Embed Size (px)
Citation preview
October 2006
Reforming Intellectual Property Rights, Bio-safety Management and the Bt Cotton Seed Industry in China: Who Benefits from Policy Reform? ∗
Ruifa Hu Center for Chinese Agricultural Policy, Institute of Geographical Sciences and Natural
Resource Research, Chinese Academy of Sciences Jia 11, Datun Road, Beijing 100101 CHINA
Carl Pray Department of Agricultural, Food, and Resource Economics
Rutgers University, 55 Dudley Road, New Brunswick, NJ 08901-8520 USA [email protected]
Jikun Huang
Center for Chinese Agricultural Policy, Institute of Geographical Sciences and Natural Resource Research, Chinese Academy of Sciences
Jia 11, Datun Road, Beijing 100101 CHINA Tel: 86-10-6488-9833; Fax: 86-10-6485-6533
[email protected] (Corresponding author)
Scott Rozelle
Department of Agricultural and Resource Economics University of California, 1 Shields Ave., Davis, CA 95616 USA
Cunhui Fan Center for Chinese Agricultural Policy, Institute of Geographical Sciences and Natural
Resource Research, Chinese Academy of Sciences Jia 11, Datun Road, Beijing 100101 CHINA
Caiping Zhang
Department of Agricultural, Food, and Resource Economics
Rutgers University, 55 Dudley Road, New Brunswick, NJ 08901-8520 USA
∗ The authors acknowledge th e financial support from the National Natural Science Foundation of China (70325003, 70333001 and 70021001) and International Food Policy Research Institute .
Reforming Intellectual Property Rights, Bio-safety Management and the Bt Cotton Seed Industry in China: Who Benefits from Policy Reform?
Abstract
China has been one of the leaders in agricultural biotechnology research and the adoption of transgenic plants. Despite this, critics argue that Chinese biotechnology policies could be improved to provide more benefits to farmers and more incentives to companies for greater research. The objective of the paper is to examine if policy changes could improve the welfare of producers and consumers and others in the cotton industry. The paper first reviews recent changes in laws and policies that affect China’s plant biotechnology sector—IPR legislation, changes in bio-safety regulation and seed industry reforms. Next, using a primary data set collected from more than 1700 plots from a sample of farmers in northern China in 1999, 2000 and 2001, we econometrically estimate the effect of changes to intellectual property rights (IPR), bio-safety regulation and seed industry reform on farmer pesticide use and yields. The results are used as the basis of a simulation exercise designed to measure the size and distribution of the benefits (and costs) of reform. We find that improvements to the IPR environment, enhanced enforcement of China’s bio-safety management regulation and greater commercialization of the seed industry positively affect the benefit of producers, consumers and technology providers.
Keywords: Biotechnology, Intellectual Property Rights, Bio-safety Management, Seed industry, China
1
Reforming Intellectual Property Rights, Bio-safety Management
and the Bt Cotton Seed Industry in China: Who Benefits from Policy Reform?
1. Introduction
Until recently, China’s government successfully raised the productivity of the
agricultural sector with new plant varieties and other modern inputs (Zhang and Fan,
1999; Fan, 1999). Because of a number of perceived problems with a public-sector
dominated system, since the mid-1990s, reformers have tried to encourage new
institutional approaches to develop and disseminate new varieties. Leaders have
passed a number of new laws governing intellectual property rights (IPRs). Officials
are experimenting with new bio-safety management approaches. There are new
initiatives pushing for the commercialization parts of the crop breeding system and
the seed industry (Louwaars et al., 2005). In many of these efforts, the private sector
is being encouraged to play a larger role. In the case of the cotton industry the
government has allowed joint ventures between international companies and domestic
seed firms to commercialize genetically modified (GM) cotton. While it is clear that
there are still weaknesses with China’s IPR environment, bio-safety management and
seed industry reforms, a number of studies have documented the success of China’s
GM cotton technology in increasing the productivity of farmers (Pray et al., 2001;
Huang et al., 2002 and 2002b; Jia et al., 2004; Wu and Guo, 2005; Yang et al., 2005a
and 2005b).
Despite past successes, a number of questions remain about the sustainability
of the way China is developing and extending new agricultural technology. Will
weakness in the intellectual property rights environment in China hurt the
effectiveness of the technology that is in the field? Is China’s current bio-safety
system so dysfunctional that farmers have to forgo income gains that would otherwise
2
have been created by better managed technologies? Are the seed reforms working and
are newly created seed enterprises providing farmers with high quality seeds?
The overall goal of this paper is to answer some of these questions. We seek to
do so by quantifying some of the costs and benefits of reforming China’s IPRs, bio-
safety regulations and seed industry policies. We explore the benefits of three sets of
policy reforms: a.) increasing the scope and improving the enforcement of IPRs; b.)
improving the bio-safety regulatory system in order to more effectively enforce bio-
safety regulations; and c.) reforming industrial policy to encourage firms inside and
outside of China to enter the GM seed market in order to increase competition and
raise the efficiency of technology transfer.
The main contribution of our paper is that it analyzes the effect of these
policies using microeconomic models of household and firm behavior in China.
Although the literature contains much discussion on the emergence of IPR, bio-safety
management and seed reform, there are few studies that empirically link the policies
with the production behavior of farmers. We also use the measured benefits from Bt
cotton farmers (from the regression analysis) and information from seed firms,
pesticide companies and government agencies to simulate the effects of strengthening
IPRs, improving bio-safety management and reforming the seed industry on the
magnitude and distribution of the benefits.
Because the scope of our work is so broad, we necessarily must limit its scope.
For example, the paper examines empirically the impact of biotechnology
management policies on the production of farmers (use of pesticides and yields) and
other actors and the distribution of those costs and benefits. While it is important to
consider enforcement costs, we have little information on such costs and end up
focusing on measuring the effect of policy efforts to implement IPR, bio-safety and
3
seed industry reform on the benefits/costs to producers, consumers and the
seed/research sectors. In carrying out the analysis, we recognize that our analysis is
more or less static (and does not explicitly seek to measure some of the longer-run
dynamic changes that could accrue from better IPRs). Therefore, we can not explicitly
measure the effect of better IPRs on the research strategy of seed and technology
firms (i.e., the willingness to invest more since they are able to capture more of the
benefits). In the discussion of the results, these more dynamics effects can only be
assumed.
2. Reforms in China: IPR, bio-safety management and seed industry
Intellectual property rights
Prior to the late 1990s it was legal for a seed company to reproduce a variety
of another company for the purpose of marketing the new variety. There also were no
restrictions on the use of another breeder’s variety as a parent in the development of
another variety. The new variety could be sold and marketed legally without paying
any licensing fees or royalties to the original creators of the parent varieties. The use
of varieties of other breeders as parents was a common practice in the 1980s and early
1990s for all crops (The State Council, 1997).
Despite the emergence of Plant Varietal Protection (PVP) in most sectors of
China’s agricultural sector since 1997, protection is still not very strong for the
covered crops, and the most important crop considered in this study, cotton, had been
excluded from protection until 2005. Even if it had been included, however, it would
not restrict the use of protected varieties as parents in the production of other varieties.
