“Why did a slump follow land reforms in Kerala?”, `South Asia Economic Journal’, March 2002
Abstract: The land reforms in Kerala have deservedly attracted attention not only for the boldness of the policy formulation but also for the perceived efficiency with which they have been administered compared to elsewhere. However, the additional perception that the state today boasts of a strong agricultural economy under girded by these very reforms is far from correct with paddy production, its erstwhile core, by now at a level below that in the year preceding their launching. As an explanation of this counter-intuitive outcome, it is proposed here that the slump is a case of the `Dutch disease’, with the difference that the boom has occurred offshore. Such an explanation alone, unlike earlier attempts, is able to account for the trajectory of the central variables, notably real wages and output. Kerala’s experience bears some lessons for the study of agrarian transformations.
1.i. Land reforms in Kerala: the background
Well before the Indian state of Kerala had attracted attention as representing an egregious configuration of high social development and low per capita income it had attracted attention as an area of democratic leftwing politics. In 1957 Kerala became among the earliest instances in history of the electoral victory of a communist party. Within weeks of assuming power the government of EMS Namboodiripad had introduced the Kerala Agrarian Relations Bill which as the very title suggests was based on a Marxian understanding of the economy. The precise consequences for agricultural production envisaged by the reforming government is a little difficult to divine presently even though it is conceivable that it were also motivated solely by the cause of ending feudalism per se. However, an expansion in production consequent upon land reform appears to be the prediction from almost every theoretical perspective. To run ahead a little, this was not to be in Kerala. Indeed, it is not just that agricultural production has not grown faster, it is that what was historically its core – the cultivation of paddy – has actually shrunk steadily. An explanation of this outcome is my objective in this paper.
For close to fifteen years following the formation of Kerala State in 1956 there have been a range of initiatives pertaining to the reform of agrarian relations, most of this conforming to the widely used term `land reforms’. These initiatives have received wide attention not only in India but also internationally. Raj and Tharakan (1983: 31) have provided an explanation of why this has been so: "Agrarian reform in Kerala over the last quarter of a century is generally believed to have been more far reaching and effective than elsewhere in India, though carried out within the same administrative and political framework as in the rest of the country." Internationally, perception about the effectiveness of the land reforms has also on occasion been extended to claims regarding its consequences for the growth of agriculture in the State, for an example of which see the assertion by McFarlane (1999: 123) that Kerala "… has a solid agricultural base, (thanks to earlier land reform being carried through comprehensively)". Actually, as I have already alluded to and shall soon show, this is somewhat off the mark as a description of the health of the State’s agricultural economy today.
1.ii. The record of rice production in Kerala
I confine attention solely to the production of rice. Historically pulses have not figured much in Kerala’s agricultural production structure as they do elsewhere in the country. On the other hand, the output of tapioca (cassava), which was not only historically a staple of sorts here but was also widely cultivated, has more or less behaved akin to rice. In Table 1 then are presented data on the area and output of rice production in Kerala. The choice of data points needs to be explained. 1956 is the year of the formation of Kerala as a state within India’s federal polity. 1996-97 was the last agricultural year for which data was available at the time when this study was commenced. To save the reader from distraction, without prejudice to the information content, the data for the intervening period has been provided for time points corresponding to the beginning and the middle of each decade alone. Except for the first half of the seventies which period contains the unique inflexion point for each series. Data for the entire intervening period is, however, graphed in the Chart which follows Table 1.
While there are many interesting observations that may be made with respect to the behaviour of output, I make only one. That is, by the agricultural year 1996-97 the output of rice is lower than it was estimated to have been in 1956-57. However, it is not as if rice production in Kerala has declined steadily from the very beginning of this period. Indeed the period breaks down more or less into two each of a rising and a declining trend in production, respectively. The year of the turnaround in the rice economy may be put down to 1974-75 when acreage under rice peaks. Since that date the trajectory of rice acreage in the State is inexorably downwards, and output may be seen to follow with a lag. The key events associated with the implementation of land reforms in Kerala may be put down to span the period 1959 - the year of the passing of the Kerala Agrarian Relations Bill - and January 1, 1970 the date as of which tenancy legally ceased to exist in the state. Essentially, this intervening period was the occasion of the implementation piecemeal of much land-related legislations. The sixties were a period of continuous growth of output, and this dynamic appears to have carried over into the first third of the seventies. While this trend does not by itself establish a benign role for land reforms in the subsequent history of the growth of rice production in Kerala – a history of unmitigated decline – it does provide reason to believe that the principal cause for the outcome needs be sought elsewhere. Indeed, I argue that such an explanation can be provided, and proceed to do so.
