Lessons from Three Decades of Green Revolution.pdf

7/29/2019 Lessons from Three Decades of Green Revolution.pdf

http://slidepdf.com/reader/full/lessons-from-three-decades-of-green-revolutionpdf 1/52

Lessons from Three Decades of Green Revolution

in the Philippines

Jonna P. Estudillo and Keijiro Otsuka

(November 6, 2002)

Jonna P. EstudilloSchool of Economics

University of the PhilippinesDiliman, Quezon City 1101PhilippinesTel. 632-927-9686 to 91; 632-920-5461FAX 632-920-5462; 632-921-3359

(Address beginning November 15, 2002)

K iji Ot k (corresponding a thor)



Keijiro Otsuka (corresponding author)

Abstract

This paper aims to assess the changing contributions of successive generations of

modern varieties (MVs) of rice to yield increase and stability and changes in total factor

productivity (TFP) in different ecosystems in the Philippines. We found that the yield

increase in irrigated ecosystem has been by far the highest due to the diffusion of pest-

and disease-resistant MVs, which also contribute significantly to yield stability. MV

contribution to yield increase in rainfed ecosystem has been significant but much less

while upland environment has experienced an upward trend in yield albeit slowly due to

the diffusion of improved traditional varieties and MVs suitable to adverse production

environments. The contribution of MVs cum irrigation to TFP growth is about 50

percent in in Central Luzon.



1. Introduction

More than three decades have passed since the International Rice Research

Institute (IRRI) released IR8 in 1966, the first modern variety (MV) of rice, which

spearheaded the Green Revolution in the Philippines. Since IR8 a large number of rice

varieties have been developed by rice scientists and accepted by Filipino farmers. Indeed

the Green Revolution is not a one-shot phenomenon but an evolutionary process

involving the replacements of earlier MVs by newer MVs with better characteristics

(Hayami and Otsuka, 1994).

MVs can be classified into three distinct groups based on the dates of release and

their distinct characteristics. The first-generation MVs (MV1) consisting of IR series

f IR5 t IR34 d l d b IRRI d C4 i d l d b th U i it f th



programs is to develop rice varieties that are suitable to adverse production environments,

e.g. those areas subject to drought, flooding, and poor soil quality (IRRI, 1997).2

It is important to mention that there are other changes in rice farming technologies

that have occurred over the study period, which we did not rigorously analyze as these

changes are mainly related to the development in MVs. These are irrigation expansion,

fertilizer application, herbicide use, integrated pest management, mechanization, and

direct-seeding.

Some observers of the Green Revolution casts doubt over the sustainability of

Green Revolution technologies particularly those that are developed for irrigated rice

cropping systems (FAO, 1997). Flinn and DeDatta (1984) and Cassman and Pingali

(1995) found an evidence of a downward trend in paddy yields in some long-term

experiments using constant level of inputs. Indeed, it is possible that given the declining



Needles to say, yield is an imperfect measure of production efficiency because it

is affected by the use of non-land inputs, particularly fertilizer. Yet since the productivity

of fertilizer increases with the introduction of MVs, it can be a reasonable proxy for the

sum of the direct yield effect and indirect effect of MV adoption through its effects on the

application of other yield-increasing inputs. While TFP index is a more desirable

indicator of technical efficiency, we cannot estimate it from the official statistics because

of the lack of detailed data on the amount and prices of inputs. We estimate TFP index

using the Central Luzon Loop survey data collected recurrently by IRRI from 1966 to

1999. It may be worth mentioning that TFP index has limitations: it cannot capture the

improvement in allocative efficiency and the socially desirable effects of increase in

cropping intensity owing to the introduction of MVs characterized by shorter-growth

duration and non-photo period sensitivity.



government. This data set contains information on rice production, yield, and area

planted in irrigated, rainfed, and upland ecosystems for the whole Philippines and its 15

regions from 1970 to 1997.3

MVs, however, are not grouped into early and improved

MVs in this data set.

2.1 An Overview of the Rice Sector in the Philippines

Rice has been traditionally grown in six major regions in the Philippines (see

Figure 1). Ilocos, Cagayan Valley, Central Luzon and Southern Tagalog have well-

developed irrigation infrastructure systems, while Bicol and Western Visayas are

endowed with favorable rainfall pattern. These six regions are the frontrunners of the

Green Revolution in the country.

Central Luzon is by far the largest producer of rice contributing, on average, to 17

percent of the overall rice production in the country from 1970 to 1997 (Figure 2).



Rice is planted in irrigated, rainfed, and upland ecosystems. The share in rice

production of irrigated ecosystem increased from 56 percent in 1970 to 75 percent in

1997 due to increased adoption of high-yielding and shorter-maturing MVs that are well

suited to irrigated environment (Table 1). On the contrary, the share in rice production of

rainfed ecosystem declined from 37 percent in 1970 to 23 percent in 1997 and the share

of upland ecosystem declined from 7 percent in 1970 to 2 percent in 1997.

2.2 MV Adoption

MVs were designed to have a significant yield advantage over TVs because of

their capacity to respond favorably to high fertilizer application and to utilize solar

energy effectively (Chandler, 1982).

