Agricultural mechanisation: adoption patterns and impact

Full text

Chapter 54
AGRICULTURAL MECHANIZATION: ADOPTION PATTERNS
AND ECONOMIC IMPACT
PRABHU PINGALI
Agricultural and Development Economics Division, FAO, United Nations, Rome, Italy
Contents
Abstract 2780
Keywords 2780
1. Introduction 2781
2. Trends and patterns in agricultural mechanization 2782
2.1. Power tillers/tractors 2783
2.1.1. Asia 2784
2.1.2. Africa 2785
2.2. Milling and other post-harvest operations 2787
2.3. Harvesting and threshing operations 2788
2.4. Labor substitution for “control-intensive” operations 2790
3. Impacts of agricultural mechanization 2792
3.1. Land preparation 2792
3.1.1. Aggregate area expansion 2794
3.1.2. Power bottleneck 2795
3.1.3. Total labor use 2795
3.1.4. Income transfer 2795
3.2. Mechanization of post-harvest operations 2796
3.2.1. Threshers 2796
3.2.2. Milling 2798
3.2.3. Herbicides 2798
4. Implications for mechanization policy 2799
4.1. Tractors are a poor instrument for stimulating agricultural growth 2799
4.2. Agricultural mechanization policy ought to be seen within the context of an overall agricul-
tural growth strategy 2799
4.3. The demand for motorizing power intensive operations, such as tillage and threshing, is
closely associated with the intensification of farming systems, while the mechanization of
control-intensive operations, such as weeding, is driven by rising real wages 2800
The assistance of Hideyuki Nakagawa and Annelies Deuss is gratefully acknowledged.
Handbook of Agricultural Economics, Volume 3
Edited by Robert Evenson and Prabhu Pingali
©2007 Elsevier B.V. All rights reserved
DOI: 10.1016/S1574-0072(06)03054-4

2780 P. Pingali
4.4. Promotion of small stationary machines for power-intensive operations such as milling and
pumping can have significant benefits for the poor 2800
4.5. Clearly established property rights could minimize the risk of displacement of small farmers
from their land 2801
4.6. Adoption of labor saving technology does not always imply labor displacement 2801
4.7. Public sector run tractor promotion projects, including tractor-hire operations, have neither
been successful nor equitable 2801
4.8. Alleviating supply side constraints to mechanization is important, but only where the de-
mand conditions are right and the enabling environment is in place 2802
4.9. Conservation agriculture is not a panacea for farming systems that are not mechanized
today 2802
4.10. Global integration of food and input markets can have positive as well as negative conse-
quences for small farm mechanization 2803
References 2803
Abstract
Over the past half a century developing regions, with the exception of Sub-Saharan
Africa, have seen labor-saving technologies adopted at unprecedented levels. Intensi-
fication of production systems created power bottlenecks around the land preparation,
harvesting and threshing operations. Alleviating the power bottlenecks with the adop-
tion of mechanical technologies helped enhance agricultural productivity and lowered
the unit cost of crop production even in the densely populated countries of Asia. Eco-
nomic growth and the commercialization of agricultural systems is leading to further
mechanization of agricultural systems in Asia and Latin America. Sub-Saharan Africa
continues to have very low levels of mechanization and available data indicate declining
rather than increasing levels of adoption, even among the countries that were the early
trendsetters, such as Kenya and Zimbabwe. This chapter documents the trends and se-
quential patterns in the adoption of mechanical technology, assesses the evidence on
the productivity and equity impact of mechanization, and discusses the implication for
mechanization policy.
Keywords
agricultural mechanization, adoption, labor productivity, impact
JEL classification: O13, O14, O31, O32, O38

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2781
1. Introduction
Over the past half a century developing regions, with the exception of Sub-Saharan
Africa, have seen labor-saving technologies adopted at unprecedented levels. Intensi-
fication of production systems created power bottlenecks around the land preparation,
harvesting and threshing operations. Alleviating the power bottlenecks with the adop-
tion of mechanical technologies helped enhance agricultural productivity and lowered
the unit cost of crop production even in the densely populated countries of Asia. Mech-
anization of agricultural operations was very selective and sequential; power-intensive
operations such as land preparation, threshing and milling were readily mechanized.
While operations that require more human judgment, such as weeding, continued to be
done by hand under low wage conditions.
Economic growth and the commercialization of agricultural systems is leading to
further mechanization of agricultural systems in Asia and Latin America. The advanced
countries of East Asia have a completely mechanized rice production system, while
the rapidly growing countries of Southeast Asia are moving in that direction [Pingali
(1998)]. Middle income countries of Latin America, such as Brazil, Chile and Mex-
ico, are observing a similar rapid shift to labor-saving technologies, both mechanical as
well as chemical. Conservation tillage in association with herbicide use has resulted in
significant cost savings in cereal production systems in Brazil and Argentina [Ekboir
(2000)]. Also in Latin America, vertical integration of the post-harvest processing in-
dustry is leading to the replacement of small village-based post-harvest facilities with
large-scale processing plants [Balsevich et al. (2003)].
Sub-Saharan Africa continues to have very low levels of mechanization and avail-
able data indicate declining rather than increasing levels of adoption, even among the
countries that were the early trendsetters, such as Kenya and Zimbabwe. The persistent
low levels of mechanization in relatively land abundant Sub-Saharan Africa has been a
longstanding puzzle in the literature on agricultural mechanization [Pingali, Bigot and
Binswanger (1987)]. The explanation is in the driving forces of agricultural intensifica-
tion and the incentives for increasing productivity growth. Agricultural areas facing rel-
atively inelastic demand conditions, due to low population densities and/or poor market
infrastructure, tend to persist in low intensity, low yield subsistence production systems.
The move to mechanical technologies for land preparation is not cost-effective in such
societies. Attempts to expand the area under cultivation and to modernize agriculture by
bringing tractors into such areas have consistently failed. Tractors by themselves are not
an effective tool for inducing the process of agricultural intensification and productivity
growth.
The critics of mechanization have argued that the widespread use of labor-saving
technologies has had serious equity consequences in terms of the displacement of labor
and tenant farmers. Existing evidence indicates, however, that the equity consequences
have not been as severe or as widespread as they are presumed to be. The mechanization
of power-intensive operations have had minimal equity effects even in the labor surplus
economies of Asia. The switch from manual labor to mechanical or chemical technolo-

2782 P. Pingali
gies for control-intensive operations, such as weeding, has had adverse equity effects in
low-wage countries. However, where markets have been allowed to function with mini-
mal government intervention, control-intensive operations continue to be performed by
human labor until wages rise due to increased labor withdrawal from the agricultural
sector. Serious equity consequences are invariably associated with policies that inap-
propriately promote mechanization, such as subsidized credit for tractor purchase.
This chapter documents the trends and sequential patterns in the adoption of mechan-
ical technology (Section 2), assesses the evidence on the productivity and equity impact
of mechanization (Section 3), and discusses the implication for mechanization policy
(Section 4).
2. Trends and patterns in agricultural mechanization
Agricultural operations can be grouped according to the relative intensity with which
they require power, or energy, in relation to the control functions of the human mind, or
judgment. Operations such as land preparation, transport, pumping, milling, grinding,
and threshing are power intensive, while weeding, sifting, winnowing, and fruit harvest-
ing, for example, are control-intensive operations. The shift of the source of power from
human to animal to mechanical power is dependent on the level of power intensity and
control intensity of the operation [Binswanger (1984)]. Table 1 offers a comparison of
operations according to their power intensity and control intensity and the sequence of
their transfer to the new power source.
In both land scarce and land-abundant economies the power intensive operations are
the ones to be mechanized first, while the mechanization of control intensive opera-
tions occurs much later and is closely associated with the wage rate. The sequence, in
which power intensive operations are mechanized, however, differs according to land
Tab le 1
Comparison of agricultural operations according to their power and control intensity
Nature of operation
and source of power
Low control intensity,
high power intensity
Intermediate intensity High control intensity,
low power intensity
Stationary operations Grinding, milling,
crushing
Water lifting
Threshing, wood cutting
Sifting, winnowing
Mobile operations Transport
Primary tillage
Harvesting root crops
Harvesting grain crops
Secondary tillage and
interculture
Weeding and harvesting
tea, coffee, and apples
Seeding
Source:Pingali and Binswanger (1987).

