Political Science Review Dataverse: https://doi.org/10.7910/DVN/PSGCJH. Received: May 13, 2017; revised: December 16, 2017; accepted: May 18, 2018. First published online: July 11, 2018. the sense that increases in the wealth of a politically excluded group are more likely to result in improvements in its political representation if also accompanied by strengthened incentives to invest that wealth in competing for political power or greater efficiency of collective action. To test the theoretical argument, this paper leverages a large-scale historical natural experiment: the impact of the green revolution on single-party dominance in India. In terms of its impact on agriculture, the green revolution was comparable in magnitude to the impact of the industrial revolution on manufacturing (Evenson and Gollin 2003), making it a valuable large-scale historical natural experiment through which to investigate the political consequences of transformative technological change. Integrating the theoretical argument with historical evidence, this paper contends that the spread of high-yielding variety (HYV) crops contributed to the emergence of a rising commercial class of agricultural producers with the wealth, incentives, and collective action capacity needed to seek a greater share of political representation. In a context of rural exclusion from the urban- and elite-dominated leadership of the dominant Congress party (Lipton 1977; Rudolph and Rudolph 1987), this contributed to the rise of agrarian opposition parties and played a pivotal role in India’s historical transition from single-party dominance to multiparty competition, an important historical episode of democratic deepening or democratization.1 Cumulatively, the political changes wrought by the green revolution represent a quintessential case of “political creative destruction,” with important theoretical implications for the political economy of growth and democratization. To test the argument, this paper exploits districtlevel panel data on the share of agricultural land planted with HYV crops over time, geocoded to over 1 Many scholars hold de facto alternation of power to be an important criterion for effective democracy (e.g., Przeworski 2000). From this perspective, India’s transition from single-party dominance to multiparty competition represents a case of democratization. 918 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Technological Change and Political Turnover 20,000 election races between 1957 and 1987 (cover- TECHNOLOGICAL CHANGE AS POLITICAL ing the key decades of the green revolution as well the CREATIVE DESTRUCTION decade prior).Exploiting time variation arising from the introduction of HYV crops to India in 1967,to- What links technological change to political turnover? gether with district-level variation in suitability for the As a form of "creative destruction."technological new crop technology based on the on the share of dis- change generates increases in productivity at the ex- trict land with a naturally occurring aquifer(HYV crop pense of old firms and technologies that are rendered cultivation was water intensive and required capacity obsolete by new ones(Schumpeter 1942).Acemoglu for irrigation),instrumental variables analyses show and Robinson (2006)argue that technological change that the green revolution played a pivotal role in the therefore creates "societal turbulence,"weakening po- decline of single-party dominance. litical incumbents;they then examine the implications Looking at mechanisms,quantitative tests and case for the equilibrium choice by incumbent political elites studies suggest that the new crop technology repre- to block or permit technological change.But how sented a potent force for political turnover because exactly does technological change destabilize incum- 4 it impacted multiple complementary channels.HYV bents? crops provided wealth to politically excluded agricul- Broadly,existing theories of the impact on techno- tural producers by improving agricultural productiv- logical change on politics can be divided into at least ity.HYV crops provided incentives to rural groups to three categories.First,modernization theory highlights seek political representation by intensifying the de- the role of technological modernization in contribut- pendency of agricultural profitability on government- ing to a teleological progression from tradition-bound provided inputs and subsidies.HYV crops also pro- political systems to more competitive ones character- vided collective action capacity to the rural sector by ex- ized by mass political consciousness and greater polit- erting downward pressure on crop market prices,which ical participation as well as competition (e.g.,Deutsch provided a focal point for cross-class rural collective 1961).Second,institutionalist arguments highlight the 4号 action.These economic changes created electoral con- role of technological change in the formation of ris- ditions favorable to the rise of agrarian opposition par- ing commercial interests that seek representation to ties promising subsidies for the agricultural sector and protect their wealth against sovereign expropriation greater descriptive representation of rural groups.A (e.g.,Acemoglu,Johnson,and Robinson 2005;Ansell comparison with the electoral effects of positive rain- and Samuels 2010:North and Weingast 1989).Third. fall shocks.which increased rural wealth but did not Marxist-inspired arguments highlight the role of tech- change rural incentives to seek political power or ca- nological change in contributing to collective action by pacity for collective action,suggests that a pure wealth new social forces that seek political power to advance effect would not have contributed to political turnover their class interests (e.g.,Boix 2015).Marxist and in- to the same extent. stitutionalist accounts have traditionally taken the his- The findings have important implications for the torical experience of European political development political economy of democratization.They demon as their inspiration,while modernization theorists have strate that technological change represents an impor- focused on explaining political change in contempo- tant force for political turnover in political systems rary developing countries. dominated by entrenched elites and parties.This is In spite of their differences of emphasis,the modern- consistent with literature that highlights the important ization,institutionalist,and Marxist approaches share important similarities.All three approaches can be 8 role that technological shocks play in long-run politi- cal development(e.g.,Boix 2015),including work that usefully integrated into a theoretical framework that shows that technological changes in the agricultural views competition for political power between multi- sector can have important political consequences (e.g. ple groups in society as a zero-sum contest for a valu- Acharya,Blackwell,and Sen 2016).These dynamics able prize (see,e.g.,Skaperdas 1996).2 The literature also have important implications for the political econ- on contests is useful as a theoretical device because it omy of development.They support the idea that en- abstracts away from specific institutional or historical trenched political elites possess strategic incentives to contexts,and the specific protagonists emphasized as a block or carefully manage technological change (e.g. result,highlighting instead the general idea that polit- Acemoglu and Robinson 2006;King,Pan,and Roberts ical power is divided between groups as an increasing 2013),adding to an existing literature that highlights function of their relative investments in the contest for other barriers to technological change,including inse power.Adopting the theoretical framework provided cure property rights institutions (North 1990),vested by contest models,it is clear that a technological shock economic interests (Olson 1982),and insufficiently can contribute to political turnover in at least three long political time horizons(Simmons 2016). ways. The remainder of the paper is structured as follows. First,increases in the wealth of a politically excluded It first theoretically elaborates the channels linking group due to benefiting from a new technology can in- technological change to political turnover.It then de. crease its ability to finance political mobilization and scribes the role of these dynamics in the case of the green revolution in India.It then describes the data and empirical strategy before reporting the main re- 2 The definition of a contest is"a game in which players compete for a prize by exerting effort so as to increase their probability of sults,testing mechanisms,and concluding. winning"(Skaperdas 1996,p.283). 919
Technological Change and Political Turnover 20,000 election races between 1957 and 1987 (covering the key decades of the green revolution as well the decade prior). Exploiting time variation arising from the introduction of HYV crops to India in 1967, together with district-level variation in suitability for the new crop technology based on the on the share of district land with a naturally occurring aquifer (HYV crop cultivation was water intensive and required capacity for irrigation), instrumental variables analyses show that the green revolution played a pivotal role in the decline of single-party dominance. Looking at mechanisms, quantitative tests and case studies suggest that the new crop technology represented a potent force for political turnover because it impacted multiple complementary channels. HYV crops provided wealth to politically excluded agricultural producers by improving agricultural productivity. HYV crops provided incentives to rural groups to seek political representation by intensifying the dependency of agricultural profitability on governmentprovided inputs and subsidies. HYV crops also provided collective action capacity to the rural sector by exerting downward pressure on crop market prices, which provided a focal point for cross-class rural collective action. These economic changes created electoral conditions favorable to the rise of agrarian opposition parties promising subsidies for the agricultural sector and greater descriptive representation of rural groups. A comparison with the electoral effects of positive rainfall shocks, which increased rural wealth but did not change rural incentives to seek political power or capacity for collective action, suggests that a pure wealth effect would not have contributed to political turnover to the same extent. The findings have important implications for the political economy of democratization. They demonstrate that technological