Weapons of the Weak- Everyday Forms of Peasant Resistance

by James C Scott

  A more detailed analysis of changing village composition may be found in appendix A, but for my immediate purposes the following points merit special note. Of those leaving the village, the poor tended to leave as individualsyoung men and women typically leaving to work in urban areas (as construction workers, manual laborers, factory workers, and domestic servants). Families that left Sedaka as families, on the other hand, tended to be well-off-buying land in a frontier area or being accepted as a “settler” on a government estate (rancangan). Fifteen new families have moved to Sedaka, and it is significant that they are among the poorest half of the village. As villagers put it, “People have children, but the land does not have children.” There is no niche in the village economy that could accommodate most of the population increase already born.

  LAND TENURE

  Many of the major changes in the economic life of Sedaka are reflected in the basic data on overall land tenure shown in table 4.4. Detailed comparisons may be found in tables 1 and 2 of appendix C. The discussion here is confined to only the most striking findings.

  The most obvious trend over the past dozen years has been an increase in both the number and proportion of marginally sized farms and a decline in the share of large farms. The number of households farming 3 relong or less has doubled from twelve to twenty-four. Five of those farms are less than a single [Page 101] relong, a size category that Horii could justifiably ignore altogether in 1967. Over the same period, the average size of these minuscule farms has declined to less than 2 relong, or 1.4 acres. The decline in small farm size is not a consequence of a stable class of smallholders squeezed onto less and less land; on the contrary, the proportion of village-held land that such small operators farm has actually grown nearly twofold from 7 to 13 percent. As a class, they have lost ground in part because they are so much more numerous and in part because the total paddy land farmed by villagers has diminished by nearly 10 percent (from 357.75 relong to 325 relong).

  TABLE 4.4 • Frequency Distribution of Farm Holding in Sedaka, 1967–1979

  The compression of farm size among a growing number of poor cultivators has not been accompanied by a parallel expansion of farm size by larger-scale cultivators in the village. In fact, nearly the reverse has occurred. Before doublecropping there were seventeen farmers who owned and/or operated more than 10 relong; collectively they monopolized more than 57 percent of village-held land (206.5 relong). Today, when the village is more populous, there are only thirteen such households, and they farm only 37 percent of village-held land (123 relong). The average farm cultivated by this strata shrank in the same period from over 12 relong to 9.5 relong. Thus we find a situation in which the size of both small farms and large farms has diminished, but the number of small farmers has doubled while that of large-scale farmers has declined.

  Looking at the average farm size for the village as a whole, the trend is comparable. The average paddy farm in 1967 was nearly 7 relong; now it is less than 5. This represents a farm size decline of 32 percent which, using the [Page 102] standard assumptions about the profits from double-cropping, means that over 90 percent of the potential profits have been eliminated merely by the striking reduction in farm size.12

  Sedaka has been caught in something of a demographic and structural pincers movement. Its population of farm households has grown by nearly one-third while the paddy land it cultivates has diminished by nearly 10 percent. Even if the village had not lost farm land, population growth alone would have reduced the average farm size from nearly 7 relong to 5.2. The balance of the decrease in average farm size to its present level of 4.7 can be attributed to loss of nearly 33 relong of the farm land that villagers cultivated in 1967. How has this reduction come about? It has not come about, we can be certain, because village landowners have withdrawn land from village tenants. The amount of land that village landowners rent out to their fellow villagers, usually close kin, has remained stable over the last twelve years. All of the land withdrawn from village cultivation in this period has instead been withdrawn by outside landlords, many of them quite rich, who have chosen to farm the land themselves, give it to their grown children to farm, or rent it out on long-term lease (pajak) to other large-scale operators.