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2783
endowments. In land-scarce economies the pumping operation is generally the first to
be mechanized, using diesel and electric pumps. Pumps provide a supply of water for
double or triple cropping and allow the expansion of cultivation onto marginal lands.
Mechanical mills, tillage, and transport equipment follow the adoption of pumps. In
land-abundant economies, the use of pumps is delayed until conditions of land scarcity
emerge; milling and transport operations are the first to be mechanized, followed by
tillage where its mechanization is feasible. The threshing operation, although power in-
tensive, is generally not mechanized where wages are low and harvested volumes are
small. Even when land is abundant, therefore, if agricultural production is mainly for
subsistence, threshing is the last of the power intensive operations to be mechanized.
The mechanization of power-intensive operations has taken place rapidly even in
countries with high population densities and low wages, such as India, Bangladesh, and
the Philippines [Herdt (1983);Pingali and Binswanger (1987)]. Mechanization reduced
the costs of power-intensive operations significantly as well as ensured their timely
completion. Mechanization of control intensive operations is more closely associated
with the wage rate. In land-scarce economies in which nonagricultural demand for labor
is low, operations such as weeding, interculture, and harvesting continue to be done by
human and animal power. Cultivators became prevalent in Japan during the late 1950s,
when agricultural wages rose in response to rapid post-war industrialization. It was
only in the 1970s that rice transplanters and harvest combines began to be used in Japan
[Ohkawa, Shinohara and Umemura (1965)]. In India, and the Philippines, where tractors
do tillage and transport, interculture continues to be done by hand and animal-drawn
equipment, while harvesting is done only by hand. In land-abundant economies where
market opportunities are good and the wage rates are high, many of the control-intensive
operations are transferred to mechanical power. This selective nature of the adoption of a
new source of power – was also observed in Europe and the United States [Hurt (1982)].
Numerous examples of the sequential adoption of mechanical technologies, for both the
developed as well as the developing world can be found in Binswanger (1984),Pingali,
Bigot and Binswanger (1987), and Pingali (1998).
2.1. Power tillers/tractors
The levels of adoption of mechanical equipment for land preparation vary significantly
by continent, by crop and by farming system. Figure 1 provides a comparison across
the three continents of the trends in tractor adoption since 1960 using data from FAO.
Intensively cultivated lands of Asia and Latin America have experienced significant
levels of mechanization, while adoption rates in Africa, especially Sub-Saharan Africa
are extremely low. In fact, one observes a reversal in tractor adoption among some
African countries that were thought to be early adopters. In 2002 average tractor use in
Sub-Saharan Africa was around 1.3 per 1000 hectares of cultivated land, while in South
Asia it was around 9.1 and in Latin America it was around 10.4 tractors for the same
time period. Tractor use in Sub-Saharan Africa peaked at 1.9 per 1000 hectares in 1986
and has gradually declined since then. Permanent cultivation systems under grain crops,

2784 P. Pingali
Figure 1. Number tractors per 1000 hectare of crop land – by region. Source:FAO (2005).
such as rice, wheat and maize, tend to have higher levels of mechanization relative to
less intensively cultivated systems under root and tuber crops.
2.1.1. Asia
In the case of Asia one needs to make a distinction between power tiller and tractor use.
Intensive wetland rice production systems in East and Southeast Asia have witnessed a
switch from animal drawn plows to power tillers. While four wheel tractors are com-
monly used for non-rice crops and for dryland environments. South Asia on the other
hand has relied much more extensively on four wheel tractors, although power tiller
numbers are rising in Bangladesh [Salokhe and Ramalingam (1998);Hossain, Bose and
Mustafi (2002)].
Japan and Korea led the rest of Asia in the speed and extent of mechanization and set
a pattern that other countries followed. In Japan by 1960, the mechanization of pumping
and threshing had already been completed, and the use of power tillers had just started
to take off. The number of power tillers on Japanese farms grew from 750,000 units in
1960 to 2.5 million units in 1965 [Kisu (1983)]. By 1965 there was one power tiller
for every 2 ha of crop land in Japan [Herdt (1983)]; by 1989, Japan had more than
one power tiller per hectare of riceland [Mizuno (1991)]. The Korean experience was
similar: number of power tillers rose from a little more than 1000 in 1965 to around
290,000 by 1980 [Cho (1983)]. By 1970, Korea had approximately one power tiller for
every 10 ha of riceland [Herdt (1983)], and by 1989, one power tiller for 2 ha of riceland
[APO (1991)]. The process of decollectivization in China has led to rapid mechanization
of farm operations using power tillers and other small machines. By 1992 China had

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2785
around 220 power tillers per thousand hectares of riceland [CSO (1993)]. In 1994, 71%
of the farm machinery belonged to individuals, 10% to the state, and 19% to collectives
[Salokhe and Ramalingam (1998)].
In Southeast Asia, mechanization of ricelands took off in the early 1980s. Power tiller
use rose from 26 and 14 per thousand ha of riceland in Thailand and the Philippines,
respectively, in 1980 [Herdt (1983)] to approximately 56 and 20 tillers per thousand ha
by 1990 [APO (1991)]. Indonesia, Myanmar, and Vietnam have been slower to switch
to power tillers for preparing ricelands, reporting less than one power tiller per thousand
hectares of rice lands in 1990. Although the situation has changed since then in Vietnam,
liberalization of the agriculture sector in the early 1990s has lead to a rapid increase in
the adoption of modern rice technologies, including power tillers [Pingali, Hossain and
Gerpacio (1997)]. Vietnam today, has the highest number of tractors per 1000 hectares
of cultivated land relative to other countries in Southeast Asia. For Thailand and the
Philippines, national average figure do not reflect the substantial variation in power tiller
adoption by rice environment. The irrigated rice bowl provinces of the two countries
tend to be highly mechanized, having more than one power tiller per 10 ha of land,
whereas the less favorable rice environments continue to rely on animal power. South
Asia is very diverse in terms of the mechanization of land preparation, both across
countries and within countries.
Although a superficial look at national average figures would seem to indicate that
South Asian countries continue to rely on animal draft power, the intensively cultivated
“Green Revolution” provinces, such as the Indian Punjab, tend to be on the same mech-
anization pathway as similar rice-growing regions in Southeast Asia [Pingali (1998)].
Mechanized land preparation is most advanced in Sri Lanka and least in Nepal. As
mentioned earlier, unlike Southeast Asian countries, in South Asia the mechanization
of land preparation has emphasized four-wheel tractors rather than power tillers. The
larger tractors are more conductive to the rice – non-rice crop rotations that are com-
mon in South Asia, more suitable for operating rental markets over a larger geographic
area, and more amenable to use as transport vehicles. Tractor numbers in India rose from
0.19 per 1000 hectares in 1961 to 9 per 1000 hectares by 2000, a level that is now at par
to that of other developing countries in East and Southeast Asia (Figure 1). Liberalized
agricultural equipment import policies in Bangladesh have lead to a dramatic increase in
small pump and power tiller use. The 1983–1984 Agricultural Census reported nonexis-
tence of farm machinery except irrigation equipment. The 1996 Agricultural Census, on
the other hand, reported ownership of 150,000 power tillers or tractors by 140,000 rural
households who constitute 1.2% farm households in the country [Hossain, Bose and
Mustafi (2002)]. Mandal (2002) estimates that since then around 15,000 power tillers
have been imported annually.
2.1.2. Africa
The first tractors appeared in Sub-Saharan Africa during the period between the wars.
They were used initially on settler farms and government-run farms. After 1945, how-
ever, tractors began to be used by African farmers also. Most of the postwar imports of