The availability of irrigation is by far the most important physical factor affecting

the adoption of MVs (David and Otsuka, 1994). The proportion of the country’s rice area



90 percent in irrigated ecosystem by 1979 (Table 2). Similarly, MV adoption rate in

rainfed ecosystem has increased consistently over time indicating that MVs perform

fairly well in rainfed areas with sufficient rainfall. On the contrary, MV adoption rate in

the uplands has remained marginal even up to the mid-1980s. MVs planted in the

uplands are similar to MVs for lowland areas but are more suited to drought-prone upland

environments.5

Upland rice breeding program started early with the release in of C22 in 1972,

developed by the University of the Philippines, although up to now the number of MVs

developed for the uplands are much less compared to MVs for the irrigated ecosystem.

The other upland MVs are UPL Ri5, UPL Ri7, PSBRc1, PSBRc3, PSBRc5, and PSBRc7.

These upland MVs have an average yield ranging from 2.2 tons to 3.0 tons per ha based

on repeated trials on farmers’ fields.6

These trends clearly show that Green Revolution



It is observed that unfavorable regions lagged behind the favorable regions in the

adoption of MVs during the early period of Green Revolution (Ruttan, 1977; David and

Otsuka, 1994). Regional data from the Philippine Census of Agriculture in 1971 reveal

that there was a faster rate of MV adoption in the six major rice-growing regions, where

irrigation infrastructures were well developed. MV adoption rates between irrigated and

rainfed ecosystems in the early 1970s were almost identical in Ilocos, Central Luzon, and

Western Visayas reflecting a favorable rainfall pattern in these regions.

2.3 Yield Trends and Yield Variations

Yield increase has been the major source of rice production growth in the country

since the mid-1960s when there was no further opening of new rice lands for cultivation

(Hayami and Ruttan, 1985). The adoption of MVs is a major factor responsible for yield

increase as MVs have yield potential of three to four times the yield capacity of TVs



Yield increase can be partly attributed to increased fertilizer use as yield of MVs

respond favorably to higher fertilizer application. Fertilizer application in terms of

elemental micronutrients (NPK kg per ha) has increased over time and is generally higher

in irrigated than in rainfed ecosystem, indicating that higher fertilizer application is more

profitable in irrigated ecosystem (Table 3). Fertilizer use is higher in Ilocos, Central

Luzon and Western Visayas than in other regions. One reason for the increased fertilizer

application is the decline in urea price relative to paddy price (Estudillo and Otsuka,

2001).

Fertilizer productivity in terms of yield per unit of NPK applied has remained

fairly constant from 1988 to 1997 in both irrigated and rainfed ecosystem, although the

productivity of fertilizer (or yield-fertilizer use ratio) is higher in irrigated ecosystem.

This finding is similar to the findings of Tiongco and Dawe (2002) using Loop Survey



In general, the yield of TVs is much less compared to the yield of MVs in

irrigated and rainfed environment. However, the yield of MVs and TVs in upland

ecosystem is fairly similar which may indicate that the new varieties do not have a

significant advantage in terms of increasing yield in the uplands. It is also interesting to

note that gaps in yield between ecosystems are larger for MVs than for TVs indicating

that the availability of good water supply is more important in the production of MVs

than in the production of TVs.

We found no significant differences in MV yields between the wet and dry

seasons in all ecosystems, even though MVs are not photoperiod sensitive making it

suitable for dry season cropping. Otsuka, Gascon and Asano (1994), however, found that

MV yields are significantly higher and have greater stability in the dry season on irrigated

farms in Central Luzon owing to a good supply of water and absence of typhoons.



recently released MVs have achieved higher yields in less favorable areas perhaps due to

investments in irrigation in these areas and the spread of MVs suitable to less favorable

environments. In contrast, yield instability in upland ecosystem has remained high with

no clear downward trend, reflecting the diversity of adverse production environments.

In summary, we have found an upward trend in paddy yield partly to due to a

wider diffusion of high-yielding MVs and increased fertilizer application. Paddy yields

are higher in irrigated ecosystem, which indicates that the generic yield potential of MVs

are better exploited in irrigated ecosystem. Regional yield differential in irrigated and

rainfed ecosystems has declined since the 1980s due to the spread of MVs that

incorporated resistant against pests and diseases. Paddy yields have increased in the

uplands albeit slowly and yet there has been no clear evidence of declining yield

instability.



ecosystem (IRG) and MV adoption (IRG*MV); between IRG and TV adoption

(IRG*TV); between rainfed ecosystem (RNF) and MV (RNF*MV); between the upland

ecosystem (UPL) and MV (UPL*MV); and between UPL and TV (UPL*TV); (2) season

dummy; (3) regional dummies; and (4) year dummies. We take logarithm for dependent

variables, so that the estimated coefficients show roughly percentage change in yield.8

Data on the fertilizer application are available only for irrigated and rainfed ecosystems

for 1988-97 with no distinction between the use for MVs and TVs.