change represents an important force for political turnover in political systems dominated by entrenched elites and parties. This is consistent with literature that highlights the important role that technological shocks play in long-run political development (e.g., Boix 2015), including work that shows that technological changes in the agricultural sector can have important political consequences (e.g., Acharya, Blackwell, and Sen 2016). These dynamics also have important implications for the political economy of development. They support the idea that entrenched political elites possess strategic incentives to block or carefully manage technological change (e.g., Acemoglu and Robinson 2006; King, Pan, and Roberts 2013), adding to an existing literature that highlights other barriers to technological change, including insecure property rights institutions (North 1990), vested economic interests (Olson 1982), and insufficiently long political time horizons (Simmons 2016). The remainder of the paper is structured as follows. It first theoretically elaborates the channels linking technological change to political turnover. It then describes the role of these dynamics in the case of the green revolution in India. It then describes the data and empirical strategy before reporting the main results, testing mechanisms, and concluding. TECHNOLOGICAL CHANGE AS POLITICAL CREATIVE DESTRUCTION What links technological change to political turnover? As a form of “creative destruction,” technological change generates increases in productivity at the expense of old firms and technologies that are rendered obsolete by new ones (Schumpeter 1942). Acemoglu and Robinson (2006) argue that technological change therefore creates “societal turbulence,” weakening political incumbents; they then examine the implications for the equilibrium choice by incumbent political elites to block or permit technological change. But how exactly does technological change destabilize incumbents? Broadly, existing theories of the impact on technological change on politics can be divided into at least three categories. First, modernization theory highlights the role of technological modernization in contributing to a teleological progression from tradition-bound political systems to more competitive ones characterized by mass political consciousness and greater political participation as well as competition (e.g., Deutsch 1961). Second, institutionalist arguments highlight the role of technological change in the formation of rising commercial interests that seek representation to protect their wealth against sovereign expropriation (e.g., Acemoglu, Johnson, and Robinson 2005; Ansell and Samuels 2010; North and Weingast 1989). Third, Marxist-inspired arguments highlight the role of technological change in contributing to collective action by new social forces that seek political power to advance their class interests (e.g., Boix 2015). Marxist and institutionalist accounts have traditionally taken the historical experience of European political development as their inspiration, while modernization theorists have focused on explaining political change in contemporary developing countries. In spite of their differences of emphasis, the modernization, institutionalist, and Marxist approaches share important similarities. All three approaches can be usefully integrated into a theoretical framework that views competition for political power between multiple groups in society as a zero-sum contest for a valuable prize (see, e.g., Skaperdas 1996).2 The literature on contests is useful as a theoretical device because it abstracts away from specific institutional or historical contexts, and the specific protagonists emphasized as a result, highlighting instead the general idea that political power is divided between groups as an increasing function of their relative investments in the contest for power. Adopting the theoretical framework provided by contest models, it is clear that a technological shock can contribute to political turnover in at least three ways. First, increases in the wealth of a politically excluded group due to benefiting from a new technology can increase its ability to finance political mobilization and 2 The definition of a contest is “a game in which players compete for a prize by exerting effort so as to increase their probability of winning” (Skaperdas 1996, p. 283). 919 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Aditya Dasgupta therefore result in an increased share of political power of immiserating industrialization in the development at the expense of incumbents.This corresponds,for of class-consciousness among industrial workers,but example,to arguments in the institutionalist literature the argument can be extended to the role of structural that "rising economic groups'newfound wealth im- change in encouraging class-or sector-based political proves their ability to successfully change the regime" identification and collective action more generally.By (Ansell and Samuels 2010,p.6).Central to this line of reducing the costs of political mobilization or by pro- argumentation is the presence of political credit con- viding a focal point for collective action around shared straints.In the absence of such constraints.competitors class or sectoral interests,technological change can fa- could simply borrow to finance political mobilization, cilitate collective action by politically excluded groups. a possibility sometimes ignored in literature assuming A clear requirement for the effects described above a straightforward connection between wealth and po- is that the technological shock primarily benefits a po- litical power.In the absence of fully functioning po- litically excluded or underrepresented group;a techno- litical credit markets,however,wealth constraints af- logical shock benefiting entrenched elites and incum- fect groups'ability to invest in political competition bents would have the reverse effect of reinforcing the and technological shocks to group wealth can therefore status quo.Additionally,the different mechanisms are shift the balance of political power. clearly complementary,suggesting mutually reinforc- Even when politically excluded groups possess suffi- ing comparative statics.Increases in the wealth of a po- cient wealth to mobilize politically,they may fail to do litically excluded group are more likely to result in im- so without incentives to seek political power.A second provements in its political representation if also accom- channel through which technological change can con- panied by strengthened incentives to invest that wealth tribute to political turnover is by strengthening the eco in competing for political power and greater efficiency nomic incentives of excluded groups to seek a greater of collective action.Conversely,without adequate in- share of political power,if political control is comple- centives to invest in competing for political power or mentary to the productive potential of the new technol- capacity for collective action,changes in wealth alone ogy.In the framework of contest models,a new tech- may fail to impact the existing balance of political nology can increase the value of political control to power. politically excluded groups.The nature of this comple- However,precisely because they threaten political mentarity depends upon the specific context.Institu- incumbents,politically destabilizing innovations ought tionalists,with their emphasis on the historical estab- to remain"off the equilibrium path"altogether due to lishment of limited government in Europe,focus on the strategic obstruction.Acemoglu and Robinson(2006) wealth defense motive,with commercial entrepreneurs develop a formal model of the "political replacement mobilizing to establish property rights and protect effect."which elaborates the conditions under which their growing assets against sovereign expropriation rational incumbents are likely to block or permit the (Acemoglu,Johnson,and Robinson 2005:Ansell and introduction of a new technology.In the model,ex- Samuels 2010:North and Weingast 1989).By contrast. tremely politically secure elites are more permissive of modernization theorists focused on interest group pol- innovation because they are not likely to be displaced. itics in developing countries highlight a more proactive The model also suggests that technological change attempt by mobilized entrepreneurial groups to cap- is more likely in highly competitive political systems ture government to advance policies complementary where frequent turnover means that incumbent elites to their class,sectoral or ethnic interests (e.g.,Deutsch have relatively short expected tenures regardless of the 1961:Bates 1974).By increasing the economic returns choice to permit innovation.It is in the case of moder- to political control,technological change can therefore ately entrenched elites,who are at risk of displacement provide incentives for politically excluded groups to and with considerable rents at stake,that technologi- mobilize politically. cal change is most likely to be blocked.This suggests Third.new technologies can facilitate collective ac- that.observationally.we will tend to see technological tion.In the framework of contest models,technologi- change in settings where the new technology is unlikely cal changes can improve the efficiency of collective ac- to threaten the incumbent or where its destabilizing ef- tion,changing the"function"mapping investments by fects are difficult to disentangle from preexisting high excluded groups to the probability of winning politi- levels of political turnover. cal power.Some technologies,such as cellular phones, However,in many cases,incumbents lack the ca- have contributed directly to collective action by re- pacity to block or fail to anticipate the long-run con- ducing the costs of political mobilization by rebel or sequences of technological change,providing an op- excluded groups (see e.g.,Pierskalla and Hollenbach portunity to observe "political creative destruction"at 2013).New military technologies are a classic example work.History is replete with examples of technological of innovations that can shift the balance of political innovations that had sweeping economic and political power between rival groups by improving one side's consequences that incumbents failed to block or an- effectiveness in conflict(see,e.g.,Boix 2015).Marx- ticipate.Important arguments suggest that technologi- ist arguments additionally emphasize the psychological cal revolutions such as advances in shipping,the emer- effects of technological change in highlighting shared gence of cities,or the industrial revolution,resulted in economic interests,providing a focal point for collec- economic as well as political revolutions (see,e.g.,Ace- tive action by individuals otherwise divided by cross- moglu,Johnson,and Robinson 2005;Boix 2015;Mokyr cutting cleavages.Marx prognosticated about the role 1992).Yet microlevel empirical evidence for a linkage 920
