Front Cover
 Title Page
 Table of Contents
 List of Tables
 List of maps
 Executive summary
 Descriptive overview of South Asia's...
 Productivity trends in rice-wheat...
 Sustainability issues in rice-wheat...
 Challenges for research and research...
 Back Cover

Group Title: Natural Resources Group paper ; 96-01
Title: Meeting South Asia's future food requirements from rice-wheat cropping systems
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00080104/00001
 Material Information
Title: Meeting South Asia's future food requirements from rice-wheat cropping systems priority issues facing researchers in the post-Green Revolution era
Series Title: Natural Resources Group paper
Physical Description: iii, 46 p. : ill., maps ; 28 cm.
Language: English
Creator: Hobbs, P. R
Morris, Michael L
Publisher: CIMMYT
Place of Publication: Mexico D.F
Publication Date: 1996
Subject: Cropping systems -- South Asia   ( lcsh )
Rice -- South Asia   ( lcsh )
Wheat -- South Asia   ( lcsh )
Food supply -- South Asia   ( lcsh )
Agriculture -- Research -- South Asia   ( lcsh )
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Spatial Coverage: Nepal
Bibliography: Includes bibliographical references (p. 41-45).
Statement of Responsibility: Peter Hobbs and Michael Morris.
Funding: Paper (International Maize and Wheat Improvement Center. Natural Resources Group) ;
 Record Information
Bibliographic ID: UF00080104
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 36830433

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page i
        Page ii
    Table of Contents
        Page iii
    List of Tables
        Page iv
    List of maps
        Page v
        Page vi
    Executive summary
        Page vii
        Page viii
        Page 1
        Page 2
    Descriptive overview of South Asia's rice-wheat cropping systems
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
    Productivity trends in rice-wheat cropping systems
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
    Sustainability issues in rice-wheat cropping systems
        Page 29
        Page 30
        Page 31
    Challenges for research and research organization
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
    Back Cover
        Back Cover
Full Text
0o 7 9

Sustainable Maize
and Wheat Systems
for the Poor

Meeting South Asia's Future

Food Requirements from

Rice-Wheat Cropping Systems:

Priority Issues facing

Researchers in the

Post-Green Revolution Era

Peter Hobbs and Michael Morris

Natural Resources Group
Paper 96-01


Meeting South Asia's Future

Food Requirements from

Rice-Wheat Cropping Systems:

Priority Issues Facing

Researchers in the

Post-Green Revolution Era

Peter Hobbs and Michael Morris *

Natural Resources Group
Paper 96-01 .

* Peter Hobbs is a CIMMYT regional wheat agronomist for Asia, based in Kathmandu, Nepal.
Michael Morris is a CIMMYT regional economist for Asia, based in Bangkok, Thailand. The
views expressed in this paper are the authors' and may not necessarily reflect CIMMYT policy.

CIMMYT is an internationally funded, nonprofit scientific research and training organization.
Headquartered in Mexico, the Center is engaged in a research program for maize, wheat, and triticale, with
emphasis on improving the productivity of agricultural resources in developing countries. It is one of several
nonprofit international agricultural research and training centers supported by the Consultative Group on
International Agricultural Research (CGIAR), which is sponsored by the Food and Agriculture Organization
(FAO) of the United Nations, the International Bank for Reconstruction and Development (World Bank), and
the United Nations Development Programme (UNDP). The CGIAR consists of some 40 donor countries,
international and regional organizations, and private foundations.

CIMMYT receives core support through the CGIAR from a number of sources, including the international
aid agencies of Australia, Austria, Belgium, Brazil, Canada, China, Denmark, Finland, France, India,
Germany, Italy, Japan, Mexico, the Netherlands, Norway, the Philippines, Spain, Switzerland, the United
Kingdom, and the USA, and from the European Union, Ford Foundation, Inter-American Development
Bank, OPEC Fund for International Development, UNDP, and World Bank. CIMMYT also receives non-
CGIAR extra-core support from the International Development Research Centre (IDRC) of Canada, the
Rockefeller Foundation, and many of the core donors listed above.

Responsibility for this publication rests solely with CIMMYT.

Printed in Mexico.

Correct citation: Hobbs, P., and M. Morris. 1996. Meeting South Asia's Future Food Requirements from Rice-
Wheat Cropping Systems: Priority Issues Facing Researchers in the Post-Green Revolution Era. NRG Paper 96-01.
Mexico, D.F.: CIMMYT.

Abstract: The importance of rice-wheat cropping systems in meeting present and future food
needs in South Asia is reviewed. Evidence from a number of factor productivity studies, which
analyze yield trends after adjusting for changes in levels of input use, suggests that growth in the
productivity of South Asia's rice-wheat cropping systems is leveling off and, in some areas,
declining. Some probable causes of this disturbing trend are considered, including soil-related
factors (depletion of soil chemicals, soil physical problems from puddling soils for rice and/or
repeated cultivation for wheat); problems relating to the quantity and quality of irrigation water;
continuous and intensive cereal cultivation, which has increased the incidence of pests (including
weeds) and diseases; and delayed planting of wheat following rice, a common practice in many
rice-wheat systems, which severely reduces wheat yields. Changes in the organization and
management of research, which are required to restore growth in productivity, are discussed in the
final sections of the paper.

ISSN: 1405-2830
AGROVOC descriptors: South Asia; rice; wheat; cropping systems; food productivity; food
supply; production policies; research projects
AGRIS category codes: E10
Dewey decimal classification: 338.19



iv Tables
iv Figures
v Maps
vi Acknowledgments
vii Executive Summary

1 Introduction

3 Descriptive overview of South Asia's rice-wheat cropping systems
3 Area covered by rice-wheat cropping systems
6 Agroclimatic conditions
6 Demographic characteristics of rice-wheat zones
9 Rice-wheat cropping systems

11 Productivity trends in rice-wheat cropping systems
11 Area trends
15 Yield trends
18 Impact on production
18 Input use trends
22 Trends in total factor productivity

29 Sustainability issues in rice-wheat cropping systems
29 Soil problems
30 Water problems
31 Insect, pest, and disease problems
32 Crop management problems

32 Challenges for research and research organization
32 Technical change: past, present, and future
34 Organization of agricultural research in South Asia
35 Weaknesses of current approaches to research
36 Elements needed for future research to be effective

39 Conclusions

41 References



.2 Table 1. Supply and demand projections for rice and wheat, South Asia

4 Table 2. Area under rice-wheat cropping systems in South Asia

15 Table 3. Growth in area under rice-wheat cropping sequences, South Asia

19 Table 4. Rice and wheat production in South Asia, 1950-90

21 Table 5. Fertilizer use on rice and wheat in selected districts of India

21 Table 6. Long-term trends in irrigated rice and wheat area, South Asia, 1950-90

23 Table 7. Trends in total factor productivity for rice-wheat systems, Pakistan, 1970-79 and 1980-89

25 Table 8. Trends in input, output, and total factor productivity for rice, India, 1971/72 to 1988/89

26 Table 9. Trends in rice and wheat yields, long-term trials, Bhairahawa, Nepal, 1976-90


9 Figure 1. Common rice-wheat cropping patterns of South Asia

11 Figure 2. Rice area, South Asia, 1958-91

11 Figure 3. Rice and wheat area per capital, South Asia, 1961-91

12 Figure 4. Rice area in Bangladesh, India, Nepal, and Pakistan

13 Figure 5. Wheat area, South Asia, 1958-93

14 Figure 6. Wheat area in Bangladesh, India, Nepal, and Pakistan

15 Figure 7. Rice yields, South Asia, 1958-91

16 Figure 8. Rice yields in Bangladesh, India, Nepal, and Pakistan

17 Figure 9. Wheat yields, South Asia, 1958-93

17 Figure 10. Wheat yields in Bangladesh, India, Nepal, and Pakistan

18 Figure 11. Rice and wheat yields in Punjab and Bihar States, India, 1960-90

20 Figure 12. Adoption of rice and wheat MVs, India, 1967-94

20 Figure 13. Fertilizer applied to rice in Bangladesh, India, and Pakistan, 1960-88

21 Figure 14. Rainfed and irrigated wheat area, India, 1951-88

26 Figure 15. Mean yields of RR21 and UP262 in the Terai, Nepal, 1976-90

27 Figure 16. Rice and wheat yields, long-term trials, Pantnagar, U.P., India, 1973-88

27 Figure 17. Rice and wheat yields, long-term trials, Faizabad, U.P., India, 1979-88

27 Figure 18. Shift in fertilizer response function resulting from resource degradation

28 Figure 19. Average response of rice and wheat to fertilizer, all India, 1966-92

29 Figure 20. Marginal response of rice and wheat to fertilizer, Karnal, India, 1969-90

31 Figure 21. Rice and wheat yields by irrigation source, Lahore, Pakistan, 1975 and 1985

33 Figure 22. Phases of technical change (following Byerlee)


3 Map 1. Extent of rice-wheat cropping systems in South Asia

5 Map 2. Distribution of rice-wheat cropping systems in Bangladesh

7 Map 3. Distribution of rice-wheat cropping systems in India

8 Map 4. Distribution of rice-wheat cropping systems in Nepal

8 Map 5. Distribution of rice-wheat cropping systems in Pakistan


This report has benefited from the contributions of a large number of people. Without implicating
them in any errors of fact or interpretation, we would like to acknowledge the many scientists
working in national agricultural research programs and in international agricultural research
centers whose work we have cited. We extend thanks also to the thousands of South Asian
farmers and their families for permitting so many intrusions by researchers into their everyday

Although it is not possible to list everybody, a number of individuals deserve particular mention.
Many researchers active in the Rice-Wheat Consortium provided information or made helpful
suggestions, including Prabhu Pingali, Terry Woodhead, Mubarik Ali, Sam Fujisaka, and Thelma
Paris of IRRI; Larry Harrington, Craig Meisner, and Tony Fischer of CIMMYT; and Derek Byerlee
of the World Bank. Donald Faris and I.P. Abrol, the Rice-Wheat Consortium facilitators, assisted
with the exchange of data. Tim Kelley of ICRISAT provided district-wise data on area and
production of rice and wheat in India. Mark Rosegrant, Mercedita Agcaoili-Sombilla, and Nicos
Perez of IFPRI contributed long-run cereal supply and demand projections. Pedro Aquino, Paul
Heisey, and Maria Luisa Rodriguez of the CIMMYT Economics Program tracked down data in
Mexico. Bharat Adhikhary of the CIMMYT office in Kathmandu helped prepare many of the
figures. Robert and Ellie Huke of Dartmouth University's Department of Geography prepared the

Kelly Cassaday of CIMMYT Information Services edited the report. N.S. Wagley and R.K. Singh of
the CIMMYT office in Kathmandu, Valairat Kuslasayanon of the CIMMYT office in Bangkok, and
Miguel Mellado of the CIMMYT Publications Department in Mexico assisted with different
aspects of production.

Executive Summary

South Asia's food supply is dominated by rice and wheat, which account for about 90% of the
region's total cereal production. Although growth in demand for these two crops is expected to
ease as the population gradually stabilizes, future supplies of rice and wheat are difficult to
forecast. Over the longer term, most analysts believe production increases will be hard pressed to
keep pace with even modest growth in demand, fueled by rising incomes. Researchers thus face a
formidable challenge in attempting to develop the improved production technologies that will be
needed to feed large numbers of people from a dwindling land area.

This paper has four objectives:
* to demonstrate the crucial importance of rice-wheat cropping systems in meeting present and
future food needs in South Asia;
* to present evidence suggesting that productivity growth in the region's rice-wheat cropping
systems is leveling off and in some areas may be declining;
to consider the causes of the apparent slowdowns in productivity growth; and
to discuss changes that will be required in the organization and management of research if
growth in productivity is to be restored.

South Asia's rice-wheat cropping systems, which cover 12 million hectares, are concentrated on
the Indo-Gangetic and Brahmaputra flood plains and in the foothills of the Himalayas. About 32%
of total rice area and 42% of total wheat area in the region comes under rice-wheat cropping
sequences, which can also include legumes, oilseeds, fodder crops, vegetables, and sugarcane.
Crop production is often linked to livestock production.

During the past three decades, growth in the area planted to rice and wheat in South Asia has
slowed considerably. With traditional sources of area growth nearly depleted, further expansion
in the area planted to these two crops is likely to be negligible. Future production gains therefore
will have to come mainly from yield increases. Although average rice and wheat yields rose at
about 2% per year between 1960 and 1990, evidence is accumulating to suggest that the impressive
rates of yield growth achieved earlier are no longer being sustained. In some intensively cultivated
zones, yields of rice and wheat have actually begun to decline.

Slower growth in yields is alarming, especially since the use of productivity-enhancing inputs
now seems to be approaching saturation levels. Adoption of modem varieties (MVs) is virtually
complete, and although farmers can look forward to realizing further genetic gains by replacing
their older varieties regularly with new ones, the emphasis in plant breeding is increasingly
shifting to characteristics other than yield per se (e.g., disease resistance, pest resistance, grain
quality). Fertilizer use on rice and wheat is now close to optimal in many zones, and application of
additional fertilizer is often unprofitable. Investment in irrigation has also become less attractive.
Many older irrigation schemes need extensive rehabilitation, and construction costs for new
schemes have risen. With traditional sources of productivity growth showing signs of exhaustion,
future productivity gains will have to come from elsewhere.

Evidence that growth in the productivity of South Asia's rice-wheat cropping systems is slowing
down has come from factor productivity studies, which analyze yield trends after adjusting for
changes in levels of input use. In Bangladesh and India, factor productivity continues to increase
in some areas, but in others it shows signs of leveling off or even declining. In Pakistan, factor
productivity has fallen by more than 2% per year during the post-Green Revolution period.

Causes of the slowdown in productivity growth are still poorly understood. Soil-related factors
are part of the problem. Soil chemicals can be depleted in intensively cultivated rice-wheat
systems in which nutrient extraction is not always matched by nutrient input, while soil physical
problems are caused by puddling of soils for rice and/or repeated cultivation for wheat. Problems
relating to the quantity and quality of irrigation water have also affected productivity. Declining
water tables force farmers to pump water from great depths, and many irrigated areas are prone
to salinity and sodicity problems. Continuous and intensive cereal cultivation has increased the
incidence of pests and diseases, while buildups of grassy weeds have become serious problems in
some areas. Delayed planting of wheat following rice, a common practice in many rice-wheat
systems, also severely reduces yield.

Agriculture in South Asia can be thought of as progressing through three phases of technical
change: a "Green Revolution Phase," an "Input Intensification Phase," and an "Input Efficiency
Phase." Many of South Asia's rice-wheat systems have entered the last of these three phases.
Multidisciplinary, systems-oriented research is required to develop the sophisticated, site-specific
management information needed to improve input-use efficiency in the context of ever more
intensive cropping systems in which avenues for successful technical innovation are increasingly

The need for new technologies and information has implications for agricultural research. Not
only does the organization of research need to be changed, but the management of research
programs needs to be made more efficient. Objectives need to be more clearly defined and better
prioritized, redundant programs need to be eliminated, and efficiency needs to be increased
through consolidation. Unfortunately, the need for restructuring comes precisely at a time when
donor assistance is declining and when many governments are reducing public support to
agricultural research.


Meeting South Asia's Future Food Requirements
from Rice-Wheat Cropping Systems:
Priority Issues Facing Researchers in the
Post-Green Revolution Era

Peter Hobbs and Michael Morris


As the twentieth century draws to a close, the
challenge facing agricultural researchers is
greater than ever. Global demand for food
continues to grow steadily, with population
growth adding 90 million new mouths to feed
each year. Since most of the earth's arable land
is already in use, additional food supplies will
have to be produced by increasing food
production from land that is currently being
cultivated.' However, much of this land
appears to be stretched to its limits, and in
some areas there are signs that current levels of
productivity cannot be maintained.

Nowhere is the challenge greater than in South
Asia. Long-term forecasts of the region's food
balances have been made by a number of
analysts (for a summary, see McCalla 1994). On
the demand side, a broad consensus prevails
that a slowdown in population growth is
expected to be offset by modest increases in
income levels. Opinions differ as to when
demand will eventually stabilize, but most
analysts believe that annual growth in cereal
consumption will average 2-2.5% until well into
the twenty-first century.2

There is less agreement regarding prospects for
future growth in food supplies, both for the
world as a whole and for South Asia.
Differences in projected rates of production
growth stem primarily from disagreements
about the sustainability of the natural resource
base that supports agriculture, about future
rates of productivity gains that may be
achieved through technological innovation,
and about government policies toward
agriculture. At the pessimistic end of the
spectrum, neo-Malthusians foresee a marked
slowdown in cereal production growth as a
result of rapid deterioration in the natural
resource base and depletion of land and water
resources (Kendall and Pimental 1994, Brown
1994, Brown and Kane 1994). Others are more
sanguine, predicting that the demand for food
eventually will be met, albeit with difficulty
(Rosegrant and Agcaoili 1994, Islam et al. 1992).
A few analysts are openly optimistic about
long-run global food balances, arguing that
there is little evidence to suggest that historical
rates of productivity growth are slowing and
hence no reason to assume that current rates
cannot be maintained (Plucknett 1993, Mitchell
and Ingco 1992).

