Gist of Economy Survey 2017-18 [Chapter 6]: Climate, Climate Change, and Agriculture

INTRODUCTION

• Prime Minister’s goal of doubling farmers’ incomes increasingly runs up against the contemporary realities of Indian agriculture, and the harsher prospects of its vulnerability to long-term climate change.
• The last few seasons have witnessed a problem of plenty: farm revenues declining for a number of crops despite increasing production and market prices falling below the Minimum Support Price (MSP).
• But in the medium to long term, the ghost of Malthus looms over Indian agriculture. Productivity will have to be increased, and price and income volatility reduced, against the backdrop of increasing resource constraints.
• Impact on agriculture by:
  • shortages of water and land,
  • deterioration in soil quality,
  • climate change-induced temperature increases and rainfall variability.

Why Agriculture Matters: An Irony

• Historical and cultural reasons:
  • Agriculture matters in India for deep reasons, not least because the farmer holds a special place in Indian hearts and minds.
  • The first salvo of satyagraha was fired by Mahatma Gandhi on behalf of farmers, the indigo farmers exploited by colonial rule.
  • Like in early, Jeffersonian America, history and literature have contributed to the farmer acquiring mythic status in Indian lore: innocent, unsullied, hard-working, in harmony with nature; and yet poor, vulnerable, and the victim, first of the imperial masters and then of indigenous landlords and middlemen.
  • Bollywood (and Kollywood and Tollywood) has also played a key role in creating and reinforcing the mythology of the Indian farmer.
• Economic reasons:
  • Agriculture also matters for economic reasons because it still accounts for a substantial part of GDP (16 percent) and employment (49 percent).
  • Poor agricultural performance can lead to inflation, farmer distress and unrest, and larger political and social disaffection—all of which can hold back the economy.
• Why people should move out of agriculture:
  • Economic reasons:
    • Economic development is considered always and everywhere about getting people out of agriculture and of agriculture becoming over time a less important part of the economy (not in absolute terms but as a share of GDP and employment).
    • The reason why agriculture cannot be the dominant, permanent source of livelihood is its productivity level, and hence the living standards it sustains, can never approach those in manufacturing and services.
    • That, of course, means that industrialization and urbanization must provide those higher productivity alternatives to agriculture. But this must happen along with, and in the context of, rapid productivity growth in agriculture, to produce greater food supplies for the people, provide rising farm incomes, and permit the accumulation of human capital.
  • Non-economic reasons:
    • Dr. Ambedkar warned about the dangers of romanticizing rural India.
    • He famously derided the village as “a sink of localism, a den of ignorance, narrow mindedness and communalism,” thereby expressing a deeper truth—an Indian social complement to the Lewisian economic insight—that in the long run people need to move and be moved out of agriculture for non-economic reasons.
• So the irony is that the concern about farmers and agriculture today is to ensure that tomorrow there are fewer farmers and farms but more productive ones.

Long run agricultural performance

• Agricultural growth:
  • Real agricultural growth since 1960 has averaged about 2.8% in India.
    • The period before the Green Revolution saw growth of less than 2%;
    • the following period until 2004 yielded growth of 3%;
    • in the period after the global agricultural commodity surge, growth increased to 3.6%.
  • China’s annual agricultural growth over the long run has exceeded that of India by a substantial 1.5 percentage points on average.
• Volatility of agriculture:
  • The volatility of agricultural growth in India has declined substantially over time. In particular, production of cereals has become more robust to drought.
  • But levels of volatility continue to be high and substantially higher than in China where the ups and downs have been virtually eliminated.
  • An important contributing factor is that agriculture in India continues to be vulnerable to the vagaries of weather because 52% (73.2 million hectares area of 141.4 million hectares net sown area) of it is still unirrigated and rainfed.
• Three objectives of this chapter:
  • to document the changes in climactic patterns in temperature and rainfall over the past six decades,
  • to estimate the effects of fluctuations in weather on agricultural productivity,
  • to use these short-run estimates in conjunction with predicted changes in climate over the long-run to arrive at estimates of the impact of global warming on Indian agriculture.

