Institutional and community adaptation from the archives: A study of drought in western India, 1790-1860

George C.D. Adamson

Accepted for publication in Geoforum

Abstract

The analysis of the socioeconomic effects of, and societal adaptation to, extreme climatic events in the past is an important tool in preparing for future adaptation. However, the history of social responses to climate is relatively understudied,and analyses that do exist take a broadly historical perspective, detached from current vulnerability, adaptation or resilience theory. This analysis attempts to address this empirical scarcity through an examination of drought responses in early-nineteenth century western India, particularly at the government level. The analysis reveals a sociopolitical system that was highly robust in the face of climate stress, despite high levels of vulnerability at the individual level, with hegemonic structures maintained by a system of grain dealer/moneylenders. The period witnessed a move towards market-driven drought response, whichapparently reduced adaptive capacity in the short term by allowing speculation in the grain market. However, this is likely to have been buffered in the medium term by increases in cultivation, greater market openness and the almost universal acceptance of market-drivendrought policybygovernment actors, resulting in a swift response. However, the apparent success of adaptation measures during the first half of the nineteenth century may have masked vulnerabilities to more extreme climatic events, contributing to the catastrophic Indian famines of the late nineteenth century. This bears similarities to the ‘catastrophe hypothesis’, the significance of which is discussed. The study argues for a greater focus on social responses to extreme climatic events in the past in order to address contemporary challenges.

Key words: Social vulnerability, adaptation, climate change, India, colonial

1.1 Introduction

The studyof social responses to climate variability in the past provides one method with which to understand the responses of social-ecological systems to climatic stress. A historical approach can have several benefits over analyses of more recent events. Historical studies permit analysis of the long-term effects of adaptation decisions upon social structures (cf. Cannon, 1994; Kates, 2000)and the influence of adaptive decisions on overall climatic vulnerability (Bowden et al., 1981; Pelling, 2003). A long-term context can help understand the processes and discourses that facilitate or constrain adaptive responses, both at the community and societal level (Stehr and von,Storch 1995; Vedwan and Rhoades, 2001; Endfield et al., 2004; Brázdil et al., 2005; Tompkins and Adger, 2004; Endfield, 2007). In the preparation of disaster management plans, historical studies represent place-specific comparators that can allow particularly sensitive regions to be determined, allowing for an estimation of the most vulnerable groups, individuals and geographical regions (Carter et al., 2007; Cœur and Lang, 2008). Analyses of the distant past are particularly pertinent where previously rare events are forecast to become more regular, such as in tropical monsoon regions (Carter et al., 2007; Kripalani et al., 2007; Berger et al., 2008; Turner and Slingo, 2009; Gupta et al., 2011; IPCC, 2012).

The study of social vulnerability to climatic variability in the past can permit thederivation oftheoretical constructs and universalitiesregarding social vulnerability, which can be applied to contemporary challenges.Several previous studies have attempted this, using the available empirical data. Bowden et al.’s (1981)‘catastrophe’ and ‘lessening’ hypotheses address the long-term effects of climatic adaptation on society, derived through analysis of the effects of climate on populations in the Tigris-Euphrates Valley, the Sahel and the US Great Plains.The ‘lessening’ hypothesis argues that all societies are able to respond to climatic events with a return period of less than 100 years. This adaptation, however, can mask or even increase vulnerability to major climatic events, i.e. those with a return period of more than 100 years (the ‘catastrophe ‘hypothesis’). This is similar to the law of unintended consequences in complex social-ecological systems outlined by van der Leeuw(2007: 234-235) outlines. This states that human interventions generally reduce in the occurrence of minor disturbances but can increase the risk of less frequent disturbances, generatingcatastrophic‘time bombs’. Other scholars focus on the constraints provided by climatic stress:Messerli et al. (2000),Brunk (2002) and Anderies (2006) highlight the role of complexity in allowing climatic stresses to propagate across society. Butzer (2012) echoes this approach, although he argues that social-ecological systems can exist in a highly complex and fragile state for decades, with the impacts of climate stress governed by the society’s political and environmental resilience. Boserup (1981) presents an alternative role for environmental stress, suggesting that stress can act as a trigger for sociocultural development by creating pressure to innovate and reform production processes(the ‘Boserupian escape clause’ (Coombes and Barber, 2005)).