1 Of the earth's land area, only 24% (3.2 billion hectares) consists of arable land. Of the arable land, 1.3 billion hectares
are classified as highly or moderately productive. Currently, 1.5 billion hectares are used for cropland, including
most of the highly and moderately productive land (Buringh and Dudal 1987).
2 Some analysts predict that global population growth will stabilize as soon as 30 years from now (Seckler 1994). If this
prediction is correct, growth in the demand for food would decline much more rapidly.

Depending upon the assumptions made about
future rates of growth in demand and supply, it
is possible to come up with widely divergent
forecasts about future food balances in South
Asia. However, the most likely scenario is that
despite a gradual slowing of growth in demand
for cereals, food grain production in South Asia
will be hard-pressed to keep pace with growth
in consumption requirements. This view, which
is consistent with the latest projections of the
International Food Policy Research Institute
(IFPRI), foresees South Asia facing a regional
cereals deficit averaging around 20-25 million
tons per year by 2020 (Table 1). Two key
assumptions underlying this projection are that
the expansion of irrigated area will slow
markedly and that crop response to additional
use of fertilizer will decrease. It is important
also to note the implicit assumption that current
rates of investment in agricultural research will
be maintained. Should investment in
agricultural research decline (which many
consider likely), food deficits could be
considerably larger.

Whatever happens, it is clear that during the
early part of the twenty-first century, South
Asia's fortunes will be greatly influenced by
developments in the agricultural sector,

particularly in the rice and wheat cropping
systems which account for the bulk of the
region's cereal production. South Asia's rice-
wheat systems will play a decisive role, not
only in determining whether food security can
be assured for more than one billion people,
but also in providing productive employment
and generating much-needed income for rural

This review of the challenges facing rice-wheat
cropping systems in South Asia has four
principal objectives:

* to demonstrate the crucial importance of rice-
wheat cropping systems in meeting present
and future food needs;

* to present evidence suggesting that
productivity growth in rice-wheat cropping
systems is leveling off and in some areas
appears to be declining;

to consider the micro-level causes of
stagnating productivity; and

to discuss changes that will be required in the
organization and management of research if
growth in productivity is to resume.

Table 1. Supply and demand projections for rice and wheat, South Asia

1990-2020 2020 2020 2020 2020 2020 2020
projected 1990 projected projected projected 1990 projected projected projected
1990 population rice rice rice rice wheat wheat wheat wheat
population growth production* production' demand' balance' production production demand balance
(million) (%) (000 t) (000 t) (000 t) (000 t) (000 t) (000 t) (000 t) (000 t)

Bangladesh 106.7 1.8 18,689 38,071 38,204 (132) 1,226 1,580 6,031 (4,450)
India 849.5 1.7 75,338 145,777 144,792 985 49,296 96,384 95,617 766
Nepal 18.9 1.8 2,580 3,814 3,407 407 860 1,401 1,357 44
Pakistan 112.4 2.8 3,448 6,207 5,309 898 14,413 27,463 42,914 (15,451)
South Asiab 1,147.7 1.8 101,430 197,617 197,588 29 65,780 126,817 148,121 (21,303)

Sources: Population data from World Bank (1992); rice and wheat data for Nepal from Thapa and Rosegrant (1995); and all other
data series from Rosegrant, Agcaoili-Sombilla, and Perez (1995).
a Milled rice.
b Includes data for Bhutan and Sri Lanka.

Descriptive Overview of South
Asia's Rice-Wheat Cropping

Agriculture in South Asia has a long and rich
history marked by a series of technological
breakthroughs which today allow an
exceedingly large number of people to be fed
from a relatively small area. Two cereals -
rice and wheat contribute the bulk of the
region's food supply and will continue to do s
for the foreseeable future. South Asia's rice-
wheat cropping systems, among the most
highly evolved agricultural
production systems on the
planet, are often portrayed
as highly productive
systems which can serve as
models for the rest of the
world. Yet this portrayal is
misleading, because these
cropping systems are
showing signs of declining
productivity caused by
resource degradation. What
is especially alarming is the
fact that the effects of
resource degradation are
often masked by increased
input use, which can lead to
the false impression that
productivity is increasing
when precisely the opposite
may be true.

Area covered by rice-wheat
cropping systems

The rice-wheat cropping systems of South Asia
extend across the Indo-Gangetic flood plain
and up into the Himalayan foothills. They span
a vast area stretching from Pakistan's Swat
Valley in the north to India's Maharastra State
in the south, from the mountainous Hindu
Kush of Afghanistan in the west to the
o Brahmaputra flood plains of Bangladesh in the
east (Map 1).3

Map 1. Extent of rice-wheat cropping systems in South Asia.
Source: R. Huke and E. Huke.

3 Although this paper focuses on rice-wheat systems in Bangladesh, India, Nepal, and Pakistan, the importance of
these systems extends well beyond South Asia. In China alone, rice-wheat cropping systems cover an additional
10 million hectares.

Rice Wheat sequence area, 1987-90
Total area in rice-wheat = 11,900,000 ha

* S

*=100,000 ha
=20,000 ha

The area covered by rice-wheat systems is
difficult to state with precision.4 Although the
national crop reporting services of Bangladesh,
India, Nepal, and Pakistan publish aggregate
production statistics for rice and wheat, none
publishes statistics specifically on rice-wheat
cropping sequences. Area figures therefore
must be pieced together based on subjective
estimates made at the district level of the
degree to which the two crops are planted in
rotation. The district-level estimates can then
be aggregated to arrive at national and
eventually regional estimates (Table 2). In 1991,
about 12 million hectares were under rice-
wheat cropping systems in Bangladesh, India,
Nepal, and Pakistan (Singh and Paroda 1994).
Rice-wheat systems covered about 32% of the
rice area and 42% of the wheat area in these

four countries and accounted for between one-
quarter and one-third of total rice and wheat

The relative importance of rice and wheat
varies between countries. In Bangladesh, rice is
by far the dominant food staple, accounting for
93% of total cereal production and
consumption. Wheat, traditionally a minor
crop, has grown in importance since 1970
thanks to government efforts to promote
production, but it remains of secondary
importance in all but a few areas. Rice-wheat
cropping systems cover an estimated 0.5
million ha, mostly in the northern and western
"wheat belt" characterized by more elevated
land and lighter, better-drained soils (Map 2).
While only 7% of the area planted to rice is

Table 2. Area under rice-wheat cropping systems in South Asia, 1988

Percentage Percentage
of wheat of rice
area area Rice-wheat
Total Total located located area as
rice rice on land on land percentage
Total Total Total Total + + where where of total
rice rice wheat wheat wheat wheat rice is wheat is rice+wheat
area production area production area production grown grown area
(million ha) (million t) (million ha) (million t) (million ha) (million t) (%) (%) (%)

Bangladesh 10.22 24.67 0.59 0.99 0.50 2.03 85 7 5
Indiaa 40.90 100.11 23.56 50.04 9.12 38.58 42 33 21
Nepal 1.43 3.21 0.60 0.81 0.49 1.74 85 34 24
Pakistan 2.03 4.83 7.27 13.80 1.63 6.67 22 80 17
South Asia 54.60 132.83 32.01 65.65 11.74 49.02 42 32 18

Source: Huke, Huke, and Woodhead (1994a, 1994b); Woodhead, Huke, and Huke (1994); and Woodhead et al. (1994).
Note: Since data on cropping patterns are not reported at the national level, the rice-wheat area for each country was
estimated based on district-level data from a subset of districts for which data were available. The rice-wheat area
in each district was estimated by multiplying the smaller of the district's rice or wheat area by a variable percentage
estimated by researchers familiar with the district. District-level estimates were then aggregated to derive national
estimates. The data are based on the average of the 1987 to 1989 data available in national program statistics.
a Includes Bihar, Madhya Pradesh, Haryana, Himachal Pradesh, Punjab, Uttar Pradesh, and West Bengal. Rice-wheat
area in Assam, Gujarat, Jammu-Kashmir, and Maharashtra is not included.

4 Since production statistics are not collected by cropping pattern, it is not possible to make a more precise estimate.
5 Much of the descriptive information presented in this section was obtained from a series of atlases compiled by
scientists working at IRRI, at CIMMYT, and in the national agricultural research programs of Bangladesh, India,
Nepal, and Pakistan. See Huke, Huke, and Woodhead (1994a, 1994b); Woodhead, Huke, and Huke (1994); and
Woodhead et al. (1994).

Rice-Wheat sequence area, 1987-90
Total area in rice-wheat = 495,933 ha

Map 2. Distribution of rice-wheat cropping systems in Bangladesh.
Source: Huke, Huke, and Woodhead (1994a).

located on land where wheat is also grown,
approximately 85% of the much smaller area
planted to wheat is located on land where rice
is also grown.

India's highly diverse rice-wheat cropping
systems are found on the Indo-Gangetic flood
plain, in the Himalayan foothills, and in some
irrigated zones further south (Map 3). With
approximately 9.1 million ha under rice-wheat
cropping systems, India dominates the regional
production statistics. Approximately 33% of
India's rice is grown in rotations involving
wheat, while about 42% of wheat is grown in
rotations involving rice. The states with the
largest areas under rice-wheat cropping
systems are Uttar Pradesh, Punjab, Haryana,
Bihar, Madhya Pradesh, and Himachal

In Nepal, rice dominates the cereal economy,
although as in Bangladesh wheat has made
important inroads during the past two decades.
Rice-wheat cropping systems cover an
estimated 0.5 million ha and are distributed
between the low-lying Terai and the more
elevated mid-hills (Map 4). Nearly all wheat
follows rice, whereas about one-third of the rice
crop is followed by wheat.

In Pakistan, where wheat is the primary food
staple, rice-wheat cropping systems cover
about 1.6 million ha and are concentrated in the
irrigated zones of Punjab and Sind and in the
Himalayan foothills (Map 5). Approximately
22% of the nation's wheat is grown in rotations
involving rice, whereas over 80% of the rice is
grown in rotation with wheat. Most wheat in
Pakistan is produced for domestic
consumption, whereas rice (especially high-
quality basmati rice) is a major export crop,
about half of which is sold overseas.

Agroclimatic conditions

Most rice-wheat cropping systems in South
Asia are located in areas featuring subtropical
to warm-temperate climates, characterized by
cool, dry weather during winter (when wheat
is grown) and warm, humid weather during
summer (when most rice is grown).
Temperatures are higher to the east and south,
reducing the length of the wheat season.

Annual precipitation is concentrated in the
summer months. Although total annual rainfall
tends to be greater in the east, winter rains are
often heavier in the west. Irrigation is
commonly used to supplement rainfall for both
rice and wheat and is available in many areas
from canals and/or tubewells.

Soils tend to be highly variable. The calcareous
soils of the Indo-Gangetic flood plain are
generally basic (characterized by high pH
levels), as opposed to the non-calcareous
Brahmaputra and Jamuna flood plain soils of
Bangladesh, which are more acidic. Free
calcium carbonate is present in some of the
Indo-Gangetic soils. Soil texture varies with
local topography, with coarse soils often found
on more elevated ridges and fine soils often
found in low-lying basins. Most soils are very
productive, although they require additions of
nitrogen, phosphorus and, in some areas,
micronutrients such as zinc.

Demographic characteristics
of rice-wheat zones

South Asia's rice-wheat belt is home to 600
million people, approximately two-thirds of
whom are employed in agriculture. Although
production and consumption statistics are not
collected on the basis of specific cropping
patterns, it is estimated that between 150 and
275 million people consume rice and/or wheat

Rice-Wheat sequence area, 1986-89
Total area in rice-wheat = 9,500,000 ha

Each dot = 10,000 ha



Bay of


Arabian Sea

1 Jammu and Kashmir

13 Andra Pradesh

'' 2 Himachal Pradesh 14 Karnataka
S3 Punjab 15 Goa
\ 17- 4 Haryana 16 Kerala
1 5 Rajasthan 17 Tamil Nadu
16 6 Uttar Pradesh 18 Assam
7 Bihar 19 Meghalaya
8 West Bengal 20 Nagaland
SrI 9 Madhya Pradesh 21 Manipur
10 Orissa 22 Mizoram
11 Maharashtra 23 Tripura
12 Gujarat 24 Sikkim
r Delhi 25 Arunachal Pradesh

Map 3. Distribution of rice-wheat cropping systems in India.
Source: Adapted from Woodhead et al. (1994).


Map 4. Distribution of rice-wheat cropping systems in Nepal.
Source: Huke, Huke, and Woodhead (1994b).

Map 5. Distribution of rice-wheat cropping systems in Pakistan.
Source: Woodhead, Huke, and Huke (1994).

produced in rice-wheat cropping systems.6 In
addition to the grain consumed at home,
surplus production may be marketed to
generate cash.

Population growth for South Asia as a whole
currently exceeds 2% per year. Even if this rate
slows as predicted to around 1.8% per year, the
population living in rice-wheat-areas is likely to
exceed 850 million by the year 2010. Income
levels are extremely low. In 1992, average
annual per capital GDP in South Asia ranged
from US$ 170 in Nepal to US$ 420 in Pakistan
(World Bank 1994). Incomes in rural areas which
depend on agriculture are usually even lower
than these national average figures.

In the past, agriculture represented the main
source of employment for landless laborers in
rural areas. More recently, urban industrial
development and the attendant employment
opportunities have offered the prospect of a
better life in the cities, inducing many people to
abandon rural villages in search of urban jobs.

In some areas this has resulted in a reduction in
the labor available for agricultural work and an
increase in rural wage rates.

Rice-wheat cropping systems

Although all rice-wheat cropping systems by
definition include rice and wheat, rice-wheat
systems can vary tremendously. Temporal and
spatial relationships between the two crops often
differ: rice and wheat may be grown in the same
plot in the same year, in the same plot in different
years, or in different plots in the same year. While
rice and wheat may be the only crops grown in a
given plot, frequently other crops are also
present, either associated with the rice
and/or wheat (grown at the same time) or
rotated with the rice and/or wheat (grown before
or after). Crops commonly included in rice-wheat
systems include oilseeds (mustard, rapeseed,
sunflower); pulses (grasspea, mung bean, black
gram, lentil); fodder crops (berseem clover, fodder
sorghum, or pearl millet); vegetables; potatoes;
sugarcane; and jute (Figure 1).

May Jun Jul
I -----

l;b2egume g
F . . . . .

Aug Sep

Oct Nov Dec Jan

Feb Mar Apr May

Figure 1. Common rice-wheat cropping patterns of South Asia.
Note: Substitutes for oilseeds (short-duration Brassica spp.) include potatoes, vegetables, and peas. Legumes
include mung bean and green manures (e.g., Sesbania spp.); fodder sorghum or millet can substitute for
legumes. Substitutes for wheat include berseem clover, lentils, grasspea, mustard, and rapeseed. Sunflower
can also substitute for wheat, but it is planted in late January or early February. Jute can substitute for rice.
Rice-wheat-sugarcane rotations that extend for several years are common in some areas.
6 The number of people who consume rice and/or wheat produced in rice-wheat cropping systems was estimated
based on: (a) the proportion of rice (and wheat) in total cereals consumption; (b) the proportion of rice (and wheat)
production that comes from rice-wheat areas; and (c) the degree of overlap between the rice-consuming and wheat-
consuming populations (P. Heisey, CIMMYT, pers. comm.).

