This Note examines changes in the cropping pattern of Kerala over the last 61 years. Specifically, it puts together a district-level database on agriculture in Kerala from 1956–57, when the State was formed, to 2016–17, the most recent year for which data are available. The available data were apportioned to nine composite districts according to the district boundaries of 1957–58 in order to construct a new dataset. The special features of this Note are the construction of continuous databases on land-use pattern and area and production statistics for Kerala for a 61-year period at the district level.

The first section is a discussion of the data sources; the second section discusses long-term trends in land utilisation; and the third section provides a description of changes in the area under crops, in production, and in the yield of some crops at the State and district levels and examines the diversification of cropping pattern using the Hirschman-Herfindahl index at the State level. The fourth and final section has concluding observations.

Introduction

Diversity of plant wealth and a rich marine ecosystem distinguish Kerala from other States of India. The crops grown here include various seasonal crops such as rice and vegetables, annual crops such as tapioca and banana, and perennial crops such as pepper and rubber. The cultivation of high-value plantation crops makes Kerala the State with the highest per hectare crop income in the country. In 2013–14, the value of agricultural output per hectare in Kerala was Rs 1,19,556.¹ From 1962–65 to 2005–8, almost all districts of the State consistently stood first among all districts in India in terms of crop output per hectare (Bhalla and Singh 2012).²

The State of Kerala has seen significant transformation of its agriculture over the last 61 years. The process of land reforms that was initiated by the first Government of Kerala in 1957 and completed by the end of the 1970s led to the conferring of ownership rights over agricultural land to 1.2 million tenant cultivators (Radhakrishnan 1981). From the 1970s, the time of the oil boom in the Gulf countries, increased migration and remittances from abroad have resulted in a significant transformation of Kerala’s economy. Specifically, high land prices have altered the land-use pattern in the State (Zachariah and Rajan 2004). These changes have affected the cropping pattern as well.

Different kinds of periodisation have been followed in studies of Kerala’s agricultural growth over the past 61 years. In their analysis of agricultural development in Kerala, Kannan and Pushpangadan (1988; 1990) divided the period from 1962–63 to 1985–86 into two distinct phases. An overall increase in agricultural growth (area, production, and yield) occurred in the first period, that is, between 1962–63 and 1974–75.³ The second period, from 1974–75 to 1985–86, was marked by a deceleration in cropping intensity and a shift in the cropping pattern in favour of plantation crops. In India as a whole, the period of the Green Revolution (1962–65 to 1980–83) was marked by an increase in production and yield, but there was no such corresponding phase in Kerala (Bhalla and Singh 2012).

The deceleration in cropping intensity of the period 1974–75 to 1985–86 was reversed in the next ten years, between 1984–85 and 1996–97, when output from the agricultural sector grew at about 5 per cent annually (Kannan 2011). Growth in rubber production contributed in a major way to agricultural growth in this period. After the mid-1990s, there occurred a period of stagnation with an overall absolute decline in gross cropped area (GCA) and in the area under major crops. Viswanathan (2014) analysed agricultural statistics between 1995–96 and 2009–10, and concluded that there was a deceleration in growth of area and production along with stagnant yield levels for most crops during these years.

This Note summarises the changes in the land-use pattern and cropping pattern of Kerala over the last 61 years, using agricultural statistics released by the Department of Economics and Statistics (DES). Though many earlier studies have analysed the cropping pattern of the State (Kannan and Pushpangadan 1990; Karunakaran 2014; Viswanathan 2014; Harilal and Eswaran 2017), a time-series for the entire period at the level of districts has not been put together before. This Note examines the problems in creating such a database, and then reviews the land-use and cropping pattern changes at the district level in the State. It does not investigate the reasons for the observed changes, but results from some previous studies on this subject are briefly discussed.

Sources of Data

Preparation of Agricultural Statistics in Kerala

As Kerala is constituted of regions where land revenues were permanently settled (or a non-land records State), the collection of land-use statistics here has been different from that in other States of India, where a patwari or a revenue official would maintain records of land use at the village level. Bakshi and Ramachandran (2008) illustrate the differences in the collection of land-use statistics between a State with land records, namely Tamil Nadu, and a non-land records State, West Bengal. In the case of West Bengal, data on crop area, irrigation, and land use are collected by different agencies including the Department of Land and Land Reforms and the Bureau of Applied Economics and Statistics. In Tamil Nadu, data on crop area, irrigation, and land use are maintained in the village records, where the cadastral number of the plot is recorded.

The case of Kerala is similar to that of West Bengal. Land-use and crop-area statistics in Kerala were assigned as tasks of the Bureau of Economics and Statistics after the formation of the State. The first Season and Crop Report, 1956–57 published by the Bureau presented an overview of agriculture in the State, and gave statistics on land utilisation, area, and production of major agricultural crops. Sample surveys of land plots were undertaken to generate estimates of land area under a nine-fold classification and under major agricultural crops at the district level.⁴ This report also had figures on livestock, the export of major crops, and provided descriptions of major agricultural crops.

In 1975–76, a centrally sponsored scheme titled “Establishment of an Agency for Reporting Agricultural/Crop Statistics” (EARAS/EARCS) was launched with the intention of covering the entire geographical area of Kerala in five years, by covering 20 per cent of the total area every year (GoK 2017a). The National Commission on Agriculture, 1976 recommended introducing a system of complete enumeration of land records in the permanently settled States. However, a complete enumeration has not been implemented. Presently, the EARAS wing of the Department of Economics and Statistics (DES) brings out agricultural statistics for every year by means of sample surveys.