2786 P. Pingali
tractors were financed by the fund for farm machinery allocated through the Marshall
Plan. Since then bilateral and multilateral aid has been supporting the import of tractors
for agricultural purposes. Between 1945 and 1981 there were three significant spurts in
the number of tractors in use, with intermediate periods of slow growth [Pingali, Bigot
and Binswanger (1987)].
The first spurt came around 1945 under colonial influence, and the countries that
started to promote the use of tractors during the period 1945–1955 – Zimbabwe, Kenya,
Zambia, and Malawi – can be called the first generation of tractor users. The use of
tractors in these countries spread from colonial farms to private farms owned by native
Africans. The second spurt of tractorization came between 1958 and 1970 and it can be
characterized as state sponsored mechanization in some newly independent countries
such as – Tanzania, Ethiopia, Ghana, and Cote d’Ivoire. In many of these countries
tractors were provided through cooperative farms, state farms, or tractor-hire services.
The third spurt in tractor numbers took place between 1970–1980 when oil and other
resource exporting countries, such as Nigeria, Cameroon and the Democratic Republic
of Congo tried to re-distribute the gains from resource exports to rural areas. Tractor
were purchased and provided to farmers either through subsidized credit schemes or
through state sponsored hire-schemes [Pingali, Bigot and Binswanger (1987)].
In a number of countries throughout Sub-Saharan Africa, including Burkina Faso,
Niger, Rwanda, Burundi, and Liberia, there was never an organized effort to increase
the use of tractors, and the number of tractors is very small. As mentioned earlier, Sub-
Saharan Africa has, in fact, fewer tractors per thousand inhabitants than either Asia or
Latin America (Figure 1) and the numbers are declining even in countries that experi-
enced the early boost in tractor numbers. The use of tractors was and continues to be
restricted to a small commercial farm sector.
Pingali, Bigot and Binswanger (1987) summarize the experiences of tractor projects
in Sub-Saharan Africa from 1945 until the early 1980s (Table 2). They identified thirty
tractor projects, of which seven were smallholder projects, thirteen were tractor-hire
schemes, and ten were block cultivation schemes. Tractor-hire schemes are government-
sponsored rental programs for multifarm use of equipment. Block cultivation schemes
are group of farms being managed and operated as a single unit, often with mechaniza-
tion and other modern inputs. The following conclusions can be drawn from Table 2:In
many tractor project areas no tractors can be found today. Where any tractors are still
being used, their use is inevitably associated with rice cultivation. But even these surviv-
ing tractors today are privately owned. The transition from the hand hoe to animal-draft
power, where its use is appropriate, continued to be made despite the emphasis on trac-
tors. Of the seventeen attempts to bypass animal traction for tractorization only three
succeeded, all of them associated with low-land rice cultivation. None of the block
cultivation schemes has ever been successful. For an evaluation of the performance of
tractors in Sub-Saharan Africa, see Labrousse (1971),Bonnefond (1967), and Kline et
al. (1969).
Given that the rapid spread of mechanical equipment has historically been associ-
ated with an abundance of land, why has the spread of mechanization in Sub-Saharan

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2787
Tab le 2
Experiences with tractor projects in Sub-Saharan Africa
Number of projects or areas Individual
farm enterprise
Tractor-hire
service
Block cultivation
schemes
Initial number of projects 7 13 10
Number of areas in which tractors are still
used in the original organization form in
the 1980s
None 5 None
Number of areas in which tractors are now
used under private operation
None 4 2
Number of areas that had animal traction
originally
372
Number of areas in which animal traction
is in use in the 1980s
693
Number of areas in which continued use
of the tractor is associated with rice
cultivation
1a9 None
Source:Pingali and Binswanger (1987).
aTractor-hire scheme.
Africa been slower than in countries such as India and China, where labor is abundant
and wages are low? Pingali, Bigot and Binswanger (1987) explain the above puzzle
in terms of the lack of farm level incentives for the intensification of agricultural sys-
tems and hence a low demand for substituting out of existing power sources which in
Africa is mainly human. Intensification of agricultural systems is associated with rising
demand for agricultural products, triggered by growing populations and/or improved
access to domestic or international markets. Farming communities facing an inelastic
demand for their products, tend to persist in farming practices that are of low inten-
sity and low productivity. Power requirements in such systems are low and can easily
be met by human labor. As farming intensities increase, the number of tasks that need
to be performed increases as does the energy requirement for each of the tasks, hence
the adoption mechanical technology. Pingali, Bigot and Binswanger (1987) observe a
positive correlation between the evolution of farming systems and mechanization based
on the extensive field research in Sub-Saharan Africa. The numerous failed experiences
across Africa indicate clearly that tractors cannot be used as an instrument for driving
the process of intensification. Where the demand side factors are in place, agricultural
intensification and the adoption of mechanical power occurs in Africa in a similar pat-
tern to Asia and Latin America.
2.2. Milling and other post-harvest operations
Postharvest processing operations are extremely labor-intensive and tedious to perform.
Miracle (1967) reported, for instance, that to grind a week’s supply of maize meal –

2788 P. Pingali
thirty pounds – by hand would take from eight to fourteen hours. The same opera-
tion would take half an hour with a hand mill and perhaps not more than ten minutes
with a motorized mill. The same is true of dehusking rice, crushing sugarcane, grinding
groundnuts, and ginning cotton. In most parts of the world these operations have been
transferred to stationary machines powered by water, wind, steam, and – more recently
– internal combustion engines or electricity.
Water was first used for milling, pounding, and grinding in the first century B.C. in
China, and its use for these purposes was fairly widespread between the second and
fourth centuries A.D. Water-powered milling had also been adopted in all corners of
Europe by the twelfth century A.D. Wind power was used concurrently for milling and
lifting water in Europe. With the advent of steam power, mills were transferred to this
source of power in the nineteenth and early twentieth centuries in both Europe and the
United States. By the outbreak of the U.S. civil war, steam power had almost completely
replaced horses and oxen for powering sugar and rice mills and to turn cotton gins. With
steam power, three men and a cotton gin could remove the seeds from 1000 to 4500
pounds of cotton a day, which was about a hundred times the amount they could gin
without steam power [Hurt (1982, p. 101)].
In South Asia, animals have long driven Persian wheels, sugarcane crushers, and
oil crushers, but the animals used in these operations are gradually being replaced by
diesel and electric engines. In India in 1973 the number of stationary engines for power-
intensive operations was about twenty times the number of tractors (India 1975 and
1976). In all of Asia mechanical milling of large traded quantities of rice had already
been introduced in the late nineteenth century, usually by steam engines, later by internal
combustion engines. Smaller rice mills have swept across Asia since the 1950s, and it
is hard to find villages where rice is still pounded by hand.
Mechanical mills were introduced in Africa after World War I and spread rapidly
through the continent. There is documentation of the earlier existence of water mills in
Angola and Kenya [Manners (1962);Jones (1959)]. Jones (1959) and Miracle (1967)
have reported widespread use of mechanical mills, both hand and motorized, across
Sub-Saharan Africa. Pingali, Bigot and Binswanger (1987) provide evidence from Sub-
Saharan Africa indicating that low intensity of farming is not a constraint to the adoption
of mechanized mills. This is mainly because the labor input required for milling is in-
dependent of the intensity of farming, and mills are rarely owned by the individual
households who use them. The service is provided on a charge-per-unit basis by pri-
vate entrepreneurs or village cooperatives. Since mills do not face any of the timeliness
problems associated with plows, efficient rental markets are easily established and the
cost of the equipment is spread over many users.
2.3. Harvesting and threshing operations
The mechanization of harvesting and threshing tends to follow a two phase pattern. In
the first phase, harvesting continues to be done by hand, but threshing is increasing
conduct by mechanical means. The second phase is characterized by the adoption of

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2789
harvester combines that provide for the complete mechanization of both operations.
The first phase starts even under low wage conditions, but where peak season labor
scarcity emerges around the harvesting period due to growing off-farm employment
opportunities. Peak season labor scarcity problems are aggravated in areas where two
or more crops are grown on the same field each year. In this case, the labor peak for
harvesting blends into the labor peak for land preparation and seeding (transplanting)
the next crop. The switch to harvester-combines occurs as economies grow and rising
real agricultural wages are observed.
In this connection it is useful to examine technological change in threshing operation
in early U.S. agriculture as documented by Hurt (1982). Until 1850 colonial farmers,
particularly those in New England, used the hand-held flail to thresh grain from the
heads. The flail consisted of a short wooden club attached to a long handle by means
of a piece of leather. In the midwestern states, where harvested quantities were much
larger, farmers used oxen or horses to tread the grain from the heads.
The first horse-powered threshing machine was patented in the United States in 1791,
but it was only between 1820 and 1830 that a number of small, simple, inexpensive, and
locally made hand- and horse-powered threshing machines began to appear on the mar-
ket. Most farmers, however, found threshing with the flail to be cheaper than investing
in a threshing machine, because the work could be done during the winter, when there
was an abundance of cheap farm labor. It was only with the advent of contract thresh-
ing operations that mechanical threshing became profitable. Threshing machines were
owned by an entrepreneur, who sent a thresher with an itinerant crew from farm to farm.
Although contract threshing imposed an immediate cash expense on the farmer, it did
free him/her from the capital investment necessary to purchase a threshing machine and
enabled him/her to get the crops to the market before prices fell.
Steam-powered threshing machines preceded steam-powered tractors by more than
ten years. In less than a decade steam had almost entirely replaced horses for power. By
1880 the Bureau of the Census estimated that 80% of the grain in the principal wheat-
producing states was threshed by steam-powered machines. Steam-powered threshing
machines were followed in the 1930s by tractor-powered harvester-combines. The adop-
tion of threshers in Europe followed the same pattern. In 1950 agriculture in Japan was
in the early stages of mechanization, with many small pedal threshers and power tillers.
This happened in response to rising wages caused by a rapidly growing industrial sec-
tor. By 1960 there were one thresher per 2.5 ha and one power tiller per 12 ha in Japan
[Herdt (1983)].
The use of mechanical threshers did not emerge in South Asia and parts of Southeast
Asia until the late 1960s. this is not surprising, since wages were low, capital costs
high, and harvested volumes small. But where the green revolution raised wages and
increased harvested volumes, small threshers were rapidly adopted in Indian Punjab,
the Philippines, and Central Thailand as soon as efficient designs were available. By the
early 1980s the new threshers were penetrating into other South Asian regions [Wal ker
and Kshirsagar (1981)]. As in the United States in the nineteenth century, these threshers
are owned by private entrepreneurs who thresh on a contract-hire basis.