...........:* :


Ice! ......'~'~`';;"~''` `'~'`~~~ ~

h h e.
:::;:;: :::::1:.::.:..:.::.i~ ~:::1 1: ::177 ijii

Because of differences in numbers of crops and
planting patterns, the intensity of rice-wheat
systems can vary. Cropping intensity is often
close to 200% in simple rice-wheat systems and
can reach 300% or more in areas where land is
scarce and the demand for food strong. Three-
crop rotations typically include two short-
duration rices followed by wheat, or a short-
duration rice followed by a short-duration
vegetable, potato, or oilseed planted before
wheat. Two factors have been particularly
important in enabling farmers to cultivate rice,
wheat, and sometimes another crop on the
same piece of land within the same year. First,
irrigation has been instrumental in facilitating
intensification. Irrigation breaks the traditional
dependence on rainfall which in areas
characterized by a single monsoon period
restricts farmers to one cereal crop per year (or
one rainfed crop and one additional crop
grown immediately following the monsoon
using residual soil moisture). Second, many
modern varieties (MVs) of rice and wheat are
short-duration, non-photosensitive varieties
bred to mature more quickly than traditional
varieties (TVs), which means they can be
planted later or harvested earlier and thus fit
more easily into intensive multicrop rotations.7

Almost all wheat varieties grown in rice-wheat
zones consist of semidwarf MVs. Wheat MVs
are preferred not only because they yield well
and resist diseases, but also because they have
a shorter growing cycle compared to most TVs.
Rice varieties grown in rice-wheat zones are
more variable. Although semidwarf MVs of
rice have been extensively adopted in many

areas, use of tall TVs is still common in some
rainfed lowland and deep-water production
environments where MVs may not be
particularly well adapted. In some areas where
livestock fodder is economically important, rice
TVs may be preferred because of their superior
fodder quality. Aromatic basmati rices (most of
which are TVs) are grown in some areas
because of their high market price, which more
than offsets relatively low yields.

Farming systems in the rice-wheat belt tend to
be characterized by a high degree of
complexity and by a reliance on external
inputs. In marginal production environments,
lack of resources and limited technological
options often foster the emergence of low-input
farming systems in which rice and wheat
cropping are well integrated into the larger
household food economy. In favored
production environments, farmers tend to have
a more commercial orientation, and use of
external inputs (especially inorganic fertilizer)
is usually much higher. Everywhere, animals
play a vital role, consuming the output of crop
production (fodder, grain, straw); providing
food and other products for the farm
household (milk, meat, leather); sustaining the
productivity of the land (organic fertilizer
residues); and serving as a source of power (of
fuel for cooking, of traction for tillage and
transportation). The many links between
cropping, animal husbandry, and other
income-generating activities complicate the
process of technology design, because changes
to one part of the system carry implications for
other parts.

7 The term "modem varieties" (MVs) as used in this paper refers to semidwarf varieties of rice and wheat developed
since 1960. "Traditional varieties" (TVs) refers to older varieties, including many varieties that have never been
worked on by a formal plant breeding program. As Byerlee (1994) has pointed out, the term "modem variety" is
something of a misnomer, since some MVs are now more than 30 years old. However, it is preserved here to
maintain consistency with other publications on the subject. The term "high-yielding varieties" (HYVs), which is
often used to refer to the same varieties, is equally inaccurate, since many semidwarf varieties were bred for
characteristics other than yield potential.

Trends in Rice-Wheat Cropping

Area trends

During the past three decades, the area planted
to rice in South Asia has increased fairly
steadily, at a rate of about 0.7% per year
(Figure 2). The expansion in rice area has been
stimulated by a number of factors, not the least
being the introduction of rice MVs beginning in
the mid-1960s, which in many countries led to
pronounced growth in the area sown to rice.
While rice area has grown fairly steadily, the
sources of growth have changed through time.
During the late 1950s and early 1960s, rice area
was expanded by bringing previously unused
land under cultivation or by planting rice in
areas previously used to grow other crops.
During the 1970s and 1980s, expansion in rice
area was fueled mainly by intensifying the

54 oo o
S52 o
o o
50 o
48 o
46 o o
42 Area = 0.0002t3- 0.0151t2+ 0.6727t + 43.35
R2 = 0.94
1958 62 66 70 74 78 82 86 90

Figure 2. Rice area, South Asia, 1958-91.
Source: IRRI (various years).

cropping pressure on land that was already
under cultivation, often following the
conversion of rainfed land to irrigated land. In
recent years, expansion of the land surface used
for agriculture has all but ceased, and the
proportion of cultivated land that is irrigated is
approaching 100% in many heavily populated
zones. As opportunities for further expansion
dwindle, growth in rice area is expected to
slow and eventually to cease altogether, which
already seems to be happening in northwestern
India (Chaudhary and Harrington 1993). With
population growth continuing at well over 2%
per year, the historical decline in area planted
per capital is likely to accelerate (Figure 3).

The overall trend for South Asia is reflected in
rice area growth patterns within the region's
major rice-producing countries. In India, rice
area grew steadily during the 1960s and 1970s,
with the total area planted to rice increasing

1961 1971 1981 1991

Ei Rice

IEI Wheat

Figure 3. Rice and wheat area per capital,
South Asia, 1961-91.
Source: FAO (various years).

SBecause of considerable year-to-year variability in the data, area and yield growth rates calculated for specific periods
tend to be extremely sensitive to the choice of years (i.e., shifting the starting or stopping point by a single year often
greatly influences the results). For this reason, in discussing national and regional area and yield trends, we have
tried to avoid mentioning average annual growth rates. To facilitate interpretation of the data presented in the
figures, trend lines have been specified as third-degree polynomial functions, a flexible form which allows for two
inflection points. Other specifications (e.g., linear, log-linear, piecewise linear) sometimes provide a better fit, but for
the sake of consistency, the same specification was used throughout.

from about 34 million hectares in 1960 to just
over 40 million hectares by 1980 (Figure 4b).
Since 1980, growth in rice area in India has
slowed, with the 1990s showing virtually no
growth at all.9 In Pakistan, the pattern has been
similar: rice area grew steadily during the
1960s and 1970s before slowing during the
1980s and 1990s (Figure 4d). Bangladesh also
experienced an early period of rapid growth in
rice area during the 1960s and 1970s, followed

4a. Rice area, Bangladesh, 1958-91
10.5 0 0 00
S10.0 0 0
S0 0
E 9.0 o
I 0
8.5. o

1958 62 66 70 74 78 82 86 90

4c. Rice area, Nepal, 1961-91

" 1.4-

0o 0

1961 65 69 73 77 81 85 89

Figure 4. Rice area in Bangladesh,
Source: IRRI (various years).

by a slowdown during the 1980s and 1990s.
However, in Bangladesh the slowdown has
been less dramatic because of the large effect of
boro rice, a crop whose importance was greatly
stimulated by the rapid expansion of irrigation
infrastructure beginning in the mid-1970s
(Figure 4a). In Nepal, rice area continues
growing at its long-term rate of about 1% per
year and thus far shows no sign of slowing
(Figure 4c).

4b. Rice area, India, 1958-92
c 40.0- o
00 0
S38.0- 0 0
36.0- 00-0
< 0

1958 62 66 70 74 78 82 86 90

4d. Rice area, Pakistan, 1958-93
a 1.9-
1.7 -
1.5 00 0
Q 1.3
1,1 0

1958 62 66 70 74 78 82 86 90

9 Not unexpectedly in a country as large as India, the pattern of growth in rice area has varied considerably between states
and sometimes even between districts within states. Numerous studies based on state- and district-level data have shown
that the Green Revolution technologies made an initial dramatic impact in the irrigated tracts of northwestern India before
spreading to less favorable production environments to the east and south. The sudden acceleration in growth in area
evident in the national data presented here was often much more pronounced at the state and/or district level.

India, Nepal, and Pakistan.

Wheat area in South Asia grew in a fashion
similar to rice area during the past three
decades, with a sudden acceleration in growth
following the introduction of MVs and a
subsequent and even more pronounced -
deceleration during the 1980s and 1990s
(Figure 5). Prior to the introduction of MVs,
wheat area in South Asia had been expanding
slowly. Beginning in the mid-1960s, the
introduction of wheat MVs (along with
investment in irrigation) stimulated a surge in
the rate at which wheat area increased. This
rapid expansion continued for nearly two
decades; only in the late 1980s did the rate of
growth slow to earlier levels. As in the case of
rice, growth in wheat area per capital has
declined steadily (Figure 3).

The overall pattern of growth in wheat area for
South Asia was reflected in the patterns of
growth in wheat area within most of the
region's major wheat-producing countries. In
India, wheat area had been stagnant prior to

Figure 5. Wheat area, South Asia, 1958-93
Source: FAO (various years).

the mid-1960s. Following the introduction of
wheat MVs, wheat area expanded dramatically
beginning in 1967 (Figure 6b). This growth
spurt ceased almost as suddenly as it had
started; since 1983, growth in wheat area in
India has slowed markedly.10 The pattern is
similar in Nepal: after remaining flat through
the late 1960s, wheat area suddenly expanded
beginning in 1967, growing rapidly for nearly
two decades before abruptly slowing in the late
1980s (Figure 6c). In Bangladesh, where wheat
was not widely grown prior to the 1970s, wheat
area experienced a decade of dramatic growth
before slowing sharply during the 1980s and
1990s (Figure 6a).1 In Pakistan, where wheat is
the most important food crop, wheat area
continues to expand at its long-term rate, and
the expansion shows few signs of slowing
(Figure 6d).

Across the region, the general picture has been
one of an acceleration in the area planted to rice
and wheat, followed by a pronounced
slowdown in the rate of expansion. Although
the timing of these events differed slightly
between countries (and although in one or two
cases the slowdown has yet to occur), generally
the spurt in growth was concentrated in the
period following the introduction of Green
Revolution technologies (MVs, fertilizer,

To what extent were changes in the area
planted to rice and wheat (considered
separately) mirrored by changes in the area
under rice-wheat cropping sequences?
Although no official statistics are published on
rice-wheat sequences, estimates have been
made for selected years for a geographic

10 As in the case of rice, the sudden acceleration in growth in wheat area evident in the national data presented here
was often much more pronounced at the state and/or district level.
1 An important factor underlying the sudden increase in wheat area in Bangladesh was the release of the variety

information system under development by a
consortium of national and international
agricultural research programs.12 During the
periods for which data are available (these
periods vary from country to country), the area
under rice-wheat cropping sequences
expanded more rapidly (in percentage terms)
than the areas planted to each crop (Table 3).
This is hardly surprising, considering that

6a. Wheat area, Bangladesh, 1968-93

0.7 o o

0.6 0 00 00
S0 0 0
c 0.5
o 0.4

E 0.3
a 0.2-
< 0o Area = -0.1662t3+ 6.0714t2-
0.1 30.229t + 139.67
1968 72 76 80 84 88 92

6c. Wheat area, Nepal, 1961-93


0.6 oo00

0.5 0 0

0.4 00o

0. Area= -0.0154t3 + 0.8088t2 +
0.1 co0 6.1988t + 91.234
R2= 0.98
1961 65 69 73 77 81 85 89 93

growth in rice and wheat area (considered
separately) has occurred during a period when
the amount of arable land used for agriculture
in South Asia has barely increased. In other
words, farmers have expanded their rice and
wheat area primarily by increasing the
cropping intensity of land already under
cultivation. Frequently farmers have achieved
this objective by introducing a second cereal

6b. Wheat area, India, 1958-93
O9 n.,

c 20.0
c, 16.0
< 14.0
in n

1958 62 66 70 74 78 82

6d. Wheat area, Pakistan, 1958-93

86 90

7.5 o
6.5 0
6.0 o0 0
5.0 0 Area = -0.0146t3 + 0.8204t2 +
0 88.332t + 4623
4.5 O R2=0.96
1958 62 66 70 74 78 82 86 90

Figure 6. Wheat area in Bangladesh, India, Nepal, and Pakistan.
Source: FAO (various years).

12 Since governments in South Asia do not report official statistics on cropping patterns, trends in the area under rice-
wheat rotations must be estimated based on trends in the area planted to rice and trends in the area planted to wheat.
In this report, the area under rice-wheat rotations in Bangladesh, India, Nepal, and Pakistan was estimated by
multiplying national data on the areas planted to rice and wheat by coefficients obtained from surveys done in
representative districts on the proportion of rice area that is also planted to wheat and the proportion of wheat area
that is also planted to rice. Until official crop production statistics reported by each country include the area under
specific cropping rotations, such ad hoc approaches will be necessary.

o o oo


0 0 o Area = 0.0547t4 4.8574t + 133.98t2 -
oo 807.26t + 14121
R2= 0.97

crop into the cropping pattern. Because rice is
normally grown in the warm, humid kharif
season and wheat in the cool, dry rabi season,
the introduction of a second cereal crop often
results in rice-wheat combinations.

Looking to the future, it is unclear how long
rice-wheat area can continue to spread.
Prospects for bringing new land under
cultivation are limited; in fact, the physical
supply of crop land is declining in South Asia
as more and more surface area is converted to
industrial or residential use.13 Adoption of
land-saving technologies will allow growth in
cultivated area to continue for some time, even
on a shrinking physical land base, but sooner
or later the rate of expansion is bound to slow.
The outlook is not promising, especially
considering that cultivated area is expanding at
a rate that lags considerably behind the
population growth rate.

Table 3. Growth in area under rice-wheat
cropping sequences, South Asia

Rice- Rice-
wheat wheat
area area Annual
(million (million growth
ha) ha) (%)

Bangladesh 1960-63 1987-90 1960-63 to 1987-90
0.05 0.50 8.6

India 1959-62 1986-89 1959-62 to 1986-89
3.97 9.53 3.2

Nepal 1975-76 1991-92 1975-76 to 1991-92
0.26 0.43 3.0

Pakistan 1970-71 1987-88 1970-71 to 1987-88
1.02 1.38 1.7

Source: Calculated from data provided in Huke, Huke,
and Woodhead (1994a, 1994b); Woodhead, Huke, and
Huke (1994); and Woodhead et al. (1994).

Yield trends

With growth in rice and wheat area apparently
flagging, attention is increasingly being
directed to the second determinant of
production: yield. What has been the history of
growth in rice and wheat yields in South Asia?

Rice yields in South Asia increased noticeably
following the introduction of MVs. In districts
where MVs made their greatest impact, the
effect was so pronounced that some observers
described it as a "yield takeoff" (Plucknett
1993). The upturn in rice yields was evident
even at the aggregate level; beginning in the
late 1960s, average rice yields for the region as
a whole began to rise steadily after having
remained flat for many years (Figure 7).

Yield growth in individual countries has often
differed from the regional pattern. In India, rice
yields rose steadily from 1967 to 1987 before
leveling off (Figure 8b). Although the
slowdown in yield growth in India has
occurred partly because more farmers grow
low-yielding TVs that nevertheless command
price premiums for their superior quality


1 2.0- 062
1.5 00

> 1.0

0.5- Yields = 0.00003t3 0.0003t2 + 0.0165t + 1.4268
R2= 0.93
1958 62 66 70 74 78 82 86 90

Figure 7. Rice yields, South Asia, 1958-91.
Source: IRRI (various years).

13 This phenomenon is already evident in China, where hundreds of square kilometers of rice and wheat land are lost
every year (Wehrfritz 1995).

(e.g., basmati), rice yields have stagnated even
where no such varietal substitution has
occurred. In Pakistan, the pattern has been
similar. Rice yields experienced a sharp burst in
growth beginning in 1967 before slowing to a
barely positive long-term trend (Figure 8d). As
in India, varietal substitution has contributed to
slower growth in yields in Pakistan, although it
does not explain it entirely. In Bangladesh, the
picture is a bit different. Rice yields in
Bangladesh began to rise during the early 1970s
following the introduction of MVs. However,
unlike elsewhere in the region, in Bangladesh
the rate of yield growth has not slowed
appreciably (Figure 8a). The continuing strong
growth in rice yields in Bangladesh almost
certainly has resulted from the increasing
importance of high-yielding boro (winter) rice,

8a. Rice yields, Bangladesh, 1958-91
0 00
" 2.0
1.5 000 0 00

1 1.0

0.5 Yields = 0.0000007t3 0.0026t2 + 0.0431t + 1.44
R2 = 0.93

1958 62 66 70 74 78

8c. Rice yields, Nepal, 1958-92



. 1.0

82 86 90

0.51 Yields = 0.000006t3 0.0023t2 + 0.0217t + 1.87
R2 = 0.42
1958 62 66 70 74 78 82 86 90

which is grown under irrigation and which
benefits from high levels of solar radiation and
cooler temperatures during the growing season.
In Nepal, rice yields have shown a modest
upward trend characterized by considerable
year-to-year variability (Figure 8c).

Wheat yields in South Asia have followed a
growth pattern similar to that of rice yields,
increasing sharply following the introduction of
MVs and rising steadily ever since. From 1966
to 1990, average wheat yields for the region as a
whole more than doubled, rising from just over
1.0 t/ha to about 2.3 t/ha (Figure 9).

As in the case of rice, wheat yield growth in
individual countries has differed from the
overall regional pattern. In India, where the

8b. Rice


- 2.0


-> 1.0

yields, India, 1958-92

0.0 1
1958 62 66 70 74 78

82 86 90

8d. Rice yields, Pakistan, 1958-93

2.5o oo o oooo
c 2.0

1.5 0oooo
> 1.0
0.5 Yields = -0.00001t3- 0.0013t2 + 0.1004t + 0.98
1958 62 66 70 74 78 82 86 90

Figure 8. Rice yields in Bangladesh, India, Nepal, and Pakistan.
Source: IRRI (various years).

