1 Introduction

T

We begin, in the next section, by outlining some of the mechanisms through which both these types of emissions affect climate and discuss how these changes are likely to affect the lives of the poor. Although there is considerable uncertainty about the extent to which climate will change, even the most conservative projections lead us to expect substantial reductions in crop yields, rises in the sea level and an increased occurrence of flood and drought conditions. India’s poor remain highly dependent on agriculture and, in the absence of creative policies that insure rural incomes and provide alternative livelihoods, increased poverty may well accompany changes in physical geography.

In Section 3, we present results of a survey of 10 villages a ffected by the Kosi flood of August 2008 in northern Bihar. A lthough the flood itself was attributed to negligence in the maintenance of the Kosi dam and its embankments, it provides a useful case study of the impact of an unexpected event on a rural and largely poor population. Using household survey data we find that with the help of transfers from local communities and the state, flood-affected households showed remarkable resilience in the immediate aftermath of the flood. Their long-term prospects however remain grim. Very little of the land in these villages is under cultivation, unemployment is pervasive and even mandated state programmes such as the National Rural E mployment Guarantee Scheme (NREGS) have remained largely inactive. Although these effects are clearly not limited to the poor, the poor are particularly vulnerable to these changes b ecause of their limited access to insurance and credit.

We conclude with some suggested directions for policies to combat climate change and its effects on the Indian population.

Given that some amount of global warming seems inevitable, these policies should combine mitigation with adaptation. A comprehensive policy agenda will include rethinking embankments, enacting and enforcing laws that limit the burning of rubbish, biofuels and plant residues, developing new crop varieties that can be used profitably in areas which undergo physical changes, controlling urban air pollution, and introducing social insurance programmes that can help local populations deal with shocks i nduced by climate change. Many of these programmes will provide benefits that go far beyond their effects on climate change. They require, however, a state committed to constructive change.

2 Projected Climate Changes and Their Effects

The average global temperature has risen by about 0.8 degrees Celsius from the pre-industrial level. The projected hazards of continuing anthropogenic climate change include decreased crop yields, the disappearance of mountain glaciers and snowpacks, more extreme weather events such as floods, droughts and storms, increased coastal flooding, and species extinctions. We discuss each of these and their likely impact on the poor.

Crop Yields

Climate change is projected to reduce net cereal production in south Asian countries by 4% to 10% by 2100 under the most conservative scenario of the Intergovernmental Panel on Climate Change (IPCC) (Cruz, Harasawa et al 2007). Several studies have found a negative correlation between crop yields and poverty rates. Time series data from India for the period 1958-94 indicate that increased yields lowered poverty almost one for one (Datt and Ravallion 1998). Household data in the 1983-99 period suggested that agricultural productivity growth was responsible for at least four-fifths of the 75% growth in real agricultural wages in that p eriod (Eswaran, Kotwal et al 2008). A recent survey of the literature on Indian economic growth concludes that “agricultural productivity would have to continue to increase for improving the living standards of the rural poor” (Kotwal, Ramaswami et al 2009).

We now have some evidence linking particle emissions to yields. These emissions result in reduced solar radiation reaching the ground, especially in winter (Chameides, Yu et al 1999), and the deposit of smoke particles on leaves obstructs photosynthesis ( Bergin, Greenwald et al 2001). It is estimated that the brown cloud formed as a result of these emissions reduced the mean a nnual rice harvest in India by nearly 4% in the period 1966-84 and by nearly 10% in 1985-98 (Auffhammer, Ramanathan et al 2006).

Glacier and Snowpack Decline

Glaciers the world over are thinning and shrinking as the planet warms, and glaciers in the Himalayas are receding faster than anywhere else. If the earth keeps warming at the current rate, Himalayan glaciers are likely to disappear altogether in 25 years (Cruz, Harasawa et al 2007). In the absence of glaciers, rivers in the Indo-Gangetic plain will become much more seasonal, threatening the rabi crop as well as domestic and industrial water supplies in the non-monsoon months. In addition, more precipitation will fall as rain rather than snow and the greater water run-offs will increase flooding.