[Table 1 goes about here]
It may be noticed that I have not here presented data on the behaviour of yield. Actually, yield has grown more or less without faltering right through the four decades since 1956. However, I do not consider increasing yield in the context of declining production, and very likely employment, a significant achievement at all. In any case, we can see that the increase in yield over the twenty-five years starting from the agricultural year 1971-72 was only marginally higher than the increase in yield over the fifteen year period ending 1970-71. So while there might even be a prima facie case for arguing that land reforms have contributed to a decline in agricultural production there is not even a prima facie case for these having contributed to any increase in yield. In this paper, therefore, I do not discuss yield any further.
[The Chart goes about here]
The experience with rice production in Kerala is easily seen as the outcome of developments in a small open economy. While openness as an idea is easily understood, the reference to size is in the sense in which it is used in trade theory whereby the country’s external terms of trade are given, so that its producers are price takers. I consider the latter a point worth stressing in the context of Kerala. Equally, while openness may be easily understood as a concept, its implication for the trajectory of Kerala’s economy has mostly been overlooked. This has led scholars, on the one level, to often focus on the wrong variable and, at another one, encouraged the belief that the state has always been in control of the economy.
2. Collapsing output: A suggested explanation
2.i. The `Dutch disease’: Theory
The `Dutch disease’, so named to describe the decline of manufacturing in the Netherlands after natural gas was discovered there, has now come to be recognised as a distinct case in open-economy macroeconomics. A phenomenon common to the developed and the developing economies, it refers to the co-existence within the traded-goods sector of an economy of booming and lagging, or progressing and declining, sub-sectors. In many cases, the booming sector has been of an extractive kind – such as minerals in Australia, natural gas in the Netherlands and North Sea oil in the United Kingdom, and the sector that is placed under pressure is the traditional manufacturing sector. For this reason the resulting condition has been referred to as ‘de-industrialisation’. However, the sequence of events is generic and is applicable to situations where the booming sector is not extractive, such as the displacement of older industry by technologically more advanced ones, or even to the case of a boom that is occurring offshore. Indeed the results from the theoretical analysis of the `Dutch disease’ may be profitably applied to the study of the effects of booms arising from a variety of exogenous shocks in a small open economy. To run ahead a little, it is these latter cases that are particularly relevant to Kerala.
Perhaps the most elegant formalisation of the `Dutch disease’ is due to Corden and Neary (1982) and I depend upon it closely here. The analysis proceeds by a decomposition of the effects of a boom on the functional distribution of income and the size and profitability of the manufacturing sector. I concentrate only on the latter set of effects. Consider a small open economy producing two goods traded at exogenously given world prices and a third non-traded good the price of which adjusts to clear the market. Label the two traded goods `energy’ and `manufactures’ and the non-traded good `services’ even though a range of possibilities exists. Equally, while there may be many sources of a boom in the traded good sector consider the specfic case of a one-shot Hicks-neutral improvement in technology. The model is made to work by recognising two effects of the boom, namely the resource movement effect and the spending effect. The boom in the energy sector leads first of all to an increase in the marginal products of the mobile factors employed there. This draws resources from other sectors. This in turn gives rise to adjustments in the rest of the economy. It is this drawing of resources into the booming sector that is described as the resource movement effect. If the booming sector uses relatively few resources that can be drawn from elsewhere in the economy this effect must naturally be small and the main impact of the boom must be due to the spending effect. Within this model the spending effect works via the higher real income resulting from the boom leading to extra spending on services raising their price and leading to further adjustments. Naturally, the impact of the spending effect depends upon the marginal propensity to consume services. Note that while some of the spending from increased real income would automatically fall on the traded goods also, notably manufactures, their price cannot rise since these are set in world markets. This has a major impact on the re-alignment of production in the economy.