As is shown in Table 4, the yield of MVs in irrigated ecosystem has been

significantly higher and has increased over time compared to the yield of TVs in rainfed

ecosystem (the control). The coefficient of the variable IRG*MV is positive and

significant and its value has increased from 0.61 in the 1970s to 0.67 in the 1980s, which

means that the yield advantage of MVs in irrigated ecosystem over TVs in rainfed



rainfed ecosystem has increased significantly from the 1970s to the 1980s. Yield of MVs

compared to yield of TVs in rainfed ecosystem is higher by 21% in 1970-79 and by about

28% in 1990-97. On the contrary, the yield of both MVs and TVs planted in the upland

ecosystem is significantly lower than the yield of TVs planted in rainfed ecosystem. It is,

however, interesting to note of the increase in the coefficients of the variables UPL*MV

and UPL*TV, which indicates that the yields of MVs and TVs in the uplands have come

closer to the yields of TVs in the rainfed environment. This might be due to diffusion of

improved TVs and MVs that are suitable for varied upland environments.

Yields in all regions are significantly lower than those in Central Luzon partly due

to lower levels of fertilizer application and perhaps due to a less extensive government

farm extension services. However, there are some indications that the regional yield gap

has declined or at least has remained the same over time as shown in the changes in the



terms of kg yield per kg of NPK applied). The level of fertilizer application in the wet

season is not significantly different from that of the dry season (the control) and fertilizer

application is higher in Central Luzon (the control) than in any other regions in the

country. Despite a significant increase in fertilizer use over time due importantly to

decreased fertilizer prices, year dummies in the fertilizer productivity function are all

insignificant, suggesting that the rice farming system is sustainable.

Overall, our regression results show that MVs have contributed to a significant

yield increase in irrigated ecosystem and to a limited extent in rainfed but not in the

uplands. There are, however, indications that rice yields have increased in the upland

ecosystem as well perhaps due to increased adoption of the improved TVs suitable to

marginal areas. Upland rice, however, accounts for a decreasing and small share of rice

production in the Philippines. Thus, it seems rational to think that the best strategy to



irrigated and favorably rainfed rice-growing areas, so that the Loop data is particularly

useful for the analysis of the evolution and impacts of the Green Revolution, which took

place mainly in favorable rice-growing ecosystems.11

The Loop survey is conducted regularly almost every four years beginning in the

wet season of 1966. There were surveys in 1966/67, 1970/71, 1974, 1979/80, 1982,

1986/87, 1990/91, 1994/95, and 1998/99. The survey years 1970 and 1974 are

considered bad years due to heavy typhoon damage. The wet season survey was

followed by the dry season survey except in 1974 and 1982. The original sample size of

the Loop survey in the wet season of 1966 was 92 respondents. This number, however,

declined due to attrition over time caused by retirement, refusal of interview, and absence

during the survey visit. The sample was expanded in 1979 to include new samples thus

increasing the total sample size to 149 farmer respondents. In the latest survey in 1999,



slightly to 1.5 in 1986-87 to 1998-99. Availability of irrigation water and the adoption

of shorter-duration MVs explain the increase in cropping intensity. Shorter maturity

period allows some farmers in rainfed areas to catch the late monsoon rain in order to

grow a second crop of rice in the months of October and November. It is important to

mention that we did not compute cropping intensity index in the regional data because

rice area is expanding even between two adjacent seasons in one crop year. Cropping

intensity index measures the number of rice crops that can be grown in one year so that

ideally this index should be computed when rice area is constant in one cropping

season.The entire rice area was under traditional varieties in 1966 when the first survey

was conducted (Table 5). The more popular traditional varieties were Inano, Binato,

Tjeremas, Raminad and Intan.12

Farmers were quick to adopt MVs such that by 1970, 66

per cent of rice area of sample farmers was planted with MV1. The speed of MV



Sample farmers quickly adopted MV3 and 61 percent MV3 adoption rate was

reached by 1986 just a year after the release of IR64. By that time, MV1 was almost

completely extinct. IR64 has been by far the most popular rice variety since 1986

because it commands a higher price in the market due to its superior grain quality.14

Adoption rates of MVs in Central Luzon far exceeded the irrigation ratio, suggesting that

MVs can be grown profitably in rainfed environment as well. A close examination of the

adoption of specific rice cultivars grown by sample farmers shows that no single

dominant variety occupies most of the rice fields in specific years and season. This was

the case in 1966 when TVs were grown, as well as in recent years when farmers were

growing only MVs.

Overall, our data clearly show that the diffusion of MVs has been remarkably fast

and there has been no reversal back to TVs and older MVs. Varietal improvements and



the percentage of rice area planted to each of the successive generations of MVs. Since

the dependent variable ranges from zero to unity, we use the two-limit tobit estimation

procedure.

We include four sets of independent variables as follows: (1) irrigation ratio

defined as the percentage of rice area with irrigation; (2) socio-economic variables such

as age and schooling of household head, farm size, and ratio of area under share-tenancy

and leasehold-tenancy; (3) provincial dummies; and (4) year dummies. The control

variable is the ratio of area under owner-cultivation for the tenancy categories;

Pangasinan in the provincial dummies; 1970 in the MV1 adoption function; 1978 in the

MV2 adoption function; and 1986 for the MV3 adoption function. There has been a

major replacement of TVs by MV1 from 1970 to 1974; MV1 by MV2 from 1979 to

1982; and MV2 by MV3 from 1986 to 1999.