Motivation

• Why study data when there already is a burgeoning and serious body of research and analysis at the international level of the impact of climate on economic activity. The answer is threefold:
  • There is the standard worry that cross-country analysis might not apply to large, individual countries such as India, which is agrarian and is home to a great diversity of climate zones.
  • An India-specific analysis would be more granular, done at a spatially more disaggregated level than coarser cross-country analysis.
  • Third and important reason—with implications for research findings and hence policy input—has to do with data quality.
    • Nearly all the available international cross-country analysis use cross-country databases on temperature, weather, and extreme events. These databases rely on Indian data but with far fewer actual measurement points (“stations”) than available with the Indian Meteorological Department (IMD). For example, Delaware temperature data base is based on 45 weather stations in India whereas the IMD data is gridded from 210 weather stations
    • There are substantial differences in both levels and trends between the two datasets.
      • For example, IMD data record much higher average levels of temperatures than the Delaware dataset (by over 1 degree Celsius on average, in climate terms, the difference between disaster and nirvana).
      • Similarly, the IMD data shows higher levels of precipitation of about 100 millimetres on average (again a potential difference between deluge and drought) with a sharply declining trend since the 1970s unlike the Delaware data.
      • These differences suggest that any analysis of long run climate impacts could be very different across these datasets.
• Thus, armed with high quality, high resolution, temperature and precipitation data, this chapter proceeds to analyze patterns in temperature and precipitation in India, and the impact they have on agricultural productivity.

TEMPORAL AND SPATIAL PATTERNS OF TEMPERATURE AND PRECIPITATION

• Temperature pattern:
  • The broad pattern of rising temperatures post 1970s is common to both seasons, rabi and kharif.
  • The average increase in temperature between the most recent decade and the 1970s is about 0.45 degrees and 0.63 degrees in the kharif and rabi seasons, respectively.
  • There has been a rise in the number of days with extremely high temperatures, and a corresponding decline in the number of days with low temperatures.
• Rainfall pattern:
  • Between the 1970s and the last decade, kharif rainfall has declined on average by 26 millimeters and rabi rainfall by 33 millimeters.
  • Annual average rainfall for this period has on average declined by about 86 millimeters.
  • The proportion of dry days (rainfall less than 0.1 mm per day), as well as wet days (rainfall greater than 80 mm per day) has increased steadily over time.
• Thus, the imprint of climate change is clearly manifest in the increasing frequency of extreme weather outcomes.
• The spatial dimensions of changes in weather:
  • Temperature:
    • Temperature increases have been particularly felt in the North-East, Kerala, Tamil Nadu, Kerala, Rajasthan and Gujarat.
    • Parts of India, for example, Punjab, Odisha and Uttar Pradesh have been the least affected.
  • Rainfall:
    • Extreme deficiencies in rainfall are more concentrated in Uttar Pradesh, North-East, and Kerala, Chattisgarh and Jharkhand.
    • There has actually been an increase in precipitation in Gujarat and Odisha and also Andhra Pradesh.
  • Spatially temperature increases and rainfall declines seem to be weakly correlated.

IMPACT OF WEATHER ON AGRICULTURAL PRODUCTIVITY

• Estimating the impact of temperature and climate on agriculture has become an increasing focus of economic research. Many of the concerns relate to developing countries because climate impacts seem to be present disproportionately, in hotter and less rich parts of the world.
• This chapter uses disaggregated data at the district level—on temperature, weather, and crop production, yields, and prices —to answer a number of important questions.

Stark heterogeneity: Extreme versus Moderate shocks; Irrigated versus Unirrigated Areas

The present analysis yields two key findings.

1. The impact of temperature and rainfall is highly non-linear and felt almost only when temperature increases and rainfall shortfalls are extreme.
   1. A large literature focuses on the impact of a one-unit increase in temperature and a one unit decrease in rainfall on agricultural yields. The analysis in the Indian context, such marginal changes in weather have little or no impact, and that the adverse effects of weather are concentrated in the extremes.
   2. These findings have important implications for the impact of climate change on agriculture, since most climate change models predict an increase in extreme weather events.
2. These extreme shocks have highly divergent effects between unirrigated and irrigated areas almost twice as high in the former compared with the latter.
   1. Unirrigated areas – defined as districts where less than 50 percent of cropped area is irrigated — bear the brunt of the vagaries of weather.
      1. For example, an extreme temperature shock in unirrigated areas reduces yields by 7 percent for kharif and 7.6 percent for rabi.
      2. Similarly, the effects of extreme rainfall shocks are 14.7 percent and 8.6 percent (for kharif and rabi, respectively) in unirrigated areas, much larger than the effects these shocks have in irrigated districts.
   2. The literature suggests that several factors over and above the level of rainfall matter for agricultural yields.
      1. The number of dry days (defined as days during the monsoon with rainfall less than 0.1 millimetres) exerts a significant negative influence on productivity: holding the amount of rainfall constant, each additional dry day during the monsoon reduces yields by 0.2 percent on average and by 0.3 percent in unirrigated areas.