0 0


Yield = 0.00001t3 + 0.0006t2 + 0.0033t + 1.43
R2= 0.92

yield takeoff occurred in 1967, wheat yields
have since grown steadily (Figure 10b). In
Pakistan, the yield takeoff also occurred in 1967,
although the yield growth that followed was
less impressive (Figure 10d). In Bangladesh, the
picture once again is a bit different. Wheat
yields in Bangladesh accelerated sharply during
the late 1970s and early 1980s as the result of
government efforts to stimulate domestic wheat
production. However, the growth in wheat
yields ceased abruptly in the mid-1980s, and
wheat yields actually fell in subsequent years
(Figure.10a). The decline was caused by a
combination of technical factors (e.g.,
conversion of prime wheat land to irrigation for
use in boro rice production) and economic
factors (e.g., removal of government subsidies
on fertilizer). In Nepal, no clear yield takeoff is
discernible; wheat yields have shown
considerable year-to-year variability around a
modest long-term upward trend (Figure 10c).

To summarize, yields of both rice and wheat
increased substantially following the
introduction of Green Revolution technologies.
However, the timing of the initial takeoff in
yields and the subsequent pattern of yield
growth varied between countries and even
between districts within countries. First to be

2.0 0

_ 0000 0
.7 1.0

Yields = -0.00001t3 + 0.0012t2 +
0.019t + 0.73 R2= 0.98
1958 62 66 70 74 78 82 86 90

Figure 9. Wheat yields, South Asia, 1958-93.
Source: FAO (various years).

10a. Wheat yields, Bangladesh, 1968-93


. 1.0

1968 72 76 80 84 88 92

10b. Wheat yields, India, 1958-93




. 1.0

1958 62 66 70 74 78 82 86 90

Figure 10c. Wheat yields, Nepal, 1961-93





0.0 i
1961 65 69 73 77 81 85 89 93

10d. Wheat yields, Pakistan, 1958-93

2.0- 0

o 0
S1.0 oo

0.5 Yields = -0.00004t3+ 0.0019t2 +
0.015t + 0.74 R2 = 0.96
1958 62 66 70 74 78 82 86 90
Figure 10. Wheat yields in Bangladesh,
India, Nepal, and Pakistan.
Source: FAO (various years).

0 0

0o0 0

Yields= -0.0000033 + 0.0008t2 +
0.0246t + 0.72 R2= 0.98

000 0 0 0 00
0 0 0
Yields = -0.00007t3 + 0.0043t2 -
0.0648t + 1.38 R2= 0.56

affected were favorable production
environments in India and Pakistan, where
yields rose dramatically beginning in the mid-
1960s.14 Following a lag, the Green Revolution
technologies spread into less favorable
production environments in both countries, but
since adoption proceeded at a slower pace and
was less extensive in these environments, yield
increases were marginal (Figure 11). In
Bangladesh and Nepal, where the Green
Revolution technologies were relatively slow in
arriving, the increase in rice and wheat yields
became apparent only in the mid-1970s.

Impact on production

Widespread adoption of Green Revolution
technologies and the resulting expansion in
area planted to rice and wheat led to marked
increases in rice and wheat production
throughout South Asia (Table 4). In 1961, rice
and wheat accounted for about three-quarters
of total cereal production in the region; by 1991,
the share of these two commodities had risen to
nearly 90%. In some areas, the impressive
production gains enabled food supplies to grow
faster than the population, increasing
production of cereals per capital and reducing
dependence on politically undesirable imports,
both commercial imports and food aid.
Elsewhere, the production gains have not kept
pace with population growth, and per capital
production of rice and wheat have declined.

Input use trends

The yield gains achieved following the
introduction of rice and wheat MVs were made
possible partly by superior germplasm, partly
by increased use of inputs (especially fertilizer
and irrigation), and partly by interactions
between inputs. In assessing prospects for
future growth in the productivity of rice-wheat
systems, it is important to appreciate the role
played by these three key inputs in the past.

11a. Rice yields
'F 2.5
- 2.0
Z3 1.5
1960 65

70 75 80 85 90

11b. Wheat yields
3.0 Punjab (favorable
l2.5 environment)
1.0 Bihar (unfavorable
0.5 environment)
1960 65 70 75 80 85
Figure 11. Rice and wheat yields in Punjab
and Bihar States, India, 1960-90.
Source: Fertiliser Association of India (various years).

14 Throughout this report, the term "favorable production environments" is used in a general sense to denote areas
where rice and wheat yields are not unduly limited by agroclimatic constraints (e.g., moisture stress, heat stress, soil
fertility imbalances). Examples of favorable production environments would be the irrigated zones of the Indian and
Pakistani Punjabs. Conversely, the term "marginal production environments" is used to denote areas where rice and
wheat yields are constrained by agroclimatic constraints, such as the mountainous zones of Nepal. Whether or not an
environment is considered favorable or marginal for rice and/or wheat production depends not only on naturally
occurring agroclimatic conditions, but also on technical and institutional factors (e.g., the presence or absence of a
functioning irrigation system, the presence or absence of an effective fertilizer distribution network).

Improved germplasm The Green Revolution
in South Asian agriculture has been described
often. During the late 1960s and early 1970s,
semidwarf varieties of rice and wheat were
introduced into India, Pakistan, Bangladesh,
and Nepal. When grown with increased levels
of fertilizer and an assured water supply, these
MVs performed significantly better than the
older, taller varieties they replaced, leading to
substantial production increases and higher
incomes for millions of farmers who adopted
the technology.

The salient feature of the early rice and wheat
MVs achieved largely through plant
breeding was their increased yield potential.
By incorporating dwarfing genes, plant
breeders developed shorter, stronger plants
which were able to produce more seed than the
old, tall TVs and whose stems were strong

enough to support additional grain weight. A
less visible feature bred into MVs was their
resistance to the pests and diseases that limited
yields of many TVs.15 Since the original MVs
were introduced, subsequent generations of
MVs have been released, offering higher yield
potential, improved pest and disease resistance,
better grain quality, and other desirable

Rice and wheat MVs spread rapidly
throughout many of the irrigated districts
where rice-wheat cropping is concentrated.
These varieties spread more gradually into less
favorable environments, as evidenced in India
by steady growth in the proportion of the area
planted to MVs, which has continued into the
1990s (Figure 12). Adoption of MVs has been
more extensive in wheat than in rice because
certain high-quality TVs of rice, such as the

Table 4. Rice (paddy) and wheat production (million t) in South Asia, 1950-90
(three-year averages)

Rice production
South Asia

Wheat production
South Asia











Sources: Rice data from IRRI (various years); wheat data for India from Fertiliser Association of India (various
years); data for Nepal from Ministry of Agriculture (various years); data for Pakistan from Ministry of Agriculture
(various years) and PARC (n.d.); data for Bangladesh from Bangladesh Bureau of Statistics (n.d.).

15 The widely held belief that rice and wheat TVs were more resistant than MVs to diseases and insect pests is
incorrect. In fact, the opposite is true. For example, many rice TVs were highly susceptible to blast and tungro virus,
diseases that cause few problems today because of the incorporation of host plant resistance into MVs through
breeding. Similarly, stem rust was a major disease of many wheat TVs, but because of the incorporation of Sr34
resistance genes into MVs, this disease rarely causes yield losses today.
16 Most of the original semidwarf rice MVs were developed at the International Rice Research Institute (IRRI), and
most of the original semidwarf wheat MVs were developed at the International Maize and Wheat Improvement
Center (CIMMYT).

basmati varieties grown in parts of India and
Pakistan, remain quite popular. The success of
rice and wheat MVs in marginal environments
has not always been recognized, perhaps
because adoption proceeded more slowly and
because the effect on yields was less
dramatic there.

Fertilizer Since MVs express their full yield
potential when soil fertility is high, many
farmers who adopted MVs found it profitable
to invest heavily in inorganic fertilizer.17 The

12a. Adoption of rice MVs
40 --

2 30
3 0 dt ai 8 -l-ariet I .(-TVs) .
S15 ... Modern-varieties -

0 __
1967 69 71 73 75 77 79 81 83 85 87 89 91 93

use of fertilizer on rice and wheat soared
throughout South Asia following the arrival of
MVs (Figure 13). Fertilizer use was further
encouraged by a sustained decline in global
fertilizer prices, which governments were
happy to pass on in the form of lower retail
prices. As a result of the sharp increase in
fertilizer use, fertilizer application levels in
many districts are now at or above
recommended levels (Table 5).

Water control The superior performance of
MVs soon stimulated increased investment in
irrigation, since high yields were critically
dependent on assured water supplies. The late
1960s and early 1970s saw an unprecedented
surge in investment in irrigation throughout
most of South Asia. When rice and wheat MVs
were introduced in the mid-1960s, less than
one-quarter of the area planted to rice and
about one-half of the area planted to wheat in
South Asia was irrigated. By 1990, about one-
half of the area planted to rice and over three-
quarters of the area planted to wheat was
irrigated.18 Growth in irrigated rice and wheat

12b. Adoption of wheat MVs

E 15
. 10 '. Modern varieties'
0. '-;. --". .: ". ." .I '- '
1967 69 71 73 75 77 79 81 83 85 87 89 91 93

Figure 12. Adoption of rice and wheat MVs,
India, 1967-94.
Source: Fertiliser Association of India (various years).

u 1

65 70 75 80 85

Figure 13. Fertilizer applied to rice in
Bangladesh, India, and Pakistan, 1960-88.
Source: IRRI (various years).

17 Contrary to the conventional wisdom, MVs grown without added fertilizer sometimes perform better than TVs,
although the yield difference is sometimes small.
18 In interpreting these figures, it is important to keep in mind that a significant proportion of the area classified as
"irrigated wheat area" is only partially irrigated. In many areas, even though irrigation facilities are in place, water is
available only during part of the cropping season.

area continues (Table 6), reflecting both
ongoing conversion of rainfed land to irrigated
land (Figure 14), as well as displacement by
rice and wheat of other, less profitable crops.

Public and private sector investment in
irrigation infrastructure continues to fuel
expansion in irrigated area throughout South
Asia. However, the incentives to invest in
irrigation facilities have declined for at least
two reasons. First, many irrigation systems

Table 5. Fertilizer use on rice and wheat in
selected districts of India

Rice 1970-71 1990-91
Punjab (kg N/ha) 87 137
Haryana (kg NPK/ha)a 40 170
Wheat 1970-71 1990-91
Punjab (kg N/ha) 54 172
Haryana (kg NPK/ha)a 40 180
Source: Chaudhary and Harrington (1993), Gill (1992), and
Sidhu and Byerlee (1990).
a Total of nitrogen (N), phosphorus (P), and potassium (K).

Table 6. Long-term trends in irrigated rice
and wheat area, South Asia, 1950-90
Crop Irrigated area (000 ha)
and time
period Bangladesh India Nepal Pakistan Total
1949-51 120 9,844 na 967 1,091
1959-61 305 12,522 na 1,200 14,027
1969-71 958 14,207 na 1,527 16,692
1979-81 1,276 16,787 188 1,981 20,232
1989-91 2,530 19,200 350 2,087 24,186
1949-51 na 3,575 na 3,424 6,999
1959-61 na 4,170 na 3,826 7,996
1969-71 na 8,763 na 4,898 13,661
1979-81 170 15,173 na 5,374 20,717
1989-91 273 18,000 na 6,429 24,702
Source: IRRI (1991); Government of Pakistan (various years).
Note: na = not available.

have become degraded through lack of
maintenance, so a considerable portion of
irrigation investment now must be devoted to
rehabilitating existing systems, rather than to
constructing new ones. Second, further
expansion of irrigation capacity requires that
water control structures be constructed in
increasingly remote areas, which tends to be
extremely costly. With global prices of rice,
wheat, and maize nearing historic lows, the
returns to cereal production are often too low
to justify installing irrigation facilities
exclusively for grain production. Opportunities
for profitable investment still exist, but
improved impact assessment procedures will
be needed to identify them (Rosegrant and
Svendsen 1993).

Thirty years after the onset of the original
Green Revolution in South Asia, it is becoming
increasingly clear that the engine of growth is
running out of steam. Modem varieties,
fertilizer, and irrigation offer increasingly
limited prospects for raising yields in the
future, especially in areas where adoption of
these inputs is already extensive.19 Evidence
for this conclusion comes from a number of

1951 56

61 66 71 76 81 86

Figure 14. Rainfed and irrigated wheat area,
India, 1951-88.
Source: Fertiliser Association of India (various years).

19 Although future generations of MVs are unlikely to deliver yield increases which are proportionally as large as those
delivered by the first generation of MVs, regular replacement of older MVs by newer ones should help farmers to
maintain current high yield levels even as older MVs lose their resistance to diseases and pests. This benefit is
extremely important and should not be overlooked.

intensively cultivated districts in northwestern
India and northeastern Pakistan, where growth
in rice and wheat yields has slowed noticeably
during the past two decades, even though
farmers have replaced older MVs with newer
ones and have tripled fertilizer doses.20
Although MVs, fertilizer, and irrigation do
have the potential to deliver productivity gains
in areas where their use is still suboptimal (e.g.,
eastern India, much of Bangladesh and Nepal),
in the heart of the South Asian grain belt where
rice-wheat cropping systems are most
productive, these traditional sources of
productivity growth are largely exhausted.

Trends in factor productivity

When the performance of agricultural
production systems is assessed, yield trends are
often used to measure changes in overall
productivity. However, yield trends can be
misleading when the quantity and/or quality
of inputs changes significantly through time.
Changes in productivity can be assessed more
accurately using measures that take into
account changes in input use.

Total factor productivity (TFP) studies -
Changes in the productivity of agriculture are
frequently evaluated using some index of total
factor productivity (TFP). Various methods
have been proposed for measuring TFP, all of
which relate changes in agricultural output to
changes in the use of inputs (see Christiansen
1975, Christiansen and Jorgenson 1970, Ball
1985, Antle and Capalbo 1988, and Barnett et al.
1994). When TFP is observed to increase
through time, this trend is interpreted as
evidence of productivity growth attributable to

factors other than increases in the amounts of
inputs used. Such factors can include
technological change, changes in the quality of
inputs, and/or changes in the physical or
economic environment.

Originally conceived as a tool for measuring
production efficiency over short periods, TFP
indices have received new attention in recent
years as indicators of the long-term
sustainability of agricultural systems. This use
of TFP indices was first proposed by Lynam
and Herdt (1989), who pointed out that any
sustainable production system would have to
be characterized by a non-negative trend in
TFP over an extended period. Others have
subsequently built upon this idea by
attempting to broaden the range of inputs and
outputs reflected in the TFP index, for example
by including environmental and social costs
and benefits not usually considered in short-
run productivity analysis, including long-term
effects and off-site effects.21 Although this work
is still in its initial stages, important progress
has been made in thinking through conceptual
issues and in proposing innovative approaches
for expanding the conventional indices (for
example, see Steiner and Herdt 1995, Ehui and
Spencer 1990, Whitaker and Lalitha 1993, and
Harrington, Jones, and Winograd 1994).

One widely used TFP index is the chain-linked
Tornqvist-Theil approximation to the Divisia
index (or Torqvist-Divisia index), which is
relatively simple to calculate and easy to
interpret. Although the theoretical
underpinnings of the Torqvist-Divisia index
are conceptually straightforward, empirical
calculation of the index may be complicated by

20 Evidence cited by Rosegrant and Pingali (1994) suggests that similar yield stagnation has occurred in intensively
cultivated rice-rice systems.
21 Since the effects of resource degradation can for a long time be masked by technical change, TFP indices as they are
currently calculated (i.e., omitting many environmental costs) are usually inadequate as indicators of sustainability.

any number of practical difficulties, including
the accurate measurement of production inputs
and outputs (particularly for individual
commodities, since disaggregated statistics are
rarely available); the valuation of production
inputs and outputs (since market prices are
often distorted by policies and/or market
failures); and definition of the boundaries of
the system being analyzed (since TFP may be
calculated at the plot level, the farm level, the
village level, the district level, the national
level, the regional level, and so on).

Quite a few country-level studies in South Asia
have reported TFP trends for the entire
agricultural sector. Analyzing the agricultural
sector as a whole simplifies the data
requirements, since inputs and outputs need
not be disaggregated by commodity. On the
other hand, the results are less useful than
commodity-specific studies, since productivity
increases realized in one part of the agricultural
sector can conceal productivity decreases
occurring in another part. Studies of TFP
focusing on the agricultural sectors of
individual countries in South Asia have
produced inconsistent results, in part because
of methodological differences. For India,
Evenson and McKinsey (1991) found little
evidence that TFP declined between 1956 and
1984, a finding subsequently corroborated by
Pray (1991) and by Rosegrant and Evenson
(1992) in studies covering slightly different
periods. For Pakistan, Evenson and Bloom
(1993) reported evidence of a sharp decline in
TFP during the post-Green Revolution period
(1975-85), but since this study assumed a fixed
share for land, the results were biased
downwards and in fact may not have been all
that different from those reported for India.