[Figures 1 and 2 go about here]
- The pre-boom equilibrium:
Figures 1 and 2 depict the effects of the boom on the labour market and the commodity market, respectively. In the former, the wage rate (in terms of manufactures) is measured on the vertical axis and the economy’s total labour supply is given by the horizontal axis OsOT Employment in services is measured by the distance from Os while the distance from OT measures employment in the two traded goods sectors together. It is assumed that the demand for labour in each sector is a decreasing function of the wage rate relative to the price of that sector’s output. Thus LM is the labour demand schedule for the manufacturing sector and by laterally adding to this the labour demand schedule for the energy sector we obtain LT, the pre-boom labour demand schedule for the entire traded goods sector. Similarly LS is the labour demand schedule for services drawn for a given price of services. Pre-boom equilibrium is at point A where LT intersects with LS yielding an initial wage rate of w0. Figure 1, however, cannot by itself provide a complete story. Note that the location of the LS schedule depends upon the initial price of services and this, unlike the price of the traded goods, is not exogenous but determined as part of the complete general equilibrium of the model.
The determination of the initial equilibrium price of services may be illustrated via the familiar Salter diagram, as in Figure 2, with traded goods on the vertical axis and services on the horizontal one. Fixed terms of trade allow the aggregation of manufacturing and energy into a single Hicksian composite traded good. The pre-boom production possibility curve is TS. If indifference curves may be used to summarise aggregate demand in the economy the initial equilibrium is at point a and the price of services is given by the slope of the common tangent to the two curves at this point.
(b) Effects of the boom on production:
Consider the occurrence of a boom in the form of Hicks-neutral technical progress in the energy sector. To highlight the two distinct effects of the boom described earlier we analyse their consequences separately and in turn. Further, in the case of the resource movement effect we conduct the analysis in two stages. First, the relative price of services is held constant, and then it is allowed to vary to clear the market. In terms of the two diagrams here, at the first stage, the labour-demand schedule LS in Figure 1 and the price ratio in Figure 2 are held constant.
Beginning with the resource movement effect, the labour demand in the energy sector increases. Note that the effect of the technological progress is to increase profitability at a given wage rate, akin to an increase in the price of energy. Now, in Figure 1, the composite labour demand schedule LT shifts out to L’T yielding the new equilibrium B. This has associated with it a higher wage rate and lower employment in both the services and the manufacturing sectors. Concentrating on the manufacturing sector, we note that with employment having fallen from OTM to OTM’ the resource movement effect has given rise to direct de-industrialisation. Turning to figure 2, the boom raises the economy’s maximum output of traded good but not that of services. The production possibility curve shifts out asymmetrically to T’S, with OT’ representing the new maximum in the traded goods sector. The resource movement effect may be represented by a movement of production from a to b. The point b lies to the left of the point a since the shifting out of labour causes a decline in output in the services sector.
As we are isolating the resource movement effect we assume that the income elasticity of demand for services is zero ignoring the spending effect. This assumption implies an income consumption curve that is vertical through a, intersecting the production possibility curve at j. We note that at the original relative price there is excess demand for services.The relative price now increases switching demand away from the good and dampening, but not reversing, the fall in the output of services. The equilibrium point must lie somewhere between b and j on TS’ implying that the output of services is reduced due to the resource movement effect.
Turn now to consider the spending effect on its own. To isolate the resource movement effect assume that the booming sector, energy uses no labour. This translates into the co-inciding of the curves LT and LM in Figure 1, and no effect of the boom may be discerned here at the original relative price. In Figure 2 the boom displaces the production possibility curve vertically upward with point b now lying directly above point a. Provided that services are a normal good in the aggregate, at the original relative price the demand may be expected to grow along an income consumption curve such as On implying point c as indicative of demand. Once again, given the original relative price, there is excess demand for services and the price of services increases. In the new equilibrium, which must lie somewhere between j and c, the output of services is higher when compared to the original situation.