Prior to the introduction of the EARAS, land utilisation surveys would provide data on area under different types of land use and crops. Nair (1983) has noted that these surveys provided estimates of the area under cultivation of major crops at the district level and of minor crops at the State level to a “reasonable level of precision.” At present, after the introduction of the EARAS, the area under each crop and types of land use are estimated using area enumeration surveys, and the yield for each crop is calculated from crop estimation surveys. A detailed note on the estimation process followed by the DES is given in KSPB (2015).

Concerns Regarding the DES Data

The major concerns regarding the data collected by the DES are as follows: (1) comparability across time periods because of changes in administrative boundaries; (2) problems with reported forest area in land-use statistics; (3) change in the methodology of estimating the extent of crop area; (4) difference between estimates provided by commodity boards and the DES; (5) lack of data pertaining to some crops, such as vegetables and tubers; and (6) deficiencies in capturing changes in homestead farming systems.

The first difficulty mentioned above, namely comparisons across time, is due to the fact that the number of districts in Kerala has increased from seven in 1956–57 to 14 at present. The focus of the DES has been on preparation of agricultural statistics for the preceding agricultural year without taking into account changes in district boundaries that could provide a comparable time-series. Moreover, inadequate storage of data collected from investigator zones makes it impossible to generate accurate estimates that would adhere to the changes in district boundaries at a later date.

Secondly, the land-use category of forests reported by the DES pertains to the official data according to revenue records. The reported geographical area net of forest area is surveyed by the DES. One criticism of official land-use statistics is the discrepancy between forest area as reported in the official statistics and forest area on the ground. Recent studies based on satellite images have shown that the area under forests has declined substantially in Kerala, and only about 19 per cent of the State’s forest cover was found to be open (Kumar 2005). A considerable amount of forest land has been converted to crop land, but these changes are not reflected in the official statistics on land-use pattern.

Thirdly, certain revisions in the method of calculating the extent of particular crops make it difficult to put together a comparable dataset across time-periods. Some earlier studies have not used data prior to 1960, citing comparability issues (Kannan and Pushpangadan 1988). For tapioca, for instance, one study uses data from 1963–64, as statistics based on crop-cutting experiments are available from that period (Pushpangadan 1988). However, documents published by the DES use area and production data from years prior to 1960, along with area and production data after 1960. More recently, changes in the figures for stand per hectare, used to enumerate the area under certain crops, have resulted in drastic variations in the area under cultivation.⁵ For example, from 2011–12, the stand per hectare for pepper has been calculated as 1,112, as against 560 before 2011–12. There is no source of information on the exhaustive set of changes introduced till date.

Fourthly, estimates for the area under cultivation and production of plantation crops – cardamom, tea, coffee, and rubber – are provided by the respective commodity boards (Spices Board, Tea Board, Coffee Board, and Rubber Board). These boards conduct independent sample surveys to arrive at estimates that are reported by the DES in its annual statistics. Only in the case of coconut, the area and production estimates are brought out by the DES and not the Coconut Development Board (CDB). However, the CDB undertakes surveys to arrive at coconut production estimates for Kerala. The sampling methodology adopted by it is different from that of the DES. Instead of identifying clusters using plots selected from the Basic Tax Register, the CDB selects districts and blocks on the basis of the extent of area under cultivation, and panchayats on a random basis within the selected blocks. From each panchayat, 10 holdings with a minimum of 20 fruit-bearing palms are selected for estimating yield. The CDB’s estimated production of coconut was 6,661 million nuts in 2012–13, which was 15 per cent more than the DES estimate.⁶

Fifthly, the number of crops reported by the DES has changed over the years, but vegetables and tuber crops (except for tapioca), which are viewed as minor crops, have been completely excluded from production estimates. As no crop-cutting experiments are conducted for minor crops, the quantity of production of these crops is not available.

Lastly, the DES methodology does not capture dynamic changes within homesteads. This limitation is particularly seen in the case of vegetables, which are extensively cultivated in homesteads in Kerala. A study by the Kerala State Planning Board indicates that the estimates provided by the Department of Agriculture are higher by 119 per cent than the area under cultivation of vegetables reported by the DES in 2014–15 (KSPB 2015). There are three major limitations of the DES methodology used to provide area estimates for vegetables. (1) The area under vegetables taken together may be significant in a sample site, but the area under individual vegetables may be recorded as zero as it may not be significant. (2) Short-duration vegetables (planted and harvested within six months) are excluded in dryland areas, as DES investigators visit dryland clusters only twice a year. (3) The stand per hectare used to estimate the area under perennial vegetable crops (for example, drumstick) may lead to overestimating or underestimating the area under cultivation because the actual crop density may be different.

An alternative technique for estimating crop area is remote sensing. The Central Tuber Crops Research Institute (CTCRI) used remote sensing to estimate the acreage of tapioca in two districts of Tamil Nadu in 2016–17 (CTCRI 2017). Another study used mixed methods to study cropping pattern, and to highlight the issues involved in using only remote sensing in the context of Kerala (Fox et al. 2017). Kerala is characterised by an increasing spread of agroforests (homesteads and plantations), and while remote sensing data exhibit a constant aerial extent for agroforests, surveys of households indicated that agroforests were actually declining. However, data from remote sensing may be useful in estimating the actual extent of forest cover, for which the DES has so far relied solely on revenue records.