Studies of TFP focusing specifically on rice-
wheat cropping systems are less common.
Noteworthy among these has been the work by
Ali and Velasco (1994) on productivity trends in
the Punjab and Sind of Pakistan during the
1970s and 1980s. Total factor productivity in
rice-wheat cropping systems fell during each of
the two decades, with the decline in TFP
averaging more than 2% per year in both study
zones during the most recent period (Table 7).22
Ali and Velasco attribute the decline in TFP to
the deterioration of resources and argue that if
this trend continues, more and more inputs will
be required to maintain current output levels.
They conclude pessimistically that current
production practices are consuming resources
which otherwise would remain available for
future generations.

Evidence of declining productivity in
intensively cultivated rice-wheat systems also
has been reported by Cassman and Pingali
(1995), who calculated TFP indices for 1970-88
using data collected in Ludhiana District in the
Indian state of Punjab. Cassman and Pingali
found that TFP increased steadily until the mid-
1970s, when it leveled off for about a decade

Table 7. Trends in total factor productivity
for rice-wheat cropping systems, Pakistan,
1970-79 and 1980-89

Province 1970-79 1980-89

Punjab (% annual growth) -2.00" -2.90"*
Sind (% annual growth) -0.24 -2.60*

Source: Ali and Velasco (1994).
Note: *,** indicate significance at the 5% and 1% level,

2Because Ali and Velasco used an analytical framework which did not explicitly incorporate changes in land use, their
results are not directly comparable with some of the other results cited earlier and should therefore be interpreted
with caution.

before beginning slowly to drop off. Based on
extensive analysis of soils data, Cassman and
Pingali attribute the decline in TFP to changes
in the physical and chemical properties of
repeatedly flooded and dried rice-wheat soils;
these changes reduce the soils' nitrogen-
supplying capacity and inhibit the efficient
uptake of nitrogen by rice and wheat plants.

In addition to the relatively few TFP studies
that have focused on rice-wheat systems, other
studies have focused on rice or wheat
separately. No attempt will be made here to
review all of these. However, it is worth citing
one or two to provide some idea of the
accumulating evidence on productivity
changes in areas where the Green Revolution
made its greatest impact.

Sidhu and Byerlee (1991) estimated
productivity changes for wheat in the Indian
Punjab during the 1970s and 1980s. They
concluded that despite consistent
intensification of input use, productivity gains
approaching 2% per year were achieved
throughout the period of analysis. These
productivity gains can be attributed in roughly
equal proportions to the adoption of land-
saving technology (e.g., MVs, fertilizer) and the
adoption of labor-saving technology (e.g.,
machinery). Although Sidhu and Byerlee
discerned no signs that the high productivity
levels achieved by the late 1980s are in
jeopardy, they warn that future sources of
productivity gains capable of increasing TFP at
rates equal to those achieved in the past are not

Mohan Dey and Evenson (1991) examined
productivity trends in rice, wheat, and other
crops in Bangladesh from 1952 to 1989. These
authors observed that slowdowns in the rate of
yield growth were not necessarily attributable
to biological deterioration of MVs or to

environmental degradation; rather, declines in
average yields were probably inevitable as
MVs spread into marginal lands of lower
production potential. Furthermore, the decline
in yield may not have been nearly as
pronounced as had commonly been assumed.
Considering that second- and third-generation
MVs tended to mature much more rapidly than
the original MVs, average yield per day may
not have declined at all. Based on their analysis
of TFP trends, Mohan Dey and Evenson
concluded that during the period of analysis,
modest productivity gains averaging slightly
less than 1% per year were achieved in both
rice and wheat. Decomposition of the TFP
index suggested that the productivity gains
were attributable in large part to research

Partial factor productivity (PFP) studies -
Unlike TFP indices, which relate changes in
output to changes in the use of all inputs,
partial factor productivity (PFP) indices relate
changes in output to changes in the use of
individual inputs. Indices of PFP have
advantages and disadvantages: they can
provide useful information about the efficiency
with which individual inputs are used, but
they may not provide an accurate view of
trends in overall productivity, because a
change in the productivity of one factor is
frequently accompanied by offsetting changes
in the productivity of other factors.

Byerlee and Siddiq (1994) disaggregated the
effect of three factors on irrigated wheat yields
in the Punjab of Pakistan: (1) initial adoption of
MVs, (2) replacement of old MVs with newer
MVs, and (3) increasing fertilizer application
levels. After the (positive) effects of these three
factors had been accounted for, wheat yield
trends showed a significant negative residual,
suggesting that other factors were responsible
for a long-term decline in yields (and hence

productivity). This conclusion is supported by
observed trends in yields of wheat MVs, which
have remained unchanged during the past two
decades despite extensive adoption of MVs,
replacement of old wheat MVs by new MVs,
and significant increases in the amount of
fertilizer applied to wheat.

Kumar and Rosegrant (1994) similarly
disaggregated productivity trends for rice in
India from 1971/72 to 1988/89. For the country
as a whole, productivity growth averaged 1.0%
over the entire period of analysis, with market
infrastructure, research investment, the
availability of canal irrigation, and balanced use
of fertilizers representing the most important
sources of productivity growth. However,
productivity trends were found to vary
considerably between regions (Table 8).
Productivity growth was reported to be strong
in the southern region and modest in the
northern and eastern regions, but in the western

Table 8. Trends in input, output, and total
factor productivity for rice in India, 1971/72
to 1988/89
Total Total Total factor
input output productivity
Eastern 1.81* 2.17* 0.36**
Western 1.74* 0.76 -0.98
Northern 6.03* 6.79* 0.76*
Southern 1.12* 2.97* 1.85*
All India (excluding
Western Region):
1971-80 2.99* 4.30* 1.31*
1981-88 2.13' 3.10* 0.97*
1971-88 2.49* 3.52* 1.03*
Source: Kumar and Rosegrant (1994).
Note: *,** indicate significance at the 5% and 1% level,
respectively. Eastern Region comprises Assam,
Bihar, Orissa, West Bengal; Western Region
comprises Gujarat, Maharashtra, Madhya
Pradesh, Rajasthan; Northern Region comprises
Punjab, Haryana, Uttar Pradesh; and Southern
Region comprises Andhra Pradesh, Tamil Nadu,
Karnataka, Kerala.

region productivity declined by an alarming
1.0% per year. Productivity growth in rice also
was found to have declined over time, with
average annual increases smaller during the
post-Green Revolution period (1981-88) than
during the Green Revolution period (1971-80).

Micro-level evidence: long-term experiment
station data Despite methodological
inconsistencies that complicate the comparison
of empirical results, the factor productivity
studies suggest that it may be dangerous to
look at rising yields and conclude that all is
well in South Asia's rice-wheat cropping
systems. Since the onset of the Green
Revolution, average rice and wheat yields have
indeed increased throughout the region.
However, we now know that yield gains in
farmers' fields have been achieved only
through the application of ever-increasing
amounts of inputs. When changes in the
quantity and quality of inputs are taken into
account, the picture changes: yields frequently
not only stagnate, but in some cases they
actually decline.

Why has this development not been recognized
more widely by researchers? One reason is that
researchers may not have been looking in the
right places. Most rice and wheat research in
South Asia has tended to focus on the
development of technologies designed to
increase productivity in the short run, generally
one or two cropping cycles. Long-term trials -
including monitoring trials designed to track
developments in farmers' fields have been
seen as time-consuming and costly, so they
have rarely been undertaken. Only recently has
the emerging realization that potentially
important long-term processes are being
overlooked induced some researchers to extend
the time frame of their analyses in an attempt to
identify and understand factors that may be
affecting productivity over the longer term.

In the wake of this shift in perspective, evidence
of long-term productivity declines is beginning
to show up at the experimental level. A number
of soil fertility trials have examined the
behavior of rice and wheat yields over the long
run in continuous rice-wheat cropping systems.
These trials have generally examined the effects
of different combinations of fertilizer. Although
most research has focused on the effects of
major nutrients (N, P, K), some trials have
included treatments with organic fertilizers, and
a few have even considered the effects of
micronutrients, such as zinc, sulfur, and boron.
Few of these long term-trials have been
properly analyzed and written up, but even
partial results provide valuable insights into the
long-term behavior of rice and wheat yields in
rice-wheat cropping systems.

Researchers in Nepal are conducting a set of
trials designed to shed light on the long-term
performance of rice and wheat yields under
intensive two- and three-crop rotations (Giri,
Acharya, and Regmi 1994). Data from rice-rice-
wheat trials planted at the Bhairahawa research
station show that in plots treated with
recommended applications of nitrogen (N),
phosphorus (P), and potassium (K) fertilizer,
yields of both rice crops (planted early and
planted on time) clearly have declined; yields of
wheat may also have fallen slightly, although
the rate of decline is not statistically significant
(Table 9). In plots treated with farm yard
manure (FYM), declining yields have been
observed in all three crops. Since application of
FYM has increased the organic matter content of
the soils and improved their nutrient status
compared to the plots treated only with
chemical fertilizer, factors other than soil
fertility per se appear to be limiting yields. When
phosphorus is omitted from the treatment,
yields decline even more rapidly.

Also in Nepal, yield declines in wheat have
been observed in varietal trials conducted at six
research stations in the lowland Terai (Morris,
Dubin, and Pokhrel 1994). Between 1976 and
1990, yields of the check variety RR21 declined
at an average annual rate of 6.8%, while yields
of the variety UP-262 declined more slowly at a
rate of 4.3% (Figure 15). During the course of
the trial, RR21 became susceptible to leaf rust
and blight, and the difference in the rate of
yield decline experienced by the two varieties
(2.5 %) can be attributed to disease effects.

Table 9. Trends in rice and wheat yields,
long-term trials, Bhairahawa, Nepal, 1976-90
(average annual percentage change)


100-30-30 FYM 100-0-30

Rice planted on time -4.51* -7.75* -27.92**
Rice planted early -10.15** -7.74* na
Wheat planted on time -0.99 -2.56* -11.90**

Source: Calculated by the authors from data reported
in Giri, Acharya, and Regmi (1994).
Note: *,** indicate significance at the 5% and 1% level,
a Rice yields adjusted to account for varietal change
occurring in 1987.


5 4 ....
Z 3

O 2
2 --- RR21 yield trend = -6.8% per year
S .......... UP-262 yield trend = -4.3% per year
(t= i ---- i -- -------

1976 78

80 82 84 86

88 90

Figure 15. Mean yields of RR21 and UP-262
in the Terai, Nepal, 1976-90.
Source: Morris, Dubin, and Pokhrel (1994).

However, the fact that both varieties
experienced significant declines in yield despite
constant levels of management suggests that soil
fertility or other as yet unidentified factors were
depressing yields.

Trials designed to measure the long-term
behavior of rice and wheat yields under
continuous multiple cropping also are underway
in India. Experiments at Pantnagar, Uttar
Pradesh, show declining yields in intensive rice-
wheat systems when input levels are kept
constant (Figure 16) (Nambiar 1994). Yield
declines in intensive rice-wheat systems have
also been recorded in Faizabad, Uttar Pradesh,
although there is evidence of an eventual
leveling off of yields (Figures 17a, 17b).

The results of the Faizabad experiment are
particularly interesting, because they provide
clear evidence of long-term degradation in the
natural resource base. In Faizabad, yield
declines have occurred across a wide range of
fertilizer application rates, suggesting that there
has been a downward shift in the entire
fertilizer response function. Such a shift is
illustrated in Figure 18. As a result of resource
degradation, the yield level achieved at fertilizer
application rate X0 has declined from Yo to Y1.
Only by increasing fertilizer application rate to

8 Rice yield trend = -2.8% per year

6- .

c 3
2 2 Wheat yield trend = -0.5% per year
1973 75 77 79 81 83 85 87

Figure 16. Rice and wheat yields, long-term
trials, Pantnagar, U.P., India, 1973-88.
Source: Nambiar (1994).

X1 can the original yield level Yo be maintained.
Because the entire response production
function has shifted down, this effect is evident
at all levels of fertilizer application.

17a. Rice yields in long-term trials
d 8n 4.5
S 4.0 N-P-K 120-0-av
3.5 -P-K80-0-av
'-: 3.0
" 2.5 N-P-K 40-0-av
> .E 2.0
2 1.5 N-P-K 0-0-0
a > 1.0
1979 80 81 82 83 84 85 86 87 88

17b. Wheat yields in long-term trials


s ="

4.0 N-P-K 120-0-av
3.5 N-P-K 80-0-av
2.5. N-P-K 40-0-av
1.0 N-P-K 0-0-0
1979 80 81 82 83 84 85 86 87 88

Figure 17. Rice and wheat yield trends, long-
term trials, Faizabad, U.P., India, 1979-88.
Source: Woodhead (n.d.).


Xo X X2 Input
xo 1 level

Figure 18. Shift in fertilizer response function
resulting from resource degradation.
Source: Byerlee (1987).

Elsewhere in Asia, signs of long-term yield
declines in intensive cereal cropping systems
have emerged in work involving continuously
cultivated rice systems. Since this work
involves different cropping patterns and
different management practices than those
typically found in the rice-wheat belt, the
results cannot be considered directly
applicable to rice-wheat systems.23
Nonetheless, they highlight how intensive
cropping systems over time can begin to show
signs of degradation. Flinn and De Datta
(1984),. Cassman et al. (1995), and Cassman and
Pingali (1995) all report a downward shift in
the nitrogen response curve, which they
attribute to a decline in the ability of the soil to
release native nitrogen for crop growth.
Pagiola (1995) describes declines in boro rice
yields recorded by researchers from the
Bangladesh Rice Research Institute in long-
term, multiple-crop trials, even when
recommended levels of nutrients were applied.
Similar factors may be contributing to the
declines in input response observed in rice-
wheat systems.

Although these and other long-term trials
provide evidence of yield declines in
continuously cultivated rice-wheat systems, it
is not always easy to determine if declining
yields are due to soil fertility problems. The
law of diminishing returns suggests that the
rate of fertilizer application should be inversely
related to the productivity of fertilizer, but the
evidence on this issue is surprisingly
inconclusive. One easily calculated measure of
average fertilizer-use efficiency, the ratio of
grain weight to nitrogen weight, appears to
indicate that fertilizer productivity has indeed
declined. In India this ratio for rice fell from

about 60 in 1966 to less than 10 in 1992; during
the same period, the ratio for wheat fell from
about 15 to around 5 (Figure 19).

But the ratio of grain weight to nitrogen weight
is deceptive, because it assumes that yields
would be zero if no fertilizer were applied
(which obviously is not the case). More
sophisticated measures of average fertilizer-use
efficiency, such as the ratio of marginal grain
weight to fertilizer weight (marginal grain
weight is the grain weight above some base
yield achieved without fertilizer), suggest that
fertilizer productivity declines may not have
been as extensive as is widely assumed. In the
Indian state of Haryana, for rice the ratio of
marginal grain weight to fertilizer weight has
fallen since the mid-1970s, but for wheat the
ratio has increased slightly (Figure 20). Other
measures of marginal fertilizer-use efficiency
similarly fail to establish whether or not
fertilizer productivity has declined in rice-
wheat systems. Chaudhary and Harrington
(1993) divided changes in crop yields by
changes in fertilizer use rates to measure factor

, 60
.) \Rice
4 40
* 30
2 20

I 0
c 1966 68 70 72 74 76 78 80 82 84 86 88 90 92
Figure 19. Average response of rice and
wheat to fertilizer, all India, 1966-92.
Source: Fertiliser Association of India (various years).

23 In continuously cultivated rice cropping systems, soils are kept wet for most of the year; because soils are kept
permanently in a reduced condition, they tend to differ from rice-wheat soils, which are periodically dried and

productivity trends in Haryana during the
1970s and 1980s. They found that fertilizer
productivity declined for rice, but it increased
for wheat indicating that diminishing
returns associated with increased use of
fertilizer may have been offset (and in the case
of wheat overpowered) by farmers' adoption of
improved germplasm and crop management

Sustainability Issues in Rice-Wheat
Cropping Systems

If the rate of productivity declines in rice-wheat
cropping systems remains uncertain, the causes
of productivity declines are even less well
understood. Many "explanations" have been
put forward, most of them only weakly
supported by empirical evidence, and therefore
they must be regarded as unconfirmed
hypotheses. What has become clear is that the
causes of apparent declines in productivity are
often site specific, varying from one location
and one cropping system to another.