China’s PVP legislation has a research exemption that explicitly allows research
4
institutes and seed companies to use PVP varieties as parents to develop new
varieties. In addition, PVP does not give proprietary protection to genes.
In response to these loopholes in the PVP laws, research institutes and seed
companies have begun to take other actions to prevent their proprietary varieties and
novel genes from being used by other scientists without permission or at least begin to
receive royalties. Above all, the seed industry is beginning to use the patent system.
For example, Dr. Guo Sandui of the Chinese Academy of Agricultural Sciences
(CAAS) received a patent on the Bt gene that he developed. This gene (henceforth,
the CAAS gene) is being used in all of the China-produced varieties that are being sold
by a CAAS joint venture enterprise (henceforth, Biocentury—Fang et al., 2001).
Monsanto also has patents on several genes that are important in the production of
transgenic plant varieties, although it did not patent its initial Cry1Ac Bt gene in
China (henceforth, the Monsanto gene). Monsanto genes are legitimately found in the
seeds sold by two joint venture seed enterprises set up originally by two foreign-
owned life science firms, Monsanto and Delta and Pine Land (DP). In the past decade
these foreign-owned firms partners with two different provincial seed companies
(Jidai in Hebei and Andai in Anhui—henceforth, both JVs are called Jidai for
convenience). The CAAS and Monsanto patents cover genes that are inserted into
plants to make plants resistant to certain classes of insects. In particular, the patents
cover processes that create transgenic cotton varieties.
In addition, research institutes and seed firms also can try to use trademarks—
another form of IPR—to protect their technology (Louwaars et al., 2005). Biocentury
has trademark protection on its name and on the names of some components of their
technology (Fang et al., 2001). Jidai uses Monsanto’s Bollgard trademark on its Bt
5
cotton varieties as a way of trying to prevent other firms from using the name on their
varieties.
While some seed companies in the cotton industry have taken steps to protect
their seed varieties, many of their actions do not appear to have helped keep other
companies from appropriating their technology. Interviews with CAAS, Biocentury,
Monsanto and Jidai managers and research administrators reveal that few people
believe that the current system of IPRs and the way that they are being enforced
provide effective protection for the technologies that are patented and for the seed
products that have trademarks. Cotton seed firms have had little success in keeping
other firms from copying their genetic technologies (specifically the genes) or
trademarks. New entrants into the Bt cotton seed market still regularly reproduce,
backcross and market the varieties developed by both Biocentury and Jidai. In fact,
we estimate that while the varieties with the CAAS gene covered 2.25 million
hectares in 2004 (or 61 percent of the nation’s Bt cotton area—Figure 1), the
legitimate Bt cotton varieties sold by Biocentury covered only 5 percent of total Bt
cotton area in the same year. The rest of the varieties using the CAAS gene did not
have permission of or pay royalties to either CAAS or Biocentury (a type of seed that
we call illegitimate in the rest of the paper). Our estimates show that legitimate Jidai
varieties covered about the same share of sown area as legitimate Biocentury varieties.
According to the Delta and Pine Land representative in Beijing legitimate cotton seed
(that was produced by Jidai contract seed producers and marketed by Jidai
representatives and agents) covered less than 5 percent of the Bt cotton market in
2004. In spite of such a low market share, varieties with Monsanto’s gene (both
legitimate and illegitimate ones) were being planted on at least 39 percent of China’s
total Bt cotton area.
6
Regulation of plant biotechnology
In addition to the IPR reforms, since the early 1990s there have been
increasing attempts by the government to regulate the plant biotechnology industry.
In the early 1990s for the first time government policymakers began to create
legislation, establish policies and set up regulations to govern the development and
sale of GM seeds. In response to the emergence of China’s agricultural biotechnology
industry, leaders from Ministry of Science and Technology (MOST) promulgated the
first set of bio-safety regulations, the “Safety Administration and Regulation on
Genetic Engineering,” in 1993 (MOA, 2005). Following the issuance of MOST’s
regulations, the Ministry of Agriculture (MOA) issued the “Implementation Measures
for Agricultural Biological Engineering” in 1996. The MOA measures delineated the
steps a firm must take: a.) to get approval to do research on transgenic organisms; b.)
to test them outside of laboratories and greenhouses; and c.) to commercialize them.
In recent years the policy and regulatory environment have become more
stringent. In May 2001 the State Council decreed a new set of policy guidelines, the
“Regulations on the Safety Administration of Agricultural GMOs,” to replace the
early regulations issued by MOST. MOA also announced three new implementation
regulations to replace their earlier rules. The new framework, which took effect in
March 2002, greatly expanded the scope of regulation to include more detailed rules
on bio-safety management, trade and labeling of GM food products.
While fairly rigorous by developing country standards, when compared to
those in many developed countries, China’s bio-safety regulations appear to have
been implemented in a fairly loose way and cover only a limited scope—especially in
the initial years. For example, in the late 1990s if a gene being evaluated had
previously passed through the bio-safety regulatory process in the United States or
7
Canada, it was generally assumed to meet China’s food safety and environmental
regulations. New policies and regulations that have arisen since 2002, however, have
increased at least the nominal level of regulation faced by those that use GM
technology and are involved with the processing and sale of GM food crops (MOA,
1999; MOA, 2005).
Although in the literature we do not find any research that provides
quantitative evidence, the record of the bio-safety committee being able to regulate
products in the field does not appear to have been effective. Our field work suggests
that the bio-safety regulation system failed to keep seed out of the nation’s seed
markets when they have not been approved (or seed that in the rest of the paper we
call unapproved seed). Specifically, there were a number of Bt cotton varieties that
were being sold even though they were almost certainly not produced with either the
Jidai or CAAS genes (the only genes that had received approval from the Bio-safety
committee by 2001, the final year of the study period).
Reforms to the seed industry
In recent years, changes also have been occurring in the seed industry. As late
as the mid-1990s local and regional state-owned enterprise (SOE) seed monopolies
dominated China’s seed industry (Qian, 1999; Keeley, 2003; Li and Yan, 2005;
Huang et al., 1999). In total, 2700 SOEs operated in their local counties, prefectures
and provinces. In many counties only the local SOE was allowed to sell seeds of the
major crops. Regulations banned the participation of non-SOE seed firms in the
production, distribution and sale of hybrid maize and rice. In the typical case, the
county-based SOEs sold their seed through township agricultural extension agents
(which in the rest of the paper we will call a traditional, non-commercial seed sales
channel). Indeed, during the 1990s agricultural extension agents earned a large share
8
of their income from selling agricultural inputs, including seed. In addition, seed also
flowed to farmers through other traditional, non-commercial channels, such as the
cotton office (a state-designated monopoly cotton procurement agency that operated
without competition until the late 1990s) and seed production bases (which are
villages or groups of villages that have contracts with the former SOEs for the
reproduction of their seed).