Time-Series Data

Creating a comparable data series for the period 1956–57 to 2016–17 involved adjusting for administrative and other methodological changes during this period.⁷

Different agencies have compiled information on agricultural statistics in the past, based on annual reports published by the DES. A compilation brought out by the Kerala State Planning Board, Human Development Data Series – 2: Agriculture (KSPB 2009), has been used extensively for this study. Annual reports released by the DES have been used for data from 2008–9 to 2016–17. A list of sources used to create the time-series is provided in Table 1.

KSPB (2009) checked the data for errors in totalling and tabulation during the compilation exercise. However, in all the listed sources, the reported data pertained to the then existing district boundaries, which have since undergone changes. In order to make districts comparable over time, the methodology of the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) was followed: that is, to fix district boundaries in 1966 and apportion the data from later districts accordingly (ICRISAT 2015). The data for area and production were apportioned in the ratio of geographical areas carved out from older districts.

I followed this method, and used the district boundaries of 1957–58 as fixed for the entire time-series. The area and production data for each district were aggregated into nine composite districts (districts that were later divided to form 14 districts). This was used to analyse changes at the level of the district. Figure 1 shows the present district boundaries and the boundaries of the nine composite districts.

Figure 1 District boundaries of Kerala, 1961 and 2011

Source: Kerala Administrative Atlas, Census of India, 2011.

Table 2 shows the method by which the area and production of different crops at the district level were apportioned to districts according to the 1957–58 boundaries.

In this way, I constructed a database consisting of land-use statistics and crop-area statistics (for 15 crops at the State level and four crops at the district level). Unless otherwise specified, all tables and figures are from this compiled dataset. I have also used the Malayalam names for the districts of 1957–58, and, unless otherwise specified, the composite districts of 1957–58 for district-level analysis.

Trends in Land-Use Pattern

Land use in India follows a nine-fold categorisation. Here I use only four categories, namely, net sown area, total fallow land (including current fallow, which is cultivable land that was not cultivated in the previous year), area sown more than once, and area used for non-agricultural purposes.

While the actual geographical area of Kerala remained the same from 1956–57 to 2016–17, the estimated total geographical area of Kerala (reported geographical area) according to village records changed three times: in 1957–58, 1975–76, and 2005–6. As the reported geographical area has changed thrice from 1956–57 to 2016–17, only percentage shares of different types of land use in the total geographical area are examined here. For the land-use category of area sown more than once, its share in net sown area rather than its share in total geographical area is represented and analysed here.

Figure 2 plots the changes in net sown area (NSA) for Kerala as a whole, and for the nine composite districts. Net sown area as a percentage of the total geographical area of Kerala grew from 48 per cent in 1956–57 to 57 per cent in 1974–75. It then remained stagnant till 1999–2000, when it reached 58 per cent. From this level, the NSA fell to 52 per cent of total geographical area in 2016–17. From 1956–57 to 2016–17, the share of NSA in geographical area increased by 4 per cent for the State as a whole, but only three composite districts showed an increase in the share of NSA. These were the northern districts of Kannur and Kozhikode, and the southern district of Kottayam, where the share of NSA in geographical area increased by 27, 9, and 13 per cent respectively. The largest decline was in Alappuzha. The share of NSA in geographical area was very high for Alappuzha in the early years, ranging from 85 per cent of total geographical area in 1956–57 to 90 per cent in 1972–73. The share of NSA in geographical area declined by 32 per cent in Alappuzha from 1956–57 to 2016–17. The second highest decline after Alappuzha was in Ernakulam, where the share of NSA in geographical area declined by 11 per cent during this period. Overall, the decline in six districts was negated by an increase in the share of NSA in Kannur, Kozhikode, and Kottayam, which resulted in an overall increase in the share of NSA in geographical area from 1956–57 to 2016–17 for the whole of Kerala.

Figure 2 Share of net sown area in total geographical area, 1956–57 to 2016–17 in per cent

Notes: 1. The Y-axis for each graph is the percentage of total geographical area.

2. The value for 1956–57 is reported as it is, the value for 1957–58 is the average of the two preceding years, and values from 1958–59 are the average of three preceding years.

Large-scale migration of cultivators from the erstwhile Travancore region (particularly from the composite district of Kottayam) to hilly tracts (particularly present-day Idukki district, which is also part of Kottayam composite district) and to the erstwhile Malabar region (particularly present-day Wayanad district, which is a part of the composite districts of Kozhikode and Kannur) resulted in major changes in land-use pattern (George and Chattopadhyay 2001; Joseph 2002).⁸ This migration started in the 1920s on account of increased population pressure on land and a depression in the prices of agricultural commodities (Tharakan 1984). It continued into the 1970s, until most of the wasteland in the erstwhile Malabar region was brought under cultivation. The extension of cultivation continued during the period of land reforms, when migrants who were tenants in the erstwhile Malabar region became owners of their leased-in land (Joseph 2002). These processes show up clearly in the compiled dataset: there was an increase in net sown area in the composite districts of Kannur and Kozhikode, which were a part of the erstwhile Malabar region, and in the composite district of Kottayam, which contains the hilly tracts of present-day Idukki district, in the years 1956–57 to 1975–76.