  If we examine the distribution of tenure categories in 1967 and again in 1979, the pattern is one of relative stability. A partial exception is the emergence, for the first time, of five households of pure noncultivating landlords in the village. But two of these landlords rent out tiny plots because they are too poor even to finance a crop and see it through to the harvest. The rest are mostly old men who can no longer cultivate and, in one case, a man whose debts have temporarily forced him to rent out his land to meet his creditors’ demands. The only other notable change has been a modest decrease in the proportion of village farmers who are pure tenants (from 44 to 35 percent). Far more striking, however, has been the decline in farm size for these pure tenants. They remain, as in 1967, by far the most numerous tenure category, but they had cultivated an average of 6.1 relong in 1967; by 1979, their average farm size had declined to 4.1 relong. If the three tenants who now cultivate more than 7 relong are removed, the average holding of the remaining tenants falls to 3.3 relong. As in the Muda region as a whole, the loss of land is largely confined to the class of pure tenants. Collectively, while their absolute numbers have grown from twenty-three to twenty-six, they have lost more than one-fifth of the land they cultivated before double-cropping. The farm size of the smaller class of pure owner-operators, by contrast, has remained small but stable over the same period.

  CHANGES IN TENANCY

  [Page 103]

  The erosion in the position of tenants has been closely related to important changes in the form of tenancy since the beginning of double-cropping. One key change, as noted earlier, was the virtual disappearance of what were called paddy rents (sewa padi) and their replacement by what are known as cash rents (sewa tunai).13 In 1967, Horii found that just under half (48 percent) of the land rental contracts in Sedaka were cash denominated rather than tied to paddy.14 By 1979, however, paddy-denominated rents were conspicuous by their absence, and over 90 percent of all tenancy contracts now provided for stipulated cash rents.

  TABLE 4.5 • Classification of Tenancy Agreements in Sedaka by Timing of Rental Payment, 1967, 1979

  Before 1970, the payment of cash rent before planting season was quite rare, as shown in table 4.5. By 1979, however, prepayment of rent had become quite common and was approaching one-half of all tenancy agreements. Most of this shift occurred in the first two years of double-cropping, when the initial profits from the new irrigation scheme combined to give most tenants the means to pay, and hence landlords the possibility of requiring, rents in advance. For poorer tenants, the burden was considerable. Exactly how many villagers were unable to raise the cash and lost land in this fashion is, as we shall see, a matter of lively dispute-a dispute that basically follows class lines.15

  [Page 104]

  The second change in tenancy has eliminated the possibility of renegotiating rent after a poor harvest. The local term for negotiable rents is “living rent” (sewa hidup) to denote their flexibility, as opposed to “dead rent” (sewa mati), which is rigidly enforced. Non-negotiable rents include all tenancies in which rent must be paid before planting as well as a number of post-harvest-rent tenancies that are rigidly enforced. This is particularly the case with post-harvest rents between landlords and tenants who have no kinship tie.

  TABLE 4.6 • Classification of Tenancy Agreements in Sedaka by Negotiability of Rents, 1967, 1979

  As table 4.6 shows, tenancies with inflexible rents were, before double-cropping, the exception, whereas by 1979 they had become common enough to constitute a majority of all tenancy agreements. A tenant paying average-to-high rents must harvest, depending on his production costs, anywhere from seven to eleven gunny sacks of paddy in order simply to break even.16 Average yields in Sedaka are roughly thirteen to fourteen gunny sacks and o
ccasionally drop as low as seven to eight. Under the older system of tenancy, the actual rent would typically be reduced to compensate, at least in part, for harvest losses. Now when the rice crop, which even with irrigation is still subject to the caprice of nature, is damaged, the tenant must absorb the entire loss and still come up with full rent in advance of the coming season. While the tenant will in almost every case continue to rent the land, the cost in a bad year is likely to be severe in terms of belt tightening, short-term migration to seek work, and new debts. The landlord-tenant relationship has in the meantime been transformed in such cases into a rigid, if unwritten, impersonal contract. The social bond once implicit in the adjustment of rents to reflect the actual situation of the tenant [Page 105] has given way to tenancies along commercial lines, in which no quarter is expected or given.

  Despite the land hunger in Sedaka, which is reflected in declining farm size, the average level of paddy-land rents has not increased dramatically. They averaged M$71 per relong in 1967 and by 1979 had climbed to M$ 112, a rate of increase that was slightly below the increase in rice production costs generally.17 For many tenants, especially those closely related to their landlord, land rents are no more burdensome a share of the harvest proceeds now than they were in 1967. But the fact that rent increases have been relatively modest by no means implies that landlords have suffered proportionally. The reverse is the case, for, with the onset of double-cropping, all landlords instantly received apermanent windfall doubling of their annual rental income, thanks to the irrigation scheme.