We used the age of household head as a proxy variable for experience in rice

farming. The coefficient of this variable is not significant in the equations for MV1 and

MV2 but highly significant in the equation for MV3. Presumably, MV3 requires more

timely application of inputs, which the older and more experienced farmers can better

handle than the younger farmers. The coefficient of schooling of head is not significant

in any of the MV adoption functions indicating that years of formal schooling is not an

important determinant of the adoption of MVs.

In the early period of Green Revolution, there were beliefs that the new rice

technology was biased in favor of large farmers who have better access to formal credit

markets (Lipton and Longhurst, 1989). To fully exploit its yield potential, MVs need

application of fertilizer and chemical inputs in higher amounts, which require large cash

outlays that small subsistence farmers may not be able to secure. The regression results



MV3 than owner farmers, judging from negative and significant coefficient of share-

tenancy dummy in the 1986-99 regression.

The coefficients of year dummies (which were included in the regression runs but

not shown in Table 6) in the MV1 adoption function are all positive but not significant

indicating that the spread of MV1 was so fast and widespread even within a short period

of time. Similarly, there appears to be no significant diffusion lag in the adoption of

MV2 across years, which indicates easy availability of the more superior MV2. Year

dummies for 1987, 1990, 1991, 1994, and 1995 are all positive and highly significant in

the MV3 adoption function, which means that over time there was a significant increase

in the ratio of area planted to MV3. It appears, however, that the spread of MV3 was

slower than MV2 partly because IR42, a popular MV2 in the market due to its improved

grain quality, remained highly adopted even when a number of the more improved MV3



typhoon damage in 1970 and 1974 on the yield of MV1. A major yield boost was gained

with the diffusion of MV2 with stronger pest resistance, which was achieved between

1979 and 1982.15

The yield of MV2 reached, on the average, about 4.0 tons per ha in

1979 to 1982. Otsuka, Gascon, and Asano (1994) confirmed that the yield gain

associated with MV2 adoption is statistically significant.

It appears, however, that a plateau in rice yield had been reached in the mid-1980s

when MV3 became popular. According to Hayami and Otsuka (1994), “the stagnation in

yield indicates that yield potential based on conventional breeding methods opened up in

the mid-1960s has largely been exhausted by the mid-1980s”. The genetic improvement

of MV3 was partly on account of improved grain quality that enabled farmers to earn

higher revenues from similar yields obtained from MV2. However, a detailed

examination of yield data by production environment reveals potential contribution of



It is interesting to find that paddy yields in Central Luzon were significantly lower

during the dry season in 1967 than in the wet season in 1966. This is essentially due to

photoperiod sensitivity of TVs. Paddy yields tend to become higher during the dry than

wet season thereafter because MVs are photoperiod insensitive. In addition, the amount

of elemental NPK applied in the dry season is about 30% higher compared to the wet

season. There is also a rising trend in yield during the dry season from 1980 to 1998,

which indicates an unexplored potential to further increase yields through good water

control and higher application of inputs.

We estimate rice yield function in three periods of time in 1970-74 when MV1

replaced TVs; in 1979-82 when MV2 largely replaced MV1; and 1986-99 when MV3

replaced MV2 (Table 8). We use a multi-variate regression model in which the adoption

of MVs were included as explanatory variables along with the four sets of independent



yield of MV2 in 1979-82 is 36 percent higher compared to MV1 on rainfed farms. The

coefficient of the variable MV2*IRG is not significant indicating that the additional

effect on yield of MV2 planted in irrigated farms appears close to zero. Note, however,

that MV2 was so widely grown in irrigated areas from 1979 to 1982 (Table 5) that

coefficient of irrigation is likely to reflect the effect of MV2 as well. Overall, it is likely

that the adoption of MV2 effectively doubled yield.

It appears that yield did not increase further with the adoption of MV3, judging

from the insignificant coefficient of MV3. Nor there has been evidence of a significant

increase in value of output per ha in 1986-99 associated with the adoption of MV3, which

is supposed to be characterized by better grain quality. It is, however, important to point

out that irrigation dummy is insignificant, either, which indicates that MV3 is equally

high-yielding between irrigated and rainfed conditions. This is likely due to the



Year dummies (which were included in the regression runs but are not shown in

Table 8) have positive and significant coefficients in the yield function in 1979-82 and

1986-99, which indicates that there has been a significant increase in yield over time.

Yield increase is particularly notable in 1979-82 with the release of insect and disease-

resistant MV2. NPK function (which was estimated but are not shown in the tables)

shows that sample farmers have used increased amounts of NPK per ha over time.

Fertilizer use is also significantly higher on irrigated farms compared to rainfed farms

and significantly more fertilizer is applied on smaller farms. Similarly, the more

educated farmers used significantly more fertilizer than the less educated ones and older

farmers used significantly more fertilizer than the younger ones. Fertilizer use was also

significantly lower on the sharecropped and leasehold farms than on owner-operated

farms. Farmers in Nueva Ecija have higher rates of fertilizer application while those in



3.2 Changes in Total Factor Productivity

We compute the total factor productivity (TFP) index in order to measure the

contribution of successive generations of MVs to technical efficiency of production (See

the Appendix for a detailed description of the methodology). TFP is an effective measure

of long-run sustainability of a technology (Lynam and Herdt, 1989). As shown in the

lower panel of Table 6, the TFP index is lower than 100 from 1967 to 1974 which means

that the growth of output is lower than the aggregate growth of input. It is therefore clear

that MV1 did not contribute much to enhancing technical efficiency in rice production.