Since the mid 1990s the laws and policies that govern the seed industry have
changed in such a way that a commercial and competitive seed industry has begun to
evolve (Keeley, 2003; Li and Yan, 2005). Among other parts of the legislation, the
law makes it clear that any entrepreneur that has access to the required minimum
amount capital and facilities can sell seed. Private companies are allowed to sell seed
(including any variety of GM or non-GM cotton) that was bred by public breeding
institutes. With the passage of this legislation, the legal protection of the monopoly
positions of county, prefectural and provincial seed companies was formally
removed.1
As the reforms began to be implemented, commercial seed distribution
channels for seeds opened along side the networks through which agricultural
extension agents (and personnel in cotton offices and seed production bases) had
traditionally sold SOE-marketed seed to farmers. New sources of investment in the
industry have emerged. For example, domestic entrepreneurs have invested in private
seed firms. Some of the traditional SOEs commercialized and have transformed
1 The evolution of the seed industry continued after th e late 1990s. In 2000 the government pas sed a new seed law that for the first time legally defined a role for the private sector. All firms —private, quasi-commercialized SOEs and traditional SOEs —were allowed to apply for permits to sell seed in any jurisdicti on. Measures also were put into place that allowed firms to have their seeds certified at the provincial level which would entitle t hem to sell seed in any county in the province. By late 2001 nine companies had permits to sell seed anywhere in th e country. For the first time, it became feasible for national companies to establish t heir own distribution and retail networks. At the other end of the spectrum thousands of small seed companies opened up to supply local needs.
9
themselves into a commercial firm. Although still few in number, foreign firms have
begun to invest in the seed industry.
In the cotton seed industry—especially in the part of the industry that is
involved in the creation and marketing of GM cotton—the government’s recent policy
efforts appear to have been effective in encouraging the development of a commercial
seed industry. Specifically, there have been three fundamental shifts in the structure of
the cotton seed industry: the appearance of large commercial seed companies that
operate at the regional or national level; the rise of private foreign firms (although
they still play a somewhat limited role); and the emergence of small, private cotton
seed firms. Despite these shifts, field visits make it clear that the traditional seed
producers (the local, county-owned SOE seed enterprise) and their distribution
channels (the agricultural extension system, cotton offices and seed production bases)
are still active.
Because of the complexities of the Bt cotton seed industry and the large
number of actors, in Figure 2 we provide a schematic diagram of the various actors in
the Bt cotton seed industry today. One source of Bt cotton seed is obtained by farmers
after it flows through the commercial seed sector. Another source is obtained after it
flows through the traditional, non-commercial seed sector. Farmers also save Bt
cotton seed and use non Bt varieties. Together Bt cotton seed from commercial seed
companies, non-traditional entities, saved seed and all non Bt cotton seed account for
all of the seed used by farmers in China today.
3. Policy shifts and the impact on productivity
In this section, we seek to answer three questions: a.) Will stronger IPRs
provide greater incentive to do research and increase productivity of cotton
10
production? b.) Will more effective bio-safety regulation enforcement benefit China’s
farmers? c.) Will reforms to the seed industry help or hurt producers?
Data
To help determine whether or not stronger IPR, more effective bio-safety
regulations and better functioning seed markets affect the welfare of farmers, we seek
to calculate the benefits and costs of Bt cotton that have been enjoyed/born by farmers
with data that we collected ourselves. To measure these effects, in the first part of our
analysis we use a data set on Bt cotton, the main GM crop being cultivated in China.
The first set includes information on costs and returns of 282 households that
produced cotton in villages that were producing Bt cotton in 1999, 407 households in
2000 and 366 households in 2001.2 These data were collected in five cotton producing
provinces in northern China (Hebei, Shandong and Henan) and eastern China (Jiangsu
and Anhui). During the three years of the survey (which to the extent possible
surveyed the same households over time) also collected data on 1716 plots of the
sample households. Such a data set allows us to assess the producer benefits and costs
for different Bt cotton varieties.
The survey instrument for each household included a number of different
blocks. First, there were sections that covered the basic information of the household:
the characteristics of the farm household, household head and the nature of the farm
2 While it is true that we have used data from a subset of these households in other papers, there, in fact, are many dimensions of the data that we use in this paper that make the data unique and special. First, in our previous paper that was published in Science (Huang et al. , 2002a) we used only 1 year of data. In 2000 and 2001 we visited our sites for second and third rounds of data collection. During the second and third rounds we followed the spread of Bt cotton and included households that were in two new Bt cotton -producing provinces that had not been included in the Science paper. Second, although the sample coverage was the same that we used in a paper that we published in the Plant Journal ( Pray et al., 2002), in this paper we draw on an entirely new block of our survey that was designed to examine the IPR/bio-safety/seed system environment in which the Bt cotton farmers are producing. None of the data that are used to create our measures of seed categories —measures of the legitimacy of varieties; measures of whether varieties were approved or unapproved; me asures of whether seed came through traditional, non -commercial channels or not —have ever been used in any other analysis.
11
itself. Another section was used to elicit information about cotton inputs and output.
These data were collected by plot.
One of the unique parts of the data comes from our inquiries about the source
of the Bt cotton seed that the farmers were planting. With these data we are able to
gain an understanding about the importance of IPR, bio-safety regulation and seed
reform. In the first part of the question, we asked farmers if they bought their seed or
if they used saved seed. If the household purchased seed, we asked for the variety
name, the source of the seed and its price. Nearly all farmers had no difficulty in
telling us this. We used this information, along with a number of assumptions, to
identify the nature of the seed (its legitimacy; whether or not its genetic material has
been officially approved; whether it came from a commercial seed firm or a
traditional, non-commercial channel). In other words, the information from this
section was used to categorize the seed used by the farmer into on the boxes in Figure
2. Based on our data, it is clear that since China began to commercialize Bt cotton, the
nation’s cotton farmers have obtained seed from a wide range of sources (Table 1). It
also is clear that the nature of China’s seed industry has been changing over time. In
particular, the sample farmers report that the major source of Bt cotton seed is some
type of seed company (as opposed to a traditional non-commercial channel). Forty-
eight percent of the sample used seed purchased from some seed company (column
1). It is likely that in the early 1990s, this number was nearly zero. However, despite
the shift, there were still many vestiges of the traditional system. Specifically, 23
percent of farm households purchased seed from government-run agricultural
extension stations and 2 percent of farmers purchased seed from the government-run
12
cotton office. Seven percent bought seed from a seed production bases.3 Finally, 20
percent of farmers use their own saved seeds.4
While the respondents were able to state precisely from which source their
seed came, it was more difficult for them to identify the exact company which
produced their varieties. For example, farmers that purchased seed from the
commercial seed companies typically could distinguish between “foreign seed”
varieties that were using the Monsanto gene (or Jidai varieties)and those that were
developed with the CAAS gene (“not foreign seed” or Biocentury varieties).
However, the same farmers often were not sure whether the seed company from
which they bought their seed was actually selling legitimate seed or not (that is if they
were actually produced and distributed by either Jidai or Biocentury or their
authorized partners and/or dealers). From our field work, we discovered that the
legitimacy of seed could be largely determined by looking at the price that farmers
paid for their seed and whether or not the seed was delinted and/or treated.