Aditya Dasgupta therefore result in an increased share of political power at the expense of incumbents. This corresponds, for example, to arguments in the institutionalist literature that “rising economic groups’ newfound wealth improves their ability to successfully change the regime” (Ansell and Samuels 2010, p. 6). Central to this line of argumentation is the presence of political credit constraints. In the absence of such constraints, competitors could simply borrow to finance political mobilization, a possibility sometimes ignored in literature assuming a straightforward connection between wealth and political power. In the absence of fully functioning political credit markets, however, wealth constraints affect groups’ ability to invest in political competition and technological shocks to group wealth can therefore shift the balance of political power. Even when politically excluded groups possess sufficient wealth to mobilize politically, they may fail to do so without incentives to seek political power. A second channel through which technological change can contribute to political turnover is by strengthening the economic incentives of excluded groups to seek a greater share of political power, if political control is complementary to the productive potential of the new technology. In the framework of contest models, a new technology can increase the value of political control to politically excluded groups. The nature of this complementarity depends upon the specific context. Institutionalists, with their emphasis on the historical establishment of limited government in Europe, focus on the wealth defense motive, with commercial entrepreneurs mobilizing to establish property rights and protect their growing assets against sovereign expropriation (Acemoglu, Johnson, and Robinson 2005; Ansell and Samuels 2010; North and Weingast 1989). By contrast, modernization theorists focused on interest group politics in developing countries highlight a more proactive attempt by mobilized entrepreneurial groups to capture government to advance policies complementary to their class, sectoral or ethnic interests (e.g., Deutsch 1961; Bates 1974). By increasing the economic returns to political control, technological change can therefore provide incentives for politically excluded groups to mobilize politically. Third, new technologies can facilitate collective action. In the framework of contest models, technological changes can improve the efficiency of collective action, changing the “function” mapping investments by excluded groups to the probability of winning political power. Some technologies, such as cellular phones, have contributed directly to collective action by reducing the costs of political mobilization by rebel or excluded groups (see e.g., Pierskalla and Hollenbach 2013). New military technologies are a classic example of innovations that can shift the balance of political power between rival groups by improving one side’s effectiveness in conflict (see, e.g., Boix 2015). Marxist arguments additionally emphasize the psychological effects of technological change in highlighting shared economic interests, providing a focal point for collective action by individuals otherwise divided by crosscutting cleavages. Marx prognosticated about the role of immiserating industrialization in the development of class-consciousness among industrial workers, but the argument can be extended to the role of structural change in encouraging class- or sector-based political identification and collective action more generally. By reducing the costs of political mobilization or by providing a focal point for collective action around shared class or sectoral interests, technological change can facilitate collective action by politically excluded groups. A clear requirement for the effects described above is that the technological shock primarily benefits a politically excluded or underrepresented group; a technological shock benefiting entrenched elites and incumbents would have the reverse effect of reinforcing the status quo. Additionally, the different mechanisms are clearly complementary, suggesting mutually reinforcing comparative statics. Increases in the wealth of a politically excluded group are more likely to result in improvements in its political representation if also accompanied by strengthened incentives to invest that wealth in competing for political power and greater efficiency of collective action. Conversely, without adequate incentives to invest in competing for political power or capacity for collective action, changes in wealth alone may fail to impact the existing balance of political power. However, precisely because they threaten political incumbents, politically destabilizing innovations ought to remain “off the equilibrium path” altogether due to strategic obstruction. Acemoglu and Robinson (2006) develop a formal model of the “political replacement effect,” which elaborates the conditions under which rational incumbents are likely to block or permit the introduction of a new technology. In the model, extremely politically secure elites are more permissive of innovation because they are not likely to be displaced. The model also suggests that technological change is more likely in highly competitive political systems, where frequent turnover means that incumbent elites have relatively short expected tenures regardless of the choice to permit innovation. It is in the case of moderately entrenched elites, who are at risk of displacement and with considerable rents at stake, that technological change is most likely to be blocked. This suggests that, observationally, we will tend to see technological change in settings where the new technology is unlikely to threaten the incumbent or where its destabilizing effects are difficult to disentangle from preexisting high levels of political turnover. However, in many cases, incumbents lack the capacity to block or fail to anticipate the long-run consequences of technological change, providing an opportunity to observe “political creative destruction” at work. History is replete with examples of technological innovations that had sweeping economic and political consequences that incumbents failed to block or anticipate. Important arguments suggest that technological revolutions such as advances in shipping, the emergence of cities, or the industrial revolution, resulted in economic as well as political revolutions (see, e.g., Acemoglu, Johnson, and Robinson 2005; Boix 2015;Mokyr 1992). Yet microlevel empirical evidence for a linkage 920 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Technological Change and Political Turnover between technological change and political turnover is tors could exploit directly by means of tube-wells (Rud surprisingly scarce.This paper aims to fill the gap by ex- 2012).This was due to the fact that.for biological rea- amining the political consequences of one of the most sons,HYV crop cultivation required steady and inten- important technological transformations impacting the sive application of water;in the presence of controlled developing world in the twentieth century:the green irrigation,HYV crops delivered yield increases of up to revolution 200%,while in the absence of irrigation they delivered lower vields than did traditional strains(Frankel 1975) HYV crops were distributed technocratically.with the THE GREEN REVOLUTION AS A technical aid of international agencies such as the Ford HISTORICAL NATURAL EXPERIMENT and Rockefeller foundations as well as major domes- tic agricultural research centers and universities.Pub- HYV crops,highly productive"dwarf"cultivars devel- lic distribution of HYV seeds occurred initially,be- oped by Norman Borlaug and scientists working in fore countrywide roll-out,via the Intensive Agricul- Mexico,revolutionized agricultural production across tural Districts Program(IADP),which targeted the 16 the developing world between the 1960s and 1980s,a model'districts viewed as most agriculturally advanced transformation commonly known as the green revo- and equipped with irrigation infrastructure to take ad- lution (Evenson and Gollin 2003).Under ideal con- vantage of the new seeds. 4 ditions,the new HYV crops,such as the IR8 vari- In districts impacted by the green revolution,HYV ety of rice and Mexican modern varieties of wheat, crop cultivation revolutionized agricultural produc- could achieve yields more than double that of tradi- tivity,resulting in a transition from subsistence to tional varieties(Frankel 1975,23).The green revolu- commercialized agriculture.This benefited an increas- tion reached India in 1966-67 when HYV seeds.at first ingly wealthy "middle peasantry"comprised of typi- primarily wheat and rice (but later other varieties as cally lower-and intermediate-caste farmers with suffi- well),were finally adapted to Indian conditions and saw cient landholdings and the economies of scale needed 4r元 widespread uptake across the country over the follow- to take advantage of the new crop technology,a ing two decades. group that,according to various estimates,repre- Two features make the green revolution in India a sented roughly 30%of the Indian population(Brass & likely candidate for a case of political "creative de- 1980).Not only was the middle peasantry increas- struction."First,HYV crops benefited a group,in this ingly wealthy,but,as in nearly all countries affected case agricultural producers,that was excluded from ef- by the green revolution,increasingly dependent on fective representation within the dominant Congress government-controlled economic policies.HYV crop party.Second,the new crop technology provided this cultivation required intensive usage of inputs,includ- politically excluded group with wealth,incentives to ing fertilizer and pesticide,electrification for operating seek political representation,and collective action ca- tube-wells,and equipment for mechanized cultivation. pacity in precisely the manner that the theoretical dis- All of these inputs depended wholly or in part on gov- cussion has suggested could result in political mobiliza ernment provision and subsidization (see e.g.,Harriss tion and turnover.Integrating the theoretical argument 1977,256).Marketing crop surpluses required infras- with historical research,this paper contends that the tructure