Spurious Linkages between Extreme Temperatures and Farmer Suicides

A recent paper, published by the PNAS (Proceedings of the National Academy Sciences of the United States) and authored by Tamma A. Carleton, titled “Climate Change and Agricultural Suicides in India” claims that “temperature during India’s main agricultural growing season has a strong positive effect on annual suicide rates.” Using state-level data for 1967 to 2013, the author suggests that an increase in 1°C temperature in a single day can cause 70 suicides. It also claimed that the evidence leads to the conclusion that it is the damage to crops by extreme temperatures that leads to economic hardship and suicide.

Regrettably, the paper has received widespread uncritical coverage in the Indian media.

We consider these claims to be baseless. These claims are a consequence of the uncritical use of data, bad assumptions, flawed analysis and unacceptable neglect of the existing literature on global warming and Indian agriculture as well as farmer suicides. Taking the conclusions of the paper at face value would lead, we strongly believe, to dangerously incorrect policy measures. Such conclusions also divert from the study of the real challenges that global warming, and extreme temperatures in particular, poses for Indian agriculture.

The paper is marked by several serious errors. The paper:

• Incorrectly uses suicide data
• Wrongly identifies extreme temperatures for crop production,
• Wrongly identifies only kharif as the relevant agricultural season in which to consider extreme temperatures in, and
• Wrongly identifies the relevant crops.

As a result, the meaning of the correlation that the author claims to find between extreme temperatures and suicides is unclear. The manner in which the paper analyses the link between extreme temperatures and crop production is wrong.

The signatories to this press note have themselves conducted a detailed study of the impact of extreme temperatures on crop production in Karnataka, one among several such studies conducted by other responsible Indian and foreign authors. No such study provides any corroborative evidence for the dramatic conclusions of this paper.

The authors of this press note have also submitted a formal comment on the paper to the PNAS, where the original paper is published.

Uncritical use of data

The paper uses state-level data on suicides, data that includes both urban and rural suicides. How can urban suicides be included in an analysis of agricultural suicides? The paper also sets aside the fact that the suicide data, taken from the National Crime Records Bureau, has separated farmer suicides from those of other occupational categories only after 1995 and the inconsistency in data prior to that year. Suicide data are gathered from police records so there is likely to be underreporting.

Bad Assumptions

The paper, in attempting to show that temperature increase affects agricultural yields does not directly examine physical yield but only considers their monetary value, based on 1960-65 prices. It is widely recognized in the climate change literature, that the impact of extreme temperatures on crops and their economic consequences should not be confused with each other. It is well-established practice to consider physical yield as the first direct impact of increased temperatures and lowered monetary income a consequence that is also affected by a host of other economic and policy factors.

Further, and more damagingly, the author does not analyse individual crops but only considers a basket of such crops including rice, wheat, sorghum, sugar, maize and millet. Cotton, closely associated to farmer suicides wherever it is grown, is a notable omission as are a host of other cash crops.

The author, surprisingly and wrongly, considers extreme temperatures only during the kharif season. She completely ignores the rabi season, despite clear evidence to the contrary from ICAR and other authoritative research from India and across the world. Rabi crops like wheat are in fact most sensitive to extreme temperatures as has been well-established by research. Despite the fact that the predominant wheat-growing season is rabi, the paper includes wheat in its analysis but does not include the impact of rabi temperatures.

The paper also considers temperatures above 20 degrees Celsius as extreme temperatures. This is flatly contradicted by what is known of the temperature dependence of crop production. In general, temperature ranges of 20 degrees to 29 degrees Celsius are known to be beneficial for crop growth. Every crop has a specific temperature threshold, ranging from 33°C to 38°C, above which a negative impact on yields is possible, and extreme temperatures in the literature refer to temperatures above these corresponding thresholds.

Flawed analysis

As a consequence of the errors explained above, the results on the negative effect of temperature on crop yields are not tenable. Of the six crops pooled, rice is mainly a monsoon crop, wheat is a winter crop, and sugarcane is a 12 to 18 month crop. How can the July-September “growing season” temperature explain changes in the combined yields of these crops?

The author finds a strong positive coefficient of proportionality when aggregate State level deaths from suicide are related to the total heat exposure in the kharif growing season alone. The quality of data on suicides is in doubt, and the definitions of temperature threshold and growing season are incorrect. The author refers to robustness checks of her results but all these checks retain the erroneous assumptions listed above. Without further study, it is not clear how we interpret the observed coefficient that she claims to find.

We are astonished to find such a poor analysis based on incorrect understanding of data and wrong analysis should find its way into policy discourse on climate change and agriculture. There is much excellent literature being produced on this important subject. The dissemination of such important and scientifically valid research is hampered by the publication of flawed papers such as the one in question.

The text of this blog is from the press note circulated by the authors.