A review of literature on famines in the historical record has been used by Fraser (2007) to derivea framework to understand how systems become vulnerable to food insecurity. The resultantmodelpresents societies as existing within a three-dimensional vulnerability ‘space’, consisting of agroecological capacity, livelihood entitlements and institutional flexibility. Vulnerability to famineis increased when the agro-ecological aspect of a system becomes more fragile, institutional capacity weakens, or livelihood options reduce. In practice, this framework is manifested infour features,evidenced by all societies in the immediate run-upto a famine(Fraser,2006; echoed in Orlove, 2005):

  1. High proportions of the population in poverty, or large increases in the number of people in poverty,
  2. Increasing reliance on fragile and specialized agricultural systems,
  3. Lack of responses designed to protect against climatic extremes at a greater extent than those previously encountered,
  4. Weak institutional response at the start of the event.

Recently there have been calls within the literature for a greater research focus on the responses of societies to historical climatic stress (Dearing et al., 2010; Pfister, 2010; Butzer, 2012), including within the Intergovernmental Panel on Climate Change Fourth Assessment Report (Carter et al., 2007: 138). However, empirical data on social responses to climate variability in the past remains relatively thin. With a few exceptions (cf. Bowden et al., 1981; Messerli et al., 2000; Endfield and Tejedo, 2004; Orlove, 2005; Fraser, 2006; 2007; Glaser et al., 2006; Endfield, 2007; McLeman et al., 2008; Butzer, 2012; Endfield, 2012), the majority of existing studies relate to very recent events, with the more-distant past (i.e. before living memory)generally overlooked. Studies that do exist, such as the wide literature on famine history(cf. Bhatia, 1991; Fagan, 2000; Davis, 2001), are generally detached from vulnerability, adaptation or resilience approaches, reflective of the differing epistemological standpoint between historians and contemporary development/adaptation research. This empirical scarcity has resulted in an element of particularity within theoretical understandings of the role of climate on society in the long term, which limits the applicability of these theories to contemporary climate change research. In order to develop our understanding of long-term climate vulnerability a substantial increase in theoretically-grounded historical case studies is necessary.

This paper adds to the cannon of literature on social responses to climate in this past, through an analysis of social vulnerability and adaptation to climatic extremes in western India (the colonial Bombay Presidency) during the period 1790-1860. It wouldbe simplistic to argue that this periodis directly analogous to today. However,certain factors render it of interest. The period experienced repeated meteorological droughts (Kallis, 2008; Adamson and Nash, 2014) and significant sociopolitical upheavals, sometimes occurring concurrently (Mooley and Pant, 1981).The region exhibited a high level of social vulnerability to climate, which is still seento some extent today[1](Meyer et al., 1998; Karl and Easterling, 1999; O’Brien et al., 2004; Brenkert and Malone, 2005; Malone and Brenkert, 2008). Global export markets and a market-driven drought response policy developed during the study period;thisbears some similarities with the current dual vulnerability of the region to climate change and globalizing forces (O’Brien et al., 2004).The years 1790 to 1860 are also particularly important as they immediately precede a period of heavy faminesduring the late-nineteenth and early-twentieth centuries (including 1868, 1877, 1899 and 1918). These havebeen studied in relative detail (cf. Srivastava, 1968; McAlpin, 1983; Klein, 1984; Bhatia, 1991; Davis, 2001; Hall-Matthews, 2005; Burgess and Donaldson, 2010; Roy, 2012); however, social vulnerability in the decades leading up to these events has not been adequately addressed, a particular oversight given the postulations within Bowden et al.’s (1981) lessening and catastrophe hypotheses.

The study area adopted spans the central-west coast of the Indian peninsular (Figure 1). Until recently little climatic information was available for this region before the late-nineteenth century. However, recent advances in dendroclimatology (Borgaonkar et al., 2010) and documentary climate reconstruction (Walsh et al., 1999; Sontakke et al., 2008; Brohan et al., 2012; Adamson and Nash, 2014; Nash andAdamson,2014) have begun to address this issue, allowing the effects of climate variability to be more fully addressed[2].Thisstudy relies on English-language records, particularly contemporary newspapersand the personal and government records of the British East India Company (EIC), the imperial trading organisation that served as the Government of an increasingly large portion of the Indian subcontinent during the early-nineteenth century[3](Table 1).A focus entirely on colonial records provides limitations, which are addressed throughout the analysis. However, several studies point towards the importance of institutional actors in driving the ability of a society to adapt, both in relation to historical (Orlove, 2005; Fraser, 2006; Butzer, 2012) and contemporary societies (Adger and Kelly, 1999; Tompkins and Adger, 2004; Adger et al., 2009).Sources pertaining tothe governing elite therefore provide strong data for the analysis of institutional adaptive decision-making,although the village-level voice is rather more poorly represented.