Crop impacts

• Crops grown in rainfed areas— pulses in both kharif and rabi—are vulnerable to weather shocks while the cereals—both rice and wheat—are relatively more immune.
• Have the impacts changed over time?
  • In the last decade for which data is available (2004-2014), the impact of rainfall shocks in yields remains unchanged, but the effect of temperature shock increases threefold (relative to the first decade).

IMPACT ON FARM REVENUE

• Extreme temperature shocks reduce farmer incomes by 4.3% and 4.1% during kharif and rabi respectively, whereas extreme rainfall shocks reduce incomes by 13.7% and 5.5%.
• These average effects mask significant heterogeneity, with the largest adverse effects of weather shocks being felt in unirrigated areas.
• Which direction farm revenues should move in –
  • • on the one hand, these shocks reduce yields,
    • but on the other, the lower supply should increase local prices.
  • The results here clearly indicate that the “supply shock” dominates – reductions in yields lead to reduced revenues.
• Other studies:
  • Existing studies for India typically analyse the impact of weather shocks on the productivity of individual crops.
  • For example, Swaminathan shows that 1 degree Celsius increase in temperature reduces wheat production by 4 to 5%.
  • Studies found similar effects for 11 African countries – a one degree increase in temperature reduces revenues by 6% on average.
  • A study by the IMF finds that for emerging market economies a 1 degree Celsius increase in temperature would reduce agricultural growth by 1.7%, and a 100 millimetres reduction in rain would reduce growth by 0.35 percent.
• Impact of climate change on agriculture performance in the long run:
  • Impact of temperature increase:
    • Climate change models as one developed by the Intergovernmental Panel on Climate Change (IPCC), predict that temperatures in India are likely to rise by 3-4 degree Celsius by the end of the 21st century.
    • These predictions combined with studies imply that in the absence of any adaptation by farmers and any changes in policy (such as irrigation), farm incomes will be lower by around 12% on an average in the coming years.
    • Unirrigated areas will be the most severely affected, with potential losses amounting to 18% of annual revenue.
  • Impact of rainfall decline:
    • Climate change models do not have unambiguous predictions on precipitation patterns.
    • But if the observed decline in precipitation over the last three decades (of over 86 millimetres) is applied to the estimates, it is found that in unirrigated areas, farm incomes will decline by 12% for kharif crops, and 5.4% for rabi crops.
  • Impact of increase in number of dry days:
    • Models of climate change predict an increase in the variability of rainfall in the long-run, with a simultaneous increase in both the number of dry-days as well as days of very high rainfall.
    • If the observed increase in the number of dry days over the past 4 decades are applied to the short-run estimates, this channel alone would imply a decrease in farm incomes by 1.2 percent.
• Overall the analysis suggests at least three main channels through which climate change would impact farm incomes:
  • an increase in average temperatures,
  • a decline in average rainfall and
  • an increase in the number of dry-days.
• All three are likely to be correlated, and therefore the total impact of climate change will not be the simple sum of these individual effects. The three effects that are identified in could be mildly additive.
• Taking these correlations into account, farmer income losses from climate change could be between 15 percent and 18 percent on average, rising to anywhere between 20 percent and 25 percent in unirrigated areas.
• The results in this chapter stand in contrast with similar studies both globally and in India. For example:
  • Deschenes and Greenstone, find mild and even positive effects of climate change on agricultural profits in the United States.
  • Kumar et al find that rice yields in unirrigated areas will only marginally be affected in the long run. Their estimates are based on climate change models that predict an increase in the average amount of rainfall.
• It is possible that these estimates overstate the true impact of climate change:
  • The estimates in this chapter are derived using short-run variations in weather, and farmers may not be able to adapt to such fluctuations in the short-run.
  • In the long-run, however, they may be able to change technologies or alter the crops they grow in response to sustained increases in temperature and changes in precipitation.
  • Further it is possible that irrigation networks might expand, mitigating to some extent the adverse impacts of climate change.