52

Up to half of the glacier decline is thought to have occurred as a result of upper atmospheric heating from the black carbon p articles in the south Asian brown cloud (Ramanathan and C armichael 2008). In addition, deposits of these soot particles in snow and ice accelerates melting. This implies that India has the ability to slow the melting of glaciers and snow by reducing p articulate emissions. We return to this point below.

Extreme Weather Events

A warmer climate is predicted to bring more extreme weather. The period 1951-2000 has witnessed an increase in the magnitude and frequency of high intensity rain in India and a decrease in the frequency of moderate rain (Goswami, Venugopal et al 2006). The record rainfall and consequent flooding in Mumbai in July 2005 was an example of this. The concentration of rainfall in a few events will tend to reduce groundwater recharge and a ccentuate droughts in water-stressed regions. India is predicted to reach a state of water stress by 2025 in which per capita water availability falls below 1,000 cubic metres per capita. An increase in cyclone intensity of 10-20% for 2-4 degrees Celsius of warming is predicted for south Asia and adjoining regions (Cruz, Harasawa et al 2007). These cyclonic storms will particularly affect the poor in coastal areas like Orissa and Bengal.

Sea Level Rise

Global warming has been raising the sea level because warm w ater has greater volume. In addition, there is the melting of Greenland and West Antarctic ice packs. A recent study that takes into account both thermal expansion of the ocean and ice-pack melt suggests that the likely range of the rise by 2100 is 0.8 to 2 metres (Pfeffer, Harper et al 2008). This will lead to the permanent displacement of millions of people in coastal areas in India, about 3 million for a 1-metre rise and more than double that number for a 2-metre rise (Dasgupta, Laplante et al 2007).

Species Extinctions

The IPCC report predicts that 30 to 40% of all species could go extinct if the temperature were to rise by another 2 to 3 degrees Celsius. Apart from its intrinsic value, biodiversity provides p ollinators and species that prey on pests. Species extinctions may result in a fundamental rearrangement of ecosystems and the effects of this on agriculture are unpredictable.

Positive Feedbacks and Loss of Control

In discussing the effects of rising temperatures above, we have ignored the role of various positive feedbacks that may come into play and amplify global warming (Lenton, Held et al 2008). The terrestrial biosphere which is now a net sink for carbon dioxide (absorbs CO2) may become a net source as temperatures rise. Methane frozen in permafrost in the Arctic may be released as the permafrost thaws, leading to still higher temperatures, more methane releases, and a process of runaway global warming that we can no longer influence by addressing our own emissions. There are other such dangers that have been identified but about which too little is known to say just how great a temperature increase is required to take the system past a point of no return. It is

august 1, 2009 vol xliv no 31

CLIMATE CHANGE NEGOTIATIONS

Figure 1: The Flooded Areas and Surveyed Villages

important to note in this connection that these feedbacks are not captured in the predictions of the climate models. This is why it is falsely reassuring to assume, as do many commentators, that if the concentration of CO2-equivalent can be held below some t arget level such as 450 parts per million, then warming will be held to 2 degrees Celsius above pre-industrial temperatures. If some of these positive feedbacks come into play at lower concentrations, it may be impossible to prevent much larger temperature increases.

3 A Case Study of the Impact of a Disaster

This section discusses the Kosi flood of 2008 in Bihar, how it a ffected the poor and how the government provided relief.

Background to the Survey

We have discussed how frequency of floods and other natural disasters is on the increase as a consequence of climate change. In this section we summarise survey data on household and state responses to a major flood in northern Bihar in 2008. The flood a ffected communities that are largely agricultural and extremely poor by national standards. Unlike many coastal areas and s ettlements near rivers, most of this population had not experienced a flood of this magnitude for several decades. To the extent that climate change will result in unexpected changes in local

Economic & Political Weekly

EPW

g eography, the impact of the flood on these households may be i ndicative of the future impact of climate change on the poor.