A distinct feature of the changes in land use in Kerala is the expansion of non-agricultural area in both absolute and relative terms. Figure 3 shows the increase in non-agricultural area for Kerala as a whole and for the nine composite districts. From 1956–57 to 2016–17, the area under non-agricultural use increased for Kerala as a whole, and for eight of the nine composite districts. Palakkad was the only district to show a small decline of 1 per cent of area under non-agricultural use as a share of geographical area. The area under non-agricultural use as a share of geographical area in Kerala was 11 per cent in 2016–17, up from 5 per cent in 1956–57. A reduction in barren and uncultivable land was seen in all the districts. In Kerala as a whole, the share of barren and uncultivable land fell from 5 per cent of geographical area in 1956–57 to a negligible share of geographical area in 2016–17.

Figure 3 Share of non-agricultural area in total geographical area, 1956–57 to 2016–17 in per cent

Notes: 1. The Y-axis for each graph is the percentage of total geographical area.

2. See note 2 for Figure 2.

One unexplained feature of the compiled database, as is evident from Figure 3, is the sudden jumps in the estimates of non-agricultural area for Kerala as a whole and for the composite districts in 1975–76 and 2005–6. There was a change in the reported geographical area of Kerala in 1975–76 when the EARAS scheme was begun. The absolute area of land under non-agricultural use in Kerala in 1975–76 was 12 per cent lower than the absolute area of land under non-agricultural use in the previous year. In 2005–6, the reported geographical area of Kerala changed again. The absolute area of land under non-agricultural use in 2005–6 was 14 per cent lower than the absolute area of land under non-agricultural use in 2004–5.

However, the area under non-agricultural use increased consistently even after these changes. The average annual rate of growth (simple average of growth rates) of absolute area under non-agricultural use from 2006–7 to 2016–17 was 2 per cent. Among the districts, Alappuzha and Thiruvananthapuram had the highest share of non-agricultural land in 2016–17, with 15 per cent of geographical area classified as non-agricultural land. The highest growth in this period of area under non-agricultural use was in Kozhikode and Thiruvananthapuram, at an annual growth rate of 3 per cent.

Migration and remittances have been discussed as a major driver of changes in land use in Kerala (Prakash 1998; Zachariah and Rajan 2004). The absolute number and proportion of migrants in the total population grew steadily from the 1970s onwards. The population of migrants increased by 76 per cent, from around 1.4 million in 1998 to around 2.4 million in 2014 (Zachariah and Rajan 2016). A major effect of remittances from abroad has been large-scale speculative investment in land and an escalation of land prices (Prakash 1998).

Total fallow land is the extent of agricultural land left uncultivated in the survey year (current fallow) or prior to that (other fallow). According to the DES, land is considered fallow if it is not cultivated for a period less than five years. Figure 4 shows that total fallow land as a share of total geographical area in Kerala was 4 per cent in 1956–57 and 3 per cent in 2016–17. While total fallow land declined steadily up to 1969–70, it increased from 1969–70 to 2016–17.

Figure 4 Total fallow land (current fallow and other fallow) as a share of total geographical area, 1956–57 to 2016–17 in per cent

Notes: 1. The Y-axis for each graph is the percentage of total geographical area.

2. See note 2 for Figure 2.

Fallow land as a share of geographical area remained stagnant or increased in all districts except Alappuzha from 2000–1 to 2016–17. Alappuzha had the highest share of fallow land (8 per cent of total geographical area) in 2000–1. This share fell steadily up to 2016–17, when the area left fallow was only 3 per cent of total geographical area. However, the reduction in fallow land did not result in a corresponding increase in the share of net sown area (NSA) or non-agricultural land. Instead, the land-use category of cultivable waste showed a higher-than-average increase during this period. The share of cultivable waste in Alappuzha increased from 2 per cent to 8 per cent of geographical area from 2000–1 to 2016–17. This was not observed in any other composite district.

Kannur and Kozhikode districts showed the highest increase in the share of NSA in geographical area, and also the biggest decline in the share of fallow land in geographical area. The share of fallow land in geographical area decreased by 6 per cent in Kannur and by 2 per cent in Kozhikode from 1956–57 to 2016–17. In Thrissur, the share of fallow land in geographical area increased from 1 per cent in 1956–57 to 5 per cent in 2016–17.

Area sown more than once is a measure of the intensity of cultivation. Area sown more than once in Kerala increased from 19 per cent of net sown area in 1956–57 to 28 per cent in 2016–17, with a peak in 2006–7 when 40 per cent of net sown area was area sown more than once (Figure 5). It is not clear what resulted in this increase in 2006–7. However, policies promoting rice cultivation were initiated in 2005–6 (Thomas 2011). These policies combined with an expansion in the cultivation of seasonal vegetables may have led to a peak in area sown more than once in 2006–7. An earlier peak in area sown more than once was in 1974–75 (37 per cent), coinciding with an expansion in the area under cultivation of rice.

Figure 5 Share of area sown more than once in net sown area, 1956–57 to 2016–17 in per cent

Notes: 1. The Y-axis for each graph is the percentage of net sown area.