Legitimate Jidai and Biocentury seed was always delinted and treated; legitimate seed
also was nearly always sold for a fixed price, about 45 yuan per kilogram. As a
consequence, when farmers reported that they bought seed for less than 30 yuan per
kilogram, even when they called the seed 33B (a common name for the seed sold by
Jidai), we assume that they were buying illegitimate seed. Likewise, when the farmer
said that he/she purchased loose and fuzzy seed (which showed that it was not
3 Seed production bases are typicall y villages or groups of villages (or sub-villages) that have contracts with some local seed comp anies or government extension centers to replicate seed. In some instances, these villages form themselves into quasi -companies under the leadership of a leader of the village. They then sell a share of the seeds that they produce directly to farmers. 4 While there is a general pattern throughout China, our data shows that there are substantial differences across space. For example, 55 percent of farmers in Henan and Jiangsu purchased seed from seed companies. In contrast, only 36 percent of househol ds in Hebei buy seed from commercial seed companies. Purchasing seed from extension agents also varies among provinces, from about 48 percent in Anhui to none in Hebei. Finally, the use of saved seed also differs among regions.
13
delinted seed/and not pink which showed that it had been treated with fungicide) even
if he/she said the seed was from Jidai or Biocentury, we also assume that the seed was
illegitimate (since all legitimate seed was delinted and treated).
Using the data from our survey (under the assumptions about the source of the
seed), we are able to take the first step in (descriptively) analyzing the effect of IPR,
bio-safety and seed system reforms on farm-level pesticide use and yields. To do so,
we first divide the cotton seed used by farmers into 9 types (Figure 2): legitimate Jidai
(“foreign seed” that was high priced, delinted and treated); illegitimate Jidai (“foreign
seed” that was low-priced or fuzzy or non-treated); legitimate Biocentury (“not
foreign seed” that was high priced, delinted and treated); illegitimate Biocentury (“not
foreign seed” that was low-priced or fuzzy or non-treated); seed from a seed company
that was unapproved (seed that was clearly not approved by the bio-safety
committee—“seed that did not contain Monsanto or CAAS gene”); seed that came
from the Agricultural Extension Station; County Cotton Office; or seed production
base; seed that was saved by the farmer and reused; and non-Bt cotton seed. After we
make such divisions, the use of pesticides and yields of each of the varieties can be
compared (Table 2). In doing so, several patterns appear from the descriptive statistics
that are consistent with the fact that IPRs, bio-safety management and seed industry
reform matters. First, in each case, legitimate varieties (either Jidao or Biocentury)
outperform the illegitimate varieties (row 3 vs. 4; row 5 vs. 6). This is true for both
pesticide use (e.g. legitimate Jidai uses only 20 kg/ha while illegitimate Jidai uses 23;
and legitimate Biocentury uses 15 kg/ha while illegitimate Biocentury uses 20 kg/ha)
and yields. In addition, our data also show that approved varieties (all of the Jidai and
Biocentury varieties—rows 3 to 6) use less pesticide than unapproved varieties (row
7). The legitimate Jidai and legitimate Biocentury varieties also have higher yields
14
than unapproved varieties. Finally, when farmers use seeds from commercial seed
companies (row 2), they also use less pesticides and have yields than when the
purchase seed from traditional, non-commercial channels (row 8).
Although based on such descriptive findings, it appears as if stronger IPR,
more effective bio-safety management and additional seed industry reform would lead
to lower pesticide use and higher yields, we have only looked at simple correlations. It
is possible that there are other observable factors that are confounding our results.
Therefore, in the next section we use regression analysis to formally test these
relationships. To facilitate the test, we use the descriptive findings (from Table 2) and
debates in the literature to generate a set of testable hypotheses that can help us
analyze the effect of IPR shifts, bio-safety regulation and seed reform on the
performance of farmers. There are three basic hypotheses:
Hypothesis 1 (effect of strengthening IPRs): There are legitimate and illegitimate seeds that are related to two different seed companies (Jidai and Biocentury). The presence of illegitimate seed in our paper is symptomatic of the failure of China’s IPR system. Legitimate seed should use less pesticide and generate higher yields than illegitimate seed. Hypothesis 2 (effect of bio-safety management): There are varieties sold by seed companies that contain genes that have been approved by the bio-safety committee (those of Jidai and Biocentury) and there are varieties sold by seed companies that do not contain the genes that have been approved by the bio-safety committee (henceforth, unapproved varieties). The presence of unapproved seed in our paper is symptomatic of the failure of China’s bio-safety system. Approved seed should use less pesticide and generate higher yields than unapproved seed. Hypothesis 3 (effect of seed industry reform): There are varieties sold by seed companies (Jidai—legitimate and illegitimate; Biocentury—legitimate and illegitimate; and other seed companies that sell unapproved varieties). There are varieties sold by non-commercial entities other than seed companies—the agricultural extension system; the cotton office; and seed production bases (henceforth, traditional channels). The presence of traditional channels in our paper is symptomatic of incomplete reform. Seed from commercial channels should use less pesticide and generate higher yields than seed from traditional, non-commercial channels.
15
3.2 Identifying the differences in the impacts—the econometric approach
To understand the net effect of seeds from different sources (and test our
hypotheses), we adopt a multivariate production function approach using the pooled
data (for three years) from five provinces. Our ultimate objective is to estimate the net
effect on farm households of their Bt cotton adoption decisions. To do so, we need to
hold the effect of other inputs and household and farm characteristics constant.5 To
achieve this objective we use a two-equation analytical framework, one for pesticide
use and the other for yields.
While conceptually our approach is fairly straightforward, in the course of the
analysis we have to address several statistical complexities. First, to identify the yield
effects of Bt seeds and pesticide use, traditional production analysis is not appropriate,
since Bt seeds and pesticides do not act like traditional, yield-increasing inputs.
Instead, we use a damage control model that accounts for the way that insect-resistant
seeds and pesticides abate that damage of insects (Lichtenberg and Zilberman, 1986;
Huang et al., 2002b; Carrsco-Tauber and Moffitt, 1992; Babcock et al., 1992; Saha et
al., 1997). Second, even using a damage control function, the estimation is
complicated by the fact that pesticide use is determined simultaneously with yields.
In econometric terms this means there is a possibility that the coefficient on the
pesticide variable in the damage-abatement production function is subject to
endogeneity-induced bias. Consequently, in estimating the two equations we need to
5 In addition to the in strumental variables (in the pesticide use equation —described below) and the characteristics of Bt cotton seed (described above), in th e estimation of the pesticide use and yields we also included measures of household characteristics ( Age and Education Level of the household head in years; whether or not anyone in the household was a Village Leader or received trainin g the use of Bt cotton—Bt Cotton Training ) and miscellaneous factors (dummy variables for Year effects; Provincial effects; and Weather effects—which equal 1 if farmers stated that their communities/fields were abnormally affected by drought or floods or 0 if not). In the yield equation, we also included measures of conventional inputs applied to cotton ( Labor in days per hectare; Fertilizer in kilograms per hectare; and Other Inputs in yuan per hectare). For a complete description of the empirical approach, see Huang et al. (2002b).