The Kosi flood of 2008-09 began with the breach of an embankment just north of the border in Nepal on the morning of 18 August 2008. The embankment had been in place for nearly half a century and silt trapped within it had raised the river bed until it was 4 metres higher than the surrounding land. When the breach occurred, the westward loop previously taken by the river was cut off, flooding the area between the breach and the river Ganga, 150 km to the south. The post-breach direction was in fact the course of the river before the embankment was built in the 1960s.

The map in Figure 1 shows both the old course to the west and the north-south course followed by the river waters after the breach.

In February 2009, we surveyed 28 households in each of 10 v illages along the new course of the river. The map in Figure 1 shows the surveyed villages along with both the old course to the west and (in darker lines) the north-south course followed by the river waters after the breach. Almost all the villages affected by the flood are in four districts: Supaul, Madhepura, Saharsa and Khagaria. Supaul, the northernmost district in this region was flooded within hours of the breach. The villages along the new path of the Kosi were devastated, but the corridor within which these villages lie is fairly narrow because the waters were flowing rapidly at this point. Most of the lasting damage to land here was through the dumping of alluvial sand carried by the water. In contrast, when the waters reached villages in the districts of Madhepura and Saharsa they had slowed and spread over much larger areas and many of these continue to remain waterlogged. In Khagaria, there were two sources of flooding; the Kosi waters flooded the northern part of the district and the Ganga overflowed into the fields in the south about a week before the Kosi breach.

Given the small number of villages that we were able to study, we decided not to pick them randomly from the affected area. Instead, to examine the role of geography on household behaviour and flood damage, we chose roughly equidistant villages starting from the north-eastern corner of Supaul district. Our northernmost village is 22.5 kilometres from the point of the breach in Nepal. We will refer to these villages by numbers 1-10, with smaller numbers closer to the point of the breach.1

There is variation across the surveyed villages in both the source and the timing of the flood. Village 1 was flooded on the night of 18 August 2008 and villages 2-4 the following day. V illages 5-8 were affected at varying times between 21 and 25 August 2008. These villagers could have had some information about the flood although it appeared that there was considerable uncertainty about the course of the flood. Our last two villages provide an interesting comparison with the rest. Village 9 is l ocated next to the embankment of the Kosi and was flooded by 13 July, over a month before the breach occurred. Katghara, our southernmost village, is on the banks of the river Ganga and was flooded by 15 August, three days before the Kosi breach. Both these villages are frequently flooded and their inclusion allows us to study how households with some experience of floods b ehave differently from others. These areas were also excluded from the area formally defined as the Kosi disaster region, and, as seen in the data presented below, the state responded to the needs of these two villages very differently from the rest.

Characteristics of Surveyed Villages

Table 1 presents data from the 2001 Census of India on mean characteristics of the surveyed villages and other villages in the four districts. The data on population are from the village-level Primary Census Abstract and those on amenities are from village directories. The characteristics of our sample seem fairly representative of the region as a whole, although a sample of 10 villages is too small to make statistical comparisons.

Table 1: A Comparison of Surveyed and Other Villages Using Village Data from the 2001 Census

We conducted both a village and a household survey in each of the 10 villages in our sample. The purpose of the village survey was to record the occupational and caste structure of the village, the total number of migrants and fatalities, and changes in the pattern of land-use caused by the flood. It also enabled us to check that we had a reasonably representative sample of households. Village information was obtained from a group of households gathered in a public space in the village, often next to the primary school. They helped us map the village and estimate the number of households in each hamlet. We

percentage of families were reported to be beneficiaries of various state-sponsored poverty programmes, although many of these classifications have been shown to exhibit considerable e rror. We see that our second village, which has the highest proportion of uneducated families and agricultural labourers, has lower coverage of these programmes than other areas.

We were surprised to observe very little permanent migration following the flood, in spite of high rates of unemployment in these villages. Four of the 10 villages had no land under cultivation at the time of our survey, yet most households who had left at the time of the flood were back in the village. There were a total of 33 deaths reported as being caused by the flood and about half of these occurred in a single village.2 Only two people in these villages were reported to be missing.