In, I

ca s

.a o
I 10

U I I I I 1 1
1969 71 73 75 77 79 81 83 85 87 89
Figure 20. Marginal response of rice and
wheat to fertilizer, Karnal, India, 1969-90.
Source: Chaudhary and Harrington (1994).

Soil problems

Soil problems (involving chemical, physical, or
biological factors) are obvious candidates to
blame for long-term productivity declines.
Because most rice-wheat cropping systems are
heavy extractors of nutrients, chemical
deficiencies are bound to develop after years of
continuous cropping unless steps are taken to
restore fertility levels. Although this conclusion
seems inescapable, it has been remarkably
difficult to establish clear links between soil
fertility levels and productivity declines in rice
and wheat. Deficiencies of nitrogen and
phosphorus are known to cause yield losses,
but most farmers now apply these two
elements at close to recommended levels.
Potassium, which is less commonly applied,
may be causing problems in some areas,
although often it has been ruled out as a cause
of yield declines. Micronutrients such as zinc,
boron, and manganese are often suggested as
yield-limiting factors, but trials that include
these elements rarely generate conclusive
evidence to corroborate this hypothesis.
Organic matter levels also are known to affect
soil chemistry, yet the role of organic matter in
rice-wheat cropping systems is imperfectly
understood. Field surveys make clear that
farmers are applying less manure to their rice
and wheat plots (primarily because of its
increasing value for use as cooking fuel), but
empirical evidence linking the decreased use of
manure to specific nutrient imbalances in rice-
wheat cropping systems is lacking.

Soil physical factors also can affect the
productivity of rice-wheat systems. The
puddling of rice soils leads to the breakdown of
soil aggregates, resulting in reduced pore sizes
and the formation in some soils of a plow pan.

24 Elsewhere in Asia, evidence is emerging that marginal returns to fertilizer use on rice also are decreasing rapidly in
intensively cultivated rice-rice systems (for example, see Cassman and Pingali, 1994).

Rice (three-year moving average)

Wheat (three-year moving average)

Puddling restricts water movement and affects
soil reactions in ways that are favorable for rice
but unfavorable for wheat. Puddling may be
one of the major causes for reduced wheat
yields when wheat follows rice, especially in
fine-textured soils. Linked to the problem of
puddling is that of poor rooting. Few studies
have been conducted on rooting processes in
rice-wheat systems, except in the coarse soils of
the Indian Punjab (Gajri, Arora, and Prihar,
1992). Poor rooting of wheat following rice
(because of poor soil structure, waterlogging,
and root-restricting soil layers) may be a major
cause of declining productivity in wheat and
should be included in the future research

Soil biological factors have received relatively
little attention from researchers but may also be
contributing to productivity declines in rice
and wheat. Solarization trials conducted at the
Bhairahawa station in Nepal show that the
incidence of root nematode in rice and root
necrosis in wheat are negatively correlated with
yield, indicating that underground pathogens
may be partly responsible for reducing yields
(Dubin and Bimb 1994). In the Philippines,
Cassman and Pingali (1995) hypothesize that
the main cause of declining rice yields has been
a reduction in the soil's natural ability to
provide nitrogen, which they attribute to
interactions between organic matter and soil
microbes. Although Cassman's and Pingali's
conclusions were based on experimental work
involving rice-rice systems, similar processes
could be at work in rice-wheat systems.

Water problems

Expansion in irrigated area and intensification
of cropping patterns have increased the
demand for irrigation water throughout South
Asia. In many areas, use of water to irrigate rice
and wheat has surpassed the natural ability of

the ecosystem to replenish itself. At the same
time, increased demand for irrigation water has
adversely affected the quality of water, with
additional negative effects on productivity.
Problems relating to the quantity and quality of
water have already impaired productivity in
many areas and threaten to become even more
of a limiting factor as more and more water is
diverted to non-agricultural uses.

Concern over water availability has mounted
especially fast in northwestern India (Maklin
and Rao 1991; Malik and Faeth 1993; Pandey,
Dwivedi, and Sharma 1992). During the past
decade, water tables have dropped at a rate of
0.5-0.8 m per year in the state of Haryana
(Harrington et al. 1993) and at a rate of 0.2-1.0 m
per year in the neighboring state of Punjab
(Gill 1992, 1994). Declining water tables not only
raise production costs (by forcing farmers to
pump water from greater depths), but such
rapid rates of decline raise serious questions
about the long-term sustainability of rice-wheat
systems. In some areas, farmers are already
being forced to cut back on the number of
irrigations they can apply to their rice and
wheat crops, and where water shortages are
particularly acute, they are sometimes forced to
limit plantings.

Problems relating to the quality of irrigation
water have also multiplied. Many rice-wheat
tracts in northeastern Pakistan and
northwestern India are being affected by water-
borne compounds. Tubewell water in particular
can contain salts and minerals that over time
reduce the productivity of soils. Data generated
at the Soil Fertility'Institute in Lahore, Pakistan, .
show that wheat yields are often lower in plots
irrigated by tubewell than in plots irrigated
using canal water (Figures 21a, 21b). Although
researchers still lack conclusive proof, they
suspect that tubewell water contains salt
concentrations which over the long term affect

soil productivity. This hypothesis is supported
by data from other studies showing that
tubewell water in the Pakistani Punjab contains
high levels of salts (Kijne and Vander Velde
1990, Byerlee and Siddiq 1994). Salinity and
sodicity problems are sometimes exacerbated
by poor water management practices in the
lower reaches of canal irrigation systems,
where sufficient water may not be available to
leach out salts (Kijne and Vander Velde 1990).

Weed, insect, and disease problems

Agricultural intensification in general and
continuous cropping of cereals in particular
have increased the incidence of weed, insect,

21a. Rice yields

4.3 N-P-K 150-150-150


>" 3.9


E Tubewell irrigation

' Canal irrigation

21b. Wheat yields

1975 1985
S Tubewell irrigation = Canal irrigation

Figure 21. Rice and wheat yields by irrigation
source, Lahore, Pakistan, 1975 and 1985.
Source: Kijne and Vander Velde (1990).

and disease problems in some rice-wheat
zones. Among the many weeds that negatively
affect productivity, the most damaging is
Phalaris minor, a grassy weed which poses an
especially serious problem for wheat, especially
in cooler areas (rice is not affected by Phalaris).
Long present in South Asia as a minor weed,
Phalaris has become a major problem in recent
years, particularly in continuous rice-wheat
cropping systems. If left unchecked, Phalaris
can cause crop losses in wheat of nearly 100%
(Aslam et al. 1989, Harrington et al. 1993, and
Hobbs et al. 1991, 1992). Chemical control must
be used, since weeding by hand is ineffective.
Although chemical control has proved effective
thus far, evidence is now emerging that Phalaris
is developing resistance to isoproturon, the
most extensively used herbicide in India (Malik
and Singh 1993).

Continuous rice-wheat cropping has been
accompanied in some areas by an increase in
insect pest problems. Growth in the area
planted to rice and wheat has expanded the
host environment for many insects, while
intensification of the cropping pattern has
extended the period during which the host
environment is present. The result has been a
noticeable buildup in insect populations and
the appearance of overlapping generations of
insects within the same cropping season. With
the exception of localized outbreaks, insect
problems remain relatively minor (although
damage caused by the rice stem borer can be
extensive). However, insect problems are likely
to proliferate in the future as the result of
additional expansion in rice-wheat area, further
intensification of rice-wheat systems, and the
increasing domination of a relatively small
number of popular MVs.

Continuous rice-wheat cropping also improves
conditions for the buildup of many plant
diseases, especially leaf diseases in rice and

Karnal bunt and foliar blights in wheat.
Contrary to the conventional wisdom, most
MVs carry high levels of resistance to major
diseases higher levels, in many cases, than
were present in the old TVs they replaced. Yet
despite the higher levels of resistance, diseases
remain a threat if the enhanced resistance is
offset by a lack of varietal diversification in
farmers' fields. Farmers within localized zones
sometimes all decide to grow the same MV -
usually the MV which yields the highest within
that zone. Under these circumstances, if the
resistance of a particularly popular MV breaks
down, a large area may be affected.

Crop management problems

Productivity levels in rice-wheat cropping
systems have been negatively affected by the
intensification process itself, in the sense that
management decisions taken to increase the
productivity of one crop often have negative
consequences for the productivity of other
crops in the rotation. For example, farmers who
decide to increase rice production by switching
to a longer-duration rice variety frequently
must delay planting their wheat crop following
the rice harvest (Byerlee and Husain 1992).
Delayed planting of wheat significantly
reduces potential yields and decreases the
efficiency of fertilizer uptake (Hobbs 1985,
Saunders 1990). Adoption of reduced tillage
practices can help to alleviate this problem, but
introduction of reduced tillage technology in
turn is likely to create new problems (e.g., stem
borer carryover in unplowed rice stubble)
(Aslam et al. 1989, Majid et al. 1988). Another
example of the crop management problems
that can arise in rice-wheat rotations involves
the different land preparation requirements of
the two crops, with puddling being beneficial
for rice but harmful for wheat.

Challenges for Research and
Research Organization

Technical change:
Past, present, and future

Byerlee (1992) has described a sequential
process of technical change which is helpful in
thinking about the agricultural intensification
currently taking place in South Asia's rice-
wheat cropping systems. According to Byerlee,
technical change in Asian agriculture proceeds
through three stages, distinguished by the
development and diffusion of technologies to
substitute for emerging factor scarcities:

* a Green Revolution Phase, during which MVs
first become available and, when grown with
modest levels of purchased inputs, lead to a
surge in productivity;

* a First Post-Green Revolution Phase (Input
Intensification Phase), when farmers improve
allocative efficiency by increasing the level of
use of purchased inputs, resulting in
movement toward economically optimal use
of these inputs; and

* a Second Post-Green Revolution Phase (Input
Efficiency Phase), during which farmers move
toward increased technical efficiency by
using available purchased inputs more
efficiently while adopting practices that
contribute to the sustainability of the
resource base.

These stylized stages of technical change can be
depicted diagramatically (Figure 22). During
the Green Revolution Phase, the introduction of
MVs shifts the production function upwards
(TV to MV1), increasing crop response to
complementary inputs such as fertilizer and
water and leading to a one-off surge in

productivity (A to B). Adoption of modest
levels of these complementary inputs
accompanies adoption of MVs, although
farmers for some time fail to exploit the full
benefits of the new technology and continue to
operate well below the technological frontier.
During the First Post-Green Revolution Phase,
farmers become familiar with the technology
and move along the suboptimall) production
function (B to C), using higher levels of
complementary inputs to improve the
allocative efficiency of production. Finally,
during.the Second Post-Green Revolution
Phase, farmers approach the new production
frontier (MV2) by further increasing the
efficiency with which they use inputs.
Depending on the strategy followed by
farmers, use of complementary inputs may
increase (D) or decrease (E) during this phase.

The three phases described by Byerlee are
useful in thinking about the intensification of
South Asia's rice-wheat cropping systems
because they provide important insights into
how the products and information needed
from the agricultural research system have
changed over the years and how they will
continue to change.

Figure 22. Phases of technical chan
(following Byerlee).
Source: Byerlee (1992).

During the Green Revolution Phase, production
technologies are largely input-based, and the
role of the research system need not extend
much beyond developing innovations that can
be extended to farmers in blanket fashion. This
was indeed the experience throughout many of
the areas in South Asia where the Green
Revolution technologies had their initial
dramatic impact. Modem varieties, fertilizer,
and irrigation technologies were transferred to
farmers via standardized technical packages,
which came with generic management
recommendations developed largely on
research stations.

During the First Post-Green Revolution Phase,
seeds, fertilizer, and water (irrigation) still have
the capacity to deliver productivity gains, but
the management of these key inputs must be
fine-tuned. This second developmental stage
described by Byerlee also was borne out in the
evolution of South Asia's rice-wheat cropping
systems. As MVs spread out of the relatively
homogeneous irrigated tracts of northwestern
India and northeastern Pakistan into less
favorable production zones, local crop
management research was needed to adapt the
original blanket recommendations to specialized
conditions. In addition, after the first generation
of MVs had spread throughout the region's
irrigated zones, local crop breeding efforts were
needed to adapt imported varieties to distinct
niches determined by agroclimatic conditions
and cropping patterns. Local breeding was also
instrumental for carrying out maintenance
breeding to preserve host-plant resistance to
mutating insect pests and diseases.

put efficiency) During the Second Post-Green Revolution
Input level Phase, new approaches to research are required
ge to develop the sophisticated, site-specific
management information needed to improve
input-use efficiency in the context of
increasingly complex cropping systems where

AB Phase 1 (technical breakth
BC Phase 2 (input intensificatii
CD, CE Phase 3 (increased in

avenues for successful change are limited. This
third and final stage described by Byerlee is
already evident in the more intensively
cultivated areas of the central rice-wheat belt,
where multidisciplinary, systems-oriented
teams of researchers are working with farmers
to explore innovative crop management
technologies designed to boost productivity by
simultaneously attacking problems from
several different angles. Because this research
requires input from many different disciplines,
as well as constant interaction from farmers
through diagnostic surveys and on-farm
testing, it tends to be costly to implement and
difficult to manage.

Organization of agricultural research in
South Asia

The fact that different technologies and
different types of information are needed at
different stages in the intensification process
has important implications for the organization
and management of agricultural research. It is
useful to review the recent history of research
to see whether technology development efforts
are evolving to keep pace with the changing
needs of farmers.

The initial success of Green Revolution
technologies had a profound influence on the
organization of agricultural research in South
Asia. Impressed by the rapid spread of MVs
nd by their catalytic effect on use of fertilizer
and irrigation, research managers throughout
the region moved quickly to adopt the
commodity-based strategy followed by the
international agricultural research centers
(IARCs).25 During the late 1960s and early
1970s, rice and wheat research programs were
formed in virtually every country of South

Asia. As in IARCs, highest priority was
assigned to germplasm improvement; the
primary goal was to develop new and better
MVs. In addition to plant breeders, the
institutes typically included plant pathologists,
physiologists, entomologists, and agronomists,
all of whom directly or indirectly supported
breeding activities. Even crop management
research was intended to support germplasm
improvement; agronomists were assigned
responsibility for determining the planting
dates, seeding rates, fertilizer application rates,
irrigation practices, and weed control practices
that would enable MVs to perform up to their

Many of the specialized rice and wheat
research programs were successful. Although
frequently hampered by inadequate funding
and shortages of trained staff, they managed to
develop a considerable amount of improved
germplasm. Following an initial start-up
period, locally developed modern rice and
wheat varieties began to appear in the mid-
1970s. In subsequent years, the rate of releases
increased as national breeding programs grew
in strength, and by the mid-1980s the regions'
stronger programs (e.g., India, Pakistan) were
releasing rice and wheat varieties at an
impressive rate (Dalrymple 1986a, 1986b;
Byerlee and Moya 1993). This work by national
research programs was actively supported by
the IARCs, primarily through provision of
improved germplasm and through training of
scientific personnel.

The commodity-based strategy was
particularly effective during the Green
Revolution period, when MVs of rice and
wheat were still spreading into new areas.
Although the MVs that had originally

25 In India, commodity-oriented research programs predated the formation of the IARCs, so the success of the Green
Revolution merely served to accelerate an ongoing trend.

spearheaded the Green Revolution proved to
be well adapted to irrigated zones, they were
not always suitable for marginal production
environments. Additional breeding research
was therefore needed to adapt improved
materials to local requirements. National rice
and wheat research institutes organized around
a strong core of plant breeding activities
proved effective at accomplishing this task.

However, circumstances have changed.
Productivity growth is slowing, suggesting that
the easy gains from the original Green
Revolution technologies have for the most part
been realized. Throughout large parts of South
Asia's rice-wheat belt particularly the
irrigated parts adoption of MVs is now
virtually complete, fertilizer application rates
are approaching optimal levels, and the
potential for affordable irrigation is largely
exhausted. By implication, if growth in food
production is to keep pace with projected
increases in demand, new sources of
productivity growth will have to be tapped.