16
rely on a two-stage, instrumental variable (TSLS) approach. As in Huang et al.
(2002b), we use two variables, one measuring pest pressure (Perception of Yield Loss)
and the other measuring Pesticide Price, to identify pesticide use.6
Finally, in order to control for other factors that might affect pesticide use or
yields—in addition to the source of seed, we also included several other
characteristics of the seed that farmers were able to reliably identify during the
household surveys. Specifically, we included two dummy variables: one variable is
equal to 1 if the seed was a Coated Seed (and 0 if non-coated); the other variable is
equal to 1 if the seed was a Hybrid Variety (and 0 if it was a conventional variety).
3.3 Results of regression analysis
Using our analytical approach, our specification and econometric framework
appear to perform well in the estimations of both equations. In the Ordinary Least
Squares versions of the pesticide use and yield equations (not shown for brevity), the
R-squares measures of goodness of fit were about 0.3, levels common for farm-level
production analysis. The signs on the coefficients of the control variables in both the
pesticide use and yield equations also are mostly as expected when we use our TSLS,
damage-abatement approach (Table 3). For example, in the pesticide use equation
(column 1), the Perception of Yield Loss variable indicates that farmer increases
pesticide use as pest pressure rises (row 2). In contrast, the negative coefficient on the
Pesticide Price variable means that higher pesticide prices dampen the demand for
pesticide (row 3). Farm Size also has negative sign in the pesticide use variables
indicating that there are scale economies in the use of pesticides. In the yield equation
6 These two variables can be used as Instrumental Variables (IVs) since they meet the two characteristics of an IV: they should affect the level of pesticide use; but should have no independent effect on yields, except through the use of pesticide. The Perception of Yield Loss is measured by asking farmers the extent of yield loss —stated in percentage terms —that they believe their crop would have incurred had they not sprayed with pesticides (th e variable ranging from 0 to 1). The Pesticide Price variable is constructed as the unit value of all of the pesticides applied by the farmer.
17
(column 2), the coefficients on the variables measuring traditional inputs are all
positive and their magnitudes are reasonable (Putterman and Chiacu, 1994).
More importantly, our analysis allows us to test Hypotheses 1 to 3 by
identifying the potential impact of enforcing IPRs, bio-safety regulations and reform
of the seed industry on Bt cotton pesticide use and yields (Table 3, column 1).7 The
potential effect of IPRs (Hypothesis 1) can be analyzed by looking at the differences
between Legitimate and Illegitimate Jidai and Biocentury varieties. If IPRs were
enforced, the illegitimate seed would not be available and more farmers would be
using the legitimate seed. As seen in the descriptive data in Table 2, the point
estimates suggest that this could increase yields and reduce pesticide costs. When
farmers used either type of legitimate seed (Legitimate Jidai or Biocentury), pesticide
use fell (when compared to non-Bt cotton) from between 39.77 and 41.45 kilograms
per hectare (rows 12 and 14). When using the illegitimate seed (either Illegitimate
Jidai or Biocentury), the fall in pesticide use is less (30.57 and 33.52 kilograms, rows
13 and 15) than when farmers used legitimate varieties.8 In other words, if IPR
regulations had been effective in keeping out unauthorized varieties that were being
sold as Jidai or Biocentury varieties, the use of pesticide by farmers would have been
less. The effect of enforcing IPRs is even clearer in the yield equations (Table 3,
column 2).9 The yields of the legitimate Jidai varieties were 25.7 percent above those
of the baseline conventional varieties; the yields of illegitimate Jidai varieties were
only 12.8 percent more (rows 12 and 13). The same was true in the case of the 7 In equations not show n here Bt reduced pesticide use by about 35kg/ha using the complete sample of the three years and five provinces in our sample. See Huang et al. (200 2b) for disaggregated analysis. 8 The differences are statistically significant. The F -statistic for the difference between legitimate Jidai and illegitimate Jidai pesticide use is 10.7 ( significant at the 1% level). The F -statistic for the difference between legitimate Biocentury and illegitimate Biocentury pesticide use is 2.56 (significant at the 10% level). 9 The F-statistic for the difference between legitimate Jidai and illegitimate Jidai yields is 5.37 (significant at t he 5% level). The F-statistic for the difference between legitimate Biocentury and illegitimate Biocentury yields is 2.57 (significan t at the 10% level).
18
Biocentury varieties of Bt cotton (rows 14 and 15). The yields of legitimate varieties
were 19.2 percent more than those of conventional varieties; at the same time, those
illegitimate versions of the Biocentury varieties were not statistically different from
conventional varieties.10
Our results also allow us to see the possible effect of enforcing bio-safety
regulations (Hypothesis 2)—especially in the case of yields (Table 3, rows 12 to 16).
We can see the benefits of bio-safety regulation by comparing the coefficients of the
seeds with approved Bt cotton varieties for commercialization (those associated with
the variables representing either Jidai or Biocentury varieties) and Unapproved Bt
cotton varieties. While the pesticide equation shows little difference in the reduction
of pesticide use, the yields of Unapproved Varieties are lower than the approved
varieties. Specifically, the yields of unapproved varieties are not statistically different
from conventional varieties of cotton, while the approved Biocentury and Jidai
varieties yield 19.2 to 25.7 percent higher (column 2, rows 12, 14 and 16).11 As in the
descriptive statistics, the results suggest that if there had been better bio-safety
management, and there had been no unapproved varieties, yields would have been
higher and pesticide use unchanged.
Finally, our results suggest that the reform of seed markets also have improved
the performance of Bt cotton varieties (Hypothesis 3). The pesticide reduction when
using varieties of seed sold through traditional channels (Cotton Office and Ag
10 Why is it that IPR contributes to lower pesticide use and higher yield s? In fact, we do not know for certain, since we are observing differences due to the source of the seed. One explanation is that it is because the company that pr oduces the legitimate varieties has more of an incentive to deliver more than just the genetics. They may want to package it in a seed that was: a.) produced uniformly (which in many circumstances will make it express the Bt toxin in a more consistent way) ; b.) is disease free; c.) was well maintained (e.g., thoroughly dried and stored in a cool, dry environment after harvest and processing); and, d.) etc. 11 The F -statistics for the equivalency of legitimate Jidai, legitimate Biocentury and Unapproved varieties for pesticide use is only 0.06, implying there is not statistical difference. However, the F -statistics for the equivalency of legitimate Jidai, legitimate Biocentury and Unapproved varieties for yields is 3.80 (and statistically significant that 5% l evel).
19
Extension Station—rows 17 and 18) is generally less and the yields are generally
lower than those purchased through the market (seed firms—rows 12 to 15 compared
to rows 17 and 18).12 Foreign Bt varieties (which were allowed by the seed reforms)
also yield higher than the average domestic Bt and non-Bt varieties (in their point
estimates). Statistical tests, however, demonstrate that there is no statistical difference
between the legitimate Jidai and Biocentury varieties in either their pesticide
reduction (the F-test is only 0.06) or yields (the F-test is 2.48). Hence, although we do
not know the exact mechanism why it is that the commercialization of the seed
industry induced by the seed industry reform policies lead to lower pesticide use and
higher yields, the results clearly show that seeds sold through commercialized seed
firms performed better. Moreover, when taken together, the econometric analysis
provides strong and consistent support that suggests that if China had improved
enforcement of IPRs, if leaders had implemented more effective bio-safety regulation
and if reforms had made seed markets more competitive (including allowing in more
foreign competition), pesticide use would have been significantly lower and yields
higher.