Figures 2a-2c (p 55) examine changes in land use as a function of the distance of the village from the breach in Nepal. The figures are based on hamlet-level information and each hamlet is labelled by number of the village in which it lies.

Most of the land in the four northernmost villages is uncultivable because of large deposits of alluvial sand from the river bed. As we look further from the breach, we see some variation across hamlets within the same village in the proportion of land that is not in use. As we move from southwards, the problem of sand deposits is replaced by that of waterlogging. The fraction of w aterlogged farm land peaks at village 7 and then drops off rapidly. The very different landscapes that have resulted from the flood imply different types of adaptation strategies will be required in making the land productive again. There was, up until the time of our survey, no active state intervention along these lines.

We now turn to an analysis of our 280 household surveys which allow for a much more detailed study of how households were affected by the flood.

Household Responses

We have information on a total of 1,655 members of the 280 households we surveyed. We asked for information on all members that were part of the household either before or after the flood. In terms of religion and caste composition, 17% of the total households are Muslim and most of the others are Hindu castes classified as Other Backward Castes by the state and national

Location and Food Supply

We asked respondents for a list of all individuals who were either part of each household before the flood, or had joined the household after the flood. For each of these individuals in our list, we asked about their current whereabouts and obtained both current and retrospective information on their employment and earnings. We also asked households whether they moved out of the village on the day the flood waters reached their village and their location and source of food on each of the 20 subsequent days. Consistent with the findings from our village survey, we find stability in household composition. Of all 1,655 household members, 96% were present in the village both before and after the flood, 3% were there before but not after and under 1% moved into households after the flood. This was in spite of large-scale evacuation from the northern villages following the flood.

Table 3 provides information on household locations and food sources. We present summary statistics separately for the four northernmost villages, then the next four, and finally the last two, which were not officially part of the Kosi disaster. We find that within 20 days of the flood, half of the families in the worsthit areas were living and eating in relief camps and about 10% of families from the first eight villages who were residing in other locations received food from households and small organisations in nearby areas.

Table 3: Household Locations Following the Flood (Percentage of total households)

Employment and Earnings

Figures 3a and 3b (p 56) plot smoothed densities of days worked and average daily earnings for the month preceding our survey in February 2009 and for February 2008. These figures only include those with positive earnings in both time periods. We find that in spite of large employment effects, wages appear to be relatively inflexible. Median wages for those with positive earnings in both years are Rs 70. The big change is in median days worked which go from 22 days to eight days and one-third of those working in February 2008 were unemployed at the time of our survey.

Patterns in employment and earnings across villages are shown in Table 4 (p 56). Not surprisingly, the largest employment e ffects are in the four northern villages, where most land remains uncultivated.

To examine whether savings were used to smooth consumption, we asked households for the value of their bank and cash savings on 1 August 2008 and on the date of the survey. Only a small fraction of households had any bank savings – 12% in 2008

Figure 3a: Average Days Worked (2008 and 2009)

.06 .04 .02 0 Average days worked, Feb 2009 Average days worked, Feb 2008 Density

0 10 20 30 40 Occupation days February 2008

Figure 3b: Average Daily Earnings (2008 and 2009)

.015 .01 .005 0 Density Average daily earnings, Feb 2009 Average daily earnings, Feb 2008

0 200 400 600 800 Earnings per day, February 2008

and 10% in 2009. Most saving was in the form of cash at home which appears to have been used to cope with the large income shock. Of all households, 51% reported non-zero cash savings in 2008 and 24% in 2009.

Table 5 describes changes in selected assets between 2008 and 2009 by village categories. In the first four villages, the fraction of households with cash savings fell from 54% before the flood, to 8% six months after the flood. These families also began with lower total savings than those in other villages. For all villages taken together, mean bank savings between the two dates (for those with positive savings) went from Rs 28,000 to Rs 4,912 and cash savings declined from and average of Rs 4,462 to Rs 743.