2790 P. Pingali
Mechanical threshers are still rare in African agriculture. Threshers are not yet prof-
itable in Africa because the harvested quantities per person are small. Therefore there
are two conditions under which threshers would be profitable in Africa: improvements
in access to markets or seed varieties leading to increases in harvested output and rapidly
rising wages, which would increase incentives for adopting labor-saving technology.
Harvester-combines have been in use in East Asia and in Malaysia for several decades
now and are emerging in Thailand, but their spread has been limited in the rest of Asia by
low harvest wages and small plot sizes. Although harvest wages are relatively low com-
pared with those in developed countries of East Asia, they are rising, and the prospects
are that over the next decade mechanization of the harvest operation will be demanded
in Southeast Asia.
Small harvesters are seen as an intermediate step in the transition to harvester-
combines in much of Southeast Asia. In the absence of land consolidation and the
re-design of riceland to form large contiguous fields, the prospects for large-scale adop-
tion of the harvester-combines are limited. Small harvesting machines have been used in
Malaysia since the mid-1980s; they began emerging in Thailand in the early 1990s and
are expected to be gradually available commercially in other Southeast Asian countries
[Pingali (1998)].
2.4. Labor substitution for “control-intensive” operations
As discussed earlier, the adoption of labor saving technologies for control intensive op-
erations, such as planting, weeding, etc., is only profitable as wages rise. In the 19th and
the first half of the 20th century mechanical alternatives were the only means available
for labor savings for these operations. However, since the later half of the last century
chemicals, such as herbicides, changes in crop establishment practices, such as direct
seeding, and knowledge intensive crop management practices, such as integrated pest
management, have emerged as alternatives to mechanical equipment for control inten-
sive operations. The following is a discussion of the attempts to seek labor savings for
control intensive operations in rice systems in Asia.
Rice transplanting and weeding operations are both control intensive as well as re-
quire high levels of labor, especially female labor, during a narrow window of time in
the crop cycle. Several attempts have been made to mechanize control-intensive opera-
tions in rice production systems since the Green Revolution [IRRI (1986)]. Mechanical
transplanters, weeders, and fertilizer applicators were some of the technologies released
by the rice research systems in Asia. With the exception of mechanical transplanters
(machine-driven) in East Asia, most attempts at mechanizing crop establishment and
crop care activities have failed. The failures can be attributed to the higher cost of
using the mechanical technology relative to the alternatives available, as discussed be-
low.
Since 1970 the use of mechanical rice transplanters spread rapidly across Japan,
Korea, and China (Taiwan), as wages rose. The use of transplanters in Japanese rice
production took off in 1970, and by 1989 there were 2.2 million transplanters in use

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2791
nationwide [Mizuno (1991)], approximately one transplanter for every 2 ha of arable
land. In Korea by 1989, approximately 66% of the area planted to rice was mechani-
cally transplanted [Chang (1991)]. Taiwan has seen similar rapid growth in transplanter
use since the late 1970s [Peng (1983)].
The alternative to mechanical transplanting, as a means of saving labor in crop estab-
lishment, is to broadcast pregerminated rice seedlings into the field. Direct seeding is
not generally possible in the temperate environments, however, because the cold spring
temperatures are not conducive to seedling establishment and growth in the field. Rice
seedlings are grown under controlled temperatures conditions and transplanted into the
field when they are old enough to resist cold stress; by this time, a few weeks in the
season, temperatures are warmer.
Mechanical transplanting has not taken off in the humid tropics of Southeast Asia,
even under high-wage conditions, such as Malaysia, because direct seeding is possi-
ble. Direct seeding eliminates the transplanting operation because germinated seed is
broadcast onto prepared (puddled) paddy soils. Most of the irrigated rice in Malaysia is
wet seeded. Wet and dry seeding methods are becoming increasingly popular in other
rice-producing countries of southeast Asia. Even though chemical weed control costs
are higher for direct-seeded rice, the savings in labor cost have been more than the ad-
ditional cost of herbicide. Empirical studies in the Philippines and Vietnam indicate
that farmers who use wet seeding relied more heavily on herbicides than farmers who
transplanted. Similarly, in Peninsular Malaysia, 95% of farmers who direct-seeded their
rice applied herbicides, at an average herbicide cost of $30/ha. On the other hand, only
46% of farmers who transplanted used herbicides, with average expenditure on herbi-
cides being $4/ha [Moody (1994);Pingali (1998)]. In low wage South Asian countries
manual transplanting of rice, using mainly female labor, persists.
Mechanical weeders were introduced into rice systems in the mid-1970s, but their
record of adoption was very poor, primarily because herbicides – a substantially cheaper
source of weed control – were available. While high-wage countries have used herbi-
cides for several decades, recent trends indicate increasing use in the so-called low-wage
countries of Asia (FAOSTAT). The Philippines, Thailand, Malaysia and India more than
doubled their herbicide imports in the 1980s [Pingali (1998)]. In these countries, herbi-
cide has been adopted in association with rice direct seeding, which in turn was adopted
in response to rising wages to replace the traditional transplanting system.
Attempts to enhance the efficiency of chemical fertilizer use through mechanical
technologies allowing deep placement of fertilizer have also generally not succeeded.
The long-term decline in global fertilizer prices has reduced farmers’ interest in improv-
ing fertilizer use efficiency and investing in equipment to do so. The poor performance
of fertilizer placement technology has been observed even in the high-income countries
of East Asia. Given the proliferation of nonmechanical options for improved crop man-
agement, the future for the adoption of machines for control-intensive operations seems
rather bleak.

2792 P. Pingali
3. Impacts of agricultural mechanization
The productivity impact of the switch to mechanical technologies for agricultural oper-
ations is measured in terms of changes in yields, labor savings, area expansion (in terms
of increases in cropping intensities), and quality of enhancement of the marketed output.
The equity impact, on the other hand, is measure in terms of labor displacement and in-
come distribution effects, particularly for the landless labor households and for women.
The productivity and equity impacts of mechanization vary depending on the power
intensity of the operation that is being mechanized, the region’s land and labor endow-
ments, and the country’s level of economic development. Since the mechanization of
power-intensive operations has been well under way throughout Asia, several studies
have documented the impact, and evidence from these studies is presented below. Few
studies of control-intensive operations and quality enhancement technologies are avail-
able because the introduction and adoption of these technologies has been sparse. The
power-intensive operations considered here are land preparation, threshing, harvesting
(small harvesters), and milling (small mechanical mills). The evidence presented below
indicates that, for power-intensive operations, the productivity benefits of mechaniza-
tion consist mainly of labor savings, and the equity implications are minimal, even in
labor-abundant, low-wage economies.
3.1. Land preparation
The movement from using animal-drawn plows to tractors or power tillers is considered
efficient if yield per hectare increases and/or if labor hours required for land prepa-
ration per hectare are reduced. Yield increases are possible only when mechanization
improves tilling quality. However, the available evidence indicates that generally no
significant yield difference exists between animal draft and tractor tillage. Herdt (1983)
found that yield differences between animal draft and tractor farms were negligible af-
ter differences in fertilizer use were considered (Table 3). This is consistent with results
from South Asia, where more than 50% of farms using tractors had significantly higher
yields, but in almost all cases these higher yields were associated with greater fertil-
izer use [Binswanger (1978)]. If we find no yield differences between animal draft and
tractor farms, we must conclude that the transition to tractor-drawn plows is rarely mo-
tivated by improvement in tillage quality. Area expansion and/or labor saving must be
the driving forces for such a transition.
Pingali, Bigot and Binswanger (1987) reviewed 24 studies on labor use by operation
on farms relying on animal draft power and farms relying on tractors in Asia. They
found changes in labor use by operation, in total labor use, and shifts in the levels of
labor use between operations. Twenty-two of the 24 studies reviewed reported lower
total labor use per hectare of crop production for tractor farms compared to animal draft
farms. Twelve studies reported reductions in labor use of 50% or more.
The greatest reduction in labor use was for land preparation, with all studies reporting
reduction in labor input exceeding 75%. It is instructive to consider cases in which the