1.2 A note on terminology adopted

Adaptation terminology within climate change literature is inconsistent and occasionally contradictory (McLaughlin and Dietz 2008). The terminology here derives from social-ecological systems theory (cf. Berkes and Folke, 1998; Anderies et al., 2004; Anderies, 2006; Janssen, 2006; Glaser et al., 2012),and differentiates between the response of social-ecologicalsystems to climatic change – categorized by robustness versus fragility – and the response of individuals and communities – represented by vulnerability, adaptive capacity and resilience. Here vulnerability is adopted underthe definition provided by Adger (2006: 268), that of “state of susceptibility to harm from exposure to stresses associated with environmental and social change and from the absence of capacity to adapt”. Vulnerability is hence a function of adaptive capacity, defined in this case as an individual or community’s ability to evolve or change its response in relation to stress (Adger2006: 270). Vulnerability is also affected by the degree of environmental exposure,as well as an individual’s entitlements, i.e. capital, social relations and institutional participation (Bohle et al., 1994). The terms fragility and robustness are used here to responses to climate stresses at the systems level (Anderies et al., 2004). Individual vulnerabilities may exist within a robust system, although they may also contribute to systemic fragility. Resilience, a term adopted infrequently in this analysis, refers to the ability of small groups or communities to “cope with external stresses and disturbances as a result of social, political and environmental change” (Adger 2000, 347).Adaptive capacity is further categorised under this definition as the facility of actors to manage resilience (Walker et al., 2004).

2.1 Droughts in western India, 1790-1860

Ten periods of widespread meteorological drought (i.e. deficient rainfall across the region) occurred in western India during 1790 to 1860 (Adamson and Nash,2014; Adamson,in press). These occurred in 1790-92, 1803, 1812, 1823-24, 1833, 1838, 1845, 1847-48, 1850 and 1855. With the exception of 1847-48, 1850 and 1855,all were accompanied by reports of food price rises, widespread migration, and death and debility due to starvation. Detaileddiscussions of the socioeconomic effects of each of these droughts are provided in Adamson (in press), and the meteorological conditions associated with them in Adamson and Nash (2014).An outline of the droughts is discussed below.

The first major drought to be recorded occurred from 1790 to 1792.Materials from this period are limited, due to the restricted extent of EIC territories. However, those that exist (predominantly the records of the government committees created to address grain shortages) report widespread crop failures, including in the islands of Bombay (Mumbai) (BL IOR/F/4/428/10490 24 December 1790), Poona (Pune) and Broach (Bharuch)[4] (NLS MS.13695 8 April 1804). The principal recorded impact was one that was reported in most major droughts during the period:market shocks in the form of sustained price rises for all food grains. These were apparently long lasting, with grain prices in Poona described in a contemporary newspaper in 1793 as “100 per cent dearer than they were three years ago”(BLNC MC1112 19 October 1793).The following drought in 1803 was concurrent with Second Anglo-Maratha War, which was fought in the Deccan from 1802 to 1804. Famine conditions, likely to be associated with both deficient rainfall and the effects of the war (Adamson,in press), were reported in the Deccan, with people dying “daily and in very great numbers” by October 1804 (BL MSS Eur F175/7 1 October 1804). Significant depopulation was reported, ostensibly amounting to around half of the population of the Konkan and western Deccan,although no exact figures are available (BL MSS Eur F175/8 5 December 1804)[5].

The drought of1812 was experienced most strongly in the region of Gujarat. Reports are available from a British Lieutenant stationed at Porbandar on the Kathiawar Peninsular, and from a report submitted to the Bombay Literary Society by the EIC Resident at Baroda (Vadodara) James Carnac. Here adeficient monsoon in 1812 wasprecededby a locust swarm and significant in-migration from the adjacent province of Marwar (present-day southern Rajasthan) due to crop failures in 1811 (Carnac, 1819: 297). “Almost all Cattle”, camels and horses were reported to have died in Porbandar in September of 1812,and grain prices rose considerably. Rice that was sold was described as“unserviceable … [with] a stench like a common sewer” (BL MSS Eur D666/1)[6]. The reported human impact includes death from starvation and disease and the voluntary sale of individuals or theirchildren into slavery, with the most severe consequences of the drought reported amongst the immigrantMarwaree population(Carnac, 1819: 298; BLNC MC1112 4 December 1813). Mortality at Ahmedabad was estimated at “a hundred thousand souls”andover 500 deaths per day were reported at Baroda, (Carnac, 1819: 300), although it is possible that these numberswere exaggerated[7].