Where will future productivity gains come
from? Traditional sources of productivity
growth will continue to be important.
Germplasm improvement (involving
conventional plant breeding methods and/or
emerging biotechnology techniques) no doubt
will be instrumental in further increasing the
yield potential of rice and wheat. Hence the
current strong interest in developing
commercially viable hybridization
technologies, in altering the architecture of rice
and wheat plants to increase their ability to
bear grain, and in developing transgenic
species incorporating genes conferring
resistance to important diseases and insect
pests.26 But germplasm improvement work

alone probably will not be enough to carry the
day. Although future gains from plant breeding
are likely to be significant, it probably will not
be possible to repeat the progress achieved
during the past 30 years. For this reason, it
seems likely that future growth in productivity
in South Asia's rice-wheat cropping systems
will come increasingly from adoption of
improved management practices designed to
increase the efficiency of input use and/or to
arrest (or reverse) the degradation of soil and
water resources (Byerlee and Pingali 1995). In
other words, information and knowledge, rather
than simply inputs, will become increasingly

Weaknesses of current approaches to

The shift in emphasis from input-based
technologies to knowledge-based technologies
suggests that changes will be needed in the
organization of research. Most research
organizations, both national and international,
are still organized along lines that reflect the
needs of an earlier period. Scientists continue to
place inordinate emphasis on controlled
experimentation designed to develop packages
of recommendations for distribution to farmers.
Meanwhile, insufficient attention is being paid
to understanding the increasingly complex
problems that threaten to undermine the
sustainability of rice-wheat cropping systems
and to designing flexible solutions that farmers
can adjust to their own particular circumstances.

This idea is hardly original. A few far-sighted
research administrators long ago realized that
research based on experiment stations and
organized along strict commodity lines cannot
effectively address the complex problems which

26 The productivity gains from germplasm improvement work have by no means been exhausted. For example, the so-
called "super rices" currently being developed at IRRI are expected to deliver yield gains of 12-15% compared to the
best commercial varieties.

are bound to arise in the post-Green Revolution
period. For years, these visionaries have argued
that more effective interdisciplinary
collaboration between researchers and closer
links with farmers will be needed if increasingly
complex problems are to be diagnosed properly
and solved. In several South Asian countries,
advocates of these views were sufficiently
influential to bring about the establishment of
systems-oriented research programs, usually in
the form of cropping systems/farming systems
units lodged within the commodity-based
institutes. However, these units generally failed
to make a noticeable impact (Tripp 1991). Their
lack of success can be attributed to several

* Cropping systems/farming systems divisions
often were formed as independent units
staffed by scientists seconded from
established disciplinary departments.
Threatened by a loss of personnel and
operating funds, managers of the established
departments sometimes attempted to impede
the transfer of resources. Cropping systems/
farming systems divisions thus found
themselves isolated from the rest of the
research system and denied access to much-
needed technical and financial support.

Incentives to link commodity-oriented
research institutes through collaborative
projects were usually lacking. Every institute
maintained a separate agenda, and many
chose to establish their own farming systems
units. Expertise needed to unravel problems
related to the productivity of the overall
farming system may have existed in other
institutes, but for all intents and purposes it
was unavailable.

Individual scientists tended to avoid
involvement in systems-oriented research
because such work was rarely given adequate

recognition. Professional advancement was
generally tied to achievements in more
strategic research areas (that is, in theoretical
or methodological research). Applied
problem-solving research such as that
envisioned for farming systems teams was
considered less prestigious.

* Interdisciplinary research units rarely had
strong links to the extension service, making
propagation of research results difficult. It
did not help that the extension services often
faced full agendas of their own and rarely
had incentives to collaborate with

* Systems-oriented research was supposed to
involve farmers, but in many instances
farmers were not actively consulted during
the planning, testing, and evaluation of new

Elements needed for future research
to be effective

South Asia's agricultural research programs
today stand at a crossroads. Established during
an earlier era when funding was plentiful, most
of the region's research institutes have been
slow to adapt to the fiscal austerity of the 1990s.
Even as the research agenda has grown and
broadened, raising new and increasingly
complex technical challenges, research
administrators have failed to take the necessary
steps to improve priority-setting, streamline
operations, increase accountability, and
r,';ivtie staff. At the same time, they have
failed to develop the base of political support
needed to ensure continuing funding for
research (Byerlee and Pingali 1995).

If reform of the organization and management
of research is overdue, what would a re-
engineered research system look like? Without

trying precisely to describe an "ideal" system
(which would in any case be impossible, since
institutional requirements differ between
countries), it should be possible to describe
some of its salient characteristics.

1) Expanded focus Research programs
should continue conducting problem-
focused "downstream" research designed to
increase the efficiency of farmers' resource
use in the short run, but this kind of
traditional, applied research will have to be
complemented by more strategic "upstream"
research designed to shed light on factors
affecting the long-run productivity of major
cropping systems (e.g., processes
contributing to resource degradation).
Added emphasis must be given in particular
to diagnostic research aimed at identifying
problems that constrain growth in
productivity; describing the pace, incidence,
and severity of these problems; unraveling
their causes; and predicting their
consequences. If this is to be accomplished,
more effort will have to be invested in
monitoring what goes on in farmers' fields,
both in terms of cropping practices and in
terms of changes in the natural resource base
upon which agriculture depends.

2) Farmer participation Far too many
scientists in South Asia spend most of their
time on the research station, rarely venturing
forth to observe what is happening in
farmers' fields or to solicit direct feedback
from farmers. Researchers need increased
input from farmers for a more accurate
diagnosis of problems and more effective
design of potential solutions. Once problems
have been correctly identified and alternative
prototype solutions proposed, farmer
participatory adaptive research can help to
tailor the solutions to specific farming

systems and farmer circumstances.
Strengthening links to farmers thus will
speed the adaptation of new technologies to
local conditions, leading to more rapid and
possibly more complete adoption.

3) Multidisciplinary perspective -
Collaboration among disciplines must begin
at the diagnosis stage and extend to research
planning, implementation, and evaluation.
Individual scientists can assume
responsibility for separate components of a
research project (which is often desirable to
take advantage of disciplinary expertise), but
results must be discussed and evaluated
with colleagues from other disciplines.
Achieving effective multidisciplinary
collaboration will require significant changes
in the structure of the professional incentive
system, which at present disproportionally
rewards disciplinary work of a more
theoretical nature, as opposed to
multidisciplinary work which is more
problem-solving in nature.

4) Intercommodity alliances As cropping
systems continue to intensify,
intercommodity alliances will become
increasingly important in the research
system. We now know that farmers in the
rice-wheat belt take decisions based on a
sophisticated understanding of the
productivity of their entire household
economy (including both agricultural and
non-agricultural activities). Gone are the
days when researchers could investigate
production technologies affecting one crop
without regard to the implications of these
technologies for other crops in the system,
for livestock, and even for non-agricultural

5) Improved links between researchers and
extension specialists As crop
management practices become ever more
complex, additional effort will be needed to
ensure that effective mechanisms exist to
speed the flow of technology from
researchers to extension specialists to
farmers, as well as the opposite flow of
information about farmers' technology needs
and about the performance of new
technologies in farmers' fields. At present,
researchers and extension specialists often
work in virtual isolation from one another.
Researchers must come to understand that
their responsibility extends beyond the
development of new technologies and that
they have a duty to participate actively in the
technology transfer process, a reality that
must be reflected in the structure of
professional incentives and rewards. At the
same time, extension agents must be
encouraged to serve as a two-way conduit
capable of effectively conveying information
in both directions between researchers and

6) Regional and international collaboration -
Regional and international collaboration is
needed to facilitate the exchange of ideas,
eliminate redundant research, and increase
efficiency. Recent work has shown that
technological spillovers between countries
and/or between regions within countries are
often extensive, particularly in the case of
plant breeding research (Byerlee 1995,
Maredia and Byerlee 1996). Given that
research activities currently tend to be
distributed among a large number of state or.
district research institutes, this suggests that
considerable cost savings could be achieved
by concentrating research efforts within a
relatively small number of adequately
staffed and well-equipped institutes.

An example of the gains that can be achieved
through regional and international collaboration
is provided by the Rice-Wheat Consortium
(RWC), a partnership linking four public
national agricultural research systems (NARSs)
and five IARCs which share the goal of
increasing productivity in South Asia's rice-
wheat cropping systems. Collectively managed
by means of a joint Steering Committee, the
RWC is working to improve the efficiency of rice
and wheat research by coordinating the activities
of consortium members. Representatives from
RWC member organizations convene regularly
to identify and prioritize important research
problems and to devise collective strategies for
their solution. Responsibility for specific topics is
assigned to the institution or institutions best
equipped to address them, the idea being to
exploit each institution's comparative advantage
while avoiding wasteful duplication of effort. In
addition to coordinating the design of research,
the RWC serves as an effective mechanism for
speeding the dissemination of relevant research
results via conferences, publications, and

Within the RWC, teams of researchers are
concentrating on several themes which are
illustrative of the "third-phase" problems
described by Byerlee. For example, one team is
working to develop innovative methods of land
preparation-cum-stand establishment designed
simultaneously to reduce tillage costs, improve
soil structure and fertility, facilitate stand
establishment, and speed turnaround between
crops (e.g., minimum or reduced tillage, surface
seeding of wheat, and direct sowing of rice).
Another team of researchers is exploring
integrated pest management (IPM) strategies
designed to reduce pest-induced crop losses
through use of resistant varieties, judicious
management of rotations, elimination of pest
habitat, and limited application of chemical
pesticides. A third team is working to develop

water management practices designed to
increase water-use efficiency, to moderate or
reverse chemical imbalances caused by water
use (e.g., salinity, sodicity), and to lower
irrigation costs. Other research themes
addressed by the RWC include soil fertility
management, disease management, and weed

One important lesson emerging from the
RWC's activities is that research on complex
"third-phase" problems requires new and
innovative approaches to research planning. As
more and more demands are placed on the
pool of funds available for research, national
funding agencies and international donors are
becoming increasingly reluctant to provide
"open checkbook" support for diversified,
multicomponent, long-term research programs.
Instead, they tend to favor clearly focused,
limited-duration projects that can be easily
monitored arnd whose outputs are readily
defined. Given this change in the funding
environment, research planning is necessarily
becoming more costly and more time-
consuming. With research issues growing
increasingly complex, considerable effort is
now required to ensure that individual research
projects designed to focus on a specific problem
area (e.g., "direct seeding of wheat using
mechanical seed drills") remain appropriately
articulated within a larger research theme (e.g.,
"improved crop establishment technologies for
wheat following rice"). Effective management
of research teams is becoming more important,
given the need to ensure that all relevant
disciplines are appropriately represented and
that individual responsibilities of participating
researchers are well defined and clearly
understood. Parenthetically, it should be noted
that the management function, which requires
time and effort, must be recognized and
appropriately rewarded in the allocation of
research funds.


The rice-wheat systems of South Asia enjoy a
well-deserved reputation of being among the
most productive cropping systems in the
world. Thanks to the wealth of their natural
resources, the industriousness of their
inhabitants, and the effectiveness of modern
crop production technologies, they have played
a vital role in helping to avert the massive food
shortages that many observers were predicting
as recently as 20 years ago. But while the
remarkable success achieved in feeding an
exponentially increasing population provides
grounds for optimism about the future of
agriculture in South Asia, the achievements of
the past should not generate a false sense of
security. As this paper has tried to show,
troubling signs are beginning to emerge that
the successes of the past will be difficult to
duplicate and that it may be difficult even to
sustain current levels of productivity unless
steps are taken soon to safeguard tomorrow's
food supplies.

Farmers in South Asia have asserted for some
time that the productivity of rice-wheat
cropping systems is declining, but researchers
and policy makers have been slow to pay them
heed. To this day, many people (including
many agricultural specialists) deny that a
problem even exists. Their complacency is
perhaps understandable, because the warning
signs are not always obvious. Many farmers
have been applying increasing amounts of
inputs to maintain yield levels, so it is not
always easy to tell whether or not productivity
has declined. And with a few significant
exceptions, researchers have not paid sufficient
attention to monitoring long-term productivity
trends, not even in experimental plots.

Evidence has started to accumulate, however,
suggesting that farmers may be right and that

there is indeed cause for concern. Throughout
South Asia, growth in rice and wheat yields has
slowed in recent years, particularly in
intensively cropped zones where MVs,
fertilizer, and irrigation have been used long
and extensively. In areas where rice and wheat
yields continue to rise, often the higher yields
are achieved by increasing the application of
costly inputs, which masks declines in the rate
of growth of total factor productivity. Declines
in the rate of total factor productivity growth
appear to have been brought about by
degradation of the natural resource base and
the attendant negative ecological effects
associated with intensive production systems
- declining soil fertility, declining quantity
and quality of groundwater, buildups of soil
toxicities, and increasing populations of pests.
These troubling developments are also being
observed in experimental plots, where signs of
long-term yield stagnation and natural
resource degradation are becoming

The faltering productivity of many of South
Asia's rice-wheat cropping systems points to an
alarming conclusion: agricultural researchers
are losing ground in the struggle to devise
lasting solutions to the increasingly complex
problems confronting farmers. Many policy
makers do not appear to be particularly
alarmed by this development, expressing
confidence that researchers will always be able

to devise technological solutions every time a
new problem appears. Others are less
confident, arguing that some of the resource
degradation that is now being observed may by
irreversible, so that by the time the extent of the
problem is fully appreciated, the damage mav
be irremediable (and even if the damage can be
reversed, doing so may turn out to be
prohibitively expensive).

Agricultural research will have to become more
effective if the countries of South Asia are to
achieve sustainable economic growth while
ensuring food security for all. What is needed,
however, is not simply more of the same
research as has been carried out in the past -
not even the same research done better. The
commodity-based approach, which was so
successful in developing the input-based
technologies that spearheaded the Green
Revolution, is proving inadequate for
developing the knowledge-based technologies
needed during the post-Green Revolution era.
Given the extremely heavy demands being
placed on an already overburdened resource
base, a "business as usual" approach is unlikely
to generate the technological innovations
needed to elicit additional production from
South Asia's vital rice-wheat cropping systems.
In fact, without dramatic changes in the way
research is organized and carried out, it may
not even be possible to sustain current high
levels of productivity.


Ali, M., and L.E. Velasco., 1994. Intensification of
induced resource degradation: The crop
production sector in Pakistan. Los Bafos,
Philippines: International Rice Research
Institute (IRRI). Mimeo.

Antle, J., and S. Capalbo (eds.). 1988. Agricultural
Productivity: Measurement and Explanation.
Washington, D.C.: Resources for the Future.

Aslam, M., A. Majid, P.R. Hobbs, N.I. Hashmi, and
D. Byerlee. 1989. Wheat in the Rice-wheat
Cropping System of the Punjab: A Synthesis of
On-farm Research Results, 1984-1988. PARC/
CIMMYT Paper No. 89-3. Islamabad, Pakistan:
Pakistan Agricultural Research Council

Ball, V.E. 1985. Output, input, and productivity
measurement in U.S. agriculture, 1948-79.
American Journal of Agricultural Economics 67:

Bangladesh Bureau of Statistics. Various years.
Yearbook of Agricultural Statistics of Bangladesh.
Dhaka, Bangladesh: Bangladesh Bureau of

Barnett, V., A.E. Johnston, S. Landau, R.W. Payne,
S.J. Welham, and A.I. Raynor. 1994.
Sustainability: The Rothamsted Experience.
Chapter 9 in V. Barnet, R.W. Payne, and R.
Steiner (eds.), Agricultural Sustainability in
Economic, Environmental, and Statistical Terms.
London, U.K.: John Wiley.

Brown, L.R. 1994. Facing food insecurity. Chapter
10 in State of the World 1994. New York, New
York: W.W. Norton.

Brown, L.R., and H. Kane. 1994. Full House:
Reassessing the Earth's Population Carrying
Capacity. The Worldwatch Environmental
Alert Series. New York, New York: W.W.

Buringh, P., and R. Dudal. 1987. Agricultural land
use in space and time. In M.G. Wolman and
F.G.A. Fournier (eds.), Land Transformation in
Agriculture. New York, New York: John Wiley.

Byerlee, D. 1992. Technical change, productivity, and
sustainability in irrigated cropping systems of
South Asia: Emerging issues in the Post-Green
Revolution Era. Journal of International
Development 4: 477-496.

Byerlee, D. 1994. Modern Varieties, Productivity, and
Sustainability: Recent Experience and Emerging
Challenges. Mexico, D.F.: CIMMYT.

Byerlee, D. 1995. On the comparative advantage of
international agricultural research: Exploiting
economies of size to generate global spillovers.
Paper prepared for the workshop, Integration of
Research Efforts of ICRISAT with NARSs and
Other International Agricultural Research
Institutions, 14-16 December, 1994, International
Crops Research Institute for the Semi-Arid
Tropics (ICRISAT), Hyderabad, India.

Byerlee, D., and T. Husain (eds). 1992. Farming
Systems of Pakistan: Diagnosing Priorities for
Agricultural Research. Lahore, Pakistan:

Byerlee, D., and P. Moya. 1993. Impacts of International
Wheat Breeding Research in the Developing World,
1965-90. Mexico, D.F.: CIMMYT.

Byerlee, D., and P. Pingali. 1995. Agricultural
research in Asia: Fulfillments and frustrations.
In D. Headley and G. Peters (eds), Agricultural
Competitiveness, Market Forces, and Policy Choices.
Proceedings of the XXII Conference of the
International Association of Agricultural
Economists, 22-29 August 1994, Harare,
Zimbabwe. Aldershot, U.K.: Gower.