On the contrary, TFP index since 1979 exceeds 100 suggesting that MV2 has had a

significant efficiency increasing effect. The value of TFP index from 1986, however, has

remained fairly similar to those in 1980 and 1982 indicating that the productivity impact

of MV3 is fairly equal to that of MV2. Also, the values of TFP index in the dry seasons



but dry season dummy has positive and significant coefficient. The latter results can be

taken to imply that MV3 contributed to improvement of TFP under the rainfed condition

and during the dry season. It is, however, difficult to verify such effects directly because

of the quick and complete dominance of MV3, unless finer classification of MV3 into

subgroups with different traits is made. Socio-economic variables such as age and

schooling of head and tenancy do not appear to have a significant impact on technical

efficiency while provincial dummies indicate that farmers in Nueva Ecija are more

technically efficient due probably to their greater access to government extension

services.

In brief, our findings show that there has been an increase in TFP during the

diffusion of MV2 followed by stagnation in productivity in irrigated rice farming system

beginning in the early 1980s. The contribution of improved MVs and irrigation to long



First, the contribution of MV to yield increase in irrigated ecosystem has been by

far the highest, which demonstrates that the potential of Green Revolution technology can

be exploited to the fullest extent in irrigated ecosystem. MV contribution to yield

increase in rainfed ecosystem is significant but much less so. Second, yield increased

dramatically only with the development and rapid diffusion of pest- and disease-resistant

MV2. Indeed, while the yield advantage of MV1 over TVs was negligibly small on the

farmers’ fields in Central Luzon, the advantage of MV2 over MV1 was decisively large.

The contribution of MV2 and irrigation combined to technical efficiency in irrigated rice

farming system is about 50 percent. Third, marginal impacts of newer MVs have

diminished and fallen to insignificant levels in recent years, according to the analysis of

yield data in the Philippines as a whole, whereas there is an indication that MV

contributed to improvements of yield and productivity under the rainfed conditions and



conditions has been largely exhausted. On the other hand, unfavorable areas, particularly

upland areas, are heterogeneous, so that a variety of site-specific technologies suitable for

particularly environments are required for the enhancement of the productivity.

Furthermore, it is scientifically more difficult to develop new varieties for upland areas.

Thus, it is doubtful whether it pays to conduct serious research for upland areas at the

international research centers, whose mandate is to create “global” or “regional” public

goods. Whether site-specific research for upland areas conducted by PHILRICE has

achieved sufficiently high payoffs needs to be examined. Finally, it may be worth

emphasizing that in order to achieve the further improvement of productivity in rice

farming in the Philippines, a new breakthrough is needed which goes far beyond the

improved resistance to pests and diseases.



APPENDIX: Estimation of Total Factor Productivity

In this appendix, we would like to explain how to estimate total factor productivity

(TFP) index for each observation in each cropping. First of all, we use the following

Tornqvist formula to estimate the total factor productivity index (TFP) cross-sectionally:

ln (TFPit/TFP0t) = [ln (Qit) - ln (Q0t)] - (1/2) Σ j (sijt + s0jt) [ln (Xijt) - ln (X0jt)] ,

where subscript “i”, “t”, “0”, and “j” denote i-th farm, t-th period, “average” farm, and j-

th input, and respectively; Q refers to quantity of output; s stands for factor share (Σ j sij =

1); and X corresponds to factor inputs. Factor share of land is estimated from residual,

i.e., gross value of production minus the sum of explicit and imputed costs of capital,

labor, and current inputs. It is well known that the above formula can be derived if the

underlying production function is constant-returns-to-scale translog form and competitive

equilibrium holds (Diewert 1976).



where ln (Q0), s0j, and ln (X0j) are averages over m seasons, i.e.,

ln (Q0) = (1/m) Σt ln (Q0t),

s0j = (1/m)Σt s0jt ,

ln (X0j) = (1/m)Σt ln (X0jt) .

Thus, (TFP0t/TFP0) measures productivity of t-th period in comparison with the average

productivity over all cropping seasons.

Finally, we estimated TFP of i-th farm in t-th period ( TFPit) using the estimated

cross-sectional TFP index and time-series TFP index in the following way:

TFPit = (TFPit/TFP0t) x (TFP0t/TFP0).



REFERENCES

Barker, Randolf and Robert Herdt (1985), The Rice Economy of Asia. Washington, DC:

Resources for the Future.

Cassman, K.G. and Prabhu Pingali (1995), ‘Intensification of Irrigated Rice Systems:

Learning from the Past to Meet Future Challenges’, GeoJournal 35, 299-305.

Caves, D. W., L. R. Christensen, and W. E. Diewert (1982), ‘Multilateral Comparison of

Output, Input, and Productivity Using Superlative Index Numbers’, Economic

Journal 92, 73-86.

Chandler, Robert Jr. (1982), An Adventure in Applied Science: A History of the

International Rice Research Institute. IRRI, Los Banos, Laguna, Philippines.

David, Cristina C., Otsuka, Keijiro (1994), Modern Rice Technology and Income

Distribution in Asia. Lynn Rienner, Boulder, CO.



