Climatic factors affect many economically relevant outcomes – agricultural output, economic growth, health etc. A careful understanding of such effects is essential to the effective design of contemporary economic policies. Moreover, with global temperatures expected to rise substantially over the next century, understanding these relationships seem increasingly important.
The evidence on uneven distribution of the impact of climate change across developing and developed countries suggests that the former are more vulnerable. For poorer countries, economic development is considered to be the best insurance against climate change (Schelling, 1992). In this context, the role of development emerges in terms of its ability to augment the adaptive capacity of a country, and its population to withstand climatic impacts. However, economic growth and development could be climate sensitive. Both, the level of aggregate income and the rate of growth could differ across developed and developing countries, due to differences in their adaptive capacities. More specifically, a unit increase in temperature in the developing countries reduces aggregate income and economic growth to a greater extent, compared to their developed counterparts. Poorer countries, with their lower adaptive capacity, are more vulnerable.
The ability of governments of developing countries, in mounting effective public programmes to counteract adverse climate impacts is not clear. With poor public infrastructure and institutional frameworks many developing countries also lack the foundations to support strong markets. For example, the absence of enforced property rights will discourage many private actors (firms and households) in developing countries from adapting efficiently. All of these problems combine to suggest that developing countries will bear the brunt of climate change damage.
It has long been observed that hot countries are the ones which tend to be poor. Dell et al. (2009) explores the relationship between income, as well as its growth, and climate (temperature). They observed a significant amount of fall in income per capita (8.5%) with every degree Celsius (°C) rise in temperature. They also showed that with a 1°C rise in temperature a poor country’s growth reduces by 1.1%. Burke et al. (2015) looks at how temperature affects the economic production in wealthy and poor countries. They show that overall economic productivity is nonlinear with temperature for all countries, with productivity peaking at an annual average temperature of 13.6°C and declining strongly at higher temperatures.
These temperature effects on growth are large enough to produce a steeper relationship between temperature and income, in the absence of offsetting forces: if an extra 1°C reduces growth by 1.1 percentage points, then it would take only eight years of sustained temperature differences to explain the overall cross-sectional relationship between temperature and income observed in the world today. Despite this, there is considerable debate on whether climatic factors play any role in a country’s economic output, or is it a result of several policy variables (institutions or trade policy).
The broad objective of the present study is to examine the relationship between climate and economic production (more specifically agricultural output), at the regional level. We do this by examining whether climate has a significant influence on the GDP of various districts in India. Using a 13 year panel data for 640 districts in India, the study examines the short-run effects of climate on agricultural output between 2001 and 2013.
Evidence of climate change: India
There has been substantial evidence, both at the regional and global level, on climate change. A myriad of research models indicate rising trends in global temperatures. The first countries to be affected by climate change are the ones located close to the equator. Although these countries have historically experienced high temperatures, the climatic conditions are slated to worsen in the wake of global warming and climate change. The impact of these conditions on rainfall is, however, ambiguous.
With a predominantly sub-tropical humid climate India is also expected to witness considerable change in climatic conditions over the next few decades (conditional on several other factors as well). However, the fact that climate change is already underway is heavily debated. Whether or not long term climate change has started is a matter of further research and discussion (and somewhat less relevant here). But, the fact that over the last two decades India has experienced increase in temperatures is evident from Figure 1. We plot the annual average deviations of daily temperatures (mean, maximum and minimum), and aggregate rainfall from their long term average calculated over a period of 35 years (1981 – 2015). Although, changes in maximum temperatures and aggregate rainfall are ambiguous, the mean and minimum daily temperatures has been on the rise since the latter half of 1990s. This is consistent with our experience of increasing number of heat waves every summer and the rising toll due to extremely hot conditions.
The need to study impact of climate on income: Agricultural context
Despite technological advancements, such as improved seed varieties, genetically modified organisms, and irrigation systems, weather is still a key factor in agricultural productivity. Climate change induced by increasing greenhouse gases is likely to affect crops differently, from region to region. Decrease in potential yields is likely to be caused by shortening of the growing period, decrease in water availability, and higher frequency of extreme weather events.
Climate change and agriculture are interrelated processes, both of which take place on a global scale. Changes in climate directly affect agriculture in a number of ways including changes in pests and diseases, changes in atmospheric carbon dioxide and ground-level ozone concentrations.
With a predominantly sub-tropical humid climate India is also expected to witness considerable change in climatic conditions over the next few decades (conditional on several other factors as well).
Climatic conditions could affect agriculture in several other ways too:
- Productivity: in terms of quality and quantity of crops
- Agricultural practices: through changes of water use (irrigation) and agricultural inputs such as herbicides, insecticides and fertilisers
- Environmental effects: in particular in relation with frequency and intensity of soil drainage (leading to nitrogen leaching), soil erosion, reduction of crop diversity
- Rural space: through the loss and gain of cultivated lands, land speculation, land renunciation, and hydraulic amenities.