investments,especially in roads for transporta- green revolution not only revolutionized agricultural tion and storage facilities.Moreover,the increased out- productivity but also played a pivotal role in the rural put of HYV crops placed downward pressure on crop political mobilization that contributed to India's his. 8 market prices,making the level at which the govern- torical transition from single-party dominance to mul- ment set crop procurement prices an increasingly visi- tiparty competition. ble issue affecting farmers'profits. HYV crops were introduced to India under a"new These economic changes converted the previously agricultural strategy"adopted in the mid-1960s,as a re quiescent countryside into a hotbed of political mo- sult of a combination of exogenous shocks (Varshney bilization.The increasing wealth of cultivators made 1998.48).First,the death of India's first prime minis them capable of political mobilization independent ter,Jawaharlal Nehru,in 1964 brought to power Lal of the vertical ties of patronage that characterized Bahadur Shastri,who was more ideologically commit- Congress party dominance in rural India (see e.g., ted to improving agricultural productivity through in- Wilkinson 2007).The magnified impact of government vestment in new agricultural technologies.Second,the policies on agricultural profitability provided incen- eys Indo-Pakistani war of 1965 heightened tensions with tives for agricultural producers to mobilize politically the United States,intensifying the desire among Indian to demand subsidies.Falling prices,a grievance that res- political leaders to reduce dependence upon American onated not only with the prospering middle peasantry food aid amid the Cold War.Third,droughts leading to but with small-scale farmers that were left out of the a food production crisis in 1966 heightened concerns green revolution,provided a focal point for cross-class about national food production,spurring efforts to dis- rural collective action.Initially,growing rural demand seminate HYV crops throughout the countryside for policy influence took the form of protests at dis- Once introduced to India,HYV crops overwhelm- trict headquarters and the national capital,organized ingly spread to districts either with preexisting irriga- by farmers'associations to demand greater subsidies tion infrastructure or capacity for irrigation,particu- investment,and remunerative prices for the agricul- larly in the form of aquifers,which individual cultiva- tural sector(Varshney 1998,81). 921
Technological Change and Political Turnover between technological change and political turnover is surprisingly scarce. This paper aims to fill the gap by examining the political consequences of one of the most important technological transformations impacting the developing world in the twentieth century: the green revolution. THE GREEN REVOLUTION AS A HISTORICAL NATURAL EXPERIMENT HYV crops, highly productive “dwarf” cultivars developed by Norman Borlaug and scientists working in Mexico, revolutionized agricultural production across the developing world between the 1960s and 1980s, a transformation commonly known as the green revolution (Evenson and Gollin 2003). Under ideal conditions, the new HYV crops, such as the IR8 variety of rice and Mexican modern varieties of wheat, could achieve yields more than double that of traditional varieties (Frankel 1975, 23). The green revolution reached India in 1966–67, when HYV seeds, at first primarily wheat and rice (but later other varieties as well),were finally adapted to Indian conditions and saw widespread uptake across the country over the following two decades. Two features make the green revolution in India a likely candidate for a case of political “creative destruction.” First, HYV crops benefited a group, in this case agricultural producers, that was excluded from effective representation within the dominant Congress party. Second, the new crop technology provided this politically excluded group with wealth, incentives to seek political representation, and collective action capacity in precisely the manner that the theoretical discussion has suggested could result in political mobilization and turnover. Integrating the theoretical argument with historical research, this paper contends that the green revolution not only revolutionized agricultural productivity but also played a pivotal role in the rural political mobilization that contributed to India’s historical transition from single-party dominance to multiparty competition. HYV crops were introduced to India under a “new agricultural strategy” adopted in the mid-1960s, as a result of a combination of exogenous shocks (Varshney 1998, 48). First, the death of India’s first prime minister, Jawaharlal Nehru, in 1964 brought to power Lal Bahadur Shastri, who was more ideologically committed to improving agricultural productivity through investment in new agricultural technologies. Second, the Indo-Pakistani war of 1965 heightened tensions with the United States, intensifying the desire among Indian political leaders to reduce dependence upon American food aid amid the Cold War. Third, droughts leading to a food production crisis in 1966 heightened concerns about national food production, spurring efforts to disseminate HYV crops throughout the countryside. Once introduced to India, HYV crops overwhelmingly spread to districts either with preexisting irrigation infrastructure or capacity for irrigation, particularly in the form of aquifers, which individual cultivators could exploit directly by means of tube-wells (Rud 2012). This was due to the fact that, for biological reasons, HYV crop cultivation required steady and intensive application of water; in the presence of controlled irrigation,HYV crops delivered yield increases of up to 200%, while in the absence of irrigation they delivered lower yields than did traditional strains (Frankel 1975). HYV crops were distributed technocratically, with the technical aid of international agencies such as the Ford and Rockefeller foundations as well as major domestic agricultural research centers and universities. Public distribution of HYV seeds occurred initially, before countrywide roll-out, via the Intensive Agricultural Districts Program (IADP), which targeted the 16 model’ districts viewed as most agriculturally advanced and equipped with irrigation infrastructure to take advantage of the new seeds. In districts impacted by the green revolution, HYV crop cultivation revolutionized agricultural productivity, resulting in a transition from subsistence to commercialized agriculture. This benefited an increasingly wealthy “middle peasantry” comprised of typically lower- and intermediate-caste farmers with sufficient landholdings and the economies of scale needed to take advantage of the new crop technology, a group that, according to various estimates, represented roughly 30% of the Indian population (Brass 1980). Not only was the middle peasantry increasingly wealthy, but, as in nearly all countries affected by the green revolution, increasingly dependent on government-controlled economic policies. HYV crop cultivation required intensive usage of inputs, including fertilizer and pesticide, electrification for operating tube-wells, and equipment for mechanized cultivation. All of these inputs depended wholly or in part on government provision and subsidization (see e.g., Harriss 1977, 256). Marketing crop surpluses required infrastructure investments, especially in roads for transportation and storage facilities.Moreover, the increased output of HYV crops placed downward pressure on crop market prices, making the level at which the government set crop procurement prices an increasingly visible issue affecting farmers’ profits. These economic changes converted the previously quiescent countryside into a hotbed of political mobilization. The increasing wealth of cultivators made them capable of political mobilization independent of the vertical ties of patronage that characterized Congress party dominance in rural India (see e.g., Wilkinson 2007). The magnified impact of government policies on agricultural profitability provided incentives for agricultural producers to mobilize politically to demand subsidies.Falling prices, a grievance that resonated not only with the prospering middle peasantry but with small-scale farmers that were left out of the green revolution, provided a focal point for cross-class rural collective action. Initially, growing rural demand for policy influence took the form of protests at district headquarters and the national capital, organized by farmers’ associations to demand greater subsidies, investment, and remunerative prices for the agricultural sector (Varshney 1998, 81). 921 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Aditya Dasgupta Over time,the green revolution created electoral characteristics and ethnic composition variables that conditions favorable to the emergence and success of are largely fixed over time.This paper instead traces agrarian and regional opposition parties.The grow- the uneven rise of rural political mobilization over time ing articulation of rural "voice"in the wake of the squarely to the adoption of the new crop technology green revolution benefited agrarian and regional oppo- and its impact on the emergence of a rising class of sition parties because rural and cultivating caste groups agricultural producers with growing incentives and ca- were heavily underrepresented in the urban and elite- pacity to seek a greater share of political power. dominated leadership of the dominant Congress party (Rudolph and Rudolph 1987).By contrast,agrarian EMPIRICAL STRATEGY and regional opposition parties offered not only sub- op//s sidies but descriptive representation for excluded rural To estimate the political effects of the green revolution. groups.Gaining representation in state legislatures was in the basic specification,this paper estimates a panel important to gaining direct control over agricultural fixed effects OLS regression of the form policies.On a more quotidian level,gaining representa tion in local district party branches,district boards,local Yrdit Yrd tn +aHYVrdt Erdit, cooperative societies,and elected village councils was also important,because these district-level institutions governed the local distribution of pesticide,fertilizer, where the outcome variable Yrdit is a measure of the dominant Congress party's election performance,in re- agricultural credit,tube wells,seeds,and a variety of other inputs required for HYV crop cultivation (Brass gion r,district d,constituency i,and year t.The main outcome of interest is the