2. See note 2 for Figure 2.

The biggest expansion in area sown more than once occurred in Kozhikode, Kottayam, and Palakkad districts. Area sown more than once as a share of NSA increased by 28 per cent in Kozhikode, 25 per cent in Kottayam, and 12 per cent in Palakkad, from 1956–57 to 2016–17. While Kozhikode and Kottayam saw the biggest expansion in NSA, Palakkad had the largest share of land under rice cultivation in Kerala. There was a sharp decline in area sown more than once as a share of NSA from 1956–57 to 2016–17 in Thrissur and Thiruvananthapuram districts, with a decline of 11 and 10 per cent, respectively.

To sum up, the initial years after the formation of Kerala, from 1956–57 to 1975–76, saw an increase in net sown area and area sown more than once, and a reduction in total fallow land, mainly on account of an expansion of cultivation in the composite districts of northern Kerala. The period after 1975–76 was marked by a stagnation of net sown area and area sown more than once, and a decline in more recent years. Land under non-agricultural use increased steadily throughout this period. Migration has been a particularly important factor in driving the underlying changes in land use in Kerala.

Trends in Cropping Pattern

State-Level Cropping Pattern

The Department of Economics and Statistics (DES), Kerala, classifies agricultural crops into food crops and non-food crops (GoK 2017a). Food crops are further classified into foodgrains (including cereals, millets, and pulses), sugar crops, spices and condiments, fruit (including fresh fruit and dry fruit), tapioca and tuber crops, and vegetables. The category of non-food crops consists of oil seeds, fibre drugs and narcotics, plantation crops, and other non-food crops (such as fodder grass, green manure crops, other crops and trees, teak, and medicinal plants). I have followed the same classification in this Note.

Though the database was generated using published reports, there were minor differences between total cropped area as reported in land-use statistics and total cropped area calculated as the sum of area under cultivation of all crops. Minor discrepancies were found in 19 of the 61 years taken up for analysis. In 1961–62, 1964–65, and 1965–66, the difference amounted to 1 per cent of the total cropped area reported by land-use statistics. For these years, I have used the total cropped area from the sum of area under food crops and non-food crops.

The most conspicuous change in the cropping pattern of Kerala from 1956–57 to 2016–17 has been the shift from food crops to non-food crops. Figure 6 shows the area under food crops and non-food crops in hectares. Three distinct phases emerge from the graph: (1) from 1956–57 to 1974–75, the area under food crops as well as non-food crops increased; (2) from 1975–76 to 1994–95, the area under food crops declined, and the area under non-food crops rose to surpass the area under food crops; and (3) from 1995–96 to 2016–17, the area under food crops continued to decline and the area under non-food crops remained stagnant.

Figure 6 Area under food crops and non-food crops, 1956–57 to 2016–17 in hectares

Note: The Y-axis is area in hectares. Ha = hectares.

The three phases of changes in cropping pattern coincided broadly with changes in net sown area (except for the change in periodisation in the third period, which starts in 2001 for change in net sown area). While the first period saw an extension of cultivation in the Malabar region, which led to an increase in the area under food crops and non-food crops, price movements explain the subsequent changes in cropping pattern. Rice registered the lowest increase in average farm prices among major agricultural commodities in Kerala from the 1970s to 1996 (George and Chattopadhyay 2001). There was a shift away from cultivation of rice during this period to more remunerative crops, such as coconut. Due to increased trade liberalisation in the third phase, price variability among commodities such as rubber, cocoa, and spices increased when the introduction of new multilateral trade agreements reduced the State’s ability to safeguard cultivating farmers using quantitative restrictions (Anoopkumar 2012; Harilal and Dhanya 2015). This is a likely factor in the rapid decline in area under food crops and stagnation in non-food crops from 1995–96 to 2016–17.

A longer period of decline in the area under food crops – starting from 1975–76 and still continuing – rules out the possibility of the change being only on account of changes in sampling methodology. A change in sampling methodology was reported for pepper and cashew in 2011, which led to large changes in the area under these two food crops in 2011–12. This affected the overall area under food crops and the total cropped area for 2011–12. While similar changes may have occurred in earlier periods also (methodological changes prior to 2011 are not public), the likelihood of these changes having affected long-term trends is low.

Fifteen crops, including 10 food crops and five non-food crops, were studied at the State level to illustrate broad changes in the cropping pattern. The food crops included one foodgrain (rice), one sugar crop (sugarcane), five spices and condiments (pepper, ginger, turmeric, cardamom, and areca nut), one tuber crop (tapioca), and two fruits (banana and plantain, and cashew). The non-food crops were two oilseeds (sesamum and coconut), and three plantation crops (tea, coffee, and rubber). The analysis was restricted to these 15 crops as data for all the years were available for them. These 15 crops accounted for 85 per cent of the total cropped area in 1956–57 and 80 per cent of the total cropped area in 2016–17.

Figure 7 plots the Hirschman–Herfindahl index (H), a measure of crop concentration, for the last 61 years.

Figure 7 Hirschman–Herfindahl Index, Kerala, 1956–57 to 2016–17

The Hirschman–Herfindahl index for a particular year is defined as:

where A_i is the area of crop i,

A is the total area of all crops, and

n is the total number of crops.

Higher values of H indicate that the area under cultivation is concentrated in fewer crops, and lower values indicate crop diversification or that the crop area is distributed among a greater number of crops.