- Adaptation: organisms may become more or less competitive, as well as humans may develop urgency to develop more competitive organisms such as, flood resistant or salt resistant varieties of rice.
Needless to say that climate change has far reaching and devastating consequences in the food supply of developing economies. Consequently, it will increase the risk of food insecurity for the marginalised sections.
Agriculture typically plays a larger role in developing economies than in the developed world. In India the agricultural sector contributes roughly 20% of GDP and provides nearly 52% of employment (as compared to 1% of GDP and 2% of employment for the US). The majority of agricultural workers are drawn from poorer segments of the population (FAO 2006). Furthermore, it is reasonable to expect that farmers in developing countries may be less able to adapt to climate change due to credit constraints or lower access to adaptation technologies.
Climate change induced by increasing greenhouse gases is likely to affect crops differently, from region to region. Decrease in potential yields is likely to be caused by shortening of the growing period, decrease in water availability, and higher frequency of extreme weather events.
This implies that any changes in climatic conditions in the developing economies are expected to not only affect agriculture but also overall income. The analysis of climate impact on income is, therefore, imperative to policy making both at the regional and national levels. The very foundations of agricultural policies would fail if the government cannot recognise two things – that climate change is underway, and that it has far reaching consequences on the agricultural sectors in the developing economies.
Although the agricultural sector is slated to reduce in importance in the Indian economy as we perceive higher growth trajectories every year, its paramount importance in the food supply chain and hence its impact on key retail inflation parameters would still affect the major section of the population, and hence policy decisions for years to come. The combined factors of climate change (even if in the short run) and its impact on agriculture in the context of developing economies like India makes it necessary to quantitatively analyse this relation.
Data & Methodology
We used weather data for 640 districts in India, from the Indian Meteorological Department, between the years 1981-2015. It includes daily temperatures levels: mean, minimum and maximum. We also used estimated district level GDP data provided by a private research fi rm Indicus Analytics. The GDP data is available only from 2001 to 2013. This consists of overall GDP and several of its components (including agricultural GDP) at current and constant prices for every district in India.
As there can be large cross-sectional correlations between a country’s climate and its economic outcomes, it is difficult to distinguish the effects of the current climate from the many other characteristics potentially correlated with it. To overcome this difficulty we use a lagged dependent variable model with constant price GDP in the following specification:
Here, GDP refers to both agricultural and overall GDP at time period t/t-1 and in a given district i. The K climatic factors, on the other hand, are a range of climate metrics that we construct from the daily mean, minimum and maximum temperatures and daily rainfall. Consistent with literature we use higher order climate variables to observe non-linear effects of climate on GDP. The lagged variable GDPi, t-1 controls for all other annual variations that could have affected the GDP in time period . For the analysis we use a random effects model with a panel specification.
The climate metrics climate k,i,t are constructed from the given weather data. These variables indicate the changes in weather patterns, if any, at the district level besides accounting for adverse climatic conditions (increase in temperatures, shortage in rainfall, heat waves, dry spells etc.). Since our predominant interest is in analysing the impact of climate on agricultural GDP, we also include weather metrics for the monsoon months (June – October).
We study the impacts of climate variables on overall GDP and agricultural GDP separately. For the purpose of comparison, along with a lagged GDP variable, we also control for the aggregate annual rainfall for a given district, and then look at the effect of the temperature variables.
Effects on overall GDP
The annual average of daily mean temperature has a negative effect on the overall GDP, while the squared annual average of daily mean temperature has a positive effect. This implies that the mean daily temperatures have a negative effect on the GDP, and this effect increases as temperatures increase. So, a rise of mean daily temperatures from 30 to 35 degree Celsius would have a more negative impact on GDP than an increase from 20 to 25 degree Celsius. The annual average of daily maximum and minimum temperatures has the same effect as daily mean temperatures, with a negative effect on overall GDP which increases with increase in temperatures.
Agriculture typically plays a larger role in developing economies than in the developed world. In India the agricultural sector contributes roughly 20% of GDP and provides nearly 52% of employment (as compared to 1% of GDP and 2% of employment for the US). The majority of agricultural workers are drawn from poorer segments of the population (FAO 2006).
The average of daily minimum temperatures for the monsoon months, however, has a positive effect on GDP but concave in nature. This means that as minimum temperatures during the monsoon months increase the overall GDP increases, but at a decreasing rate. So, an increase of minimum temperatures from 20 to 25 degree Celsius will have a more positive effect on the GDP than an increase from 25 to 30 degree Celsius. All of these effects are significant at the 99% confidence interval. The results can be seen in Figure 2. In Figure 2 we plot the coefficient estimates of the temperature variables (and their corresponding squared terms in separate plots), with the 95% confidence intervals superimposed on them.