Estudillo, Jonna P., Fujimura, Manabu and Hossain, Mahabub (1999), ‘New Rice

Technology and Comparative Advantage in Rice Production in the Philippines,

1966-94’, Journal of Development Studies, 35 (5), 162-85.

FAO Regional Office for Asia and the Pacific (1997), ‘Trends of Yield and Productivity

of Modern Rice in Irrigated Rice Systems in Asia’, International Rice Commission

Newsletter, 46, 19-25.

Flinn, John 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.

Hayami, Yujiro and Masao Kikuchi (1978), ‘Investment Inducement to Public

Infrastructure: Irrigation in the Philippines’, Review of Economics and Statistics

60 (1), 70-77.

Hayami, Yujiro and Keijiro Otsuka (1992), The Economics of Contract Choice: An



of High-yielding Rice Varieties in South Asia. Johns Hopkins University Press,

Baltimore.

Herdt, Robert (1987), ‘A Retrospective View of Technological and Other Changes in

Rice Farming’, Economic Development and Cultural Change 35, 329-349.

Herdt, Robert and Celiz Capule (1984) Adoption Spread and Production Impact of

Modern Rice in Asia. Agricultural Policy Research Committee, Tokyo.

Hossain, Mahabub (1988) The Nature and Impact of the Green Revolution in

Bangladesh. International Food Policy Research Institute, Washington, D.C.

International Rice Research Institute (1997), Sustaining Food Security Beyond the Year

2000. IRRI, Los Banos, Laguna, Philippines.

Kikuchi, Masao, Atsushi Maruyama and Yujiro Hayami (2002), ‘Investment

Inducements to Public Infrastructure: Irrigation in the Philippines and Sri Lanka



Otsuka, Keijiro, Fe Gascon, and Seki Asano (1994), ‘Second-Generation MVs and the

Evolution of the Green Revolution: The Case of Central Luzon, 1966-90’,

Agricultural Economics 10 (3), 283-295.

Pingali, Prabhu., Hossain, Mahabub, Gerpacio, Roberta (1997), Asian Rice Bowls: The

Returning Crisis? CAB International, Wallingford, UK, in association with the

International Rice Research Institute.

Philippine Census of Agriculture, 1971. Department of Agriculture, Manila, Philippines.

Philippine Census of Agriculture, 1991. Department of Agriculture, Manila, Philippines.

Highland Rice Production in the Philippine Cordillera (2000), Philippine Rice Research

Institute in collaboration with the International Institute of Rural Reconstruction,

Munoz, Nueva Ecija, Philippines.

Rice Statistics Handbook, 1970-97 (2000), Philippine Rice Research Institute, Munoz,



Table 1

Shares of Rice Production by Ecosystem, Philippines, 1970-97

Ecosystem 1970 1980 1990 1997

Irrigated 56 60 71 75Rainfed 37 37 28 23

Upland 7 3 1 2Total 100 100 100 100



Table 2

Adoption of Modern Rice Varieties by Ecosystem in the Philippines, 1970-97

Year Irrigated Rainfed Upland2

Percentage of Area Planted with MVs

1

1970 66 43 01971 70 54 01972 73 55 01973 77 61 01974 80 64 01975 80 62 0

1976 84 65 101977 86 69 161978 89 74 10

1979 90 78 81980 88 78 131981 89 79 131982 93 82 181983 92 83 19



Table 3

Yields and Fertilizer Use by Ecosystem in the Philippines, 1970-97

Year Irrigated Rainfed Upland Irrigated Rainfed Irrigated Rainfed

Yield

(tons/ha/season)

NPK use1

(kg/ha/season)

Fertilizer productivity

(yield/NPK use)

1970 1.97 1.40 0.88

1971 1.88 1.43 0.83

1972 1.88 1.27 0.83

1973 1.82 1.21 0.71

1974 1.89 1.31 0.72

1975 2.11 1.35 0.73

1976 2.03 1.36 0.941977 2.17 1.46 0.90

1978 2.27 1.47 0.97

1979 2.50 1.45 0.92

1980 2.54 1.70 1.06

1981 2.52 1.70 1.16

1982 2.63 1.76 1.07

1983 2.57 1.58 1.01



Table 4

Rice Yield and Fertilizer Use Function in the Philippines, 1970-97

Variables ln (yield/ ha) ln(NPK/ha)

ln (yield/ NPK)

9

1970-79 1970-76 1980-89 1990-97

IRG1

0.29** 0.13**(12.61) (5.07)

IRG*MV2

0.61** 0.56** 0.67** 0.68**(25.41) (21.47) (34.59) (30.70)

IRG*TV3 0.39** 0.37** 0.40** 0.37**(16.41) (14.30) (20.90) (16.78)

RNF*MV4

0.19** 0.16** 0.25** 0.24**

(7.97) (6.31) (13.01) (11.15)UPL*MV

5-0.22** -0.12* -0.19** -0.06**

(-6.41) (-1.74) (-9.37) (-2.42)UPL*TV6 -0.46** -0.44** -0.46** -0.25**

(-18.87) (-16.64) (-22.90) (-9.89)Dry season -0.11** -0.07** -0.08** -0.07** 0.01 -0.06**