election performance of the 1965). Congress party in state assembly elections,defined in In addition to defeating the Congress party at the terms of percentage point vote share and an indica- state level,newly emergent agrarian and regional op position parties played a pivotal role in the Janata tor for winning the seat as well.The focus is on state elections because,as discussed,the decline of single- coalition,which defeated the Congress party in na- party dominance first occurred at the state level,which tional elections for the first time in India's history in 1977 By the 1980s,opposition parties drawing support laid the groundwork for the later emergence of multi- from rural and cultivating caste groups had established party competition nationally.The variable HYVdr is a multiparty competition in many of India's states,and district-level measure of HYV crop adoption,defined as the share of district-level agricultural land planted also played an important role in the Janata Dal coali- tion,which defeated the Congress party for the sec- with HYV crops. ond time in parliamentary elections in 1989,ushering All specifications include district,yrd,and region- in India's transition to multi-party competition at the year,in,fixed effects.A panel fixed effects regression of this form estimates the effect of changes within dis- national level. The outburst of agrarian political mobilization is tricts over time in HYV crop adoption on the Congress widely recognized as a key component of the dominant party's performance in elections,while also partialing Congress party's long-run decline (e.g.,Rudolph and out any time-based shocks and trends common to con- Rudolph 1987).Prominent arguments trace this rural stituencies within regions.3 Region-year fixed effects political mobilization to the structural bias in favor of ensure that the effects of HYV crop adoption are iden- the rural majority in democracies (e.g.,Varshney 1998). tified from over-time comparisons only between dis- Others focus on the ethnic dimension of mobilization tricts located within geographically and linguistically by lower-caste groups in opposition to the upper-caste comparable regions.To account for the fact that HYV bias of the Congress party's leadership (e.g.,Jaffrelot crop adoption is measured at the district level while 2003).Both arguments have important synergies with election outcomes are observed at the assembly con- those advanced in this paper.The presence of robust stituency level (multiple assembly constituencies fit elections provided the institutional conditions through contiguously within a district),for over-time correla- tion of errors within districts,and for regional spatial which rural interests could assert themselves;in their absence,technological change may have contributed correlation in the diffusion of HYV crops,all analy- to political change by conflictual means,as some ob- ses report robust standard errors adjusted for two-way servers feared contemporaneously(Frankel 1975).The clustering within districts and by region-years. The election data are based on reports produced by rural-urban divide also corresponded to a conflict be- tween cultivating"backward caste"groups and upper- the Election Commission of India,which were scraped by Jensenius (2013)for elections from 1961 onward caste groups.Caste associations provided an important This paper extends this election dataset back in time by basis for collective rural action,and the platforms of scraping the data for elections held in 19574 The data agrarian and regional opposition parties often fluidly combined ethnic as well as sectoral appeals to these on HYV crop adoption is based on annual district-level overlapping politically excluded groups. crop coverage and production statistics published by L However,electoral institutions and ethnic rivalries cannot explain the timing of the Congress party's de- 3 The regions include:north (including the Hindi-speaking states cline,as elections and ethnic rivalries existed prior to as of Bihar,Haryana,Madhya Pradesh,Uttar Pradesh,and Rajasthan olus Punjab),south (Andhra Pradesh,Karnataka,Tamil Nadu),east well as following the green revolution.In the empirical (West Bengal and Orissa).and west (Gujarat and Maharashtra). analysis,unit fixed effects absorb electoral institutional Table A2 in the Online Appendix lists the elections in the dataset. 922
Aditya Dasgupta Over time, the green revolution created electoral conditions favorable to the emergence and success of agrarian and regional opposition parties. The growing articulation of rural “voice” in the wake of the green revolution benefited agrarian and regional opposition parties because rural and cultivating caste groups were heavily underrepresented in the urban and elitedominated leadership of the dominant Congress party (Rudolph and Rudolph 1987). By contrast, agrarian and regional opposition parties offered not only subsidies but descriptive representation for excluded rural groups.Gaining representation in state legislatures was important to gaining direct control over agricultural policies.On a more quotidian level, gaining representation in local district party branches, district boards,local cooperative societies, and elected village councils was also important, because these district-level institutions governed the local distribution of pesticide, fertilizer, agricultural credit, tube wells, seeds, and a variety of other inputs required for HYV crop cultivation (Brass 1965). In addition to defeating the Congress party at the state level, newly emergent agrarian and regional opposition parties played a pivotal role in the Janata coalition, which defeated the Congress party in national elections for the first time in India’s history in 1977. By the 1980s, opposition parties drawing support from rural and cultivating caste groups had established multiparty competition in many of India’s states, and also played an important role in the Janata Dal coalition, which defeated the Congress party for the second time in parliamentary elections in 1989, ushering in India’s transition to multi-party competition at the national level. The outburst of agrarian political mobilization is widely recognized as a key component of the dominant Congress party’s long-run decline (e.g., Rudolph and Rudolph 1987). Prominent arguments trace this rural political mobilization to the structural bias in favor of the rural majority in democracies (e.g., Varshney 1998). Others focus on the ethnic dimension of mobilization by lower-caste groups in opposition to the upper-caste bias of the Congress party’s leadership (e.g., Jaffrelot 2003). Both arguments have important synergies with those advanced in this paper. The presence of robust elections provided the institutional conditions through which rural interests could assert themselves; in their absence, technological change may have contributed to political change by conflictual means, as some observers feared contemporaneously (Frankel 1975). The rural-urban divide also corresponded to a conflict between cultivating “backward caste” groups and uppercaste groups. Caste associations provided an important basis for collective rural action, and the platforms of agrarian and regional opposition parties often fluidly combined ethnic as well as sectoral appeals to these overlapping politically excluded groups. However, electoral institutions and ethnic rivalries cannot explain the timing of the Congress party’s decline, as elections and ethnic rivalries existed prior to as well as following the green revolution. In the empirical analysis, unit fixed effects absorb electoral institutional characteristics and ethnic composition variables that are largely fixed over time. This paper instead traces the uneven rise of rural political mobilization over time squarely to the adoption of the new crop technology and its impact on the emergence of a rising class of agricultural producers with growing incentives and capacity to seek a greater share of political power. EMPIRICAL STRATEGY To estimate the political effects of the green revolution, in the basic specification, this paper estimates a panel fixed effects OLS regression of the form, Yrdit = γrd + τrt + αHYVrdt + εrdit, where the outcome variable Yrdit is a measure of the dominant Congress party’s election performance,in region r, district d, constituency i, and year t. The main outcome of interest is the election performance of the Congress party in state assembly elections, defined in terms of percentage point vote share and an indicator for winning the seat as well. The focus is on state elections because, as discussed, the decline of singleparty dominance first occurred at the state level, which laid the groundwork for the later emergence of multiparty competition nationally. The variable HYVrdt is a district-level measure of HYV crop adoption, defined as the share of district-level agricultural land planted with HYV crops. All specifications include district, γrd, and regionyear, τrt , fixed effects. A panel fixed effects regression of this form estimates the effect of changes within districts over time in HYV crop adoption on the Congress party’s performance in elections, while also partialing out any time-based shocks and trends common to constituencies within regions.3 Region-year fixed effects ensure that the effects of HYV crop adoption are identified from over-time comparisons only between districts located within geographically and linguistically comparable regions. To account for the fact that HYV crop adoption is measured at the district level while election outcomes are observed at the assembly constituency level (multiple assembly constituencies fit contiguously within a district), for over-time correlation of errors within districts, and for regional spatial correlation in the diffusion of HYV crops, all analyses report robust standard errors adjusted for two-way clustering within districts and by region-years. The election data are based on reports produced by the Election Commission of India, which were scraped by Jensenius (2013) for elections from 1961 onward. This paper extends this election dataset back in time by scraping the data for elections held in 1957.4 The data on HYV crop adoption is based on annual district-level crop coverage and production statistics published by 3 The regions include: north (including the Hindi-speaking states of Bihar, Haryana, Madhya Pradesh, Uttar Pradesh, and Rajasthan plus Punjab), south (Andhra Pradesh, Karnataka, Tamil Nadu), east (West Bengal and Orissa), and west (Gujarat and Maharashtra). 4 Table A2 in the Online Appendix lists the elections in the dataset. 