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2793
Tab le 3
Summary of studies comparing rice yields on farms that used animal or hand for land preparation with farms
that used machinery
Author Area Comparison Reported
yield
(t/ha)
Fertilizer
(urea)
(kg/ha)
Adjusted
yield
(t/ha)
Pudasaini Nepal (without
pumps)
Bullock vs
tractors
1.7161.7
2.1 164 1.4
Pudasaini Nepal (with
pumps)
Bullock vs
tractors
2.1 214 2.1
2.3 264 2.1
Sinaga West Java,
Indonesia
(1979–80 wet
season)
Animal vs
tractors
4.9 323 4.9
4.9 323 4.9
Sinaga West Java,
Indonesia (3
seasons, 1978–80)
Manual vs
tractors
3.8 285 3.8
3.9 308 3.8
Tan and Wicks Nueva Ecija,
Philippines (1979
wet season)
Water buffalo vs
tractors
2.6892.6
4 129 3.8
Anuwat Central Thailand
(irrigated–
transplanted)
Bullock vs
tractors
2.6322.6
2.8482.6
Anuwat Central Thailand
(rainfed
broad-cast)
Bullock vs
tractors
0.230.2
0.220.2
Alam Bangladesh Bullock vs
power tiller
1.5n.a.1
.5
Deomampo
and Torres
Central Luzon,
Philippines
Before vs
after tractor
and tillers
2.2572.2
2.6792.1
Antiporta and
Deomampo
Philippine
provinces
Animals vs
tractors
tillers
2.6862.6
2.8 117 2.5
Source:Herdt (1983).
labor used for land preparation was reduced by 50% or more and trace the effect on
other operations. Consider weeding first. Of the 14 cases with 50% or higher reduction
in labor for land preparation, only two reported a reduction in weeding labor greater
than 25%. Ten reported reduction in weeding labor smaller than 25%, and two reported
increases in weeding labor relative to farms relying on animal traction. The situation
for harvesting is quite similar. Of the 14 cases with 50% or higher reduction in labor

2794 P. Pingali
for land preparation, only one resulted in an equal reduction in harvesting labor, nine
reported labor reduction less than 25%, and one found increased labor requirements.
These results indicate that labor savings resulting from the transition to tractors are
confined mainly to land preparation, where one observes a substantial reduction in labor
peak. However, where aggregate area expansion effects exist (including an increase in
cropping patterns), one could expect a rightward shift in the demand curve for weeding
and harvesting labor, despite the lower per hectare requirements. Rice cropping inten-
sities have increased significantly from the joint mechanization of land preparation and
threshing, especially in humid tropical Southeast Asia [Juarez and Pathnopas (1983)].
With the introduction of modern rice varieties and irrigation infrastructure in Southeast
Asia, two to three crops of rice can be grown on the same plot of land per year. Because
the first crop is usually harvested during the rainy months, the danger that grain would
spoil was very high if threshing was not done soon after harvest. Also, fields had to be
cleared of the previous crop before irrigation water could be released. Without the use
of power tillers and threshers, high intensity cropping could not have been sustained in
much of Southeast Asia.
What are the equity consequences of a shift to tractors/power tillers from animal-
drawn plows for land preparation? The answer to this question depends on the answers
to four related questions. First, did tractor owners expand the size of their operation by
displacing tenant farmers? Second, was there a power bottleneck around land prepara-
tion before tractors arrived? Third, has total labor use on tractor owners’ farms increased
or decreased since tractors arrived? Finally, has there been a net transfer of income from
tractor owners to agricultural labor? Each of these questions will be examined in turn.
3.1.1. Aggregate area expansion
Aggregate area expansion occurs only when private farms extend into fallow or re-
claimed land or when cropping intensity is increased on a given plot of land. Where
private area expansion occurs at the cost of other farmers (mainly through the displace-
ment of tenant farmers), there are no benefits from aggregate area expansion; rather,
adverse equity consequences are observed. Any inference on the aggregate expansion
in agricultural output as a result of an increase in cultivated land area brought about
by tractor use would require additional information on where the additional area comes
from.
Lockwood et al. (1983) found in Faisalabad, Pakistan, that 70% of the area expan-
sion on tractor farms came from tenant displacement. The remaining 30% came from
increasing area rented in, leading to further displacement of tenants. Eighty-eight of
the original 105 tenants lost their land when the landowners bought their first tractor.
The average size of tenant farm declined from 4.4 to 3.4 ha. This study confirms the
earlier findings in the same area of McInerney and Donaldson (1975).Jabbar, Bhuiyan
and Bari (1983) reported that 81% of power tiller owners in Mymensingh, Bangladesh
increased their cultivated area by expanding into previously rented-out land.

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2795
Where uncultivated or fallow land was not available, the promotion of tractors for
land preparation led to high levels of tenant displacement. The exception as noted earlier
is the growth in rice cropping intensity owing to mechanization, which did not lead to
tenant displacement and resulted in positive employment benefits [David and Otsuka
(1994)].
3.1.2. Power bottleneck
Mechanization of the land preparation can lead to significant productivity gains with
minimal equity effects in areas with a power bottleneck during land preparation. Such
bottleneck could occur because of shortage of either labor or draft animals. Labor short-
ages are most common in sparsely populated areas and in areas with good nonfarm
employment opportunities. Peak season labor shortages have also occurred with an
increase in rice cropping intensities, as discussed previously, around the harvest op-
eration for one crop and land preparation for the next. Shortage of draft power may
occur in sparsely or densely populated areas (in the latter case, shortages occur because
maintenance of draft animal is relatively expensive because of the high cost of fodder).
Significant productivity benefits accrue to the alleviation of labor and power bottle-
necks, through mechanization. The adoption of tractors (power tillers) in areas with a
power bottleneck during land preparation could lead to general improvement in income
levels.
3.1.3. Total labor use
It was observed earlier in this section that tractor use reduces labor use per hectare for
land preparation, but leads to an increase in area cultivated by tractor farms. Where
uncultivated or fallow land is available, or where cropping intensities increase, the use
of tractors (or power tillers) for land preparation could lead to an increase in labor
employment.
3.1.4. Income transfer
Income distributional effects of mechanized land preparation depend on the nature of
final demand and the extent of output growth. Consider first areas with an inelastic de-
mand for the food that is produced. These are typically small, closed economies, where
neither exports nor imports of food (in this case, rice) are allowed. In such economies
the total demand for food is determined only by domestic demand, and any increase in
output will lead to a decline in price.
Where final demand is inelastic, aggregate area expansion caused by mechanized
land preparation results in a transfer of income from tractor-owning households to
landless-labor households. This happens for two reasons: (i) expansion into previously
uncultivated areas (or an increase in cropping intensities) results in increased employ-
ment opportunities, and (ii) real price of food output declines as a result of output

2796 P. Pingali
expansion, and since the landless labor are net purchasers of food, the effect of this
price decline is analogous to increasing their income. The direction of income transfer
is reversed where the opportunities for aggregate area expansion are limited. Here, trac-
tor farms expand into land previously cultivated by tenants and the result is net labor
displacement rather than an increase in aggregate output. Where final demand is elastic
(open economies, or economies with large domestic demand), aggregate area expan-
sion does not lead to a decline in prices, but a limited amount of income transfer to
labor is observed as a result of expanded employment opportunities. If opportunities for
aggregate area expansion are not available, the effects are the same as discussed in the
preceding paragraph.
3.2. Mechanization of post-harvest operations
Next to land preparation, harvesting, threshing and milling are the most arduous op-
erations in rice production. Consequently, where mechanical technologies for these
operations exist and can be profitably used, they tend to be adopted fairly rapidly.
Small mechanical mills, for instance, spread spontaneously across the world without
any government program promoting them. What have been the productivity and equity
consequences of mechanizing post-harvest operations?
3.2.1. Threshers
The late 1970s and 1980s have seen the rapid spread of mechanical threshers in parts
of Southeast Asia and parts of India [Duff (1986);Walker and Kshirsagar (1981)]. In
any given area, the private profitability (efficiency) of using a mechanical thresher over
hand beating and animal or tractor treading is determined by yield benefits of mech-
anized threshings, marketable surplus generated on the farm, and by labor wages and
availability during the harvesting–threshing period.
Proponents of the thresher technology usually argue that the mechanical thresher
presents a significant increase in realized yields due to: (i) a more complete thresh-
ing of grain than manual or treading techniques; (ii) a reduction in losses caused by
repeated handling of both threshed and unthreshed materials; and (iii) an increase in
cropping intensity resulting from a lower turnaround time with mechanical thresher
use. On-farm experiments comparing manual and mechanical threshing have shown
that mechanical threshers reduce grain loss by 0.7 to 6% of total yield [Toquero and
Duff (1985)]. However, there have been no studies of actual farmer thresher use to see
if such savings are observed in practice. Cropping intensity effect has been discussed
above, and to the extent that threshers contribute to sustaining high cropping intensities,
thereby allowing more timely completion of operations they have a positive effect on
yields.
The existing evidence indicates that the primary motivation for mechanical thresher
use is the labor saving benefits. The adoption of a portable axial-flow thresher in the
Philippines resulted in a decline in threshing labor requirements from 7.69 labor days