  In Sedaka, as in the rest of Kedah, all but a very few tenancy agreements are oral contracts struck up informally between the two parties. This pattern is in violation of the Padi Cultivators’ Ordinance of 1955 (reenacted in 1967), which provides for the registration of tenancy contracts and sets maximum rents (as a proportion of the average harvest) for each soil class. Needless to say, most rents in Sedaka exceed that ceiling today as they did in 1967. No serious effort has been made to enforce the act against what would surely be the stiff opposition of the landowning class, which forms the core of the ruling party in the countryside. For tenants renting from close relatives, the act is unnecessary, as they typically have fairly secure tenure and pay flexible, concessionary rents. For those who pay market rents, requesting registration would be an invitation to immediate dismissal and in any case would be pointless inasmuch as registration does not prevent the landlord from resuming cultivation himself or, in practice, from insisting on illegally high rents under the table.

  Average rent levels are, however, quite misleading when what we have in Sedaka and in Kedah generally is a highly fragmented land rental market with rents that ran the gamut in 1979 from M$200 per relong to virtually zero.18 [Page 106] The major “fault lines” in the market for tenancies lie along the dimension of kinship, as table 4.7 clearly shows. The disparity in rent levels reflects a bifurcated tenancy system in which a substantial share (42 percent) of those who rent in land are the beneficiaries of concessionary terms from their parents or grandparents. Within this protected market, seasonal paddy-land rents are substantially below those prevailing outside. Rents between non-kin are, as would be expected, by far the highest, with rentals between more distant kin falling roughly midway between the two extremes. Despite the commercialization brought about by the green revolution, land tenure is still dominated by kinship tenancies, which form fully two-thirds of all land rentals, as they do in Muda as a whole, and which provide some shelter from the full effects of a thoroughly capitalist market for land. Concessionary kinship tenancies between close kin, however, represent only 40 percent of land farmed by tenants in Sedaka, since the average size of the plot rented from close kin is smaller than the plot rented to non-kin.

  TABLE 4.7 • Rental Rates for Tenancies Classified by Degree of Kinship between Landlord and Tenant in Sedaka, 1979

  Kinship is not the only social tie that mitigates the terms of tenancy. When landlord and tenant both live in the village, the rent is typically below what the market will bear. Thus the rent paid by an unrelated tenant to his landlord outside the village averages M$ 129 per relong a season, while rent to an unrelated [Page 107] landlord inside the village averages only M$105.19 The economic impact of tenancy within Sedaka is also apparent in the timing of rent payments. Threequarters of tenants who have some relation (but not a parent-child relation) to their landlord are permitted to pay after the harvest, while less than half of those renting from relatives outside the village are given this concession.20 All this is palpable evidence, if one needed it, that the village is a community in modest but significant ways. Unfortunately, the land available under such terms is in no way equal to the land hunger and demographic pressures it confronts.

  Outside the somewhat sheltered customary market for tenancy between kin and neighbors, the winds of a competitive market blow briskly. The average rent for land rented to non-kinsmen is one indication of this competition. Another indication is what might be termed the Haji landlord market for rents. As noted earlier, many members of the well-to-do Haji landowning class are highly commercial operators who in many cases have branched out into other related enterprises such as tractor rental, rice milling, and trucking. There are in fact nine villagers who rent land from Hajis to whom they are not closely related. The average rent for these nine tenancies is M$140 (per relong, per season), which is well above even the mean rent for all tenants who rent from outside landlords with whom they have no kinship relation. Though none of these Hajis operates on anything like the scale of Haji Broom, they are a key segment of a small but powerful landowning class operating on strict commercial principles.