(-7.44) (-4.20) (-7.17) (-5.60) (0.72) (-2.55)CAR

7-0.22** -0.22** -0.24** -0.23** -0.59** 0.34**



Table 4 (Continued)

Rice Yield and Fertilizer Use Function in the Philippines, 1970-97

Variables ln (yield/ ha) ln

(NPK/ha)

ln (yield/

NPK)9

1970-79 1970-76 1980-89 1990-97

Southern Mindanao -0.19** -0.23** 0.06 -0.05 -0.62** 0.56**

(-4.71) (-5.08) (1.90) (-1.39) (-8.75) (6.53)Central Mindanao -0.27** -0.24** 0.13** -0.005 -0.74** 0.60**

(-6.61) (-5.34) (4.12) (-0.13) (-9.84) (6.71)ARMM

8-0.22** -0.27** -0.08** -0.15** -0.81** 0.47**(-5.42) (-5.93) (-2.54) (-3.90) (-11.38) (5.52)

CARAGA -0.72** -0.73** -0.37** -0.17** -0.97** 0.73**(-17.40) (-15.84) (-11.32) (-4.17) (-13.55) (8.46)

Constant 0.38** -0.43** 0.50** 0.58** 4.45** -3.51**(9.08) (10.06) (15.86) (14.44) (64.60) (-41.02)

R-squared 0.66 0.67 0.76 0.67 0.52 0.37

No. of Observations 1565 1043 1679 1185 570 565



1970-79 1970-76 1980-89 1990-97

Southern Mindanao -0.19** -0.23** 0.06 -0.05 -0.62** 0.56**(-4.71) (-5.08) (1.90) (-1.39) (-8.75) (6.53)

Central Mindanao -0.27** -0.24** 0.13** -0.005 -0.74** 0.60**(-6.61) (-5.34) (4.12) (-0.13) (-9.84) (6.71)

ARMM8

-0.22** -0.27** -0.08** -0.15** -0.81** 0.47**(-5.42) (-5.93) (-2.54) (-3.90) (-11.38) (5.52)

CARAGA -0.72** -0.73** -0.37** -0.17** -0.97** 0.73**(-17.40) (-15.84) (-11.32) (-4.17) (-13.55) (8.46)

Constant 0.38** -0.43** 0.50** 0.58** 4.45** -3.51**(9.08) (10.06) (15.86) (14.44) (64.60) (-41.02)

R-squared 0.66 0.67 0.76 0.67 0.52 0.37 No. of Observations 1565 1043 1679 1185 570 565

Note: Year dummies are included in the regression model but are not shown in thistable. Numbers in parentheses are t-values. ** means significant at 1% level, *

means significant at 5% level.1 IRG refers to the dummy for irrigated ecosystem.

2 IRG*MV refers to the interaction term IRG and dummy for the adoption of modern varieties (MVs) of rice.



41

Table 5

Irrigation Ratio, Cropping Intensity, and Adoption of Modern Rice Varieties in Central Luzon, the Philippines, 1966-99

Variable 1966 1967 1970 1971 1974 1979 1980 1982 1986 1987 1990 1991 1994 1995 1998 1999

W1

D W D W W D W W D W D W D D W

Sample size 92 17 62 13 58 149 81 136 120 64 108 56 100 56 46 79

Irrigation ratio 60 100 60 100 53 71 100 72 68 100 58 100 61 100 100 65

(% area)

Cropping 1.1 1.2 1.6 1.5 1.5 1.5 1.5

intensity

Adoption of MVs (% area)

TV2

100 94 34 8 27 0 0 0 0 0 0 0 0 0 0 0

MV13

0 6 66 92 73 8 9 3 1 2 2 3 0 0 0 0

MV24

0 0 0 0 0 92 91 97 38 6 8 15 6 0 0 0

MV35

0 0 0 0 0 0 0 0 61 92 90 82 94 100 100 100

1 W refers to the wet season and D refers to the dry season.2 TV refers to traditional rice varieties.3 MV1 refers to the first generation of modern rice varieties.

4 MV2 refers to the second generation of modern rice varieties.



Table 6

Determinants of the Adoption of Modern Rice Varietiesin Central Luzon, the Philippines, 1970-99, Tobit Estimation

MV11

MV22

MV33

(1967-74) (1978-82) (1986-99)

Intercept -0.08 0.65** -5.69**(-0.10) (3.40)

(-9.07)

Irrigation ratio 0.04 0.06* 0.34**(0.50) (2.12)

(4.00)Ln age of head 0.05 0.04 1.21**

(0.26) (1.05)

(8.51)Ln schooling of head 0.004 -0.005 0.31

(0.11) (-0.34)(0.81)

Ln Farm size -0.05 -0.06** -0.21**

(-0.66) (-2.67)(-3.74)

% area under share-tenancy -0.07 -0.06 -0.72**(-0.43) (-1.01)



43

Table 7

Yield of Modern Rice Varieties and Total Factor Productivity Trendsin Central Luzon, the Philippines, 1966-99

Variable 1966 1967 1970 1971 1974 1979 1980 1982 1986 1987 1990 1991 1994 1995 1998 1999

W1

D W D W W D W W D W D W D D W

Yield (tons/ha/season)

