Farmer Suicides in India Part 3: State-Level Distributions

June 17, 2012

farmersuicides_thumbnail.jpg

By The Sanhati Collective

In two previous articles (Part 1 and Part 2), we have highlighted the phenomenon of farmer suicides in India. Data collected by Sanhati from various issues of Accidental Deaths and Suicides in India (available from the website of the National Crime Records Bureau) show that between 1995 and 2010, more than 253,000 farmers have committed suicides across India. We have argued that the unprecedented wave of farmer suicides, which largely continues unabated as we write this note, is the result of acute agrarian distress sweeping across large swathes of rural India. The agrarian distress, in turn, is the result of the adoption of neo-liberal policies.

Our argument about the link between neoliberal policies, agrarian distress and farmer suicides was informed by analysis of both aggregate level policy changes and detailed case studies. In the first part of our series on farmer suicides, we have argued that, at the aggregate level, adoption of neoliberal policies relating to availability of institutional credit in rural areas, procurement of farm output, minimum support prices, import of agricultural products, declining subsidies for fertilizers, kerosene, electricity, etc., and neglect of irrigation facilities have led to declining profitability of agricultural production. Rising essential expenditures for health care, educations, transportation and food, have thus, pushed farmers increasingly into indebtedness. With the decline of institutional credit, this has meant that farmers would seek credit from the new moneylenders and the private micro finance institutions (MFIs), both of which charge very high interest rates and resort to regular intimidation for collection of payments. Rising indebtedness and mounting payment pressure makes farmers extremely vulnerable to shocks. A crop failure or a major sickness in the family or a life event can push already indebted farmers to the wall and induce them to take their lives. By analyzing two case studies – cotton production in the Vidarbha region of Maharashtra, and coffee & spice production in Wayanad district of Kerala – we complemented the aggregate level analysis with finer details of the matter.

Readers of Sanhati and activists from various regions of the country have asked us about the spread of farmer suicides across the Indian states. While we had indicated the distribution of various aspects of farmer suicides across major Indian states in our previous articles, both in the first and second parts, we would like to take this opportunity to do a more thorough state-level analysis. In this third article in our series on farmer suicides in India, we would like to address questions like the following: which states have witnessed the largest wave of farmer suicides? Which states have witnessed the highest intensity of suicides?

Why is a state-level analysis useful? In our opinion, it provides activists, academics and policy makers with valuable information about the geographical distribution of farmer suicides, both its absolute magnitudes and its intensity. Mass political activities, or informed pro-people policy interventions (if that is still possible in contemporary India), around this issue can use such data for strategic and tactical purposes. Since political, administrative and economic resources are limited, such data can be used to prioritize sorely needed actions.

To facilitate such processes, this article will use a series of charts to summarize our findings about the state-level distribution of farmer suicides in India. To construct these charts we use the previously collected farmer suicide data from the National Crime Records Bureau and population data – total state population, male population in states, population of male and female cultivators across states – from the Census of 2001. We hope to update this analysis as soon as relevant population data from the 2011 Census is released by the Registrar General of India.

In this article we will present rankings of states according to the following measures: (a) total number of farmer suicides, (b) total number of farmer suicides as a ratio of the total population of states, (c) suicide mortality rates (SMR) of farmers (i.e., total number of farmer suicides as a ratio of the total number of farmers in a state), (d) ratio of farmer to non-farmer suicide mortality rates. Each of these measures are useful in their own right, and as we explain below, throw light on some aspect of the problem of farmer suicides.

Our preferred measure is the fourth one: ratio of farmer to non-farmer suicide mortality rates. So, we will start with that. But before we present the data and the ranking of the states we would like to point to the most important aspect of this article: no matter which measure is used, a group of states always emerge at the “top” of the rankings. This makes the rankings we report in this article extremely robust.

This has two very significant implications. First, no matter how one twists and turns the evidence, it is difficult to escape the conclusion that acute agrarian distress is driving the unprecedented wave of farmer suicides in India since the mid-1990s (we will explain why the ratio of farmer and non-farmer SMR for any states is a good indicator of agrarian distress in that state). Second, the phenomenon is concentrated, in the main, in about 10 states, and hence, these states cry out for immediate and massive intervention.