  Nowhere is the competition for land more apparent than in the dramatic transformation of long-term leasehold or pajak tenancy over the past decade. It was possible in 1967 to view leasehold tenancy as part and parcel of customary [Page 108] tenure between relatives. Horii found three cases of leasehold tenancy in Sedaka, all of which were between relatives, and the average rent charged was well below the standard for cash rents at the time.21 He concluded justifiably that such multiseason tenancies with rent paid in advance were not instances of a “landlord’s hard-hearted demand for advance payment” of a large sum but rather “an expression of mutual aid between kin… a kind of subsistence credit system between related landowners and tenants.”22

  By 1979, leasehold tenancy had become a thoroughly commercial transaction reflecting the new profits available to some landowners and tenants under doublecropping. The number of cases of pajak tenancy had grown from three to seven; the amount of land involved had doubled from 10 to 20 relong and, most striking, the mean pajak rental, adjusted to a seasonal basis, worked out to M$142 per relong, well above even the mean rents between nonrelatives. If anything, the high rent is an understatement of current pajak rents, since a number of the existing agreements were struck well before 1979, when lower rents prevailed, and since one would also expect that lump-sum, long-term rents such as these would be discounted at something approaching the current rate of interest and result in lower, not higher, seasonal rents. The current levels of pajak rents are best indicated by the most recently concluded agreement, which allows the tenant (Tok Omar, #37) to cultivate 3 relong for two seasons in return for a rent of M$ 1, 110-the equivalent of a seasonal rent per relong of $185.

  If, as most observers agree, highly commercial pajak tenancies are becoming increasingly common in Muda as a whole as well as in Sedaka, the impact on access to land is ominous.23 What leasehold tenancy along these lines implies, as we have noted earlier, is that it will no longer be predominantly the poor who rent in land but the wealthy commercial operators with the ready capital to bid for such tenancies. An example of the obstacles that confront even middle peasants who attempt to compete in this market is the case of Rokiah (the wife of Mat Buyong, #34) who was offered a lease of 2.5 relong for four seasons from her brother at a lump-sum rent of M$ 1,600. To raise this amount, all the gold jewelry in the house
was pawned and $500 in loans were taken from two shopkeepers at the usual high interest. Now her brother has announced that he wants to tack another year (M$800) onto the lease, before the initial term has [Page 109] expired. Unless she scrambles to find the additional M$800, she risks losing the land to another tenant. Even Lazim, one of the largest farm operators in the village, was hard-pressed to raise the M$3000 pajak rent for the 4 relong he leased (for seven seasons) from Fadzil. Despite both the long-run attractiveness of the contract (on a per season basis the rent per relong was only M$107) and the fact that Fadzil allowed him to delay by one season the payment of a M$ 1,000 balance, much of the initial sum had to be borrowed from a nearby Chinese trader. It is thus no coincidence that not a single leasehold tenant in the village comes from among the poorest third of its households, while it is precisely these households that most desperately need land to farm. It is also indicative of the commercial nature of such leaseholds that they are now always written and notarized even when the parties are closely related.24

  Four villagers have already lost land they previously rented when the owner insisted on shifting to a long-term lease and they were unable to come up with the necessary cash. Nearly 18 relong have passed out of village cultivation in this fashion. More worrisome still are the reports villagers receive almost daily at the market or from relatives about larger blocks of land passing into the hands of well-heeled, commercial tenants, many of whom they say are Chinese who already own tractors and in some cases shares in a combine-harvester syndicate. Thus a villager in nearby Sungai Bujur has rented 8 relong for fifteen seasons, at a rent of M$15,000, to a Chinese shopkeeper and tractor owner. On a somewhat more modest scale, Tok Kasim (#60) in Sedaka has leased out 4 relong for ten seasons in order to raise the M$6,000 down payment for a taxi his son now drives. The tenant is the Chinese tractor owner for whom Tok Kasim works as a broker, lining up ploughing jobs and collecting fees.25 He says he would have liked to rent to a Malay but no one he knew had the cash when he needed it. Coupled with these reports is daily evidence of the competition for farming land by outside commercial entrepreneurs. Hardly a week passes according to Lebai Pendek (#73), a well-to-do landowner in Sedaka, that he is not asked by the Chinese shopkeepers with whom he trades if he would be willing to lease land to them.