We see that while both the resource movement effect and the spending effects on their own cause an increase in the relative price of services, their impact on the output of services is asymmetrical. The former tends to reduce output while the latter tends to raise it, and there is no presumption as to which will dominate. However, the point of this somewhat elaborate exercise, is to bring out that while there may be some ambiguity regarding output response in the services sector there is none of this whatsoever in the case of the manufacturing sector. We can see this from Figure 1 directly. Allowing for the rise in the price of services the labour demand schedule for that sector must shift outwards. At the new equilibrium G we find the wage rate still higher at w2 and the employment in manufacturing still lower at OTM’’. At the end of the analysis we see that the boom gives rise to direct de-industrialisation reflected in the lowering of output from OTM to OTM’ and indirect de-industrialisation reflected in the further lowering of output from OTM’ to OTM’’. The former is caused by the resource movement effect alone while the latter is caused by the rise in the price of services resulting from a lower output due to the resource movement effect and the higher demand from the spending effect. Since the manufacturing sector’s employment unambiguously falls the same must be true of output within that sector.
2.ii. Agricultural decline in Kerala as the `Dutch disease’
The decline of rice production in Kerala may be plausibly seen as similar to the case of a traded goods sector in a small open economy being constricted by a boom, the quintessential case of the `Dutch disease’. The standard version has concentrated on the case of a booming natural resource sector exerting a squeeze on the manufacturing sector, the boom itself having been caused by a fresh discovery or technological progress in the extraction of the natural resource. However, as I have already indicated, the formal structure of the model is consistent with many alternative interpretations concerning both the structure of the economy and the source of the boom. Within the specific version that we have just worked through this only requires that `manufacturing’ is replaced by `agriculture’ and that the boom in the `domestic energy’ sector be replaced by a boom `offshore’. In the very same Figure 1 then the labour demand for the energy sector may be replaced by demand for domestic labour to service this offshore boom, we may call it labour export. Indeed the specific reality of Kerala was that starting the first half of the seventies there has been an expansion in the demand for emigrant labour following the rise in real income in the Arabian Gulf region following the four-fold hike in oil prices in late 1973. So, exactly as in the Corden and Neary model recounted here, it has even been an energy sector that has boomed, though not the domestic one! I find such an application of the standard model quite appealing. Neither the assumption of full employment equilibrium in the original version nor the feature of a single economy-wide wage-rate need lead to our baulking at its use. The role of full employment in the model is to ensure that the resource-movement effect always bites, in that every outflow of labour from a line of production, or ‘sector’ in the language of the model, automatically leads to a decline in output. This is not necessarily such a horrendous assumption to make. The observation in an economy of aggregate unemployment does not by itself guarantee the free flow of labour between lines of production within an economy. If the hubris that allows for the appeal to the idea that `nature abhors a vacuum’ is foolhardy in the context of the industrial economies, to ignore the highly segmented labour markets of traditional agricultural labour markets is even more so. In particular, I refer to a long-term feature of the labour market in Kerala, prevalent into the seventies, that not all occupational boundaries disappeared fast enough as wages altered. Thus any outflow of labour traditionally engaged in agriculture does not necessarily lead to this being filled by labour inflow from among the unemployed elsewhere within the economy.
2.iii. The evidence
Kerala has had a long history of migration even to the Arabian Gulf region where Indian clerical staff were employed in the multi-national oil companies as early as the nineteen forties. However, a significant break in this pattern appeared in the first-half of the seventies when following the favourable terms-of-trade shock the policy makers of this region deployed the increased (windfall) income on expansion of their economies. The resulting boom led to a unprecedented demand for manual labour which was met through labour import. Indians were a large part of this, and among the earliest migrants were unskilled labourers from Kerala. Continuous data on migration from India, by region and occupational category, is not available for the early part of this phase and whatever evidence there is points to some waxing and waning according to local policy changes in the destination countries. However, what is beyond doubt from the scanty data that we have access to - apart from lore in Kerala, which there is no reason for the economist to ignore – is that after 1973 there is a very significant shift in the nature and magnitude of migration from Kerala. There is now a distinct shift in destination and type of migrant, with the emergence of a migration to the Gulf of manual labour. The best conspectus of the evidence – even though from disparate sources - of migration to the Gulf by region and occupational category is to be found in Nair (1983) to which work I refer the reader. In the scrutiny of my hypothesis regarding the decline of agriculture in Kerala I take a heightened emigration of manual labour to the Gulf starting the first-half of the seventies to be a widely accepted fact and look for contingent evidence within the rice economy.