CONCLUSIONS AND POLICY IMPLICATIONS

• Based on newly compiled weather data and a methodology that has not been applied to Indian data so far, this chapter estimated the impact of temperature and precipitation on agriculture. The main findings are as follows:
  • A key finding—and one with significant implications as climate change looms—is that the impact of temperature and rainfall is felt only in the extreme; that is, when temperatures are much higher, rainfall significantly lower, and the number of “dry days” greater, than normal.
  • A second key finding is that these impacts are significantly more adverse in unirrigated areas (and hence rainfed crops such as pulses) compared to irrigated areas (and hence crops such as cereals).
  • Applying IPCC-predicted temperatures and projecting India’s recent trends in precipitation, and assuming no policy responses, give rise to estimates for farm income losses of 15 percent to 18 percent on average, rising to 20 percent-25 percent for unirrigated areas.
• The policy implications are stark. India needs to spread irrigation – and do so against a backdrop of rising water scarcity and depleting groundwater resources especially in North India.
  • In the 1960s, less than 20 percent of agriculture was irrigated; today this number is in the mid-40s.
  • The Indo-Gangetic plain, and parts of Gujarat and Madhya Pradesh are well irrigated.
  • But parts of Karnataka, Maharashtra, Madhya Pradesh, Rajasthan, Chattisgarh and Jharkhand are still extremely vulnerable to climate change on account of not being well irrigated.
  • India pumps more than twice as much groundwater as China or United States.
  • There is 13 percent decline in the water table over the past 30 years.
• Solution:
  • Technologies of drip irrigation, sprinklers, and water management—captured in the “more crop for every drop” campaign—may well hold the key to future Indian agriculture and hence should be accorded greater priority in resource allocation.
  • The power subsidy needs to be replaced by direct benefit transfers so that power use can be fully costed and water conservation furthered.
  • Agricultural research:
    • There is need to embrace agricultural science and technology with renewed ardor.
    • Swaminathan urged that anticipatory research be undertaken to preempt the adverse impact of a rise in mean temperature.
    • Agricultural research will be vital in increasing yields but also in increasing reliance to all the pathologies that climate change threatens to bring in its wake: extreme heat and precipitation, pests, and crop disease.
    • The analysis shows that research will be especially important for crops such as pulses and soyabean that are most vulnerable to weather and climate.
  • Insurance:
    • Climate change will increase farmer uncertainty, necessitating effective insurance.
    • Building on the current crop insurance program (Pradhan Mantri Fasal Bima Yojana), weather-based models and technology (drones for example) need to be used to determine losses and compensate farmers within weeks (Kenya does it in a few days).
• In thinking about agricultural policy reforms, it is vital to make a clear distinction between two agricultures in India:
  • There is an agriculture—the well-irrigated, input-addled, and price-and-procurement-supported cereals grown in Northern India—where the challenge is for policy to change the form of the very generous support from prices and subsidies to less damaging support in the form of direct benefit transfers.
  • Then there is another agriculture (broadly, non-cereals in central, western and southern India) where the problems are very different:
    • • inadequate irrigation,
      • continued rain dependence,
      • ineffective procurement,
      • insufficient investments in research and technology (non-cereals such as pulses, soyabeans, and cotton),
      • high market barriers,
      • weak post-harvest infrastructure (fruits and vegetables),
      • challenging non-economic policy (livestock).
• How this will happen given that agriculture is a state subject is an open political economy question:
  • The Hirschmanian bottom-up forces of “voice” and “exit” along with benevolent and-strategic top-down planning and reforms will all have to play a key part.
  • The cooperative federalism “technology” of the GST Council that brings together the Center and States could be promisingly deployed to further agricultural reforms and durably raise farmers’ incomes.

Click here to download Gist of Economic Survey 2017-18- Chapter 6