Byerlee, D., and A. Siddiq. 1994. Has the Green
Revolution been sustained? The quantitative
impact of the seed-fertilizer revolution in
Pakistan revisited. World Development

Cassman, K.G., S.K. De Datta, D.G. Olk, J.M.
Alcantara, M.I. Samson, J.P. Descalsota, and
M.A. Dizon. 1995. Yield decline and nitrogen
economy of long-term experiments on
continuous, irrigated rice systems in the
tropics. In R. Lal and B.A. Stewart (eds.),
Sustainable Management of Soils. Michigan:
Lewiston Publishers, CRC Press Inc.

Cassman, K.G., and P.L. Pingali. 1995.
Extrapolating trends from long-term
experiments to farmers' fields: The case of
irrigated rice systems in Asia. In V. Barnett.,
R. Payne, and R. Steiner (eds.), Agricultural
Sustainability: Economic, Environmental, and
Statistical Considerations. London, U.K.: John
Wiley. Pp. 63-84.

Chaudhary, M.K., and L.W. Harrington. 1993. The
Rice-Wheat System in Haryana: Input-Output
Trends and Sources of Future Productivity
Growth. Mexico, D.F.: C.C.S. Haryana
Agricultural University Regional Research
Station (Karnal) and CIMMYT.

Christiansen, L.R. 1975. Concepts and
measurement of agricultural productivity.
American Journal of Agricultural Economics 57:

Christiansen, L.R., and D. Jorgensen. 1970. U.S.
real product and factor input, 1929-67.
Review of Income and Wealth 16:19-50.

Dalrymple, D. 1986a. Development and Spread of
High Yielding Wheat Varieties in Developing
Countries. Washington, D.C.: U.S. Agency for
International Development.

Dalrymple, D. 1986b. Development and Spread of
High Yielding Rice Varieties in Developing
Countries. Washington, D.C.: U.S. Agency for
International Development.

Dubin, H.J., and H.P. Bimb. 1994. Soilborne
diseases of wheat in warmer areas of South
Asia: An update. In D.A. Saunders .a: i G.P.
Hettel (eds.), Wheat in Heat-stressed
Environments: Irrigated, Dry Areas, and Rice-
Wheat Farming Systems. Mexico, D.F: UNDP/
ARC/BARI/CIMMYT. Pp. 353-360.

Ehui, S.K., and D.S.C. Spencer. 1990. Indices for
Measuring the Sustainability and Economic
Viability of Farming Systems. Resource and
Crop Management Program Research
Monograph No. 3. Ibadan, Nigeria: Resource
and Crop Management Program,
International Institute of Tropical
Agriculture (IITA).

Evenson, R.E., and E.A. Bloom. 1993. Research
and productivity in Pakistan agriculture. In
A.S. Haider, Z. Hussain, R. McConnen, and
S.J. Malik (eds.). Agricultural Strategies in the
1990s: Issues and Policies. Islamabad,
Pakistan: Pakistan Association of
Agricultural Social Scientists. Pp. 152-182.

Evenson, R.E., and J.W. McKinsey. 1991.
Research, extension, infrastructure, and
productivity change in Indian agriculture. In
R.E. Evenson and C.E. Pray (eds.), Research
and Productivity in Asian Agriculture. Ithaca,
New York: Cornell University.

FAO (Food and Agriculture Organization of the
United Nations). Various years. AGROSTAT
PC. Rome, Italy: FAO.

Fertiliser Association of India. Various years.
Fertiliser Statistics. New Delhi, India:
Fertiliser Association of India.

Flinn, J.C., and S.K. De Datta. 1984. Trends in
irrigated-rice yields under intensive
cropping at Philippine research stations.
Field Crops Research 9: 1-15.

Gajri, P.R., V.K. Arora, and S.S. Prihar. 1992.
Tillage management for efficient water and
nitrogen use in wheat following rice. Soil and
Tillage Research: 1-16.

Gill, K.S. 1992. Research on Rice-Wheat Cropping
Systems in the Punjab. Punjab Agricultural
University Monograph. Ludhiana, Punjab,
India: PAU.

Gill, K.S. 1994. Sustainability issues related to
rice-wheat production system in Asia. In
R.S. Paroda, T. Woodhead, and R.B. Singh
(eds.), Sustainability of Rice-Wheat Production
Systems in Asia. Bangkok, Thailand: Food
and Agriculture Organization of the United
Nations. Pp. 36-61.

Giri, G.S., G.P. Acharya, and A.P. Regmi. 1994.
Results of a long-term rice-wheat soil
fertility experiment in the Terai of Nepal. In
D.A. Saunders and G.P. Hettel (eds.), Wheat
in Heat-stressed Environments: Irrigated, Dry
Areas and Rice-Wheat Farming Systems.
Mexico, D.F.: UNDP/ARC/BARI/
CIMMYT. Pp. 309-316.

Government of Pakistan. Various years.
Agricultural Statistics of Pakistan. Islamabad,
Pakistan: Government of Pakistan.

Harrington, L.W., P. Jones, and M. Winograd.
1994. Operationalizing sustainability: A total
factor productivity approach. Paper
presented at the Land Quality Indicators
Conference, June 1994, Centro Internacional
de Agriculture Tropical (CIAT), Cali,

Harrington, L.W., S. Fujisaka, M.L. Morris, P.R.
Hobbs, H.C. Sharma, R.P. Singh, M.K.
Chaudhary, and S.D. Dhiman. 1993. Wheat
and Rice in Karnal and Kurukshetra Districts,
Haryana, India: Farmers' Practices, Problems
and an Agenda for Action. Mexico, D.F.: HAU,
ICAR, CIMMYT, and IRRI. Pp. 1-54.

Hobbs, P.R. 1985. Agronomic practices and
problems for wheat following cotton and
rice. In R.L. Villareal and A.R. Klatt (eds.),
Wheats for More Tropical Environments: A
Proceedings of the International Symposium.
Mexico, D.F.: CIMMYT. Pp. 273-277.

Hobbs, P.R., G.P. Hettel, R.P. Singh, Y. Singh,
L.W. Harrington, and S. Fujisaka. 1991. Rice-
Wheat Cropping Systems in the Tarai Areas of
Nainital, Rampur, and Pilibhit Districts in
Uttar Pradesh, India: Diagnostic Surveys of
Farmers' Practices and Problems, and Needs for
Further Research. Mexico, D.F.: ICAR/

Hobbs, P.R., G.P. Hettel, R.K. Singh, R.P. Singh,
L.W. Harrington, V.P. Singh, and K.G. Pillai.
1992. Rice-Wheat Cropping Systems in Faizabad
District of Uttar Pradesh, India: Exploratory
Surveys of Farmers' Practices and Problems, and
Needs for Further Research. Mexico, D.F.:

Huke, E., R. Huke, and T. Woodhead. 1994a. Rice-
Wheat Atlas of Bangladesh. Los Bafios,
Philippines: IRRI/CIMMYT/BARC. Pp. 1-

Huke, R., E. Huke, and T. Woodhead. 1994b. Rice-
Wheat Atlas of Nepal. Los Baiios, Philippines:

IRRI (International Rice Research Institute).
Various years. World Rice Statistics. Los
Bafios, Philippines: IRRI.

Islam, N., M.W. Rosegrant, S.K. Jayasuriya, and
M. Agcaolili. 1992. Food situation and
outlook for Asia: Issues, prospects, and
policy implications. Washington, D.C.:
International Food Policy Research Institute
(IFPRI). Mimeo.

Kendall, H.W., and D. Pimentel. 1993. Constraints
on the expansion of the global food supply.
Paper prepared for the Rene Dubos Forum
on Technological Prospects for
Sustainability, November 1993, Bellagio,

Kijne, J.W., and E.J. Vander Velde. 1990. Salinity
in Punjab watercourse commands and
irrigation system operations: The imperative
case for improving irrigation management
in Pakistan. Lahore, Pakistan: International
Irrigation Management Institute (IIMI).

Kumar, P., and M.W. Rosegrant. 1994.
Productivity and sources of growth for rice
in India. Economic and Political Weekly 29(53):

Lynam, J., and R. Herdt. 1989. Sense and
sustainability: Sustainability as an objective
in international agricultural research.
Agricultural Economics 3(4): 381-398.

Majid, A., M. Aslam, N.I. Hashmi, and P.R.
Hobbs. 1988. Potential use of minimum
tillage in wheat after rice. In A.R. Klatt (ed.),
Wheat Production Constraints in Tropical
Environments. Mexico, D.F.: CIMMYT. Pp.

Maklin, M.C., and M.V. Rao. 1991. Rice and
wheat production and sustainability of the
irrigated rice-wheat cropping system in
India. World Bank Internal Report. New
Delhi, India: The World Bank.

Malik, R.K., and S. Singh. 1993. Phalaris minor
resistance against Isoproturon. Hissar, India:
Department of Agronomy, Haryana
Agricultural University. Mimeo.

Malik, R.P.S., and P. Faeth. 1993. Rice-wheat
production in Northwest India. In P. Faeth
(ed.), Agricultural Policy and Sustainability:
Case Studies from India, Chile, the Philippines
and the United States. Washington, D.C.:
World Resources Institute. Pp. 1-35.

Maredia, M.K., and D. Byerlee. 1996. Efficiency of
research investments in the presence of
international spillins: Wheat research in
developing countries. American Journal of
Agricultural Economics (forthcoming).

McCalla, A.F. 1994. Agriculture and food needs
to 2025: Why we should be concerned. Sir
John Crawford Memorial Lecture, presented
27 October 1994, at International Centers'
Week, Consultative Group on International
Agricultural Research, Washington, D.C.

Ministry of Agriculture (Nepal). Various years.
Agricultural Statistics of Nepal. Kathmandu,
Nepal: HMG Ministry of Agriculture,
Department of Food and Marketing
Services, Agricultural Statistics Division.

Ministry of Agriculture (Pakistan). Various years.
Agricultural Statistics of Pakistan. Islamabad,
Pakistan: Ministry of Agriculture.

Mitchell, D.O., and M.D. Ingco. 1993. The world
food outlook. Paper prepared for the Rene
Dubos Forum on Technological Prospects
for Sustainability, November 1993, Bellagio,

Mohan Dey, M., and R.E. Evenson. 1991. The
Economic Impact of Rice Research in
Bangladesh. Gazipur, Bangladesh:
Bangladesh Rice Research Institute.

Morris, M.L., H.J. Dubin, and T. Pokhrel. 1994.
Returns to wheat breeding research in
Nepal. Agricultural Economics 10: 269-282.

Nambiar, K.K.M. 1994. Yield sustainability of
rice-rice and rice-wheat systems under long-
term fertilizer use. In Resource Management
for Sustained Crop Productivity. New Delhi,
India: Indian Society of Agronomy.

Pagiola, S. 1995. Environmental and Natural
Resource Degradation in Intensive Agriculture
in Bangladesh. Environment Department
Papers, Environmental Economics Series,
Paper No. 17. Washington, D.C.: The World

Pandey, R.K., B.S. Dwivedi, and A.K. Sharma
(eds). 1992. Rice-Wheat Cropping System:
Proceedings of the Rice-Wheat Workshop Held
on 15-16 October, 1990, at Modipuram, Meerut.
Modipuram, Meerut, India: Project
Directorate for Cropping Systems Research.

PARC (Pakistan Agricultural Research Council).
n.d. Statistical Bulletin on Wheat in Pakistan.
Islamabad, Pakistan: Social Science Division,

Plucknett, D.L. 1993. Science and Agricultural
Transformation. IFPRI Lecture Series No. 1.
Washington, D.C.: International Food Policy
Research Institute.

Pray, C.E. 1991. Determinants of output and
productivity growth in eastern India:
Implications for future governments. New
Brunswick, New Jersey: Department of
Agricultural Economics, Cook College,
Rutgers University. Mimeo.

Rosegrant, M.W., and M. Agcaoili. 1994. Global
and regional food demand, supply, and
trade prospects to 2010. Paper presented at
the roundtable meeting, Population and
Food in the Early Twenty-First Century:
Meeting Future Food Needs of an Increasing
World Population, 14-16 February,
International Food Policy Research Institute,
Washington, D.C.

Rosegrant, M., M. Agcaoili-Sombilla, and N.D.
Perez. 1995. Global Food Projections to 2020:
Implications for Investment. Food, Agriculture,
and the Environment Discussion Paper 5.
Washington, D.C.: International Food Policy
Research Institute.

Rosegrant, M.W., and R.E. Evenson. 1992.
Agricultural productivity and sources of
growth in South Asia. American Journal of
Agricultural Economics 74: 757-61.

Rosegrant, M.W., and P. Pingali. 1994. Sustaining
rice productivity growth in Asia A policy
perspective. Journal of International
Development 6(6): 665-688.

Rosegrant, M.W., and M. Svendsen. 1993. Asian
food production in the 1990s: Irrigation
investment and management policy. Food
Policy 18(1): 13-32.

Saunders, D.A. 1990. Crop Management Research
Summary of Results. Wheat Research Centre
Monograph No. 5. Nashipur, Bangladesh:

Seckler, D. 1994. Trends in World Food Needs:
Zero Growth in the 21st Century. Center for
Economic Policy Studies Discussion Paper
No. 18. Winrock International Institute for
Agricultural Development. Pp. 1-17.

Sidhu, D.S., and D. Byerlee. 1991. Technical
change and wheat productivity in post-
Green Revolution Punjab. Economic and
Political Weekly 26(52): A159-A166.

Singh, R.B., and S. Paroda. 1994. Sustainability
and productivity of rice-wheat systems in
the Asia-Pacific region: Research and
technology development needs. In R.S.
Paroda, T. Woodhead, and R.B. Singh (eds.),
Sustainability of Rice-Wheat Production
Systems in Asia. Bangkok, Thailand: Food
and Agriculture Organization of the United
Nations. Pp.1-36.

Steiner, R., and R. Herdt. 1995. Measuring
sustainability at the farm level: A total social
factor productivity approach. In V. Barnett,
R. Payne, and R. Steiner (eds.), Agricultural
Sustainability in Economic, Environmental, and
Statistical Terms. London, U.K.: John Wiley.

Thapa, G.B., and M.W. Rosegrant. 1995.
Projections and Policy Implications of Food
Supply and Demand in Nepal to the Year 2020.
Policy Analysis in Agriculture and Related
Resource Management, Research Report
Series No. 30. Kathmandu, Nepal: Winrock

Tripp, R. 1991. The limitations of on-farm
research. In Tripp R. (ed.), Planned Change in
Farming Systems: Progress in On-Farm
Research. New York, New York: John Wiley.
Pp. 247-257.

Wehrfritz, G. 1995. "The Bad Earth: China's
Looming Grain Crisis." Newsweek Magazine,
May 15, 1995. Pp. 8-14.

Whitaker, M., and S. Lalitha. 1993. Quantifying the
Relative Productivity and Sustainability of
Alternative Cropping Systems. Resource
Management Program Economics Group
Progress Report 115. Hyderabad, India:
International Crops Research Institute for
the Semi-Arid Tropics (ICRISAT).

Woodhead, T., R. Huke, and E. Huke. 1994. Rice-
Wheat Atlas of Pakistan. Los Bafos,
Philippines: IRRI/CIMMYT/PARC.

Woodhead, T., R. Huke, E. Huke, and L.
Balababa. 1994. Rice-Wheat Atlas of India. Los
Baios, Philippines: IRRI/CIMMYT/ICAR.

Woodhead, T. (no date.) Rice-wheat literature
review. Los Bafos, Philippines: International
Rice Research Institute (IRRI). Mimeo.

World Bank. 1992. World Development Report 1992.
Oxford, England: Oxford University Press.

World Bank. 1994. World Development Report 1994.
Oxford, England: Oxford University Press.

New Papers from the Natural Resources Group

95-01 Land and Livelihoods: Patterns of Rural Development in Atlantic Honduras
D. Buckles and G. Sain

96-01 Meeting South Asia's Future Food Requirements from Rice-Wheat Cropping Systems: Priority
Issues Facing Researchers in the Post-Green Revolution Era
P. Hobbs and M. Morris

96-02 Soil Fertility Management Research for the Maize Cropping Systems of Smallholders in Southern
Africa: A Review J.D.T. Kumwenda, S.R. Waddington, S.S. Snapp, R.B. Jones, and M.J. Blackie

ISSN: 1405-2830

International Maize and Wheat Improvement Center
Centro Internacional de Mejoramiento de Maiz y Trigo
Lisboa 27, Apartado Postal 6-641, 06600 M6xico, D.F., M6xico

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs