Concept for Land Reform on Java

By:
Roy Prosterman
Robert Mitchell
Rural Development Institute
May 2002

At the outset we note that there are a number of important issues
that the Government should deal with in the area of land relations.
Many of these is sues fall within the scope of MPR Provision No. IX
of 9 November 2001 “On the Agrarian Renewal and Management
of Natural Resources.” The present paper addresses just one of
those issues, although it is an issue that many Asian societies have
found to be critically important: the issue of redistributive land
reform using existing cropland.
Land reform programs in Asia that were implemented effectively have provided
a number of benefits: increased crop production, improved nutrition and
improved social status for poor rural households, a foundation for sustained
and inclusive economic growth, reduced social unrest and instability, better
environmental management, a reduced flow of desperate rural families to the
cities, and improved access to credit for the new landowners. Land reform can
be one of the most targeted and successful kinds of anti-poverty programs. In
particular, land reform can provide meaningful financial resources and food
sources and an important safety-net (even when only a small area is distributed
to each family) to the very poor and landless.
Unfortunately, there has not been any effective or widespread land reform in
Indonesia. However, new possible approaches to land reform in Indonesia
could create widespread benefits.
This paper is divided into three sections: (1) an analysis of the rural land reform
conducted in Indonesia during 1960 - 2000; (2) an analysis of present
landholding patterns; and (3) an analysis of possible programs for further land
reform in Indonesia. Our subject here is redistributive land reform:
transmigration programs are not discussed.
1 This paper is prepared under the Land Law Initiative funded by the United States Agency for
International Development, grant no. 497-G-00-01-00031-00.
2
1. Land reform in Indonesia during 1960 - 2000
Although the Indonesian land reform made some progress during 1960 - 1965,
the activity did not result in the redistribution of much agricultural land and did
not affect many agricultural families, either on Java or off Java. Since 1965
there has been almost no land reform activity.
a. Extent of Indonesian land reform
During the period 1960 - 2000, BPN data shows that the Government of
Indonesia redistributed 850,128 ha under the land reform program.2 Of that
amount, 339,227 ha was on Java.3 As Table (a) shows, this represents only
3% of cropland in Indonesia and only 6% of cropland in Java as of the 1960’s.
By any standard, the land reform program in Indonesia has affected a very
small percentage of cropland.4
Table (a). Cropland distributed in Indonesian land reform.
Cropland
redistributed
Total cropland
in use
Redistributed land
as % of total
cropland in use
Indonesia 850,128 ha 26,000,0005 3%
Java 339,227 ha 5,800,000 6%
Source: BPN Directorate of Land Reform and Land Tenure, March 2000;
FAOSTAT available at http://apps.fa.org/page/collections?subset
=agriculture (visited May 2002).
2 The total amount acquired was 1,470,690 ha, which means that only 58% of the land acquired
was ever distributed.
3 BPN Directorate of Land Reform and Land Tenure, March 2000.
4 At present we do not have data regarding the percentage of cropland that was sawah and the
percentage of cropland that was tegalan. However, there is no reason to presume that all, or
even a majority of the redistributed land was sawah. To the extent landowners were allowed to
select which land to place in the redistribution land fund, they would presumably select the least
valuable land, which might be tegalan or other land that was less valuable than sawah.
5 According to FAO data, this total included 18,000,000 ha of arable land and 8,000,000 ha of
permanent cropland. The comparable FAO figures for 1998 are 17,000,000 ha of arable land
and 13,000,000 ha of permanent cropland. See FAO Production Yearbook, 1999, Table 1.
These contemporary FAO figures are roughly comparable to the figures of 17.14 million ha for
cropland and 16.54 million ha for agricultural estates from Indonesian statistical sources we cite
below. Based on this correspondence, we have estimated in Table (a) that the arable cropland
on Java in the 1960’s was approximately the same as at present (5.14 million ha), and that the
permanent crop area was comparable for agricultural estates (620,000 ha), with almost all of
the growth of that figure off of Java. This yields the estimate for Java, totaling 5,800,000 ha.
3
Chart (i). Javanese cropland distributed in land reform to
date.
6% of
Javanese
cropland was
redistributed
However, the fact that only roughly 6% of cropland was distributed on Java is
not the most important shortcoming of the land reform program. Even more
important was the fact that so few families benefited from the program.