Table 4: Employment Patterns Across Villages: February 2008 and 2009

We finally turn to the role of state-sponsored relief and development schemes in helping households adapt to the fall in their incomes and assets. We examine three different types of government interventions: employment through the NREGS, the distribution of staple grains through the public distribution system (PDS), and food and cash transfers through the emergency relief fund that was set up following the flood.

Table 6 (p 57) summarises our results on government relief and stabilisation programmes. We find that although 77% of households had heard of the NREGS, very few had benefited from it. We asked for total days worked under the programme before and after A ugust 2008. About 8% of the households in our sample had received benefits under the programme and on average, this group of beneficiaries had engaged in 21 days of work before A ugust 2008. Since August 2008, only four days of work were r eported in the entire area, two days each by two households of village number 2. Most likely, these were days before the occurrence of the flood. On the basis of our survey, it therefore appears that the NREGS was completely ineffective as a stabilisation scheme in providing work to the rural poor at a time when they had lost their usual source of employment.

We find that the PDS did provide rice and wheat at lower prices than the market, but only to a few families. During the seven days preceding the survey, 17% of households bought either rice or wheat from the PDS. From the table we see that the outreach of the PDS seems to be most limited in the villages where food scarcity was the greatest. For families that did have access to the PDS, the prices paid were about half of those prevailing in the open market.

Finally, we turn to government transfers in cash and kind. We listed all transfers received by the household over two time periods, July 2007 to June 2008 and July 2008 to the survey date. These transfers include pensions, relief payments and all other allowances. We find that while 20% of families received transfers

Table 5: Changes in Household Savings

in the first period, 80% received them after July 2008. In the year before the flood, transfers were most generous in villages 9 and 10, where 45% of families received some transfers. The modal transfer during 2007-08 was Rs 1,200 and this was received by about one-fifth of the families in villages 5-10, but by no family in villages 1-4.3 In the following six months, the largest transfers were received by households in the first four villages and a large fraction of families received exactly the same amount, Rs 5,840. The last two villages, which were not officially part of the Kosi disaster affected area, received no compensation and even previous transfers seem to have been suspended or delayed. Since our survey was completed in February, it is possible that these were subsequently paid out.

We can summarise our results from the Kosi survey as follows. In the immediate aftermath of the flood, households seemed r emarkably resilient – there were very few fatalities or missing people and most individuals returned to their villages after most of the relief camps closed in November 2008. The state relief a pparatus seemed effective, both in terms of getting large numbers into relief camps and ensuring that cash and grain transfers

august 1, 2009 vol xliv no 31

CLIMATE CHANGE NEGOTIATIONS

EPW

floods and droughts with new crops that are suited to their changed soil composition and rainfall patterns. It can also bring in new agricultural technologies that provide farmers with i ncentives to give up practices such as the burning of crop residues which aggravate the brown cloud problem.

In addition to the above approaches, are standard urban air pollution control policies. Emissions from transport and congestion can be reduced through investments in electric railways, public transport, and traffic control. Road pricing to deal with congestion externalities is now technically feasible and has been implemented in Singapore as well as London and Stockholm.

Finally, a word on the link between international and domestic policies. In a nutshell, the problem is that Indians in general and the poor in particular have a strong objective interest in getting the world to sharply reduce greenhouse gas emissions, but little ability to do so since we are responsible for only 3% of world emissions of carbon dioxide. This is made worse by the fact that the true scale of the problem is known to only to a handful of people in this country. As a consequence, Indian government policy on c limate change so far has not been motivated by domestic d emand, but by external pressure. This is deeply ironic since India is in much greater immediate danger from climate change than are the affluent countries that are pressuring it to act. The government

Notes

1 We first chose the village in Supaul that was c losest to the breach and yet accessible. We used district maps from the district census handbooks of these areas to select villages 2 through 10 to be 13 kilometres apart (as the crow flies) according to the scale given there. Subsequently, looking at digitised maps such as the one shown in Figure 1, these earlier distances do not appear to be a ccurate.