4. Identifying the magnitude and distribution of Bt cotton benefits
In the previous section, the empirical analysis demonstrated that users of Bt
cotton, especially those varieties that were approved by the bio-safety committee and
bought legitimately by farmers through the market, mostly used less pesticide and had
higher yields than seed that came through other sources. However, based on the
12 The F -statistics (for both pesticide reduction and yields) in the test of e quivalency between legitimate Jidai, legitimate Biocentury, unapproved seed from seed companies and seed sold by Extension agents and Cotton office area all large and stat istically significant. The seed production bases coefficients are not different, however.
20
regression analysis, the magnitudes of the benefits are still unclear. Furthermore, we
do not know from the regression analysis the distribution of the benefits. In response
to this possible criticism, we pursue two objectives in this section. First, we want to
identify (under a set of plausible assumptions on supply and demand elasticities—
which we call the baseline scenario) which groups of farmers benefit from the
adoption of Bt cotton and the size of those benefits. Second, we use the baseline
scenario as the basis of a simulation analysis to examine how the size and distribution
of the benefits would have differed if China’s IPRs were strenthened, bio-safety
regulations were more effectively enforced and the seed industry policy environment
improved.
4.1 Identifying net gains
The first step in identifying the net benefits and the distribution of the gains
from Bt cotton is to make assumptions on the value of pesticide savings, yield gains,
imputed labor costs and prices of the different types of seeds. We also need values of
behavioral parameters, such as supply and demand elasticities. To produce estimate of
these parameters, we rely mostly on two sets of information. First, we can calculate
the increases in yields and savings from pesticides for seeds from different seed
sources using the results from the regression in Table 3. When multiplying the net
gains (or savings) by the price of the pesticide and the price of output (which also
come from our household data), we come up with a value of yield changes (the
changes in yields times the changes in prices) by seed source and value of pesticide
savings by seed source (Table 4, columns 1 and 2). The household data also can
provide us with estimates of the imputed value of the labor cost reductions and change
in value of seeds paid by households by seed source (Table 4, columns 3 and 4).
21
Second, we also need to make assumptions about a set of demand and supply
elasticities for cotton in China. The baseline estimates of elasticities of supply (0.5)
and demand (1.0) for cotton are the same as those used in CAPSIM, a policy
simulation model that was created by the authors and one that has been used in many
analyses (Huang and Rozelle 2003—Table 5, column 1, rows 1 and 2). In order to be
assured that our results are not unduly influenced by our choice of supply and demand
elasticities, we also run a number of variants of the baseline analysis with alternative
values of the elasticities (Table 5, columns 2 to 5, rows 1 and 2). In the subsequent
discussion we refer only to the baseline results because, although the use of
alternative elasticities affects the division of estimated benefits between consumers
and producers, the total welfare change is generally insensitive to changes in the
assumptions about the supply and demand elasticities.
Using a standard methodology, the consumer-producer surplus model
developed by Alston et al. (1996), and adding to this analysis an approach that uses an
industrial profits model following the work of Moschini and Harvey (1997), we can
measure the benefits of Bt cotton (Table 5). According to our analysis, we find that
consumers of cotton capture some of the benefits of the extension of Bt cotton. In fact,
according to our analysis, about one-third of the benefits accrue to consumers. The
benefits to consumers come mainly from the fact that the price of cotton fell after the
adoption of Bt cotton. Interestingly, the results of this analysis differ from that we did
using 1999 data (Pray et al., 2001) since consumers did not benefit in the late 1990s.
Before 2000, the government artificially kept the price of cotton high, despite rising
supply, and so consumers did not benefit.
Despite falling cotton prices, farmers are the major beneficiaries of Bt cotton
(Table 5, row 4), a result that is driven primarily by the reductions in cost of
22
production of farmers. In the baseline analysis producers gain more than 2500 million
yuan which is more than 60 percent of the total welfare gain to society (row 4 versus
row 3). In contrast, seed companies and the providers of the genes continue to capture
only a small portion of the gains (Table 5, rows 6 to 8).13 Finally, insecticide
companies are the major losers (rows 9). Their sales are estimated to have fallen by
1,930 million yuan. If we assume their profit rates were about 5 percent, this would
translate into reduced profits of 96 million yuan.
Hence, from the analysis it can be seen that the big winner in the initial
commercialization of Bt cotton has not been the seed or biotechnology firms; they
captured only a small share of the benefits. If the small size of the benefits which are
captured by the seed and biotech companies continues, especially given that new Bio-
safety regulations are increasing the time and costs of commercializing new GM
crops, there may be little incentive for commercial companies to conduct biotech
research in China in the future (which we base on our logical deductions and not from
the numbers in our analysis). In fact, we may already be seeing evidence of this. For
example, Monsanto has pulled out of Jidai, the joint venture that started with DP and a
local seed company to sell Bt cotton seed. Currently, Monsanto only collects royalties
from the enterprise being run by Jidai partnership. In addition, Monsanto currently has
put on hold any plans to use its new double gene construct—Bollgard II—in any of
the varieties that are being used in China. Finally, Biocentury lost so much money
13 The benefits of the seed companies were calculated by multiplying the average profit of the companies in our sample of seed companies times the total amount of B t cotton seed purchased by farmers. The benefits to the companies are the average royalties per kg that seed companies in our sample reported paying to Jidai or Biocentury times the amount of legitimate seed that their affiliates sold.
23
during its first decade of operation that it recently allowed itself to be radically
restructured under new management.14
4.2 Size and distribution of reforms: simulation analysis
In this section, we come back to the three questions posed at the beginning of
this paper: a.) If IPRs could actually be enforced, how much would the welfare of
farmers, consumer and other actors change? b.) If the nation could improve its
enforcement of biosafety regulations, how would this affect the size and distribution
of benefits from Bt cotton? c.) Finally, how would the benefits from Bt cotton shift if
the structure of the seed industry were to change? Using our set of relatively simple
assumptions (but assumptions that we believe are reasonable enough to help us
illustrate our point), we seek to answer these questions with policy simulations that
allows us to compare the benefits from the three proposed policy shifts to the benefit
schedule that would be realized under the baseline scenario.
4.2.1 Scenario 1: Strengthen IPRs
In scenario 1 we simulate the effect of reforms to China’s IPRs. In this
scenario the government is assumed to broaden the scope of the Plant Variety
Protection Act to include cotton and to more effectively enforce IPRs on plant
varieties. The main effect of the reforms in terms of our analysis would be to
eliminate illegitimate varieties. From the view point of the producers in China, we are
assuming that all farmers that purchased cotton seeds from Jidai or Biocentury would
have purchased varieties that had been managed and sold by the companies
themselves (i.e., all of the seed was legitimate). In other words, in this scenario there
would be no illegitimate Jidai and Biocentury varieties and sales by Jidai and
14 Interview with Yashang Yang, Beijing, May 2005.
24
Biocentury would equal the total volume of seed needed to supply all farmers that
planted Bt cotton in 2001 (both legitimate and illegitimate seed).