The decline in overall GDP due to increases in maximum and mean daily temperatures could be attributed to various factors – falling agricultural productivity due to destruction of crops, lower worker productivity due to unconducive conditions, etc. But given the overwhelming importance of agriculture in India, the reduction in its productivity is a major reason.
The effect of minimum temperatures, however, is the opposite. The exact reasons for this needs further analysis and research. Although, a sufficient increase in the range of temperatures could actually lead to a point of inflexion where the curvature changes entirely, and we see a negative effect of any increase in minimum temperatures thereafter.
Effects on agricultural GDP
The annual average of daily mean and maximum temperatures have a similar effect on agricultural GDP as the annual average variables had on overall GDP. The impact of these variables is negative on agricultural GDP, and this effect increases as the temperature values increase. The average of daily mean and minimum temperatures in monsoon, however, has a positive effect on agricultural GDP with a decreasing slope. So, as mean and minimum daily temperatures increase in the monsoon months agricultural productivity increases, but the rate at which it increases falls as the temperature values keeps increasing.
Figure 3 depicts these relations. Here as well, we plot the coefficient estimates of the climate variables (and their squared terms) with their respective 95% confidence intervals superimposed upon them.
The exact mechanisms through which the climate variables affect the agricultural output or the reasons behind a positive effect on agricultural productivity of a rise in mean and minimum temperatures is subject to further research. However, the negative effect of a rise in annual average mean temperatures on agricultural output is fairly clear. And given that mean temperatures are rising in India over the last two decades gives meaning to this relation.
We have already seen that daily mean and minimum temperatures in India are on the rise, at least over the last 20 years. And now, we have evidence of a negative effect of rising mean temperatures on both agricultural and overall GDP. These findings imply that if this climate change through temperature rise is a longer term change in India then the effect can have devastating consequences in the future, if climate change cannot be factored into policy decisions.
The random effects coefficient estimates for the effect of annual average mean temperatures on agricultural GDP is roughly -0.11 log points, which means that a one degree Celsius increase in daily mean temperatures would, on an average, decrease agricultural output by 11%. As a result of this the overall GDP would reduce by 28%. This is a rather alarming situation for all sections of the population – be it environmentalists worrying about long term climate changes being underway in India, or economic policy makers who are bent on reaching target growth levels of 8% per annum. If the changes in weather are ignored by policy makers the resulting fall in food supply due to lower agricultural output would lead to inflationary pressures, and wreak havoc in the Indian economy for years to come.
Under these circumstances the need of the hour is a sustainable development scheme, keeping in mind the environmental consequences of policy decisions. A thorough cost benefit analysis of every major real time investment in factories, mines, motorway and railroad expansion, which requires large swathes of land area (often leading to clearing forest space), is necessary. Despite the marginal short run negative impact on investment of such schemes, the economy would still perform better in the longer run, if no further weather changes occur.
The global community also needs to come forward and analyse the consequences of climate change. The Paris climate change conference was a good start, but for effective outcomes each and every economy needs to move forward in coalition. A greater awareness of the precise impacts of climate change on economic outcomes serves a better purpose in convincing various nations to adopt policies in tune with their impacts on climate. Hence, the analysis of climate change impact on GDP (and other economic outcomes) seems to be of paramount importance.
The analysis of climate impact on income is, therefore, imperative to policy making both at the regional and national levels. The very foundations of agricultural policies would fail if the government cannot recognise two things – that climate change is underway, and that it has far reaching consequences on the agricultural sectors in the developing economies.
The objective of this short discourse on climate change impact on GDP was to explore possible relations between the two. In the process we observed evidence of definite climate change that has happened in the last 20 years in India. We do not know whether these changes are going to be sustained over a longer period of time or not. But the negative impact of even the slightest increase in temperatures, on national income, indicate the necessity of providing sufficient importance to even minor changes in weather patterns. The ability of weather changes in undermining economic growth is fairly clear from the empirical analysis of the relation between the two. Therefore, policy makers need to understand these effects precisely.
Our analysis shows that overall increase in annual average temperatures have a negative effect on income, and this effect is increasing in temperatures. However, any increase in monsoon temperatures seem to enhance production. We also have evidence of rising temperatures in India. Under the light of these facts, it is logical to conclude that the growth potential of India might not be completely realised owing to weather changes. There might be other factors too but weather change is definitely one of them, and an important one. Although, further analysis can reveal more pressing facts, the preliminary results are substantial enough to ask for immediate action.
Burke, M., Hsiang, S. M. & Miguel, E. (2015). Global Non-Linear Effect of Temperature on Economic Production. Nature, 527, pp. 235 – 239.
Dell, M., Jones, B. F., &Olken, B. A. (2012). Temperature Shocks and Economic Growth: Evidence from the Last Half Century. American Economic Journal: Macroeconmics, Vol. 4(3), pp. 66 – 95.
Schelling, T. C. (1992). Some Economics of Global Warming. American Economic Review, Vol. 82 (1), pp. 1 – 14.