The drought of 1824waspreceded by a deficient monsoon in many regions, particularly in the Deccan (Adamson and Nash,2014). A Grain Committee was created in 1824 to address food shortages; the minutes of the Committee reporta total failure of the rabi[8](winter) crop in Poona in 1823 (BL IOR/F/4/919/2586725 November 1823), coupled with widespread water shortages (BL MC1112 1 November 1823). A grain “scarcity” was reported after the deficient monsoon of 1824 (BLNC MC1112 9 October 1824; BLNC MC1112 16 October 1824; DAGM GEN 27 October 1824), with migration into Bombay from Kathiawar and reports of the selling of children (BL MSS Eur D888/1 23 September 1824[9]).Widespread famine was, however, apparently avoided due to heavy rainfall in October (BL MSS Eur F88/425 5 October 1824; BLNC MC1112 9 October 1824; BLNC MC1112 16 October 1824; DAGM GEN 27 October 1824)[10]. Substantially elevated grain prices were reported until at least the monsoon of 1825, particularly in Gujarat, with severe water shortages and cattle deaths reported in Bombay and the Deccan (BLNC MC1112 3 July 1824; BL MSS Eur D888/1; BL MSS Eur F88/426)[11].

The 1833drought waspreceded by 3-4 years of localised deficient monsoon, with crop failures reported throughout western India, particularly aroundPoona (DAGM REV 3/407 4 September 1830; 14 August 1831; 30 September 1831; DAGM REV 8/476 30 July 1831; 15 March 1833)[12]. In November 1832 newspapers and government records reported grain prices in the Deccan at three times the average (BL IOR/F/4/1439/56798 7 November 1832; BLNC MC1112 17 November 1832). The Deccan therefore suffered most severely following the deficient monsoon of 1833 (DAGM REV 34/501 15 July 1833).During this season, in contrast to 1824, late-monsoon rains were especially poor (Adamson and Nash, 2014), with government records noting a heavy impact on the harvest (DAGM REV 34/501 16 October 1833). Newspaper reports suggestthat grain prices remained elevated in until the monsoon of 1834 (BLNC MC1112 6th August 1834), again coupled with a severe water shortage (BL MC1112 12 April 1834).

The most severe societal effects of the drought of 1838 were once again reported in Gujarat, particularly the Kathiawar Peninsular. Contemporary newspaper and government reports detailed almost total crop failures in Bombay (BL MC1122 3 October 1838), Gujarat (BLNC MC1112 19 January 1839; DAGM REV 17/1101 28 September 1839; DAGM REV 17/1101 14 September 1839) and Nashik (UBL C I3/O75/2 10 November 1838), with substantial deaths of cattle. 1838 was, however, the last year during the study period when widespread food insecurity was reported. A localised famine was recorded in English-language newspapers aroundPoona and Ahmednagar in 1845 (BL SM73 2 February 1846) – during which time it was reported that the population survived on “jungle trees” (BLNC C1114 2 February 1846) – and further widespread droughts were reported in 1847-48, 1850 and 1855. However, reports of socioeconomic disruption decreased substantially during this period, and the predominant complaints were of water shortage rather than food stress and starvation. The possible reasons for this will be discussed.

2.2 Social Vulnerability in the Bombay Presidency

Due to the generally illiterate nature of the rural population, little ofthe archival information consulteddirectly relates to everyday village life. Nevertheless,combining archive material with existing published studies (using both English- and Indian-language materials) permitsa general overview of the rural economy. Broadly, western India contained a hegemonic social structure, maintained by a mixture of economic and moral-economic agreements (Hardiman, 1996; Vasavi, 1998). At the base of this structure were rural labourers and servants, whose upkeep was dependent upon payment in kind from cultivators (ryots), (Fukazawa, 1972;McAlpin, 1983;Guha, 2004).Cultivators were themselvesoften indebted to village-level moneylenders (sahukarsof the Baniyacaste) (Kaiwar, 1994; Hardiman, 1996). These sahukarsaccepted the majority of the payment for loans in grain, and therefore also formed the principal grain dealers. The flow of capital through the social hierarchy was tied closely to the grain trade, with capital, predominantly in the form of grain, moving from village sahukars, through urban graindealers, to the major urban bankers who financed the ruling classes (Hardiman, 1996). This system, coupled with a network of indigenous tax collector/administrators (Mamletdars)[13], allowed the ruling classes (originally the Maratha confederacy and later the EIC) to maintain their hegemony without the need for force. Direct feudal relationships (the zamindarsystem) existed in parts of Gujarat,although these were not favoured by the colonial government (Kaiwar, 1994).