To make all other conditions equal to that of the baseline year, the same
assumptions used to calculate the net benefits in Table 5 (column 1) are made. For
example, seed prices do not change from the baseline scenario. Farmers also are
assumed to continue to be able to save and plant their own-saved seed at the rates
observed in recent years in China (about 20%). In addition, all supply and demand
elasticities, yield gains, pesticide reductions and profits and royalties are the same as
those used in the baseline. Beside improvements to IPR, we assume that bio-safety
regulations are imperfectly enforced and seed markets function poorly (both as were
assumed in the baseline).
Our simulation results demonstrate the large benefits that would accrue to
farmers from IPR reform (Table 6, rows 1 and 2). Under the IPR reform farmer
benefits nearly double, rising by 2 billion yuan over the baseline case. The main gains
accrue from the fact that producers purchase seeds that have higher yields and require
lower levels of pesticides. Because stronger IPRs would not allow the emergence of
illegitimate or unapproved seed, farmers also would enjoy yields that were higher (the
difference between the sizes of the coefficients on the legitimate and
illegitimate/unapproved varieties in the regression analysis).
According to our analysis, IPR reform will bring significant gains to all actors
involved with Bt cotton, especially China’s producers. In particular, consumer welfare
would go up by nearly 1 billion yuan due to lower cotton prices (to 2,296 million from
1,302 million). Because seed sales rise, the profits of Jidai and Biocentury also would
increase by 350 million yuan (419-69) as would the royalties of the gene providers,
Monsanto and CAAS (rising to 280 million yuan (229+51) from 44 million (27+17)).
25
But while all other actors—consumers, seed companies and gene providers—benefit,
the gains of farmers outweigh the gains of all other actors combined. Hence, the
familiar story that farmers benefit much less than life science firms from new
biotechnology products (that would be spawned by IPR reform) is not supported by
our simulation results. IPR reforms, however, also could have longer term dynamic
effects (although we do not model them explicitly here): the changes could provide
seed companies and life science firms with more cash and greater incentives to
increase their research, which possibly could lead to even greater benefits in the
future.
4.2.2 Scenario 2: Biosafety enforcement
Following the same simulation strategy as in scenario 1, we seek to understand
the effect of increased regulation of bio-safety, holding all other assumptions the same
as in the baseline (i.e., imperfect IPRs and poorly functioning seed markets). In this
scenario we assume one change: instead of allowing unapproved Bt cotton varieties in
the fields of farmers due to weak enforcement of bio-safety regulations, new efforts to
enforce the regulations are able to suppress the use of non-approved varieties. The
impact of suppressing unapproved varieties will depend on whether Jidai and
Biocentury would have been able to supply enough seed of varieties with the
approved ones to replace those with the unapproved varieties. In scenario 2a, we
assume that the companies can not supply the seed and farmers have no alternative
but to use non-Bt cotton. This is a possible scenario because Jidai has not had
permission from the biosafety committee to sell seed in most provinces in which the
unapproved varieties are being grown; in addition, Biocentury (although having
permission to sell their seed in all provinces) was still having management problems
which actually made it difficult for them to meet their actual demand in 2001 (not to
26
mention being able to meet the higher demand of the alternative scenario). In scenario
2b, we assume that Jidai and Biocentury were able to supply farmers with enough
legitimate seed (with approved genetic material) to replace all of the illegitimate seed.
The basic objective of the scenario, whether there is zero replacement of varieties
with unapproved material with approved material (scenario 2a) or complete
replacement (scenario 2b), is to assess the effect better bio-safety management on Bt
cotton production and consumption.15
When there is no substitution between the unapproved varieties and the
approved varieties (scenario 2a), enforcing biosafety regulation can be seen to
actually impose a cost on farmers (Table 6, rows 1 and 3). In scenario 2a, the
enforcement of biosafety regulations eliminates unapproved varieties, but farmers
have to revert to conventional cotton varieties. Although the yields of unapproved
varieties were lower and levels of pesticide use higher than varieties with approved
varieties (that is, illegitimate varieties were less effective than legitimate varieties),
the yields of illegitimate varieties are higher than those of conventional, non-Bt cotton
and pesticide use levels are much lower. This means without being able to substitute
into a viable alternative, bio-safety management, in fact, can reduce producer welfare
(by 671 million yuan —2586 minus 1915—column 1). With less cotton available,
prices rise and consumers also get hurt slightly (1302-957=1345). Since in scenario
2a, the suppliers of the unapproved seed disappear and Jidai and Biocentury do not
sell any replacement seed, the profits of Monsanto and CAAS are unchanged.
15 In order to carry out the simulation, we assume that the return to the unapproved varieties from our sample for 2001, which accounted for about 20 percent of China’s Bt cotton area, either would revert to the returns of conventional cotton (which means the returns would fall by the size of the coefficient on the unapproved variable in the regression in Table 2) or it would rise by an amount based on the differences between the size of the coefficients on the Jidai/Biocentu ry legitimate varieties and the coefficient on the unapproved variet y.
27
When replacement seed from Jidai and Biocentury replaces the seed that was
contained the unapproved varieties (Scenario 2b), farmers gain about 300 million
yuan over the baseline because they produce yields that are higher and use less
pesticides than when they use approved Bt cotton seed (2884-2586—Table 6, column
1, row 1 versus row 4). The gain to consumers is 139 million yuan (1441-1302—
column 2, row 1 versus row 4). The gains to the seed and gene supply industry also
are substantial. Essentially, they triple their profits.16 It is important to note that while
these gains are not inconsequential, they are considerably smaller than the gains from
the IPR scenario. The gains in this case, however, are much smaller than the gains
from stronger IPRs, mainly because despite having 20 percent of cotton area in
unapproved varieties, the area in approved but illegitimate varieties is larger.
4.2.3 Scenario 3: changes in seed industry structure
Finally, in Scenario 3 we seek to measure the magnitude and distribution of
the benefits to the reform of China’s seed industry. To do so, we make the same
assumptions as in the baseline, except in the first case (Scenario 3a) we allow all
farmers to buy their seed from the market and do not allow them to buy seed from
non-market sources (e.g., agricultural extension stations or cotton offices). The logic
of our reform scenario is that if we eliminate local monopolies which make farmers
buy seed from local agents (that often do not sell farmers the best quality Bt cotton
seed), yields would rise and pesticide levels would fall. In the scenario this is exactly
what is simulated since cotton yields are lower and pesticide levels are higher when
farmers use seeds from the two non-market sources. In our sample 25 percent of the
fields were planted with seed from agricultural extension stations and cotton offices.