922 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Technological Change and Political Turnover FIGURE 1.Mean HYV Crop Adoption and Change in Congress Seat Share A.HYV Crop Adoption B.Congress Party Decline 回 0.000to0.126 國 -79.0to-33.39 口 0.128to0.170 ■ -33.39to-19.88 口 0.170to0.245 口 -19.88to-6.76 0.245to0.341 -6.76to8.25 0.341to0.727 8.25to52.0 Notes:Panel A:Districts shaded according to quintile of mean share of agricultural land under HYV crop cultivation between 1967- 1987.Panel B:Districts shaded according to quintile of pre/post-1967 percentage point change in the Congress party's seat share in state assembly elections. 91707.50109 the Indian government,based on agricultural sample adoption between 1967 and 1987 Panel B displays over- surveys conducted by local district-level officials.This time change in the Congress party's election perfor- data was compiled into a district-level panel dataset mance,computed by subtracting the average percent- covering 270 districts across India's major states be- age of seats won in elections after the introduction of tween 1957 and 1987 by Sanghi et al.(1998).Connect- HYV crops in 1967(inclusive)from the average per- ing the political and agricultural data yielded a panel centage of seats won before 1967 dataset linking district-level HYV crop adoption data A concern regarding a panel fixed effects OLS ap- to constituency-level electoral data for over 21,000 proach is that incumbent politicians may have strate- state assembly election races across India's major states gically manipulated the diffusion of HYV crops over between 1957 and 1987 Figure 1 displays district-level time,especially as agricultural policy became increas- maps of the data.Panel A displays mean HYV crop ingly politicized as a result of rural political mobiliza- tion.The direction of the bias implied by such dynamic targeting is unpredictable,depending upon whether 5 The districts covered in the dataset span all of India's major states the Congress party targeted supporters,opposition ar- except for Assam and Kerala,and account for over 85%of India's land area.To account for administrative splits of districts over time. eas,or swing voters (Dixit and Londregan 1996).To all data are aggregated to the level of 1961 district boundaries.During address endogeneity in the diffusion of HYV crops, the period under analysis,India experienced three rounds of redis- this paper utilizes an instrumental variables identifica- tricting of assembly constituency boundaries:in 1961,1967 and 1976. tion strategy.An instrumental variables identification This paper therefore utilizes the 1956,1961,1967,and 1976 reports of the Delimitation Commission to name match constituencies to con strategy also eliminates downward bias in coefficient temporaneous districts.It then utilizes the Administrative Atlas of estimates resulting from measurement error,which is India to match contemporaneous districts to 1961 boundaries. likely to be sizable,as a result of sampling error in the 923
Technological Change and Political Turnover FIGURE 1. Mean HYV Crop Adoption and Change in Congress Seat Share A. HYV Crop Adoption 0.000 to 0.126 0.128 to 0.170 0.170 to 0.245 0.245 to 0.341 0.341 to 0.727 B. Congress Party Decline −79.0 to −33.39 −33.39 to −19.88 −19.88 to −6.76 −6.76 to 8.25 8.25 to 52.0 Notes: Panel A: Districts shaded according to quintile of mean share of agricultural land under HYV crop cultivation between 1967– 1987. Panel B: Districts shaded according to quintile of pre/post-1967 percentage point change in the Congress party’s seat share in state assembly elections. the Indian government, based on agricultural sample surveys conducted by local district-level officials. This data was compiled into a district-level panel dataset covering 270 districts across India’s major states between 1957 and 1987 by Sanghi et al. (1998).5 Connecting the political and agricultural data yielded a panel dataset linking district-level HYV crop adoption data to constituency-level electoral data for over 21,000 state assembly election races across India’s major states between 1957 and 1987. Figure 1 displays district-level maps of the data. Panel A displays mean HYV crop 5 The districts covered in the dataset span all of India’s major states except for Assam and Kerala, and account for over 85% of India’s land area. To account for administrative splits of districts over time, all data are aggregated to the level of 1961 district boundaries.During the period under analysis, India experienced three rounds of redistricting of assembly constituency boundaries: in 1961, 1967, and 1976. This paper therefore utilizes the 1956, 1961, 1967, and 1976 reports of the Delimitation Commission to name match constituencies to contemporaneous districts. It then utilizes the Administrative Atlas of India to match contemporaneous districts to 1961 boundaries. adoption between 1967 and 1987. Panel B displays overtime change in the Congress party’s election performance, computed by subtracting the average percentage of seats won in elections after the introduction of HYV crops in 1967 (inclusive) from the average percentage of seats won before 1967. A concern regarding a panel fixed effects OLS approach is that incumbent politicians may have strategically manipulated the diffusion of HYV crops over time, especially as agricultural policy became increasingly politicized as a result of rural political mobilization. The direction of the bias implied by such dynamic targeting is unpredictable, depending upon whether the Congress party targeted supporters, opposition areas, or swing voters (Dixit and Londregan 1996). To address endogeneity in the diffusion of HYV crops, this paper utilizes an instrumental variables identification strategy. An instrumental variables identification strategy also eliminates downward bias in coefficient estimates resulting from measurement error, which is likely to be sizable, as a result of sampling error in the 923 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Aditya Dasgupta TABLE 1.Pretreatment Determinants of HYV Crop Adoption Dependent variable: HYV Intensity IADP (1) (2) (3) (4) (5) (6) Pre-INC 0.001 0.0000 (0.0004) (0.001) Pre-Irrigation 0.52* 0.24* (0.03) (0.06) Aquifer 0.15** 0.01 (0.02) (0.03) Constant 0.20* 0.13# 0.17** 0.05 -0.003 0.05* (0.03) (0.01) (0.01) (0.04) (0.02) (0.02) Observations 270 270 270 270 270 270 Notes:Unit of observation is district.Outcomes:HYV Intensity is mean district-level HYV crop adoption,1967-1987 IADP is an indicator for whether a district was selected for the Intensive Agricultural Districts Program.Explanatory variables:Pre-INC is pre-1967 average Congress seat share.Pre-irrigation is share of agricultural land with access to irrigation in 1966.Aquifer is the share of district land with a naturally occurring aquifer.Analysis estimated by OLS. p<0.1:*p<0.05;*p<0.01 4号 agricultural surveys from which the official statistics on via district and region-year fixed effects,which also ab- HYV crop adoption are obtained,as well as the aggre- sorb time invariant omitted variables as well as region- gated level,the district,at which HYV crop adoption is specific time shocks.The second stage regression is & measured. the main OLS specification.The exclusion restriction To construct an instrument for HYV crop adoption, requires that,conditional upon covariates,areas with this paper utilizes the fact that HYV crops,the cultiva- greater aquifer coverage saw a larger over-time decline tion of which required intensive and steady dosages of in the Congress party's election performance after 1967 water,delivered yield improvement only in conditions only as a result of differentially higher rates of HYV of access to controlled irrigation.As Table 1 illustrates, crop adoption over time and not for other reasons.The for this reason,after their introduction,HYV crops plausibility of this assumption is discussed and tested overwhelmingly disseminated to districts with preexist- following the main results. ing irrigation infrastructure or capacity for irrigation in To measure district-level aquifer coverage,this pa- the form of aquifers.The first stage regression takes the per utilized historical groundwater maps published in form: the National Atlas of India.based on information from the Geological Survey of India.The advantage of uti- HYVrdr yrd+in+BAquiferrd x Post967+vrdu. lizing historical maps,first published in 1977 is that they provide information about aquifers that was available where Aquiferrd x Post967 is the instrument,a cross- contemporaneously as the green revolution was occur- ring.The maps contain detailed local outlines of three sectional measure of the share of district land with a aquifer types in order of depth:(i)fairly extensive thick naturally occurring aquifer interacted with a dummy aquifers occurring beyond 150 meters,(ii)aquifers with variable that"switches on"'for all districts with the in- limited extent occurring between 100 meters and 150 troduction of HYV crops to India from 1967 onward. meters,and (iii)aquifers with restricted extent occur- This paper uses aquifer coverage as the cross-sectional ring up to 100 meters.To quantify district-level aquifer suitability measure since aquifers are naturally occur- coverage,the groundwater maps were first geocoded to ring and more clearly exogenous than preexisting irri- shapefiles of 1961 district boundaries.GIS software was gation infrastructure..Lower-order terms are absorbed utilized to generate polygons corresponding to aquifers and to compute the share of each district's surface area intersecting with an aquifer of any type.A snapshot of 6 As Table 1 illustrates,aquifer coverage was also less subject to ex- the coding process is depicted in Figure Al of the On- plicit government targeting by government programs,like the IADP. line Appendix. intended to disseminate HYV crops.Though less clearly exogenous, a possible advantage of preexisting irrigation coverage as the cross- sectional suitability measure is that it is more strongly predictive of REGRESSION RESULTS HYV crop adoption than is an instrument based on aquifer cover- age,suggesting a bias-variance tradeoff (Conley,Hansen,and Rossi A simplified version of the OLS regression analysis is 2012).Table A4 in the Online Appendix reports the main results us- /:sony ing preexisting irrigation coverage as an alternative cross-sectional visualized in Figure 2,which depicts a scatter plot of suitability measure in the instrument.The results are substantively change in Congress party vote share and seat share comparable. from the pre-to the post-1967 HYV crop introduction 924