Farmer versus Nonfarmer SMRs

One of the best ways of getting a quantitative handle on the phenomenon of farmer suicides and compare it across Indian states is to look at the ratio of farmer to nonfarmer SMR (i.e., farmer SMR divided by non farmer SMR). This ratio measures the difference in the suicide mortality rates between farmers and non farmers in each state. When this ratio is significantly larger than unity it implies that the SMR for farmers is larger than the corresponding SMR for non farmers. Thus, this measure gives a natural cut-off point that can allow us to link farmer suicides with agrarian distress. For states which record a farmer-nonfarmer SMR ratio that is much larger than unity one can plausibly make the argument that agrarian distress, which is specific to the production and distribution conditions in agriculture, is primarily responsible for the wave of farmer suicides in those states. This is because the ratio compares SMRs for two sub-populations – farmers and nonfarmers – within the same state.

Hence, for states which record a higher-than-unity value, this ratio captures factors specific to agricultural production that drive the higher suicides among farmers while taking account of state- specific factors that might affect both farmers and non-farmers.

For instance, it is well known that Kerala and West Bengal has had historically high incidence of suicide among all Indian states. The high incidence of suicide would impact both farmers and non- farmers in the state. But by taking the ratio of farmer to non-farmer SMR for each state, this measure adjusts for such state-specific variations in the incidence of suicide.

This measure has another built-in advantage. It takes account of the variation in the share of farmer population across states. Recall that this measure computes the ratio of farmer and nonfarmer SMRs in each state. Thus, when the farmer (and non-farmer) SMR is computed for each states, one is already taking account of the different proportions of farmer (and non-farmer) populations across states. This is important because there is enormous variation in the farmer-cultivator population across Indian states. In 2001, Kerala had total population of 31841374 and cultivator population of 724155; in the same year, Himachal Pradesh had total population of 6077990 and cultivator population of 1954870.

Thus, in 2001, Kerala had the lowest share of cultivator population among Indian states (at about 2.3 percent of the state population) and Himachal Pradesh had the highest share (at about 32 percent of the state population) with other states falling in between. Hence, it is important to take account of this large variation among states of the share of cultivator population [1].

Farmer suicides might vary from year to year because of idiosyncratic reasons which are not all necessarily related to agrarian distress. To remove such short run fluctuations (or year-specific factors) from the picture, we use a 5 year average of the measure around the year 2001, i.e., the measure is constructed as the ratio of (a) the average SMR for farmers over 1999-2003 (average farmer suicides over the period 1999-2003 from NCRB divided by the population of cultivators in 2001 from the Census) and (b) the average SMR for non-farmers over the same period (average non-farmer suicides over the period 1999-2003 from the NCRB divided by the population of non-cultivators in 2001 from the Census).

farmerfig1.jpg

Figure 1 presents rankings of states according to this measure averaged over the period 1999-2003. Since a value of the measure below unity implies the absence of higher-than-normal incidence of farmer suicide, we only rank states with a score above unity. Figure 1 shows that 11 states have a score above, and they are ranked, in decreasing order, as follows: Kerala (KR), Maharashtra (MH), Chhattisgarh (CG), Karnataka (KT), Andhra Pradesh (AP), West Bengal (WB), Madhya Pradesh (MP), Goa (GA), Tamil Nadu (TN), Uttar Pradesh (UP), and Gujarat (GJ).

Kerala is a clear outlier with a score which is more than 3 times the score for Maharashtra. This is partly driven by the fact that Kerala has a very small share of farmer-cultivator population (only about 2.3 percent) compared to other states. But the high score for Kerala means that the incidence of farmer- suicide (driven by agrarian distress) is very high, something that has been only obliquely highlighted by P. Sainath’s reporting, because Kerala has seen relatively high numbers of farmer suicides even when it has only a very small farmer-cultivator population.

The next group of states, MH, CG, KT, and AP, have scores ranging between 1.67 and 2.16. In these states, the farmer SMR was about twice as high as the non-farmer SMR, providing clear evidence of acute agrarian distress. The next group is formed by three states: WB, MP and GA. For these states, the score ranges between 1.30 and 1.38, which are all significantly higher than unity giving indication of agrarian distress. The last group is comprised of three states, TN, UP and GJ, whose scores are much closer to unity. Hence, the intensity of agrarian distress might be lower in these states.