In terms of households, the Indonesian land reform program has redistributed
land to only 1,292,851 families, including 816,849 families on Java. The
average recipient received 0.66 ha throughout Indonesia and 0.42 ha on Java.
In terms of 1963 population data for “farming” households (which does not
include agricultural laborer households), the land reform program affected
approximately 11% of farming families in Indonesia and approximately 10% of
farming families on Java.6
Moreover, if we assume that during the 1960’s that for every 100 farming
households in Indonesia there were 46 agricultural laborer households (which
is the proportion that exists today for all of Indonesia) and that for every 100
farming households on Java there were 64 agricultural laborer households
(which is the proportion that exists today for Java), then the percentage of
“agricultural households” (that is, farmer households plus agricultural laborer
households) that received land in Indonesia was 7% and the percentage on
Java was 6%. These percentages are very low, and represent a very small
proportion of the rural families that were landless or land-poor in the 1960’s.
But the results of the land reform are even less impressive if one considers that
the most vulnerable group -- seasonal agricultural laborer families -- were
largely excluded from the 7% of all Indonesian agricultural families that
received land. Such families were the lowest priority of the eight priority
categories of recipients described in Article 8 of PP No. 224 of 1961 “On
Implementation of Redistribution of Land and Compensation.” It is thus
reasonable to assume that only a very small percentage of the 5,600,000
6 Almost all land reform in Indonesia occurred prior to 1965. According to data of the National
Statistics Bureau [BPS], in 1963 there were roughly 12,236,000 farming households in
Indonesia (excluding East Timor, Maluku and Irian Jaya), of which approximately 7,935,000
farming households were on Java (“farming” households are those that control some cropland).
BPS 1993 Agricultural Statis tics, available at www.bps.go.id/census/ agh.shtml (visited March
18, 2002).
4
agricultural laborer households (which we estimate existed in the 1960’s)
received land in the Indonesian land reform program.7
Table (b). Families receiving land in Indonesian land reform.
Families
receiving
land
Total farming
families in
1963
(excluding
laborer
families)
Families
receiving
land as % of
farming
families
Estimated
total
agricultural
families
(including
laborer
families)
Families
receiving
land as % of
estimated
total
agricultural
families
Indonesia 1,292,851 12,236,000 11% 17,864,000 7%
Java 816,849 7,935,000 10% 13,013,000 6%
Source: Based on BPN Directorate of Land Reform and Land Tenure, March 2000.
Chart (ii). Javanese families affected by land reform to
date.
6% of
Javanese
agricultural
families
received land
International comparisons provide some perspective. The land reform
conducted in Japan during 1948 - 1951 redistributed a total of 41% of all
cropland to 81% of non-landowning families, while land reform conducted in
Taiwan beginning in 1953 redistributed a total of 44% of cropland to
approximately 100% of non -landowning families. As table (c) demonstrates,
the Indonesian land reform affected a comparatively small proportion of the
non-landowning households:
7 The estimate is made based on current ratios of agricultural laborer households to farming
households, both for Indonesia and for Java alone. See chart (iii) below. We estimate that
roughly 5 million of these laborer households lived on Java.
5
Table (c). Land reform in various Asian countries.
Country Year begun % of cultivated
land redistributed
% of nonlandowning
households
receiving land 8
Japan 1948 41 81
Taiwan 1953 449 ±100
South Korea 1948 33 64
South Vietnam 10 1970 44 75
Philippines 1972, 1988 4 11
India (overall) n/a 1.2 11 or 28 11
West Bengal n/a 27.7 12 [ ]13
Kerala [We will add these figures to the final version.]
Indonesia 1961 3 [ ]14
Java 1961 6 [ ]15
As the numbers indicate, broad and successful land reforms have been carried
out in Japan, Taiwan, South Korea, South Vietnam, and the Indian state[s] of
8 The non -owning category includes both tenants and agricultural laborers. In Japan, Taiwan,
South Korea, and South Vietnam, nearly all non-landowning agricultural households were
tenants; there were very few households primarily dependent on some form of paid agricultural
labor.
9 These totals include 40% of tenants, on 14% of cultivated land, who acquired ownership-like
interests in land through one of the very rare instances of successful regulation of the landlordtenant
relationship – guaranteed permanent possession at fixed low rents – by a developing
country.