The H value for Kerala was 0.25 in 1956–57 and 0.23 in 2016–17. However, there was a clear period of declining H from 1956–57 to 1986–87, followed by a period of rising H from 1986–87 to 2016–17. The lowest value of H, at 0.19, was in 1986–87, the year of highest crop diversification. The three crops with the highest shares in total cropped area in 1956–57 were rice (35 per cent), coconut (21 per cent), and tapioca (10 per cent). In 2016–17, only two crops held a share of more than 10 per cent of total cropped area: coconut with 30 per cent and rubber with 21 per cent.

This observation is based on changes in the area under cultivation for 15 selected crops. The 10 food crops selected for analysis comprised 83 per cent of the area under all food crops in 1956–57. While the share of these crops increased to 87 per cent of area under all food crops in 1974–75, it declined steadily after that, and in 2016–17, the area under cultivation of these 10 food crops constituted 66 per cent of area under all food crops. In comparison, the share of area under cultivation of the five selected non-food crops in total area under all non-food crops remained at around 90 per cent from 1956–57 to 2016–17. In other words, the importance of the 10 selected food crops declined over time, and other food crops such as spices and vegetables that were excluded from the analysis gained prominence between 1956–57 and 2016–17.

The analysis of changes in cropping pattern at the State level based on 15 crops suffers from two major limitations: (1) absence of crops that would have increased or decreased in importance during the period; (2) change in the area of some crops at the local level (district or sub-district level), which may not be evident in an aggregated analysis.

The area under cash crops, such as nutmeg, vanilla, cocoa, and lemongrass, is not a part of the database as data for these crops were not available for the entire time-period. The DES classifies nutmeg and vanilla as food crops (spices and condiments), and cocoa and lemongrass as non-food crops (plantation crop, and fibre and narcotic crop, respectively). The area statistics for these crops are available for a shorter period of time. The estimated area under cultivation of nutmeg in 1977–78, the year when cultivation was first reported separately, was 2,978 hectares. This grew to 22,675 hectares in 2016–17, or 2 per cent of the total area under food crops. Vanilla was first reported as a separate crop in 2005–6, when it was cultivated across 3,273 hectares in Kerala. Though this was a negligible share of the total area under food crops, there was a drastic reduction in the area under vanilla, with only 120 hectares under cultivation in 2016–17. The area under cocoa was reported as 6,057 hectares in 1977–78, the year of its first reporting, and 14,404 hectares in 2016–17. The area under lemongrass cultivation stood at 2,332 hectares in 1976–77 and grew to 7,762 hectares in 1984–85. However, the area under lemongrass cultivation in 2016–17 was reported as 143 hectares.

These changes, though not significant at the State level, bring out some micro-level changes in cropping pattern. Settler farmers of the Malabar region (northern districts) cultivated lemongrass and cocoa in the 1960s and 1980s but abandoned these crops following a price crash (Joseph 2002). There was a shift from pepper and cardamom to nutmeg and vanilla after prices of the latter increased in the early 2000s, however, the price of vanilla crashed after the mid-2000s and farmers shifted back to other spices (Thomas 2009). Price responsiveness of area is high for commercial crops.

The database does not account for changes in the area under cultivation of vegetables, which are important crops in Kerala. KSPB (2015) states that production of vegetables in the State increased from about 825,000 tonnes in 2011–12 to about 1,354,569 tonnes in 2014–15.⁹ The rise of development programmes for vegetables initiated by the State, micro-irrigation facilities, and introduction of micro-nutrient inputs helped in increasing the yield and production of vegetables. However, even if vegetables are added to the database, they may not reflect the actual trends in cropping pattern due to issues in the sampling methodology followed by the DES.

In spite of these limitations, the analysis of cropping pattern showed the following trends. The area under cultivation of rice was around 30 per cent of total cropped area from 1956–57 to 1975–76, but fell to 7 per cent of total cropped area in 2016–17. The area under tapioca fell from 10 per cent of total cropped area in 1956–57 to 3 per cent in 2016–17. From 21 per cent of total cropped area in 1956–57, coconut increased its share to 30 per cent of total cropped area in 2016–17. In terms of share of total cropped area, rubber moved from 4 per cent of total cropped area in 1956–57 to 21 per cent in 2016–17. These four crops show the largest variation in area under cultivation in Kerala over the last 61 years.

District-Level Cropping Pattern

Changes in area, production, and yield at the level of composite districts were analysed for four important crops: rice, tapioca, coconut, and rubber.

Kerala is characterised by three elevation zones. Lowlands that run along the coastline, known for rice and coconut cultivation; the midland region consisting of valleys and slopes, which supports different types of seasonal and perennial crops; and highlands characterised by the cultivation of plantation crops (George and Chattopadhyay 2001). All three elevation zones are present in all the composite districts of Kerala, although the share of each zone varies across the districts. Based on elevation and other specific biophysical characteristics, 13 different agro-ecological zones have been identified for Kerala.¹⁰

Here, I have analysed the data for area under cultivation of four crops at the block level for the years 2010–11 to 2015–16.¹¹ The blocks are classified into one among the thirteen agro-ecological zones in Kerala. I have used these data to identify prominent agro-ecological zones of cultivation of individual crops that are discussed here.