Two strands of evidence with respect to the evolution of Kerala’s rice economy give me reason to believe in the validity of the explanation that I have proposed, it being, let me emphasise, that the origin of the agricultural decline lies in the exit of labour from the sector. The first of these is that, as seen from Table 1, area under cultivation peaks in the very first agricultural-year after the quadrupling of the international price of oil in late 1973. It has not escaped my attention though that this is altogether too neat! Therefore, I prefer to highlight the second strand of evidence that I’ve referred to above, that the contemporaneous movements of real wages and the output of paddy in Kerala since 1973 is in line with the predictions of the model of the `Dutch disease’. That is, the real wage has risen even as the output has fallen (again, as seen from Table 1). This would not have been worth mentioning for a moment if the very same relationship had not puzzled earlier researchers. Thus in the well-known compilation of a real-wage series for agricultural workers by Jose (1988) the author remarks on the unusual behaviour of these variables in Kerala which feature singles it out from the rest of the country. In most of the states of India agricultural output and real wages were found to have moved together. However, it was found that in Kerala the real wage increased while agricultural production had declined. Since Jose's study is largely empirical in its approach it is not clear from which perspective surprise was being expressed. Indeed, we see that while the study may have thrown up the apparent puzzle of rising wages in a situation of declining production within a sector, it poses no particular challenge when the same phenomenon is interpreted as a case of the `Dutch disease’, with the agricultural sector viewed as part of the larger economy. In this story the wage is endogenous, to be understood as having no independent explanatory power per se. A pointer from the model is that the wage rate is determined by economy-wide factors, something that is missed when the focus is on the relation between wages and employment in the agricultural sector alone. Studies in the latter mode exist and routinely claim to have provided an explanation of declining production by pointing out that real-wage growth has exceeded the growth of land yield, treated as a proxy for labour productivity when fixed proportions are assumed in the technology. Observe that this practice cannot by itself be considered to have provided an explanation. At best, it only succeeds in pointing out that the observed relationship between output and some key variables is in line with a version of the neo-Classical model of profit-maximising behaviour on the part of producers. By contrast, an application of the model of the `Dutch disease’ does succeed in providing an account of the origins of the decline of agriculture in Kerala, predicting accurately some of the subsequent sequences.
[Table 2 goes about here]
Even though agricultural decline within a model of the `Dutch disease’ is not wage-driven a clear trajectory of the sector-specific wage is predicted. As an outcome of labour emigration, the product wage must rise. In keeping with this, the time series of the product wage – defined as the money wage-rate divided by the product price - for paddy cultivation in Kerala is provided in Table 2. This presents us with an unambiguous picture. However, before moving on to discuss it I discuss the data construction itself and the mode in which it is presented here. Indeed these two aspects are intertwined. Note that the data have been presented discontinuously for two phases, the series being denoted `I’ and `II’, respectively. This has been necessitated by the form of the unprocessed data. First, since female labour has historically been a significant input into the cultivation of paddy an attempt has been made here to capture this in the product wage. This implied having to work with a truncated time-series, for data on the money wage-rate for female agricultural labour is available only for the period since the seventies. A second reason for an inevitable truncation came from the nature of the available official statistics for the price of paddy, also discontinuous in its availability. In the construction of the product wage the lack of a continuous price index is not necessarily a non-starter, for we may use data on the actual market price of paddy which is available. However, the available data again comes truncated, the break occurring once again in the seventies, with the data for the earlier period being the price of paddy per para, a traditional Malayali measure of volume, while for the later period it is the price of paddy per quintal, as in the standard contemporary measure of weight. This therefore determined the presentation of the data in two sections in Table 2, with the data for the period prior to 1973 being the annual average daily wage rate for male agricultural workers in terms of a para of paddy and the data for the period after 1973 being the average of the annual wage rates for male and female agricultural workers in terms of a quintal of paddy. However constituted, these are precise measures of the product wage in paddy cultivation. The data on both the money wage-rate and the price of paddy have been culled from various issues of the annual `Economic Review’ of the Government of Kerala. The source, when cited, is invariably the Government’s Department of Economics and Statistics.