10 This non-communist land-to-the -tiller program was implemented during 1970 - 1973.
Concurrently, monthly communist recruitm ent in the South dropped by 70 - 85% (although the
land reform could not affect the divisions coming down from the North) and rice production
increased by 30%. See ROY L. PROSTERMAN AND JEFFREY M. RIEDINGER, LAND REFORM AND
DEMOCRATIC DEVELOPMENT Ch. 5 (Baltimore: Johns Hopkins University Press, 1987).
11 There are two different denominators for the base number of non-landowning households in
India, with the approximately 5 million total benefiting households India -wide constituting either
11% or 28%, depending on which base number is used.
12 As in Taiwan, a part of the land reform shown was successfully regulated tenancy, on 20% of
the cultivated land.
13 We have not separately calculated the percentage of non-landowning households receiving
land in West Bengal. Relative to all agricultural households, however, it was approximately
54% (including 20% who benefited from successfully regulated tenancy).
14 It is difficult to estimate the number of families who substantially depend on farming leased
land (tenant farming), both for Indonesia and for Java alone. Nevertheless, tenant farming
appears to involve substantially less land and a smaller proportion of Indonesian farmers than
in Japan, Taiwan, South Korea or South Vietnam. See note 18 below.
15 See note 12 supra.
6
West Bengal [and Kerala]. In each case there is ample documentation of the
major positive impact of land reform on agricultural productivity, farmer incomes
and status, household nutrition, rural stability, and other key features of rural
society.
2. Inequality of Indonesian landholdings in agriculture
The inequality in access to land persists, both with respect to cropland and
pekarangan.
a. Cropland
In 1993 approximately 19,713,000 households controlled some cropland,
including roughly 9,579,000 farm households that controlled less than 0.5 ha of
cropland, and approximately 10,134,000 farm households that controlled 0.5 ha
or more.16 In addition, in 1993 there were approximately 9,054,000 agricultural
laborer households, which, by definition, controlled no cropland.17 The
proportions of landless and land-poor agricultural households are much higher
on Java than off-Java, as shown in chart (iii).18
Thus, of the total of approximately 11,465,000 agricultural families outside Java
(farming families plus agricultural laborer families), about 17% are farming
families with holdings under 0.5 ha, and another 20% are agricultural laborer
families. By comparison, of approximately 17,302,000 agricultural families on
Java, about 44% are farming families with less than 0.5 ha, and another 39%
are agricultural laborer families.
16 Of the 9,579,000 families controlling less than 0.5 ha, roughly 2.8 million families controlled
less than 1500 m2. Of the holdings smaller than 1500 m2, approximately 80% were on Java.
Land Holding Farmers Sample Census (National Statistics Bureau [BPS], 1993 Agricultural
Census), Tables 2 and 19.
17 1993 Census Report on Household Registration in the Sub -Sectors of Paddy, Secondary
Crops and Horticulture, (National Statistic Bureau [BPS], 1993 Agricultural Census), Table E.
At page 24, the report defines “labor household in agricultural sector” as “a household [in
which] one member or more is working in the agricultural sector to obtain wage/salaries eith er
in cash or in kind.” This contrasts with the farming household, which the report defines at page
23 as a household “engaged in crop farming,” which is further defined as “growing plants (such
as paddy, fruits, [etc.]) intending to be sold or to obtain income with risk.”
18 It is difficult to estimate what proportion of farming households substantially depend on
cultivating land as tenants rather than as owners, although the proportion of leased farm land
does not appear to be high. On Java, about 20% of the 5,135,000 ha of cropland is
characterized as “land originated from other parties.” Off-Java, that figure is only 11%. 1993
Agricultural Census, Land Holding Farmers Sample Census, Table 16. It also appears that
roughly 316,000 farm households on Java and 429,000 farm households off-Java may depend
entirely on leased in land. Id.
7
Chart (iii). Farmer households and agricultural laborer households
on Java and Off-Java
2,962,000
6,732,000
7,152,000
7,608,000
1,991,000 2,322,000
Off Java Java
Farmer households holding 0.5 ha or more of cropland
Farmer households holding less than 0.5 ha of cropland
Agricultural laborer households holding no cropland
Source: Land Holding Farmers Sample Census (National Statistics Bureau
[BPS], 1993 Agricultural Census), Table 1; 1993 Census Report on Household
Registration in the Sub-Sectors of Paddy, Secondary Crops and Horticulture,
(National Statistic Bureau [BPS], 1993 Agricultural Census), Table E.