Aditya Dasgupta TABLE 1. Pretreatment Determinants of HYV Crop Adoption Dependent variable: HYV Intensity IADP (1) (2) (3) (4) (5) (6) Pre-INC 0.001 0.0000 (0.0004) (0.001) Pre-Irrigation 0.52∗∗∗ 0.24∗∗∗ (0.03) (0.06) Aquifer 0.15∗∗∗ 0.01 (0.02) (0.03) Constant 0.20∗∗∗ 0.13∗∗∗ 0.17∗∗∗ 0.05 − 0.003 0.05∗∗ (0.03) (0.01) (0.01) (0.04) (0.02) (0.02) Observations 270 270 270 270 270 270 Notes: Unit of observation is district. Outcomes: HYV Intensity is mean district-level HYV crop adoption, 1967–1987. IADP is an indicator for whether a district was selected for the Intensive Agricultural Districts Program. Explanatory variables: Pre-INC is pre-1967 average Congress seat share. Pre-irrigation is share of agricultural land with access to irrigation in 1966. Aquifer is the share of district land with a naturally occurring aquifer. Analysis estimated by OLS. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗ p < 0.01 agricultural surveys from which the official statistics on HYV crop adoption are obtained, as well as the aggregated level, the district, at which HYV crop adoption is measured. To construct an instrument for HYV crop adoption, this paper utilizes the fact that HYV crops, the cultivation of which required intensive and steady dosages of water, delivered yield improvement only in conditions of access to controlled irrigation. As Table 1 illustrates, for this reason, after their introduction, HYV crops overwhelmingly disseminated to districts with preexisting irrigation infrastructure or capacity for irrigation in the form of aquifers. The first stage regression takes the form: HYVrdt = γrd + τrt + βAqui f errd × Post1967 t + νrdit, where Aqui f errd × Post1967 t is the instrument, a crosssectional measure of the share of district land with a naturally occurring aquifer interacted with a dummy variable that “switches on”’ for all districts with the introduction of HYV crops to India from 1967 onward. This paper uses aquifer coverage as the cross-sectional suitability measure since aquifers are naturally occurring and more clearly exogenous than preexisting irrigation infrastructure.6 Lower-order terms are absorbed 6 As Table 1 illustrates, aquifer coverage was also less subject to explicit government targeting by government programs, like the IADP, intended to disseminate HYV crops. Though less clearly exogenous, a possible advantage of preexisting irrigation coverage as the crosssectional suitability measure is that it is more strongly predictive of HYV crop adoption than is an instrument based on aquifer coverage, suggesting a bias-variance tradeoff (Conley, Hansen, and Rossi 2012). Table A4 in the Online Appendix reports the main results using preexisting irrigation coverage as an alternative cross-sectional suitability measure in the instrument. The results are substantively comparable. via district and region-year fixed effects, which also absorb time invariant omitted variables as well as regionspecific time shocks. The second stage regression is the main OLS specification. The exclusion restriction requires that, conditional upon covariates, areas with greater aquifer coverage saw a larger over-time decline in the Congress party’s election performance after 1967 only as a result of differentially higher rates of HYV crop adoption over time and not for other reasons. The plausibility of this assumption is discussed and tested following the main results. To measure district-level aquifer coverage, this paper utilized historical groundwater maps published in the National Atlas of India, based on information from the Geological Survey of India. The advantage of utilizing historical maps, first published in 1977,is that they provide information about aquifers that was available contemporaneously as the green revolution was occurring. The maps contain detailed local outlines of three aquifer types in order of depth: (i) fairly extensive thick aquifers occurring beyond 150 meters, (ii) aquifers with limited extent occurring between 100 meters and 150 meters, and (iii) aquifers with restricted extent occurring up to 100 meters. To quantify district-level aquifer coverage, the groundwater maps were first geocoded to shapefiles of 1961 district boundaries.GIS software was utilized to generate polygons corresponding to aquifers and to compute the share of each district’s surface area intersecting with an aquifer of any type. A snapshot of the coding process is depicted in Figure A1 of the Online Appendix. REGRESSION RESULTS A simplified version of the OLS regression analysis is visualized in Figure 2, which depicts a scatter plot of change in Congress party vote share and seat share from the pre- to the post-1967 HYV crop introduction 924 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Technological Change and Political Turnover FIGURE 2.HYV Crop Adoption and Congress Party Decline Across Districts A.Change in Congress Vote% B.Change in Congress Seat Share ● 导 8 4号元 ● ● 0.0 0.2 0.4 0.6 0.0 02 0.4 0.6 Mean Post-1967 HYV Crop Adoption Notes:Horizontal axis represents mean share of agricultural land under HYV crop cultivation between 1967-1987.Vertical axis repre- sents percentage point change in the Congress party's district-level vote share in panel A and seat share in panel B.Points represent districts.Line represents fitted bivariate OLS regression line. period against mean post-1967 HYV crop adoption.A The average share of agricultural land under HYV large,negative relationship is apparent,providing pre- crops across all district-years following the introduc- liminary evidence that districts that experienced the tion of HYV crops to India was 0.24(descriptive statis- green revolution with greater intensity experienced a tics for major variables are reported in Table Al of larger long-run decline in the dominant party's elec- the Online Appendix).This implies that,if we multi- tion performance.Table 2 reports the OLS regression ply this number by the estimated impact of HYV crops estimates,based on panel data and using assembly con- the green revolution accounted for roughly a 1.54 per- stituencies as the unit of analysis.The coefficients sug- centage point reduction in Congress party vote share gest that the green revolution played an important and 6.1 percentage point reduction in Congress party's role in the decline of single-party dominance in India. seat share from the pre-to the post-green revolution The results of the baseline specification,reported in period.These are very large estimates,accounting for columns (1)and (6),imply that a 10 percentage point roughly a third of the Congress party's observed de- MM//:sdny increase in the share of agricultural land planted with cline of 5.8 percentage points in vote share and 16.9 per- HYV crops reduced the dominant party's vote share by centage points in seat share from the pre-to the post- 0.64 percentage points and the probability of electing a HYV crop introduction period in the sample under dominant party legislator by 2.6 percentage points. analysis. 925
Technological Change and Political Turnover FIGURE 2. HYV Crop Adoption and Congress Party Decline Across Districts Notes: Horizontal axis represents mean share of agricultural land under HYV crop cultivation between 1967–1987. Vertical axis represents percentage point change in the Congress party’s district-level vote share in panel A and seat share in panel B. Points represent districts. Line represents fitted bivariate OLS regression line. period against mean post-1967 HYV crop adoption. A large, negative relationship is apparent, providing preliminary evidence that districts that experienced the green revolution with greater intensity experienced a larger long-run decline in the dominant party’s election performance. Table 2 reports the OLS regression estimates, based on panel data and using assembly constituencies as the unit of analysis. The coefficients suggest that the green revolution played an important role in the decline of single-party dominance in India. The results of the baseline specification, reported in columns (1) and (6), imply that a 10 percentage point increase in the share of agricultural land planted with HYV crops reduced the dominant party’s vote share by 0.64 percentage points and the probability of electing a dominant party legislator by 2.6 percentage points. The average share of agricultural land under HYV crops across all district-years following the introduction of HYV crops to India was 0.24 (descriptive statistics for major variables are reported in Table A1 of the Online Appendix). This implies that, if we multiply this number by the estimated impact of HYV crops, the green revolution accounted for roughly a 1.54 percentage point reduction in Congress party vote share and 6.1 percentage point reduction in Congress party’s seat share from the pre- to the post-green revolution period. These are very large estimates, accounting for roughly a third of the Congress party’s observed decline of 5.8 percentage points in vote share and 16.9 percentage points in seat share from the pre- to the postHYV crop introduction period in the sample under analysis. 925 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Aditya Dasgupta The estimates are robust to a range of additional specifications.The specifications reported in columns 38 2型 uIM ONI pue 价E)M-1snne Ker (2)and(7)additionally control for year fixed effects interacted with a variety of pretreatment variables:a measure of initial district-level Congress party seat share in 1957 a measure of the urbanization rate ac- cording the 1961 census,an estimate of pre-green rev- olution agricultural productivity per rural capita(com- puted by multiplying district crop production statistics 可 墨 型型 for the six major crops by real crop prices and divid- ing by the rural population according to the 1961 cen- sus),a dummy variable indicating historical exposure to direct British colonial rule,and a variable represent- ing the share of land exposed historically to a colonial- 可 (29.x 型型 era landlord-based land tenure system.The estimates remain substantively large and significant,suggesting AAH that the results are not driven by time trends related to these important variables. Columns (3)and (4)and(8)and(9)split the sam- 里 型型 ple into the northern and non-northern regions,to ex- amine effect heterogeneity.Scholarship on the rise of agrarian mobilization in India often focuses on lower- caste and rural movements in the Hindi belt and north- ern India (see,e.g.,Brass 1980;Jaffrelot 2003).The 可 思 型 ONI north of India is culturally distinct and,apart from Pun- jab,relatively less prosperous than other regions of the country,raising the possibility that the effects of the green revolution were perhaps unique to the spe- cific context of this region.However,the argument ad- vanced in this paper,tying political turnover to techno- logical change,is not intrinsically region specific.This is p 2型 supported by the data and analysis.The estimates are similar in both subsamples,demonstrating that the ef- fects of technological change on political turnover can exist across heterogeneous social environments. Finally,columns(5)and (10)control for state-year 程 型型 8L0ZL fixed effects to address potentially unobserved cross state heterogeneity and limit the comparisons driving the estimates to over-time changes in HYV crop adop- tion across districts within the same state.While the estimated coefficients still possess the expected signs, 型型 罩 点 the magnitude of the coefficients shrinks considerably. 