Actually, the fact that we have to work with separate series for each of the two periods is not a disadvantage, for it so happens that the point of the bifurcation divides the overall period exactly according to our interest. Note that 1973-74 is the year of the first oil-price hike that had triggered unprecedented migration to the Arabian Gulf Region from Kerala. However, as in the case of the data on acreage and production, a comment on the data points chosen is in order. Note that while for the first phase the data is presented continuously, for the second one the data has been presented only for selected points, being the beginning and the middle of each decade except of course for the first and final years for which the data was available. This is done since the time series for the first phase showed no trend while for the second phase, even though it fluctuated a bit, a very marked trend exists. The entire series are available from the author upon request.
Since the data presented is so stark in what it conveys I confine my commentary to the briefest. Over the twenty five year period since 1973, there is more than just a brisk trend in the product wage. On the contrary, for the period prior to it there has been none to speak of. In a reading of the data it may be of value to know that the product wage in 1973-74 was higher than in 1960-61, thus removing any concern that the use of index numbers might have masked the relative levels of the product wage in each half of the truncated series spanning close to four decades. But returning to the feature of greatest interest, in the twenty five years since 1973 there is a six-fold increase in the product wage! Even as the direction of the wage increase is in line with the prediction of the model considered here, its magnitude exceeds our expectation.
2.iv. Further considerations
I presume that I have been able to make a reasonable case for the view that the decline of rice production in Kerala has to do with the flight of labour from agriculture consequent upon the boom in the Arabian Gulf region. However, I would like to yet discuss three issues with a bearing on the proposed explanation. These have to do with a putative role in the migration for land reforms per se, the issue of the occupational composition of the migrants and the role of prices as opposed to labour availability in contributing to the decline in paddy cultivation.
It is plausible to suggest that the nature of land reforms in Kerala and the manner of their implementation may have created the pre-conditions for migration. In particular, we may argue that one of the consequences of land reforms in Kerala is to have created a labour market where one did not exist. While to those familiar with the situation this would be of no surprise, the creation due to land reforms of a labour market may not be an easily understood consequence and the very possibility needs to be explained. What we have had here is the destruction, due to the reforms, of an institution peculiar to Kerala, an institution which, curiously, is recognised best in terms of the nomenclature of the rights accorded, as an integral part of land reforms, to one set of participants. I refer to the arrangement whereby labouring families were historically permitted to set up a kudil or hutment on landed property upon the understanding that they participated in the agricultural activity of the janmi (landlord) and even of his or her kudiyaan (tenant) in return for the right to habitation. This institution is to be seen as a timeless arrangement central to feudalism in Kerala, and itsraison d’etre was to ensure a steady supply of labour. In the context of agricultural production, when commercialisation was less than complete, the guarantee of labour supply was of far greater importance to the functioning of the system than the fact that it was cheap, which observation is often made even as it is of dubious significance while referring to a situation where a labour market did not exist, rendering comparison with the market wage entirely hypothetical.
In a provision unique to the legislation in Kerala land reforms left the members of each kudil entitled to ten cents of land surrounding the hutment on grounds of kudikidappu avakasham (right to the homestead). This had the immediate effect of alienating the beneficiary from the existing labouring arrangement, turning all the members of the kudil into the free agents that they were not hitherto. In this sense, we may speak of land reforms having created a labour market where it had not existed. However, by the seventies this arrangement is likely to have been confined, relatively, to the erstwhile Malabar District comprising north Kerala while the decline in rice production has been pretty much uniformly spread across the State. As for migration to the Gulf the districts recording the largest migration are Malappuram, Thrissur and Thiruvananthapuram, indicating a fair spread across the State and discounting the extent to which migration was released to the specific effect of the land reforms discussed here. At the same time, it points to the very likely existence of a factor influencing events at the economy-wide level for the State. In the sudden expansion of demand for manual labour emanating from the Gulf we have precisely such a factor.