Rice

The most prominent change in cropping pattern over the last 61 years pertains to rice cultivation in Kerala. The area and production of rice increased during the initial years after the formation of the State and during the period of land reforms. Figure 8 shows that for all-Kerala and for five districts – Kozhikode, Palakkad, Thrissur, Ernakulam, and Alappuzha – the area under rice cultivation showed positive growth from 1956–57 to 1975–76. The southern districts of Thiruvananthapuram and Kollam had a distinctly lower area under cultivation of rice, and showed no growth in area from 1956–57 to 1975–76. For all districts and for all-Kerala, this period of growth between 1956–57 and 1975–76 was followed by a decline in the area of rice, which has continued up to 2016–17.

Figure 8 Area under cultivation of rice in Kerala and its nine composite districts, 1956–57 to 2016–17 in hectares

Notes: 1. The Y-axis for each graph is the area under cultivation in hectares.

2. See note 2 for Figure 2.

From 1956–57 to 2016–17, Palakkad remained the composite district with the highest extent of area under rice cultivation and the highest level of production. But even in Palakkad, there has been a decline in area under cultivation and in production of rice. From 2010–11 to 2015–16, within Palakkad, the highest average area under rice cultivation was in the agro-ecological zones of Palakkad plains and Chittoor black soil type. Along with these, the Kuttanad agro-ecological zone in Alappuzha and Kottayam, and the central midlands agro-ecological zone in Palakkad and Thrissur also had a high share of rice cultivation.

The production of rice did not show a sharp decline immediately after 1975–76 (Figure 9). The total production of rice declined in absolute terms only from 1984–85. The production of rice in Kerala in 1956–57 was 887,170 tonnes and in 2016–17, 436,483 tonnes. Today, the production of rice is concentrated in Palakkad and Alappuzha districts, which accounted for 34 per cent and 24 per cent, respectively, of total rice production in Kerala in 2016–17.

Figure 9 Production of rice in Kerala and its nine composite districts, 1956–57 to 2016–17 in tonnes

Notes: 1. The Y-axis for each graph is production in tonnes.

2. See note 2 for Figure 2.

The yield growth too has been slow in this period. The yield of rice in Kerala increased from 1,146 kg per hectare in 1956–57 to 2,547 kg per hectare in 2016–17, a growth rate of 1 per cent per annum. In 2014–15, the yield of rice in Kerala was 2,837 kg per hectare. The corresponding figures for Tamil Nadu and Punjab were 3,191 kg per hectare and 3,838 kg per hectare, respectively (GoI 2017a).

The increase in rice output from 1956–57 to 1974–75 was partly due to high prices resulting from a local shortage of rice (Panikkar 1980). Surveys suggest that the decreasing profitability of rice cultivation, especially on account of the increase in wage rates, is one reason for cultivators shifting away from rice after 1976–77 (GoK 1999). One study of factors that led to changes in the acreage allocation of rice revealed that though the prices of rice and coconut had an inelastic effect on area under cultivation of rice, both factors were statistically significant from 1960–61 to 1983–84 (George and Mukherjee 1986).

Tapioca

The area under tapioca cultivation has closely followed the pattern of rice (Figure 10). Tapioca cultivation peaked in 1975–76 and has declined since. The largest producer of tapioca in Kerala is Kollam district, which saw the highest increase in area under cultivation and production from 1956–57 to 1975–76. The southern midlands zone of Kollam and Thiruvananthapuram accounted for most of the area under cultivation of tapioca in these two decades.

Figure 10 Area under cultivation of tapioca in Kerala and its nine composite districts, 1956–57 to 2016–17 in hectares

Notes: 1. The Y-axis for each graph is the area under cultivation in hectares.

2. See note 2 for Figure 2.

Tapioca cultivation is concentrated in the southern districts of Kottayam, Kollam, and Thiruvananthapuram (Figure 11). The production of this crop has fallen relative to peak levels in the mid-1970s, with the largest decline in the composite district of Kollam, where it fell from a peak of 2,013,331 tonnes in 1972–73 to 629,211 tonnes in 2016–17. Yields of tapioca increased from 6,948 kg per hectare in 1956–57 to 36,842 kg per hectare in 2016–17. This increase in yield, however, did not result in a corresponding increase in production, as the area under cultivation fell.

Figure 11 Production of tapioca in Kerala and its nine composite districts, 1956–57 to 2016–17 in tonnes

Notes: 1. The Y-axis for each graph is production in tonnes.

2. See note 2 for Figure 2.

The period that witnessed an increase in the production of tapioca was thus the same as the years that saw an expansion of rice cultivation, that is, 1956–57 to 1975–76.¹² Low consumer demand and declining profitability after 1975–76 have been cited as possible reasons for a shift away from tapioca production (Pushpangadan 1988).

Coconut

The area under coconut cultivation grew from 1956–57 to 2016–17. In 1956–57, the acreage under coconut was the highest in Alappuzha district, a predominantly coastal agro-ecological zone. Coconut cultivation has spread across Kerala in the last 61 years, and the largest share of area under cultivation of coconut belongs to the northern midlands and Malappuram type agro-ecological zones. The growth in area was highest in the northern districts of Kannur, Kozhikode, and Palakkad (Figure 12). Production also followed a similar trend, with these three districts contributing 62 per cent of overall production in 2016–17 (Figure 13).