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2797
per ton of grain for foot treading to 0.81 labor day per ton. In Thailand, the adoption
of a large axial-flow thresher resulted in labor savings of 3.50 labor days per ton of
grain relative to the traditional threshing by buffalo treading. Consequently, large gains
in labor productivity were observed in both cases [Duff (1986)].
Whether the above labor savings actually lead to efficiency gains or not depends
on the real wage of threshing labor, the timely availability of labor and other sources
of power for threshing, and the nature of final demand. The following generalization
is possible: Mechanical threshing is economically efficient where the farmer faces
an elastic demand curve for his output and a power bottleneck exists during the
harvesting–threshing period either due to high land–labor ratios or due to opportu-
nities for non-farm rural and urban employment. It is important to remember that a
combination of elastic demand and a threshing power bottleneck is required for me-
chanical threshers to be profitable. Equity considerations in thresher adoption center
around the source of the labor that is displaced. Surveys in Nueva Ecija, Philippines,
showed that postharvest labor on mechanized farms was 25% lower than on farms in
which rice was manually threshed. Disaggregated into family and hired labor, the data
reveal that much of the labor savings came at the cost of landless households whose
labor services declined by 31% [Duff (1986)]. A similar decline in the use of hired
labor was observed in Laguna and Iloilo, Philippines [Juarez and Pathnopas (1983)].
Ahammed and Herdt (1983) used a general equilibrium model to estimate the nation-
wide employment implications of thresher use (among other production methods) in
the Philippines. They found that a 1% increase in rice production would generate em-
ployment effect of 16,000 labor years if manual threshing is used. The employment
effect would be 22% lower (12,400 labor years) if portable threshers are used. Taking
all sectors of the economy into account, substituting mechanized threshing for manual
threshing would reduce the employment-generating potential of increased rice produc-
tion by 7%. This overall adjustment depends on urban employment opportunities, the
ability of labor to migrate between regions and between sectors, and industrial poli-
cies.
A similar analysis in Thailand found that all the labor savings came from family la-
bor and not for hired labor [Sukharomana (1983)]. This is because Thailand has a very
favorable land-to-labor ratio and therefore, a very small landless labor class. Tradition-
ally, the threshing operation is done by family labor with buffalo or tractor treading. The
use of large axial-flow threshers on a contract basis has resulted in the release of family
labor for other activities.
For both family and hired labor, the thresher had a differential impact on men,
women, and children. Mechanical threshers considerably reduced the arduousness of
post-production tasks. The lighter nature of the work made it possible for women and
children to substitute for men in the threshing operation [Ebron (1984)]. Where off-
farm employment opportunities exist, thresher use can result in increased incomes of
labor households since male workers may be released for other income generation while
women and children provide threshing labor.

2798 P. Pingali
3.2.2. Milling
The use of small mechanical mills for dehusking paddy or for pounding grain into
flour is perhaps the least controversial of all forms of mechanization. Handpounding of
grains, the traditional alternative to mechanical milling, is an extremely labor-intensive
task usually performed by women. The efficiency gain in the shift to mechanical mills
comes from the resulting labor savings. Because much of the rural milling is done
piece-meal for home consumption only and because traditional handpounding is done
by female household members, switching to small mechanical mills results in substan-
tial gains in leisure time for women. Consequently, one observes the widespread use of
small rural mills on a contractual basis.
With one significant exception, small mechanical mills have increased efficiency of
food production without adverse equity effects. The exception is Bangladesh where
significant displacement of hired labor has occurred as a consequence of mechanical
mills [IBRD (1987)]. Traditional rice milling in Bangladesh is done by a foot-operated
mortar and pestle known as a ‘dheki’, which is usually owned by large landowners and
operated on a contractual basis by women from landless and low-income households.
The ratio of milling costs with the ‘dheki’ to those with the mill is about 12:1, not
counting the transport costs of bringing the rice to the mill. Thus, the mill owners can
charge much lower rates for milling than ‘dheki’ operators.
The rapid spread of mechanized milling has benefited large landowner households,
subsistence-farming households, and urban consumers including the urban poor, whose
rice prices have been reduced. Female members of large surplus farms have more leisure
time, because they no longer have to supervise hired ‘dheki’ operators. Female members
of subsistence farms who previously operated the ‘dheki’ for their home consump-
tion are relieved from time-consuming and physically-demanding labor. Nonetheless,
women from landless families who work for wages have suffered as a result of the
mills because of absent alternative remunerative employment. The policy challenge in
Bangladesh has been to find alternative employment opportunities for sustaining the in-
comes of the women of landless and low-income households, without slowing down the
growth in small mechanical mills.
3.2.3. Herbicides
Throughout Asia, the ratio between price of herbicide and wage rate has been declin-
ing steadily overtime, making herbicide use economically attractive. In Iloilo province,
Philippines, the cost of weed control by herbicide in wet seeded rice is less than one-fifth
of the cost of a single hand weeding [Moody (1994)]. Similarly, in West Java, Indone-
sia, and Mekong Delta, Vietnam, the cost of hand weeding is 3 to 5 times the cost of
herbicides. Economic analysis of weed control practices in the Philippines indicate that
the marginal benefit cost ratio associated with herbicide usage is as high as 16 [Naylor
(1996)].

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2799
Although economic and technological factors are likely to lead to increased substitu-
tion of labor by herbicides, such a substitution may result in short term adverse social
consequences. Due to the need to complete weeding within a short time window, farm-
ers traditionally relied on hired labor for weeding. As small and marginal farmers are
the main source of hired labor, their income and employment will be adversely affected
unless they can find an alternative use of their time. Increase in cropping intensity made
possible by improved technology has in the past been able to absorb most of the dis-
placed labor [David and Otsuka (1994)]. The extent to which substitution of labor by
capital in agriculture will occur without imposing welfare costs on certain groups de-
pends mainly on government policies on exchange rate, pricing of inputs and outputs
and macro-economic policies. To the extent that herbicides are made cheaper relative
to labor due to distortionary price policies, substitution of herbicides for labor is so-
cially undesirable. The social costs are of course lower in societies experiencing a rapid
withdrawal of labor from the agricultural sector.
4. Implications for mechanization policy
4.1. Tractors are a poor instrument for stimulating agricultural growth
Over the last century there have been several dozen attempts at introducing tractors into
areas that are sparsely populated, particularly in Africa, in an attempt to rapidly expand
the area under cultivation and increase production. These attempts have consistently
failed because the market infrastructure and economic incentives that induce a produc-
tion response were not present. The end result was a “boom and bust cycle”, a rapid
expansion in output that was invariably followed by a collapse in local prices and a
subsequent abandonment of the newly opened land and the tractors. Tractors ought to
be seen for what they are: a potential tool in the production process and not a driver of
economic change.
4.2. Agricultural mechanization policy ought to be seen within the context of an
overall agricultural growth strategy
A broad based strategy for agricultural growth provides incentives and infrastructure
that enables farmers to enhance productivity growth. Macroeconomic policies that tend
to discriminate against or tax the agricultural sector need to given as much attention as
agriculture sector specific policies. Policies that enhance rural infrastructure are neces-
sary for sustainable productivity growth and overall rural development. Farmers respond
to improvements in incentives and market conditions by changing production practices
and investing in new technologies, including mechanical technologies. Broadbased rural
infrastructure development also reduces the costs to private entrepreneurs that supply
technology and inputs as well as market output.