As of 1993, there were 17,145,000 ha of cropland in Indonesia (not counting
16,544,000 ha of agricultural estates), but on Java there were only 5,135,000
ha of cropland (as well as a mere 620,000 ha of agricultural estates). 19
b. Pekarangan
The cropland figures that we analyzed in subsection (a) do not include the
important category of pekarangan, which is also known as home-and-garden
land.20 There are 5,132,000 ha of pekarangan in Indonesia, of which 1,736,000
ha are on Java.21
Like the distribution of cropland, the distribution of pekarangan is very unequal.
Thus, for Indonesia as a whole, 40.28% of households have less than 100 m2
of pekarangan, 25.24% have 100 - 200 m2, 11.72% have 200 - 300 m2 and
22.76% have 300 m2 or more.22 Table (d) shows the distribution for the four
provinces of Java.
19 See Land Holding Farmers Sample Census, Table 16, and Statistical Yearbook of Indonesia
2000, Table 5.1.1.
20 As just indicated, the figures in sub-section (a) do not include land of agricultural estates,
which occupy 16,543,000 ha in Indonesia, including only 620,000 ha on Java.
21 Statistical Yearbook of Indonesia 2000, Table 5.1.1.
22 Hadi Susilo Arifin, “Study on the Vegetation Structure of Pekarangan and Its Changes in
West Java, Indonesia," doctoral dissertation for the Graduate School of Natural Science and
8
Table (d). Size distribution of pekarangan land in agricultural provinces of
Java (percentages of households that have pekarangan in the size groups
shown)
< 100 m2 100 < 200 m2 200 < 300 m2 > 300 m 2
West Java 52.29% 25.00% 8.77% 8.95%
Central Java 27.50% 27.57% 13.20% 31.73%
East Java 34.52% 25.83% 13.33% 26.31%
D.I. Yogyakarta 33.51% 17.48% 14.61% 34,40%
Source: Hadi Susilo Arifin, “Study on the Vegetation Structure of Pekarangan and
Its Changes in West Java, Indonesia,” doctoral dissertation for the Graduate School
of Natural Science and Technology, Okayama University, Japan, March 1998, at
Appendix Table 2 (citing 1995 Housing and Settlement Statistics, Indonesian
Statistics Center Bureau, 1996).
Microstudies can provide an even more specific sense of the differences in
pekarangan held by households in many parts of Java. For example, a 1988
study of the lowland rural village of Sukoharjo in Central Java found that of
1002 village families, 44 families owned no pekarangan or household plot, 347
families owned 50 m2 or less, 328 families owned 50 - 100 m2, 259 owned 100
- 500 m2 and only 24 families owned more than 500 m 2.23
Chart (iv). Percentage of families owning pekarangan in Sukoharjo
1 - 50 m2
35%
50 - 100 m2
33%
100 - 500 m2
26%
No land
4%
500+
2%
Moreover, over the years, if the village does not convert additional sawah into
pekarangan, the families use the existing pekarangan less and less for
Technology, Okayama University, Japan, March 1998, at Appendix Table 2 (citing 1995
Housing and Settlement Statistics, Indonesian Statistics Center Bureau, 1996).
23 E. Rajagukguk, “Agrarian Law, Land Tenure and Subsistence in Java: Case Study of the
Villages of Sukoharjo and Medayu” (Ph.D. dissertation, University of Washington, 1988), at pp.
202-203.
9
gardening, and the families instead build additional houses on the pekarangan
for sons and daughters who begin their own families.24
To reduce the poverty of agricultural laborer households, a land reform
program could be designed in which the Government would: (1) purchase small
amounts of sawah or tegalan located adjacent to the existing pekarangan land
in each Javanese village; (2) rezone the purchased sawah (tegalan) as
pekarangan; (3) divide the sawah (tegalan) into small plots; and (4) distribute
the new pekarangan plots to the poorest families in the village.
For example, in the village of Sukoharjo described above, if the government
decided to purchase enough sawah to distribute 200 m2 of pekarangan to the
719 households that now own pekarangan of 100 m2 or less, this would require
the government to purchase 14.38 ha.25 If the government decided to provide
200 m 2 to each of the 1002 village families, it would have to purchase 20 ha.