可 As discussed,however,this is potentially driven by the fact that the inclusion of a large number of fixed ef- fects tends to attenuate OLS estimates in the presence of measurement error (see,e.g.,Angrist and Pischke 平 型型 2008).A test of this is to compare OLS estimates with instrumental variables estimates,which are not biased downward due to measurement error or due to the po- tentially strategic distribution of HYV crops. As discussed,the intuition behind the IV strategy 型型 is to examine whether areas that saw higher rates of HYV crop adoption from the pre-to the post-1967 period,as a result of greater suitability for the new technology,also experienced a larger over-time decline d of single-party dominance.A simplified version of the analysis is visualized in Figure 3 in the form of two district-level plots.Panel A displays the absence of a re- lationship between aquifer coverage and the Congress 7 District-level data on direct colonial rule and colonial-era land tenure come from Iyer (2010). 926
Aditya Dasgupta TABLE 2. OLS Estimates: Effects of HYV Crops on Single-Party Dominance Dependent variable: INC Vote % INC Win (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) HYV − 6.41∗∗ − 6.82∗∗∗ − 9.13∗∗∗ − 2.19 − 0.57 − 25.57∗∗∗ − 23.43∗∗∗ − 29.44∗∗∗ − 19.55∗ − 7.35 (2.94) (2.48) (3.54) (4.18) (2.74) (6.14) (4.18) (7.32) (10.33) (6.11) Specification Main Controls Northern Non-North State-Year FE Main Controls Northern Non-North State-Year FE Region-Year FE Yes Yes Yes Yes No Yes Yes Yes Yes No District FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 21907 20057 9889 12018 21907 21907 20057 9889 12018 21907 Notes: Unit of observation assembly constituency-year for election years between 1957–1987. Outcomes: INC Vote % indicates Congress percentage point vote share and INC Win indicates whether Congress won seat. Explanatory variables: HYV crop adoption is district-level share of agricultural land planted with HYV crops. Standard errors adjusted for two-way clustering within districts and by region-year. ∗ p < 0.1; ∗∗ p < 0.05; ∗∗∗p < 0.01 The estimates are robust to a range of additional specifications. The specifications reported in columns (2) and (7) additionally control for year fixed effects interacted with a variety of pretreatment variables: a measure of initial district-level Congress party seat share in 1957, a measure of the urbanization rate according the 1961 census, an estimate of pre-green revolution agricultural productivity per rural capita (computed by multiplying district crop production statistics for the six major crops by real crop prices and dividing by the rural population according to the 1961 census), a dummy variable indicating historical exposure to direct British colonial rule, and a variable representing the share of land exposed historically to a colonialera landlord-based land tenure system.7 The estimates remain substantively large and significant, suggesting that the results are not driven by time trends related to these important variables. Columns (3) and (4) and (8) and (9) split the sample into the northern and non-northern regions, to examine effect heterogeneity. Scholarship on the rise of agrarian mobilization in India often focuses on lowercaste and rural movements in the Hindi belt and northern India (see, e.g., Brass 1980; Jaffrelot 2003). The north of India is culturally distinct and, apart from Punjab, relatively less prosperous than other regions of the country, raising the possibility that the effects of the green revolution were perhaps unique to the specific context of this region. However, the argument advanced in this paper, tying political turnover to technological change, is not intrinsically region specific. This is supported by the data and analysis. The estimates are similar in both subsamples, demonstrating that the effects of technological change on political turnover can exist across heterogeneous social environments. Finally, columns (5) and (10) control for state-year fixed effects to address potentially unobserved crossstate heterogeneity and limit the comparisons driving the estimates to over-time changes in HYV crop adoption across districts within the same state. While the estimated coefficients still possess the expected signs, the magnitude of the coefficients shrinks considerably. As discussed, however, this is potentially driven by the fact that the inclusion of a large number of fixed effects tends to attenuate OLS estimates in the presence of measurement error (see, e.g., Angrist and Pischke 2008). A test of this is to compare OLS estimates with instrumental variables estimates, which are not biased downward due to measurement error or due to the potentially strategic distribution of HYV crops. As discussed, the intuition behind the IV strategy is to examine whether areas that saw higher rates of HYV crop adoption from the pre- to the post-1967 period, as a result of greater suitability for the new technology, also experienced a larger over-time decline of single-party dominance. A simplified version of the analysis is visualized in Figure 3 in the form of two district-level plots. Panel A displays the absence of a relationship between aquifer coverage and the Congress 7 District-level data on direct colonial rule and colonial-era land tenure come from Iyer (2010). 926 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
Technological Change and Political Turnover FIGURE 3.Reduced Form Relationship between Aquifer Coverage and Congress Performance A.Before Green Revolution B.After Green Revolution ● 品 8 4号 ● 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 Aquifer Coverage Notes:Horizontal axis represents share of district land with a naturally occurring aquifer.Vertical axis represents percentage point Congress party seat share.Panel A represents the pre-1967 period and panel B the post-1967 period.Points represent districts.Lines represent fitted bivariate OLS regression lines. party's election performance before the introduction tween the instrument and HYV crop adoption across of HYV crops in 1967 Panel B displays the negative specifications. reduced-form relationship that emerged following the The magnitude of the IV estimates is substantially introduction of HYV crops.Table 3 reports the IV re- larger than that of the OLS estimates.This is likely gression estimates,showing that districts that experi- due in part to measurement error attenuating the OLS enced the green revolution with greater intensity due estimates:consistent with this,the IV estimates show to greater aquifer coverage saw a much larger decline greater stability when a large number of fixed effects in single-party dominance.The baseline IV estimates are introduced in columns(5)and (10)than do the OLS reported in columns(1)and(6)suggest that a 10 per- estimates.The impact on dominant party seat share is centage point increase in the share of agricultural land virtually unchanged,though the estimated impact on planted with HYV crops resulted in a 2.3 percentage vote share shrinks very slightly.This suggests that the point decline in Congress party vote share and 8.1 per- OLS results are likely biased downwards relative to the centage point loss of seat share.The sign and mag- true effect of HYV crop adoption on dominant party nitude of the estimates are robust across the inclu- decline.Another reason that the OLS estimates may sion of covariates and different subsamples of the data be biased is if HYV crops were strategically distributed, The first stage estimates show a strong relationship be- particularly in later stages of the green revolution when 927
Technological Change and Political Turnover FIGURE 3. Reduced Form Relationship between Aquifer Coverage and Congress Performance Notes: Horizontal axis represents share of district land with a naturally occurring aquifer. Vertical axis represents percentage point Congress party seat share. Panel A represents the pre-1967 period and panel B the post-1967 period. Points represent districts. Lines represent fitted bivariate OLS regression lines. party’s election performance before the introduction of HYV crops in 1967. Panel B displays the negative reduced-form relationship that emerged following the introduction of HYV crops. Table 3 reports the IV regression estimates, showing that districts that experienced the green revolution with greater intensity due to greater aquifer coverage saw a much larger decline in single-party dominance. The baseline IV estimates reported in columns (1) and (6) suggest that a 10 percentage point increase in the share of agricultural land planted with HYV crops resulted in a 2.3 percentage point decline in Congress party vote share and 8.1 percentage point loss of seat share. The sign and magnitude of the estimates are robust across the inclusion of covariates and different subsamples of the data. The first stage estimates show a strong relationship between the instrument and HYV crop adoption across specifications. The magnitude of the IV estimates is substantially larger than that of the OLS estimates. This is likely due in part to measurement error attenuating the OLS estimates; consistent with this, the IV estimates show greater stability when a large number of fixed effects are introduced in columns (5) and (10) than do the OLS estimates. The impact on dominant party seat share is virtually unchanged, though the estimated impact on vote share shrinks very slightly. This suggests that the OLS results are likely biased downwards relative to the true effect of HYV crop adoption on dominant party decline. Another reason that the OLS estimates may be biased is if HYV crops were strategically distributed, particularly in later stages of the green revolution when 927 Downloaded from https://www.cambridge.org/core. Shanghai JiaoTong University, on 26 Oct 2018 at 03:53:06, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S000305541800031X