2800 P. Pingali
4.3. The demand for motorizing power intensive operations, such as tillage and
threshing, is closely associated with the intensification of farming systems, while
the mechanization of control-intensive operations, such as weeding, is driven by
rising real wages
The need for increased energy requirements with the intensification of agricultural sys-
tem has been extensively documented in the literature [Boserup (1965);Pingali and
Binswanger (1987)]. The movement from land-extensive cultivation systems, such as
shifting cultivation systems, to land-intensive permanent cultivation systems increases
both the number of tasks performed and the intensity with which they are performed. For
operations that require high levels of power, such as tillage, human labor is gradually
replaced by animal and then tractor power. In intensively cultivated systems mecha-
nization of power intensive operations is profitable, even under low wage conditions. In
such systems, human labor continues to be used for seeding, weeding, crop care, and
harvesting. The co-existence of mechanical and human labor disappears as wages rise
due to economic growth and the increasing availability of off-farm employment oppor-
tunities. Mechanical seeders, herbicides, and harvester-combines substitute for human
labor as economies grow. The sequential adoption of mechanization, first for power in-
tensive and then for control intensive operations, is not a historical artefact, it is a farmer
response induced by the changing relative prices of factor inputs.
4.4. Promotion of small stationary machines for power-intensive operations such as
milling and pumping can have significant benefits for the poor
Small mills have spontaneously and successfully penetrated even the most remote vil-
lages around the developing world. Mechanization has released labor, invariably family
labor especially women from the arduous task of de-husking, pounding and milling
grain, often on a daily basis. Poor households benefit the most, since the released la-
bor can be reallocated for other income earning activities or for leisure. Governments
can play a catalytic role in the further spread of mills, first in promoting research and
development on mills that are more energy efficient and improve on quality; second in
providing credit and other support for rural entrepreneurs to acquire and operate rural
mills.
The adoption of small pumps is less spontaneous yet equally crucial for the livelihood
of poor rural households. Pumps help stabilize food supplies in drought prone lands, and
where enabling conditions exist, the commercialization of smallholder agriculture. The
adoption of pumps resulted in the intensification of cropping in the Indo-Gangetic Plains
that extend through Northern India and Bangladesh. Small holders were able to grow
a dry season crop of rice or vegetables exclusively for the market, hence stimulating
overall growth in the rural economy. Governments can play a similar catalytic role, as
with mills, in helping small farmers acquire pumps.

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2801
4.5. Clearly established property rights could minimize the risk of displacement of
small farmers from their land
In both land scarce as well as land abundant economies, tractor ownership is associated
with an expansion in area cultivated. In the absence of clearly established rights to land,
tractor-induced farm size expansion could come at a cost to the poor. In land scarce
economies, tractor adoption has resulted in the displacement of tenant farmers, while
in land abundant economies, traditional access rights to land have been impinged upon
as tractor farms expand into uncultivated or fallow lands. Inappropriate promotion of
mechanization, through subsidized tractor prices or cheap credit programs, tends to ag-
gravate the negative equity impacts without commiserate productivity benefits. Property
rights give small and tenant farmers the bargaining power to prevent encroachment or
to seek compensation. Formal land titles empower small farmers further by providing
them the collateral necessary for acquiring credit for the purchase of machinery and
other agricultural inputs.
4.6. Adoption of labor saving technology does not always imply labor displacement
Mechanization has often been seen as having a negative impact on agricultural em-
ployment and therefore detrimental for densely populated “labor surplus” countries. In
reality the picture is not as straightforward, whether labor displacement occurs or not
depends on: the operation, the labor market, and the policy environment. As discussed in
this chapter, mechanization of power intensive operations, water lifting, tillage, milling,
etc., have minimal labor displacement effects. The adoption of water pumps tend to in-
crease cropping intensity and hence labor requirements. Mechanized land preparation
shifts the demand for labor from land preparation to weeding and harvesting operations.
While mechanical mills, release female family labor from the arduous task of hand
pounding grain. The mechanization of control intensive operations such as weeding and
harvesting could have negative employment effects if promoted under low wage condi-
tions. However, if the adoption of labor saving technologies for these operations occurs
in response to rising wages, due to growth in the non-farm sector, then the labor dis-
placement consequences are small. Labor displacement problems are most pronounced
when mechanization policy is inappropriate and machines are promoted where they are
not required, such as low intensity farming systems, or where wage rates and the oppor-
tunity cost of labor are low.
4.7. Public sector run tractor promotion projects, including tractor-hire operations,
have neither been successful nor equitable
Public sector record as a promoter of tractor use and as a supplier or tractor services has
been poor both conceptually as well as operationally. The pervasive misconception that
tractors are a shortcut to agricultural modernization has resulted in the inappropriate
promotion of tractors in environments where private farmer decisions would not lead to

2802 P. Pingali
intensification or tractor use. Public sector tractor projects tend to displace the private
sector (or prevent its growth) as a supplier of equipment, spare parts and maintenance
services. Being donor driven, public sector projects do not build a self-sustaining system
for the long term supply of tractors and associated services. Hence the service collapses
at the end of the project. Public sector run tractor hire services are a particular case of
operational inefficiency and poor longevity. Where economic conditions are right, the
private sector has been an efficient provider of equipment and mechanization services.
The public sector can play an important catalytic role in promoting private sector sup-
ply as well as private initiatives in equipment research and development. The latter in
the particular context of unfavorable production environments and communities facing
special needs, such as AIDS affected populations.
4.8. Alleviating supply side constraints to mechanization is important, but only where
the demand conditions are right and the enabling environment is in place
Lack of or low level of adoption of mechanization is often attributed to supply side con-
straints, such as the lack of equipment and spare parts suppliers and skilled mechanics
that can provide maintenance services. However, its demand side factors, such as un-
favorable relative factor prices and market access conditions that are a more plausible
explanation of the poor spread of mechanical technologies, especially in Sub-Saharan
Africa. The private sector, where it has been allowed to operate freely and where the
enabling conditions are in place, has successfully met the demand for equipment and
spare parts, as well as for repair and maintenance. Governments ought to play a facili-
tating role in reducing the transactions costs involved, for farmers as well as for small
entrepreneurs, in the acquisition and maintenance of mechanical technologies.
4.9. Conservation agriculture is not a panacea for farming systems that are not
mechanized today
Conservation agriculture, which is generally taken to imply the systematic application
of planting without tillage, cover crops, and crop rotation, is seen by some as an oppor-
tunity for bypassing the need for mechanical power for land preparation. However, it is
completely false to link conservation agriculture with mechanization strategy. In mech-
anized conservation agriculture access to a tractor and to a no-till drill is required and
thus brings with it the same set of issues as for conventional cropping systems. In shift-
ing cultivation systems, the question of adopting conservation agriculture is moot since
the practices followed by these farmers, such as the incorporation of fallow vegetation
minimal land preparation and the use of a dibble stick for seeding are practices that
are consistent with principles of conservation agriculture. However, benefits have been
shown in hand-till and draft animal systems. No-till technology for manual and draft an-
imal planting systems have been developed that show a reduction in labor requirements
coupled with yields that are less impacted by soil moisture deficits. Which is not to say
that conservation agriculture has been proven successful in all agro-climatic zones and

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2803
associated farming systems. In some cases there is competition for plant residues and
potential conflicts between the needs of conservation agriculture and pastoral livestock
that are yet to be resolved. Farmers in remote locations who face poor market access
conditions are unlikely to find the package of practices that conservation agriculture rec-
ommends more remunerative. However, this is a problem for mechanization as a whole,
not just for conservation agriculture. For a farmer to increase expenditure on inputs re-
quires a market for increased production, with a return that justifies the increased input
expense.
4.10. Global integration of food and input markets can have positive as well as
negative consequences for small farm mechanization
Global integration of input markets implies cheaper access to mechanical and other
technologies. Technologies developed elsewhere can be more easily transferred and
adapted to country specific agro-ecological and farming system conditions. On the other
hand, the global integration of food markets exemplified by the global spread of super
markets could have more ambiguous effects for the small farmer. Modern food systems
impose high standards for quality and safety that the post-harvest equipment and han-
dling practices on small farms may not be able to meet. Scale economies may become
increasingly important in meeting the quantity and quality requirements of super mar-
kets and hence lead to the displacement of smallholders from the emerging commercial
food systems. Whether innovations in post-harvest technologies aimed at smallholders
can reverse this situation is an open question.
References
Ahammed, C.S., Herdt, R.W. (1983). “Impacts of farm mechanization in a semi-closed input–output model
of the Philippine economy”. American Journal of Agricultural Economics 35 (3), 516–525.
APO (1991). Utilization of Farm Machinery in Asia, Report of APO Multi-Country Study Mission, 19–29
June. Asian Productivity Organization, Tokyo, Japan.
Balsevich, F., Berdegue, J.A., Flores, L., Mainville, D., Reardon, T. (2003). “Supermarkets and produce
quality and safety standards in Latin America”. American Journal of Agricultural Economics 85 (5),
1147–1154.
Binswanger, H.P. (1978). The Economics of Tractors in South Asia. Agricultural Development Council and
the International Crops Research Institute for the Semi-Arid Tropics, Hyderabad, India.
Binswanger, H.P. (1984). “Agricultural mechanization: A comparative historical perspective”, Staff Working
Paper No. 673. World Bank, Washington, DC.
Bonnefond, P.H. (1967). “Problems of profitability of an agricultural development action of the SATEC
operation in Mossi”. Ministere de Plan: Office de la Recherche Scientifique et Technique Outre-Mer
(ORSTROM), Bondy, France (in French).
Boserup, E. (1965). Conditions of Agricultural Growth. Aldine Publishing Company, Chicago, IL.
Chang, J.Y. (1991). “Country papers: Republic of Korea”. In: Utilization of Farm Machinery in Asia, Report
of APO Multi-Country Study Mission, 19–29 June. Asian Productivity Organization, Tokyo, Japan.
Cho, K.H. (1983). “Other country report on the status of farm mechanization – Republic of Korea”. In: Farm
Mechanization in Asia. Asian Productivity Organization, Tokyo, Japan.