TV2

2.3 1.9 2.4 1.4 2.1

MV13

1.4 2.6 2.5 2.2 2.8 3.8 3.0 5.0 3.6 4.9 5.9

MV24

3.6 4.3 4.1 3.5 4.1 3.4 3.5 4.6

MV35

3.6 4.3 3.6 4.6 3.9 4.4 4.8 3.4

NPK6

9 17 29 59 39 62 78 63 67 88 70 103 93 125 150 143

(kg/ha/season)

TFP index7

(1966=100) 100 72 97 61 74 110 119 135 129 138 115 142 124 134 162 114

1 W refers to the wet season and D refers to the dry season.2 TV refers to traditional varieties (TVs) of rice

3 MV1 refers to the first-generation of modern rices consisting of IR5 to IR34 and C4 series.4 MV2 refers to the second-generation of modern rices consisting of IR36 to IR62.

5 MV3 refers to the third-generation of modern rices consisting of IR64 to IR74 and PSBRc series.

6 NPK means nitrogen, phosphorous, and potassium.7 TFP means total factor productivity index computed using the Tornqvist formula.



Table 8

Determinants of Yields in Central Luzon, the Philippines, 1966-99

Explanatory variables ln yield1970-74

ln yield1979-82

ln yield1986-99

ln value of production per ha

1986-99

MV11

0.22(1.20)

MV22 0.31*(1.88)

MV33

-0.13 -0.07

(-1.53) (-0.78)IRG4 0.20 0.70** 0.04 0.05

(1.05) (3.39) (0.61) (0.62)IRG*MV15 -0.07

(-0.33)IRG*MV26 -0.33

(-1.52)IRG*MV3

70.07 0.04

(0.19) (0.42)

Dry season dummy -0.39** 0.07 0.16** 0.25**(-2.29) (1.06) (3.46) (4.97)

ln age of head 0.05 0.10 0.05 0.08




means significant at 5% level.1 MV1 refers to proportion of rice area planted with the first generation of modern

varieties (MVs) of rice consisting of IR5 to IR34 and C4 series.2 MV2 refers to proportion of rice area planted with the second generation of MVs

consisting of IR36 to IR62. 3 MV3 refers to proportion of rice area planted with the third generation of MVs

consisting of IR64 to IR74 and PSBRc series.4 IRG means the proportion of rice area with irrigation.

5 IRG*MV1 refers to the interaction term between IRG and MV1.6 IRG*MV2 refers to the interaction term between IRG and MV2.

7 IRG*MV3 refers to the interaction term between IRG and MV3.



Table 9

Determinants of Total Factor Productivity in Central Luzon,

the Philippines, 1966-94

Explanatory variables 1970-74 1979-82 1986-99

MV11 0.09

(0.68)MV2

20.32**

(2.05)MV33 -0.08

(-1.10)

IRG4

0.07 0.41** -0.01(0.55) (2.21) (-0.18)

IRG*MV15

0.00(0.00)

IRG*MV26 -0.25(-1.31)

IRG*MV37

0.04(0.53)

Dry season dummy -0.29** -0.06 0.21**

(-2.49) (-1.10) (5.26)ln age of head -0.09 -0.01 -0.03

(-0.66) (-0.17) (-0.50)




means significant at 5% level.1 MV1 refers to proportion of rice area planted with the first generation of modern

varieties (MVs) of rice consisting of IR5 to IR34 and C4 series.2 MV2 refers to proportion of rice area planted with the second generation of MVs

consisting of IR36 to IR62. 3 MV3 refers to proportion of rice area planted with the third generation of MVs

consisting of IR64 to IR74 and PSBRc series.4 IRG means the proportion of rice area with irrigation.

5 IRG*MV1 refers to the interaction term between IRG and MV1.6 IRG*MV2 refers to the interaction term between IRG and MV2.

7 IRG*MV3 refers to the interaction term between IRG and MV3.



48

Cagayan Valle

Ilocos

CentralLuzon

SouthernTagalog

Bicol

WesternVisa as

Figure 1: Regional Map of the Philippines Map is courtesy of the Geographic Information System of the International Rice Research Institute.



49

02468

101214

161820

C A R

I l o c

o s

C a g a y a n

V a l l e y

C e n

t r a l L u z

o n

S o u

t h e r n

T a g a

l o g

B i c o

l

W e s

t e r n

V i s a y

a s

C e n

t r a l V i s a y

a s

E a s

t e r n

V i s a y

a s

W e s

t e r n

M i n d a n

a o

N o r t h e r n

M i n d a n

a o

S o u

t h e r n

M i n d a n

a o

C e n

t r a l M i n d a n

a o

A R R

M

C A R A G A

Figure 2

Average of Regional Rice Production Shares in the Philippines, 1970-97



Yield of MVs

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

1970 1975 1980 1985 1990 1995 2000

Y i e l d

( t o

n s / h a )

Irrigated

Rainfed

Upland

Yield of TVs

2.50

3.00

3.50

4.00

/ h a )



51

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

1970 1975 1980 1985 1990 1995 2000

Irrigated

Rainfed

Upland

Figure 4

Coefficient of Variation in Yields in the Philippines

Lessons from Three Decades of Green Revolution.pdf

Documents