3. Recommendation regarding distribution of pekarangan
The figures in chart (iii) above illustrate how impractical it would be in the
twenty-first century to implement a “land reform” program that attempted to
redistribute even as much as 0.25 ha to all or most of the landless or land-poor
families on Java (or in Indonesia as a whole). (The figures in chart (iii) also
emphasize that the most significant lack of access of land is among families on
Java.)
For example, if the government were to attempt to distribute an average of 0.25
ha to just 50% of the 14,340,000 agricultural families on Java which chart (iii)
shows to be either farmer households who hold less than 0.5 ha of cropland or
agricultural laborer households who hold no cropland, this would require
1,792,000 ha (0.25 ha x 7,170,000 families). However, with only 5,135,000 ha
of cropland on all of Java, such a redistribution would require the government
to redistribute 35% of all cropland on Java. This is clearly an impossible task.
We may consider another possible model. Assume that the government
decided to redistribute an average of 0.10 ha (1000 m2) to each of the
6,732,000 agricultural laborer households on Java. This would require 673,000
ha, which is equal to 13% of all cropland on Java.26 This is somewhat easier to
imagine than the first example, but this would nevertheless be very difficult and
would require both strong political will and large financial resources.
24 See, for example, Hadi Susilo Arifin, supra note 22, at p. 85.
25 Village families cultivated a total of 232 ha of cropland, including 224 ha of sawah. See E.
Rajagukguk, supra note 23, at pp. 159-160.
26 This figure for percentage of cropland could be reduced if the Government were able to use
some of the 620,000 ha of Javanese agricultural estate land in the redistribution program.
However, the location of the agricultural estate land is an important consideration. To the
extent that the Javanese estate land is not located near where the Javanese agricultural
laborer families presently live, the Government would have to implement a difficult and
expensive program of resettlement within Java.
10
a. One possible approach
In order to ensure that land reform benefits the most vulnerable population, the
Government of Indonesia should perhaps consider a more targeted program to
provide basic pekarangan to the poorest agricultural households on Java.
According to this concept, assume the Government acquires Javanese sawah
(or tegalan) in appropriate locations, converts such land to pekarangan and
redistributes 0.03 ha (300 m2 ) of pekarangan to the poorest 5 million
households in rural Java. Extensive published scholarly research on
pekarangan in Indonesia and comparable house-and-garden plots in other
countries shows that garden plots can make a significant contribution to the
nutrition, income, status and overall well being of agricultural laborers and other
poor rural households.27 Such plots also provide special benefits to women,
giving them a place close to home to garden and perform other economic
activities -- such as tending animals, engaging in home industries, etc. -- that
can provide them with an important source of independent income.28 On Java,
research has shown that pekarangan produces 44% of total food calories, 32%
of total proteins and 65% of fuel consumed by rural households.29 If the land is
instead used for intensive production of grain, research has found that irrigated
plots as large as 400 m2 can yield enough paddy rice to provide 40 - 60% of the
basic yearly nutritional requirements for a family of four persons.30
We thus expect that the distribution of such pekarangan plots to landless
families could produce enormous economic and social benefits, even though
the land requirements for such a land reform program would be extremely
27 See, e.g., R. Marsh, “Building on Traditional Gardening to Improve Household Food
Security,” in Food, Nutrition and Agriculture, No. 22 (FAO 1998) at 1; Anne Stoler, “Garden Use
and Household Economy in Java,” in AGRICULTURAL AND RURAL DEVELOPMENT IN INDONESIA
(Gary E. Hansen ed. 1981) at 244; Vera Ninez, “Introduction: Household Gardens and Small-
Scale Food Production,” in FOOD AND NUTRITION BULLETIN, Vol. 7, No. 3 (UN University Press,
Sept. 1985) at 2; Otto Soemarwoto, et al., “The Javanese Homegarden as an Integrated Agro-
Economic System,” in FOOD AND NUTRITION BULLETIN, Vol. 7, No. 3 (UN University Press, Sept.