A caveat is immediately in order. State-level comparisons, by construction, ignore variations within states. This might be a serious problem if there are significant district-level variations within states. For instance, a state which has a low ratio of farmer and non-farmer SMR might have districts with very high and others with relatively low ratio of farmer and non-farmer SMRs. By averaging out such district-level variations, state-level comparisons cannot take account of district-specific factors. While it is beyond the scope of this article to make district-level comparisons, it is important to keep this caveat in mind while interpreting the results reported here.

While we believe that the ratio of farmer and non-farmer SMRs is one of the best measures to compare the intensity of farmer suicides across states, activists and researchers have often used other measures too. Hence, we will now present rankings of states by three other commonly used measures: total number of farmer suicides, farmer suicides as a ratio of the state population, and farmer SMR. The important thing to note, as we have already pointed out, is that even though the rankings change a little across these alternative measures, the same set of states emerge at the top for all these measures. This means that the identification of the group of problem states (as far as farmer suicides is concerned) is robust and should be taken extremely seriously in policy debates and political work.

Total Number of Suicides across States: 1995-2010

The first alternative measure that we use is the absolute number of farmer suicides between 1995 and 2010. Figure 2 gives the ranking of the “top” 15 states according to the total number of farmer suicides reported between 1995 and 2010. The total number of suicides gives us a first quantitative approach to the phenomenon by answering the question: how large is the magnitude of farmer suicides across the various states? Just like the overall figure of 253785 farmer suicides for India convey the enormity of the problem we are dealing with, the overall numbers for states convey a similar sentiment.

farmerfig2.jpg

The states with the highest number of reported farmer suicides are, in decreasing order: Maharashtra (MH), Karnataka (KT), Andhra Pradesh (AP), Madhya Pradesh (MP), West Bengal (WB), Kerala (KR), Chhattisgarh (CG), Tamil Nadu (TN). With 50481 reported farmer suicides, Maharashtra stands in a league all by itself. Karnataka, Andhra Pradesh and Madhya Pradesh form a close second group with West Bengal, Kerala, Chhattisgarh and Tamil Nadu forming the third group to watch closely.

Total Number of Suicides: 2001-2010

Since some states were formed in the early 2000s, we also provide a ranking of states by the total number of suicides reported between 2001 and 2010. Figure 3 gives this ranking. Compared to Figure 2, there two significant changes in the rankings: (1) Chhattisgarh moves up from the 7th position to the 4th; (2) West Bengal and Kerala switch positions. This shows that the newly formed state of Chhattisgarh has forged ahead of many others in an ignominious way, and that Kerala has outstripped its sister state in more ways than one.

farmerfig3.jpg

Farmer Suicides for Every 1 Lakh Persons

It will be pointed out, and quite rightly we believe, that the rankings provided in Figure 2 and 3 need to be modified because different states are vastly different in terms of population. To understand the severity of the problem across states, the argument would go, we need to normalize the total number of farmer suicides by the size of the population. This is a valid point and so in Figure 4, we provide rankings of states by just such a measure: total number of farmer suicides between 1995 and 2010 divided by the state population in 2001. The year 2001 is used because it is a kind of “mid-point” for the period under study and thus provides a natural point to use for state population comparisons.

farmerfig4.jpg

The state with the highest number of farmer suicides for every 1 lakh persons is Sikkim. This result might be driven by the fact that the population of Sikkim is very small. Among the larger states, we observe the following ranking in terms of farmer suicides per 1 lakh population: CG, KT, KR, MH, AP, WB, TN.

Farmer Suicide Mortality Rates

We could refine the measure further. Since the share of farmers in the total population varies by states, it might be argued that we need to normalize by the farmer population rather than the total population. In fact, when we do so we get the suicide mortality rates (SMR) for farmers, a measure that is commonly used to measure mortality across different groups.

farmerfig5.jpg

In Figure 5, we present the ranking for the “top” 15 Indian states in terms of the farmer SMR in 2001. Kerala is a clear outlier with a whopping 143 suicides per 1 lakh farmers, more than 3 times the score for the next ranked state, Karnataka. The group of 4 states which follow Kerala, with farmer SMRs between 30 and 37, are (in decreasing order) KT, GA, CG and MH. The next group of 3 states have farmer SMRs close to 20: WB, TN, AP.