Figure 12 Area under cultivation of coconut in Kerala and its nine composite districts, 1956–57 to 2016–17 in hectares

Notes: 1. The Y-axis for each graph is area under cultivation in hectares.

2. See note 2 for Figure 2.

Figure 13 Production of coconut, in Kerala and its nine composite districts, 1956–57 to 2016–17 in million nuts

Notes: 1. The Y-axis for each graph is production in million nuts.

2. See note 2 for Figure 2.

The area under coconut increased significantly after 1974–75, the year when area under rice peaked. Unni (1983) shows that large tracts of rice fields were lost to other uses between 1970 and 1975. Coconut became the preferred substitute crop, arguably because it was less labour-intensive.¹³

While the area and production of coconut expanded from 1956–57 to 2016–17, yields remained stagnant. Coconut yield was 6,920 nuts per hectare in 1956–57 and 6,889 nuts per hectare in 2016–17. Between 1956–57 and 2016–17, 2014–15 was the year that reported the highest yield, at 7,491 nuts per hectare. The lowest yield during this period was in 1983–84, at 3,814 nuts per hectare. It is clear that increases in the production of coconut have come from an expansion in the area under coconut cultivation, and not from increased yield.

Rubber

Rubber is another crop for which area and production grew from 1956–57 to 2016–17. The area under rubber grew both for Kerala as a whole and for all its districts (Figure 14).

Figure 14 Area under cultivation of rubber in Kerala and its nine composite districts, 1956–57 to 2016–17 in hectares

Notes: 1. The Y-axis for each graph is area under cultivation in hectares.

2. See note 2 for Figure 2.

The area under rubber expanded from the traditional regions of Kollam and Kottayam to all districts of Kerala between 1956–57 and 2016–17 (Chattopadhyay and Franke 2006). One of the explanations for this spread of area under rubber cultivation was a shift in cultivation from large planters to smallholders, with support from the public sector (George et al. 1988; George 1999). The Rubber Board and Rubber Research Institute were set up to provide incentives to farmers and to improve the yield of natural rubber. The replanting subsidy scheme, introduced in 1957, led to the dispersion of high-yielding planting material among smallholders. From 1955–56 to 1995–96, the area of rubber under holdings less than two hectares in size increased from 25 per cent to 75 per cent of the total area under rubber cultivation.

The period of expansion of area under rubber cultivation also saw a growth in its yield. The yield of rubber is measured as production per tapped area, as not all trees are tapped in a year. The available statistics show that the yield of rubber increased from 780 kg per hectare in 1980–81 to 1,851 kg per hectare in 2009–10. In 2015–16, the yield of rubber in Kerala was 1,437 kg per hectare (GoK 2017b).

Production of rubber too grew steadily till 2012–13 (Figure 15). Rubber production in Kerala is concentrated in the southern districts of Kottayam and Kollam, and in the central district of Ernakulam. After 2012–13, the production of rubber fell steeply in Kerala as a whole and in all its composite districts. The reason for this could be the fall in rubber prices, both in national and international markets, after 2011 (GoK 2017b), leading farmers to stop tapping due to very low returns. While this decline in prices has not affected the area under rubber cultivation, it has significantly affected the production of natural rubber.

Figure 15 Production of rubber in Kerala and its nine composite districts, 1956–57 to 2016–17 in tonnes

Notes: 1. The Y-axis for each graph is production in tonnes.

2. See note 2 for Figure 2.

Conclusion

This Note on changes in land-use pattern and cropping pattern in the State of Kerala over 61 years is based on a newly constructed dataset. Using the ICRISAT (2015) methodology, data for 14 districts of Kerala formed in different years were apportioned to the nine composite districts that existed in 1957–58. Thus, I have constructed a continuous 61-year time-series on area, production, yield, and land use at the district level for Kerala. The problems of creating such a time-series have been discussed. The three main trends noted for the period 1956–57 to 2016–17 are as follows.

There is a clear shift in land use with a decline in net sown area and a rise in area under non-agricultural use. An important explanation for the rise in area under non-agricultural use is the growth in migration and increased remittances that are used for land and housing requirements.

There were three distinct phases in terms of cropping pattern: 1956–57 to 1974–75, 1975–76 to 1994–95, and 1995–96 to 2016–17. The area under food crops increased in the first phase, and decreased in the second and third phases. The area under non-food crops increased in the first and second phases, and stagnated in the third. The analysis of 15 crops at the State level indicates that the share of area under 10 selected food crops (rice, sugarcane, pepper, ginger, turmeric, cardamom, areca nut, tapioca, banana and plantain, and cashew) in area under all food crops declined, and the share of area under five selected non-food crops (sesamum, coconut, tea, coffee, and rubber) in area under all non-food crops remained stagnant from 1956–57 to 2016–17. The values of the Hirschman-Herfindahl index, used here to measure crop diversification, show that the cropping pattern was more diverse from 1956–57 to 1986–87 and less diverse from 1986–87 to 2016–17.

Finally, the Note examines changes in the area and production of rice, tapioca, coconut, and rubber, four major crops, at the district level over the last 61 years. It also analyses sub-district data for the more recent periods. The area and production of rice and tapioca fell between 1956–57 and 2016–17 for all districts and for Kerala as a whole. In the same period, there was a significant increase in the area and production of coconut and rubber. Today, the share of non-food crops is greater than the share of food crops in the total cropped area of Kerala.