1985) at 4; Keith Rosen, “Puerto Rican Land Reform: The History of an Instructive Experiment,”
73(2) YALE LAW JOURNAL 344 (1963); G.J.A. Terra, “Mixed-Garden Horticulture in Java,” in
MALAYSIAN JOURNAL OF TROPICAL GEOGRAPHY, Vol. 3 (Oct. 1954) at 36; William Thiesenhusen,
Tim Hanstad, Robert Mitchell and Erman Rajagukguk, “Land Tenure Issues in Indonesia”
(report prepared for USAID, 1997). See generally Hadi Susilo Arifin, supra note 22; Tim
Hanstad, Jen Brown and Roy Prosterman, “Larger Homestead Plots as Land Reform?
International Experience and Analysis From Karnataka, RDI REPORTS ON FOREIGN AID AND
DEVELOPMENT NO. 113 (Aug. 2001).
28 See, e.g., DAVID BORNSTEIN, THE PRICE OF A DREAM: THE STORY OF THE GRAMEEN BANK AND
THE IDEA THAT IS HELPING THE POOR TO CHANGE THEIR LIVES (1996) at 152 -154; NATIONAL
INSTITUTE OF ADVANCED STUDIES (NIAS), S TATUS OF RURAL WOMEN IN KARNATAKA (1998) at 21.
29 Soemarwoto, supra note 27; Leslie Brownrigg, “Home Gardening in International
Development: What the Literature Shows, Including an Annotated Bibliography and Inventories
of International Organizations Involved in Home Gardening Projects” (League for International
Food Education, 1985) at ii.
30 William Thiesenhusen, Tim Hanstad, Robert Mitchell and Erman Rajagukguk, “Land Tenure
Issues in Indonesia” (report for USAID, 1997) at 38 n. 66.
11
modest. If the Government distributed 0.03 ha of pekarangan land to 5 million
households on Java, it would have to acquire only 150,000 ha of land, or only
3% of Java’s cropland.
If the Government program is based upon conversion of sawah to pekarangan
and the Government pays an average of Rp. 100 million (US $ 10,000) per ha
for sawah,31 the acquisition of 150,000 ha would cost approximately Rp. 15
trillion (US $ 1.5 billion). If the Government implemented such a program over
5 years, the annual cost for the land would be Rp. 3 trillion (US $ 300 million).
Under this plan, the cost of land per family benefited would be approximately
Rp. 3 million (US $ 300), which is about 1/10th of the per family cost for the
transmigration program.32 In addition, this scheme would require conversion of
only 3% of the total cropland (exclusive of agricultural estates) on Java.33
Table (e). Summary of pekarangan land reform scheme for Java.
Number of
families
receiving
land on Java
Size of
land for
each
family (ha)
Total land
required (ha)
% of
cropland
on Java
Total cost of land Cost of land per
family
5 million 0.03 ha
(300 m2)
150,000 3% Rp 15 trillion
($ 1.5 billion)
Rp 3 million
($ 300)
This land reform proposal could be based on the following principles:
First, the land reform should give top priority to the poorest households, which
are likely to be the seasonal agricultural laborer households that work for
various farmers and have no permanent employment. In the earlier land
reform, Article 8 of PP 224 of 1961 placed this class of agricultural family at the
bottom of the priority list.
Second, the land should be distributed in small plots in order to benefit the
maximum number of agricultural families. If the land reform described here
were conducted on Java, it would potentially affect 75% of agricultural laborer
families on Java.
Third, the Government should pay market value for the sawah (or tegalan)
acquired for conversion into pekarangan. The land reform program to date has
31 Based on BPN estimates and results of RDI April 2000 fieldwork in East and Central Java,
the equivalent of $10,000 per ha is a reasonable price for sawah.
32 These costs do not include the costs of administration of the land reform program, which
should be a small fraction of the cost of acquiring the land. There would also be small costs for
extending a simple road to the new pekarangan settlement (which should adjoin the existing
settlement), and costs for other basic infrastructure. These costs would increase the total cost
of the program to a range of Rp. 4 - 5 million per family receiving land.
33 If the land reform program acquired tanah bengkok rather than privately owned sawah, the
program might be less controversial. In that case, the Government would have to provide an
alternate source for salaries of village officials. See discussion in sub -section (c) below.