A similar ranking in terms of the SMRs for male farmers is presented in Figure 6, which displays similar patterns as the one represented by Figure 5. It is common for researchers and activists to separately look at the SMR for male farmers. Hence, we report this statistic here. But it must be remembered that there is wide variation in the share of female farmers among Indian states. For instance, in 2001, the share of female farmers varied from 17-18% in Kerala and West Bengal to above 50% in Arunachal Pradesh, Himachal Pradesh and Nagaland. Moreover, the agricultural workforce in India is getting increasingly feminized with time. According to a report by the NCEUS (National Commission for Enterprises in the Unorganised Sector), between 1993- 94 and 2004-05 the percentage of female farmers had grown from 34.1% to 36.1%. Hence, the traditional approach of focusing on male farmers might no longer be very informative; one needs to look at the farmer population as a whole.2

farmerfig6.jpg

Conclusion

In this article, we have presented four measures of farmer suicides across Indian states. We have used these measures to rank states to allow us to make inferences about the severity of the problem of farmer suicides across different states. We have presented four measures, instead of only one, because of two reasons.

First, different researchers and activists have often used different measures, thus making comparisons among their conclusions rather difficult. By presenting all the commonly used measures at one place, we hope to address problems of comparability. We think this will facilitate different activists and researchers to clearly see the source of their different conclusions.

Second, and more importantly, there is a group of states which emerge to the “top” of the ranked list no matter which measure we use. This group is comprised of the following states: Kerala, Maharashtra, Chattisgarh, Karnataka, Andhra Pradesh, West Bengal, Madhya Pradesh. Since all the measures that we have used are intuitive and throw light on some aspect of the problem of farmer suicides, the group of states that emerge at the top according to all of these measures clearly provide very strong evidence of acute agrarian distress and call for immediate and decisive intervention.

Notes

1.When we use the term “farmer” we mean to refer to what the Census of India calls “cultivators”, i.e., those who carry out agricultural production predominantly on their own land rather on other people’s land. Thus, the Census divides the agricultural workforce into “cultivators” (i.e., those who are self employed) and “agricultural labourers” (those who work on other people’s land), the latter including sharecroppers and wage labourers. We choose to use the population of cultivators, rather than the sum of cultivators and agricultural labourers, from the Census because the National Crime Records Bureau records suicides of self-employed farmers.

2. At this point the reader might raise the following valid question: how do we know that the farmer SMR is higher than the non-farmer SMR? If the two are similar in magnitude, there might not be any basis for arguing for the linkage between agrarian distress and farmer suicides. We agree. That is why our preferred measure is the ratio of farmer SMR to non-farmer SMR. When this ratio is higher than unity, it provides evidence of agrarian distress. This ratio is used for the rankings presented in Figure 1. As the value of this ratio shows, there are 11 states with a value larger than 1. Hence, data skepticism cannot be sustained.

2 Comments »

2 Responses to “Farmer Suicides in India Part 3: State-Level Distributions”

  1. kk Says:
    June 18th, 2012 at 7:32 pm

    Good work.

    Just a definitional quibble regarding your interpretation of the census definition of “cultivators” and “agricultural labourers”. I think it is confusing/wrong to reduce “cultivators” to “self employed” (c-m-c’), and include all sharecroppers within the category of “agricultural labourers”.

    The Census 2001 says: “For purposes of the census a person is classified as cultivator if he or she is engaged in cultivation of land owned or held from Government or held from private persons or institutions for payment in money, kind or share.” So this clearly includes even sharecroppers (excluding the marginal ones, who till others’ land for wages (either in cash or kind or share (deferred wage)). Further, “Cultivation includes effective supervision or direction in cultivation”. This will include both “self-employed” and employers – using the more apt Marxist terminologies, it includes rich, middle and a section of poor peasantry.

  2. Deepankar Basu Says:
    July 3rd, 2012 at 11:00 am

    @ kk

    Your point is well taken. The Census category of “cultivator” would include sharecroppers too, as you have rightly pointed out.

    Our motivation for using the number of “cultivators” as an approximation for the self-employed farmers come from the fact that the National Crime Records Bureau report farmer suicides only for self-employed farmers. The Census category of “cultivator” seems to be the best we could do to approximate the total number of self-employed farmers in a state. Hence we chose to normalize by that.

Leave a comment