A response that may be encountered when the argument in this paper is made is that migration to the Gulf from Kerala has largely been from sections of its society other than the agricultural workers. This stems from a lack of understanding of both the model of the `Dutch disease’ and of the nature of migration from Kerala. First, however, the prediction from the model itself. Even in the full employment version of the model, as in Corden and Neary, the decline in the output of the traded good sector is not dependent on the extent of migration from that sector in relation to the extent of migration from the non-traded good sector. The parallel for the discussion here is that it is only necessary that some agricultural workers migrate. Outside the specific formulation in the Corden and Neary model, if a wage differential is allowed for, we can conceive of the situation where service-sector workers migrate and hitherto employed agricultural workers take their place. There is an implicit asymmetry embedded in this argument, suggesting that there may be entry into the service sector from the agricultural sector but not into the latter from a presumed pool of the unemployed. The relative confidence underlying this assumption is that, historically, the social origins of agricultural labour in Kerala has been caste specific with certain castes granted exclusive access to certain tasks such as the transplantation of paddy and the felling of coconut. However, even though the scenario of a changing labour allocation without commensurate entry into the stock of agricultural labour is entirely plausible it is not necessary for me to take resort to it here in order to push my explanation of the agricultural decline. The earliest surveys of the occupational origins of the migrants as reported by Nair (1983: 20) can be found to register ‘agricultural labour’ as a category. This can hardly be surprising when it is recognised that the initial demand in the Gulf was for labour to man construction sites and that, initially, by far the largest migration was of unskilled manual labourers from rural Kerala. If agricultural labourers did in fact migrate then we might expect the kind of behaviour observed elsewhere in this country by Thorner (1961) to have come into play here too. Members ofagricultural labour households with successful migrants exit the labour force. Moreover, even where they may not consider it infra dig to remain as labourers, the effort-reducing wealth effect of remittances on the household labour-supply decision should be taken into account. Thus it is conceivable that the outflow to the Gulf has had a multiplier effect, as it were, sucking out of the labour force in Kerala a larger potential supply than itself.
Finally, it may appear as if I have underplayed the role of prices. Certainly I have not ignored it. However, for two reasons, I have not gone beyond providing evidence of the trajectory of the product wage in paddy cultivation. First, when all of agriculture is declining, even in terms of acreage, the intra-agriculture relative price must lose some of its influence in the determination of the cropping pattern. Thus the relative price of paddy cannot be expected to account for much of the declining acreage under this crop. Secondly, even at the level of the individual crop, if the explanation proposed here is the correct one, resort to the use of inter-crop relative prices to account for the decline of paddy cultivation would leave the field hazardous with observational equivalence. For, any expansion in acreage of potential substitutes such as the coconut, ostensibly due to favourable relative price changes, is entirely consistent with the argument of generalised labour scarcity in the economy, with a cultivator’s motive for switching out of paddy into coconut production being that it is far less labour intensive than paddy cultivation. In fact, it is for the latter reason that the cultivation of coconut with hired labour is so well-suited to the observed practice of absentee land-ownership, an arrangement quite contrary to the original spirit of agrarian reform in Kerala. In any case, evidence of a relative-price shift, intra-agricultural or inter-sectoral, of the nature or strength required to support a price-based explanation of the agricultural transformation in Kerala is yet to be provided.