12
attempted to acquire land by paying less than market value, which has
generated vigorous opposition to the reform.34
Fourth, the Government must ensure that there is adequate financing for the
land reform. To date, the State has not provided adequate financing for land
reform, and the Government has paid for only a portion of the land acquired for
redistribution.35
Fifth, as part of program monitoring and administration, the Government must
ensure that it redistributes all land that it acquires.36
Sixth, the Government should not focus on excess land and absentee land, but
should instead purchase land near the village. If necessary, the law can be
amended to define the land reform as a “public purpose” so that the
Government can acquire the sawah (or tegalan) by compulsion under the laws
related to acquisition of land for public purposes. Because the new
pekarangan should be located close to the existing village, the Government
should acquire the sawah (tegalan) near the village. For this reason, it is not
necessary to acquire such land only from excess owners or absentee owners.
The Government should acquire the land located near the village, regardless of
the identity of the owner or the size of the owner’s total landholding.
Seventh, the land reform program should be focused on Java because that is
where the poorest agricultural households are most heavily concentrated. As
noted above, if the land reform program distributed pekarangan plots to 5
million agricultural families in Java, it could benefit three-quarters of Javanese
agricultural laborer families.37 Five million families would represent
approximately six times the 816,000 households on Java that benefited from
the previous land reform.
34 The original formula for compensation was reasonably adequate for the first 5 ha: it provided
for paying 10 times the annual net production, with gradually decreasing payments for larger
holdings. Annual net production was defined as half the gross production for sawah. PP No.
224 of 1961, art. 6. These payments were several times larger than the payments provided in
the Taiwan and South Korean land reform laws. E. Rajagukguk, supra note 23, at page 90.
However, an inflation rate of 650% during 1963 - 1965 caused the payment to be practically
worthless. E. Rajagukguk at page 96. Moreover, it appears likely that the Government did not
pay anything for a substantial part of the acquired land. See note 35 infra.
35 Although the 134,558 ha for which the Government paid is only 9% of the 1,470,690 ha
acquired for redistribution in Indonesia (according to data of BPN Directorate of Land
Arrangement and Land Tenure, March 2000), it is not clear how much of the total was acquired
from landowners (as opposed to state land, former land of princes, etc.).
36 Of the 1,470,690 ha the Governm ent acquired in the 1960’s, it apparently distributed only
850,000 ha, which is 58% of the total acquired. It is not clear how that undistributed land is
being used today. Perhaps the Government has granted it to plantation companies under
HGU, or perhaps state enterprises use the land.
37 This assumes that all, or almost all, of the beneficiaries are agricultural laborer families.
13
Chart (v). Javanese agricultural laborer families affected
under proposed pekarangan distribution
75% of
Javanese
agricultural
laborer families
receive land
Eighth, the land reform program must be streamlined to avoid unnecessary
processes and approvals. In the earlier land reform program, there were at
least 15 institutions that had to approve applications to receive land, which
indicates an enormously wasteful process.38
Ninth, the land reform program must be carefully monitored through
independent sample surveys and Rapid Rural Appraisal interviews39 to ensure
that it meets the program’s specific goals. In this kind of program, and to the
extent that any donor agency provides financing, it is entirely possible to use
the funding only as the program actually achieves results. According to this
approach, the financing should be released on a “progress payments” basis as
monitoring confirms that the progra m has achieved each increment of progress
in distributing land to families who are in the intended beneficiary group.40
38 E. Rajagukguk, supra note 23, at page 94.
39 In Rapid Rural Appraisal (RRA) interviews, rural interviewees are not responding to a
questionnaire, but actively participate in a semi-structured interview. The researchers use a
checklist of issues as a basis for questions, not necessarily addressing all questions in each
interview and som etimes departing from the basic questions to pursue interesting, unexpected,
or new information.
40 The Government could perhaps consider a second approach for distributing new pekarangan
land, which would be a “universal” distribution of additional pekaran gan land to rural families.
Under this approach, the Government would distribute new pekarangan plots averaging
perhaps 200 m2 to all, or nearly all, rural households on Java. Such an approach would take
into account several additional factors that might make it a viable alternative to the program just
described:
(1) This approach would eliminate the necessity of ensuring that an approach targeted to a
smaller number of families -- for example, the “5 million” suggested above -- actually reached
the poorest households and that the benefits were not diverted or “intercepted” by families who
were wealthier or more powerful.