2 It was unclear why these deaths occurred. There appears to be have been some uncertainty about whether the village would be flooded, and the w ater reached late at night, making it harder for families to escape.

3 The two families receiving transfers in the July 2007-June 2008 period received different amounts. We have left the cells for the modal transfer and the percentage of households receiving this transfer blank.

References

Auffhammer, M, V Ramanathan et al (2006): “Integrated Model Shows That Atmospheric Brown Clouds and Greenhouse Gases Have Reduced Rice Harvests in India”, Proceedings of the National Academy of Sciences, 103 (52), 19668-672.

Baland, J-M, P Bardhan et al (2008): Forests to the People: Decentralisation and Forest Degradation in the Indian Himalayas (Delhi: Indian Statistical Institute).

Bergin, M H, R Greenwald et al (2001): “Influence of Aerosol Dry Deposition on Photosynthetically Active Radiation Available to Plants: A Case Study in the Yangtze Delta Region of China”, Geophysical Research Letters, 28 (18), 3605-608.

Chameides, W L, H Yu et al (1999): “Case Study of the Effects of Atmospheric Aerosols and Regional Haze on Agriculture: An Opportunity to Enhance Crop Yields in China through Emission Controls?”, Proceedings of the National Academy of Sciences, 96 (24), 13633-13636.

Cruz, R V, H Harasawa et al (2007): Asia. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, M L Parry, O F Canziani,

has shown little interest in operationalising its own National A ction Plan on Climate Change that was released a year ago.

5 Conclusions

We began by articulating the mechanisms that link climate change to rural livelihoods. Our study of the Kosi flood highlighted the manner in which natural disasters alter long-term

o pportunities for the poor. The state was impressive in providing immediate relief but has so far been completely ineffective in f acilitating long-term adaptation. This makes us sceptical of the extent to which the state will help poor rural populations deal with the effects of climate change.

The sluggish response of the Indian government to mitigating and adapting to climate change has been accompanied by the rhetoric of helplessness. State officials have often argued that I ndia’s share of emissions is too small for anything we do to m atter. We surveyed a well-respected body of research that e mphasises the need to distinguish between greenhouse gases and particle emissions and shows that the latter are a major source of climate-related problems. Since these emissions are d irectly a ffected by Indian policies and there are some low-cost options for particle emission reductions, it is important that these opportunities be seized.

Ramanathan, V and G Carmichael (2008): “Global and Regional Climate Changes Due to Black C arbon”, Nature Geoscience, 1 (4), 221-27.

Smith, K R (2000): “National Burden of Disease in I ndia from Indoor Air Pollution”, Proceedings of the National Academy of Sciences, 97 (24), 13286-293.

Smith, K R, R Uma et al (2000): “Greenhouse Implications of Household Stoves: An Analysis for India”, Annual Review of Energy and the Environment, 25 (1), 741-63.

Somanathan, E (2009): “What Do We Expect from an International Climate Agreement? A Perspective from a Low-income Country” in R N Stavins and J E Aldy Climate Change Policy Beyond Kyoto: Implementing Architectures for Agreement (Cambridge, UK: Cambridge University Press).

Venkataraman, C, G Habib et al (2005): “Residential Biofuels in South Asia: Carbonaceous Aerosol Emissions and Climate Impacts”, Science, 307 (5714), 1454-56.

– (2006): “Emissions from Open Biomass Burning in India: Integrating the Inventory Approach with High-resolution Moderate Resolution Imaging Spectroradiometer (MODIS) Active-fire and Land Cover Data”, Global Biogeochemical C ycles, 20, GB2013.

Permission for Reproduction of Articles Published in EPW No article published in EPW or part thereof should be reproduced in any form without prior permission of the author(s). A soft/hard copy of the author(s)’s approval should be sent to EPW. In cases where the email address of the author has not been published along with the articles EPW, can be contacted for help.

august 1, 2009 vol xliv no 31