2804 P. Pingali
CSO (1993). China Statistical Yearbook. China Statistical Office, Beijing.
David, C., Otsuka, K. (1994). Modern Rice Technology and Income Distribution in Asia. International Rice
Research Institute, Los Baños, Philippines.
Duff, B. (1986). “Changes in small-farm rice threshing technology in Thailand and the Philippines”, Working
Paper No. 120. International Rice Research Institute, Los Baños, Laguna, Philippines.
Ebron, L.Z. (1984). “Changes in harvesting–threshing labor arrangements in Nueva Ecija”, Master thesis.
College of Economics and Management, University of the Philippines at Los Baños, College, Laguna,
Philippines.
Ekboir, J. (2000). “Innovation systems and technology policy: Zero tillage in MERCOSUR”, Economic Pro-
gram Paper. CIMMYT, Mexico City, Mexico.
FAO (2005). FAOSTAT database. Available at http://faostat.fao.org/faostat/default.jsp?version=int&hasbulk
=1; accessed July 2005.
Herdt, R.W. (1983). “Mechanization of rice production in developing Asian countries: Perspective, evidence
and issues”. In: Consequences of Small Farm Mechanization. International Rice Research Institute, Los
Baños, Laguna, Philippines.
Hossain, M., Bose, M.L., Mustafi, B.A.A. (2002). “Adoption and productivity impact of modern rice varieties
in Bangladesh”, Paper presented at the Workshop on Green Revolution in Asia and Its Applicability to
Africa, FASID, Tokyo, Japan, December 8–10, 2002.
Hurt, R.D. (1982). American Farm Tools: From Hand Power to Steam Power. Sunflower University Press,
Manhattan, KS.
International Bank for Reconstruction and Development (IBRD) (1987). Agricultural Mechanization: Issues
and Options. The World Bank, Washington, DC.
IRRI (1986). Annual Report. International Rice Research Institute, Los Baños, Laguna, Philippines.
Jabbar, M.A., Bhuiyan, M.S.R., Bari, A.K.M. (1983). “Causes and consequences of power tiller utilization in
two areas of Bangladesh”. In: Consequences of Small Farm Mechanization. International Rice Research
Institute, Los Baños, Laguna, Philippines.
Jones, W.O. (1959). Manioc in Africa. Stanford University Press, Stanford, CA.
Juarez, F., Pathnopas, R. (1983). “Comparative analysis of thresher adoption and use in Thailand and the
Philippines”. In: Consequences of Small Farm Mechanization. International Rice Research Institute, Los
Baños, Laguna, Philippines.
Kisu, M. (1983). “Mechanization of rice farming”. In: Farm Mechanization in Asia. Asian Productivity Or-
ganization, Tokyo, Japan.
Kline, C.K., Green, D.A.G., Donahue, R.L., Stout, B.A. (1969). “Agricultural mechanization in equatorial
Africa”, College of Agriculture and Natural Resources Research Report No. 6. Michigan State University,
East Lansing, MI.
Labrousse, G. (1971). “Historic–economic aspects of the development of mechanized cultivation of food
crops in tropical countries”. Agricultural Research Center, Bambey, Senegal.
Lockwood, B., Munir, M., Hussain, K.A., Gardezi, J. (1983). “Farm mechanization in Pakistan: Policy and
practice”. In: Consequences of Small Farm Mechanization. International Rice Research Institute, Los
Baños, Laguna, Philippines.
Mandal, M.A.S. (2002). “Agricultural machinery manufacturing and farm mechanization: A case of rural
non-farm economic development in Bangladesh”. In: Proceedings of the International Workshop on Fos-
tering Rural Economic Development through Agriculture Based Enterprises and Services, GTZ, Berlin,
Germany, 20–22 November 2002.
Manners, R.A. (1962). “Land use, trade and the growth of market economy in Kipsigis country”. In: Bohan-
nan, P., Dalton, G. (Eds.), Markets in West Africa. Northwestern University Press, Evanston, IL.
McInerney, D.P., Donaldson, G.F. (1975). “The consequences of farm tractors in Pakistan”, World Bank Staff
Working Paper No. 210. The World Bank, Washington, DC.
Miracle, M.P. (1967). Agricultural Growth in the Congo Basin. University of Wisconsin Press, Madison, WI.
Mizuno, T. (1991). “Agricultural mechanization trends in Japan”. In: Utilization of Farm Machinery in Asia,
Report of APO Multi-Country Study Mission, 19–29 June. Asian Productivity Organization, Tokyo,
Japan.

Ch. 54: Agricultural Mechanization: Adoption Patterns and Economic Impact 2805
Moody, K. (1994). “Postplanting weed control in direct seeded rice”. In: Ali, A.H. (Ed.), Direct Seeding Prac-
tices and Productivity. Malaysian Agricultural Research and Development Institute, Penang, Malaysia.
Naylor, R. (Ed.) (1996). Herbicides in Asian Rice: Transitions in Weed Management. Institute of International
Studies, Stanford University, Palo Alto, CA.
Ohkawa, K., Shinohara, M., Umemura, M. (Eds.) (1965). Estimates of Long Term Economic Statistics of
Japan since 1868. Agriculture and Forestry, vol. 9. Toyo Keizai Shinposha, Tokyo.
Peng, T.S. (1983). “Status of farm mechanization: Republic of China”. In: Farm Mechanization in Asia. Asian
Productivity Organization, Tokyo, Japan.
Pingali, P.L. (1998). “Dealing with growing labor scarcity: The adoption and impact of mechanical and chem-
ical technologies in the Asian rice economies”. In: Pingali, P.L., Hossain, M. (Eds.), Impact of Rice
Research. International Rice Research Institute, Los Baños, Philippines.
Pingali, P.L., Bigot, Y., Binswanger, H. (1987). Agricultural Mechanization and the Evolution of Faring Sys-
tems in Sub-Saharan Africa. The World Bank, Washington, DC.
Pingali, P.L., Binswanger, H.P. (1987). “Population density and agricultural intensification: A study of the
evolution of technologies in tropical agriculture”. In: Johnson, G., Lee, R. (Eds.), Population Growth and
Economic Development. National Research Council, Washington, DC.
Pingali, P.L., Hossain, M., Gerpacio, R.V. (1997). Asian Rice Bowls: The Returning Crisis? CAB Interna-
tional, Wallingford, UK.
Salokhe, V.M., Ramalingam, N. (1998). “Agricultural mechanization in the South and Southeast Asia”, Paper
presented at the International Conference of the Philippine Society of Agricultural Engineers, Los Baños,
Philippines, 21–24 April 1998.
Sukharomana, S. (1983). “The impact of farm power strategies in Thailand”. In: Farrington, J., Abeyratne,
F., Gill, G. (Eds.), Farm Power and Employment in Asia. Agricultural Development Council, Bangkok,
Thailand, pp. 139–150.
Toquero, Z.F., Duff, B. (1985). “Physical losses and quality deterioration in rice post-production systems”,
IRRI Research Paper No. 107. International Rice Research Institute, Los Baños, Laguna, Philippines.
Walker, T.S., Kshirsagar, K.G. (1981). “The village level impact of machine threshing and implications for
technology development in semi-arid tropical India”, ICRISAT Economics Program Progress Report No.
27. ICRISAT, Hyderabad, India.