(2) This approach would take into account the fact that pekarangan can provide unique
nutritional and ecological functions even for hous eholds that are somewhat better off, but that
most of the essential open area of even medium-sized pekarangan plots has been lost in
recent years due to widespread fragmentation through inheritance. Such fragmentation leads
14
b. An additional important variable: tanah bengkok
Under either of the models for land reform described in sub-sections (a) and (b)
above, there is another important variable that might significantly reduce the
amount the Government pays to acquire land for redistribution, and that might
facilitate the land reform program in other ways. That variable is the possible
use of land curre ntly designated as tanah bengkok (village salary land) as part
of the land redistributed in the land reform.41 It has been estimated that such
land may constitute 10% or more of the cropland in Central Java and East
Java, which would amount to roughly 300,000 ha of cropland. Tanah bengkok
probably exists in a very large majority of villages in those two provinces, and
such land is usually of good quality and located near the village.
If the Government undertook a land reform program of the type described in
sub-section (a), and if the Government acquired and used even one-quarter or
one-third of existing tanah bengkok (equivalent to roughly 75,000 - 100,000 ha
of existing tanah bengkok), this might significantly reduce the cost of such a
program. The appropriate price that the Government should pay for tanah
bengkok is probably the replacement of the flow of income produced by such
land currently (and used to pay village officials). The Government could pay
that price by opening a special trust account in a local bank in the name of the
village. The Government would deposit into the account an amount sufficient
to generate the same flow of income that the tanah bengkok currently
produces. It is very likely that the deposited amount -- the capitalized valu e of
the net income to be expected from the tanah bengkok taken, or perhaps the
capitalized value of the rent that a village official would have received by
leasing out such land -- would be significantly less than the full market value of
equivalent land held by a private owner.
Thus, the use of tanah bengkok might substantially cut land costs for the
program. It should also be possible for the Government to acquire such land in
a way that would minimize disruption or opposition from village officials. This
could be done by identifying tanah bengkok that village officials are leasing out
to building multiple houses on what had been a single parcel of pekarangan. See Hadi Susilo
Arifin, supra note 22, at p. 85.
(3) Virtually by definition, such a program would reach all of the rural poor on Java.
If the Government were to distribute pekarangan to all 17.3 million agricultural households on
Java, the program would require 346,000 ha (200 m2 x 17,300,000 families). This would
represent about 6.7% of the cropland on Java. While this would require more land and would
be more expensive than the alternative described in sub-section (a), it is possible that such
increased costs would be outweighed by the popular support for a universal distribution. Such
a program would require approximately Rp. 34.6 trillion (US $ 3.46 billion) for purchase of land,
or about Rp. 7 trillion a year if the Government carried out the program over 5 years. Cost of
land per family benefited would be about Rp. 2 million (US $200).
41 See E. Rajagukguk, supra note 23, at pp. 271-72. Other types of suitable non-private land
(suitable with respect to quality and location) might also be available in some circumstances.
15
(as is the case for most tanah bengkok) and then paying the official from the
trust fund an annual amount equal to the amount they receive from the lease.
The Government should also pay the existing tenant on such land some
amount in compensation for disturbing the lease agreement.
c. Contents of the land reform program
Regardless whether the Government adopts the land reform model described
in sub-section (a) above or chooses to adopt some other model, it is important
for the Government to select some actual model for land reform. To speak of
“land reform” in the abstract, without reference to an actual program, may be
counterproductive since it creates the illusion that land reform may be realized
in the near future without taking concrete steps to achieve the reform.
The Government’s response to MPR Provision No. IX of November 9, 2001
should consider formulating a land reform program that offers proposals on
each of the following issues:
(a) what kind of people should receive land;
(b) where in Indonesia will land reform be conducted;
(c) how much land will each recipient receive;
(d) approximately how much land is available for redistribution
from the various sources;
(e) what will be the method for identifying the land to be
redistributed;
(f) what will be the cost for acquiring such land (in each
category, if several categories of land are involved);
(g) how will the Government pay for the land acquired;
(h) over how many years will the Government implement the
program;
(i) who will make decisions regarding who receives land in each
village, and how will they make such decisions;
(j) will the State require recipients to pay for the land received,
and on what te rms (level of payment, over what term, etc.);
(k) will the law impose restrictions on the recipients' use of the
land or forbid the resale of the land for some period; and
(l) what will be the roles of the kabupaten governments and
village heads in the la nd reform.