It has not escaped my attention that, while the evidence presented in this paper pertain exclusively to the cultivation of paddy, through the title of this article I have suggested an agricultural decline. The latter is not an oversight, having indeed a dual motivation. The first has to do with the fact that I am concerned with land reforms, and the target area of the land reforms in Kerala was overwhelmingly the cultivation of paddy. The thrust of land reforms being the abolition of tenancy the policy invariably came to target paddy because it was in its cultivation that tenancy predominated. Not only were plantation crops such as tea, coffee and rubber excluded from the ambit of the reforms, but also, the land under coconut, which had fallen within the purview of the reforms, was seldom leased out. Since the largest number of beneficiaries of the reforms were directly or indirectly related to the cultivation of paddy its subsequent dynamics become a natural ground for the investigation of the consequences of reform. Secondly, the cultivation of paddy was historically the lynch pin of Kerala’s agriculture. At the time of launching of the land reforms, the largest number of agriculturally employed persons were directly in this activity. Therefore paddy cultivation bore the same relationship, in terms of numbers, to those whose livelihood depended on it in comparison with the rest of agriculture. Now, the substantial diminishing of this segment of production without replacement by other crops is, in my view, reasonably described as `the decline of agriculture’ as a whole. From the growth pattern in Kerala’s agriculture as outlined in Kannan and Pushpangadan (1988), we see that the area and output of an `all crop’ (read `agricultural’) index is declining after 1975-76; even though, in line with my perspective, the decline in the area and output of paddy is of course found to be much higher. While this would constitute the necessary evidence for my argument, to depend upon it in its entirety would be to take recourse to a purely statistical approach. On the other hand, I have started out with the position that historically in Kerala the cultivation of paddy had a significance far exceeding that conveyed by indicators such as acreage – even though this was hardly ever insubstantial – and am implying that the manifestation of the decline in agriculture in the State is the shrinking cultivation of paddy which being by far the most labour intensive activity would naturally have borne the brunt of labour emigration.
There are two sets of lessons to be drawn from the recent agrarian history of Kerala, one each with a bearing on approaches to the study of agriculture and on how non-agrarian factors might constrain policy initiatives such as land reforms, respectively. Among the first, the principal one is that the incentive structure can fly in the face of the outcome expected from institutional change. In this case, the abolition of tenancy as an institution and the ushering in, technically speaking, of capitalist property-relations could do little to make dynamic the agricultural production in Kerala given the drift in the incentive structure. We see that a world-view based on social property relations alone severely underestimates the autonomy of factors governing agricultural dynamism, notably technological change. Another instance of this, even though of a transformation in polar contrast to that in Kerala, is the rapid transition to a dynamic agriculture in the Indian Punjab in the mid-sixties even as the institutional structure did not alter much. The second lesson from Kerala is the overwhelming importance of the macroeconomic environment for outcomes in the agricultural sector and the constraints that this might impose. Policy interventions aimed at the agricultural sector could be nullified when independent developments contingent upon the openness of the economy offer the actors an alternative. In the emigration to the Gulf from Kerala I have pointed out one instance of this. Since, in the literature on development, the State has received a great deal of attention as an economy with successful policies it may not be inappropriate to query the myth that the policy maker there has been in command.
Corden, W.M. and Neary, J.P. (1982), Booming Sector and De-industrialisation in a
Small Open Economy, Economic Journal, 92: 825-848.
Government of Kerala, Economic Review, Thiruvananthapuram: The Planning Board.
Jose, A.V. (1988), Agricultural Wages in India, Economic and Political Weekly, 23:
Kannan, K.P., and Pushpangadan, K. (1988), Agricultural stagnation and economic
growth in Kerala: An exploratory analysis. Working Paper 227, Thiruvananthapuram:
Centre for Development Studies.
McFarlane, B. (1999), `Review of Kerala’s Economy by B.A. Prakash’, Journal of
Contemporary Asia, 29: 122-4.
Nair, P.R.G. (1983), Asian Migration to the Middle East: Emigration from Kerala.
Working Paper 180, Thiruvananthapuram: Centre for Development Studies.
Raj, K.N., and Tharakan, M. (1983), `Agrarian Reform in Kerala and its Impact on
the Rural Economy – A Preliminary Assessment’, in A.K. Ghose (ed) Agrarian Reform
in Contemporary Developing Economies. New Delhi: Sage.
Thorner, D. (1961) `The Agrarian Problem in India Today’, in Daniel and Alice Thorner
(eds), Land and Labour in India. Delhi: Asia Publishing House.