16 Seed companies gain 137 million yuan and the gene providers 87 million yuan (53 million for Monsanto plus 34 million for CAAS —columns 3 to 5, row 1 versus row 4). The seed industry’ s gain in profits (224 million yuan, 137+87) is close to the level of the producer’s gains (although farmers are still the largest single gainer).
28
If these sources of seed were shut down and replaced entirely by Jidai and Biocentury
seed, our simulation shows that farmers and consumers would gain considerably
(Table 6, columns 1 and 2, row 1 versus row 5). In fact, the producer gains are larger
than those in case of the biosafety management reform scenario (Scenario 2b).
Producer surplus increases by more than 300 million yuan (2909-2586) and consumer
surplus rises by more than 150 million yuan (1465-1302). In addition, all other
actors—seed firms and gene providers—also gain.
The importance of competition in the seed industry that is introduced by the
participation of foreign seed companies can be seen in Scenario 3b (Table 6, rows 1
and 6). If the genetic material introduced into China by Monsanto through the seeds
provided by Jidai had been kept out, China’s farmers would have only had access to
varieties that included the CAAS gene. However, since the Jidai varieties
outperformed the Biocentury varieties in terms of yields (and significantly so in a
statistical sense), farmers would have lost the benefits of using Jidai’s varieties, the
highest performing varieties in China (1965-2586—column 1, row 1 vs. row 6). While
consumers and the seed companies would also be hurt (columns 2 and 3, row 1 versus
row 6), part of the gain would be offset by the higher profits of Biocentury and CAAS
(columns 4 and 5, row 1 vs. row 6). In total, society would lose by not having access.
5. Conclusions
This paper reviews the changes that have taken place in the laws and policies
that officials use to manage China’s plant biotechnology sector—IPR legislation; bio-
safety regulations and seed industry reforms. Research institutes and seed companies
have applied for plant variety protection for hundreds of varieties that are associated
with more than 50 crops. However, according to the nation’s policy at the time that
29
this paper was written, cotton was still excluded. The use of patents to protect
biotechnology products also is growing. Moreover, China has developed a bio-safety
regulatory system which, among other activities, has approved the commercialization
of Bt cotton varieties. Large domestic and international seed firms have started to
invest in the seed industry in response to China’s new seed laws and in response to
several changes in the nation’s seed regulations.
In reviewing China’s IPRs, bio-safety regulations and seed industry reforms,
the paper has identified shortcomings in the current systems. Despite the weaknesses,
seed firms are developing new varieties at a rate greater than in the past. These
investments, however, are still extremely small compared to those of private seed
firms in many OECD countries or other developing countries (Pray and Fuglie, 2001).
Using data collected from farmers in China, our regression analysis not only
confirms what earlier studies have shown—Bt cotton uses much less pesticide and has
higher yield when compared to conventional, non-Bt cotton, it also shows that
legitimate seed from Jidai and Biocentury provides more benefits to farmers than
(illegitimate) Bt seed that was not from them. Bt cotton varieties that had been
approved by the bio-safety committee performed better than the unapproved varieties.
Furthermore, seed that come from commercial seed companies also gave better
returns than varieties from government agencies. All of this suggests that it might be
time to cease the operations of some of the antiquated government seed operations. It
also is interesting to note that Bt seed from seed companies has been shown to have
an advantage over farmer saved seed. However, even though Bt cotton varieties that
are saved by farmers require higher levels of pesticides and achieve lower yields than
seeds purchased from commercial seed companies, saved seeds still outperfom
conventional, non-Bt cotton varieties.
30
The results of our analysis have two, alternative implications. First, the
regression analysis and the policies simulations (which are based on the regressions)
support three types of policy changes because they could have benefits for farmers,
consumers, biotechnology companies and the seed companies (which use the
technology from the biotech companies). Scenario 1, which strengthens IPRs;
Scenario 2b, which strengthens enforcement and encourages the seed companies using
legitimate seed to expand; and, Scenario 3a, which forces more rapid privatization of
government seed supply, mostly lead to higher welfare for producers and consumers.
The simulation analysis, however, also shows under what conditions is it that
some policies do not lead to increased benefits for farmers—at least in the short run.
For example, if the government had prohibited unapproved varieties and had not
allowed Jidai or Biocentury to expand immediately into the areas were the
unapproved varieties were sown (scenario 2a), farmers would have been net losers
and the suppliers of technology (that is, Monsanto and CAAS) would not have gained
(and would have had no increased incentives for research). The only winners would
have been the suppliers of conventional seed and pesticides. Likewise, if the
government had not permitted international technology to be used in China’s Bt
cotton market (scenario 3b) farmers, Monsanto and Jidai would have lost revenues
while Biocentury and CAAS plus the pesticide companies would have gained.
The simulation analysis suggests that the government should press forward
with a multiple policy changes that have brought China its current successes in
biotechnology. First, gradual strengthening of IPRs is needed. In the case of China,
one such act would be found by specifically including cotton in PVP and then
enforcing PVP effectively. This would not only provide short run benefits to farmers
and companies, but it might provide incentives for more investments in research.
31
However, this paper does not provide any evidence that firms will actually do more
research. Second, continued privatization of the seed industry and opening of the seed
industry to international technology would help. Most farmers are now served by
commercial seed firms. However, according to our results, there still appears to be
gains that can be captured in areas in which the government continues to be involved
in the business of providing seed. In other words, privatization should be encouraged.
Finally, as improvements to the seed industry are being pushed, strengthening
enforcement of bio-safety regulations will also increase the benefits of technology.
Alternatively, there may be another set of policy suggestions that are implied
by the results that do not require the government to increase is efforts in
implementation and enforcement of IPRs, bio-safety management and seed industry
reform. It could be that such efforts are difficult to implement in practice and the
costs of implementation and enforcement (which are ignored in our analysis) could
offset all or part of the benefits. If the main reason that farmers do not all use the
most effective types of seeds, legitimate Jidai and legitimate Biocentury, is that they
did not know any better, it could be the most appropriate response to this study’s
findings is to publicize the results. If farmers (and local officials) know the benefits
of using legitimate Bt cotton seeds, without any additional regulation, it is possible
that they would adopt these varieties in higher numbers. Of course, if there was a
large, unexpected rise in demand (e.g., due to the revelation of new information), it is
possible that the price of seed could rise (because it would be in short supply in the
time immediate after the surge in new demand). If the price rose too much, it could
have a dampening effect on the demand by farmers. In short, maybe the best policy is
to take no action beyond publicizing the research results and then trying to facilitate
the efforts of the seed companies to meet any new demand.
32
Figure 1. Bt and non-Bt cotton area in China, 1997-2004
0
1000
2000
3000
4000
5000
6000
1997 1998 1999 2000 2001 2002 2003 2004Source: Authors' survey of provincialgovernments
1,00
0 ha Non-Bt cotton
Monsanto Bt cotton
CAAS Bt cotton
33
a “Foreign seed” contains genes that are licensed from Monsanto, a foreign life science company. b “Not foreign seed” contains genes that are licensed from a laboratory in the Chinese Academy of Agricultural Sciences (CAAS). Figure 2. Schematic Diagram of China’s Bt Cotton Seed Industry.
Bt cotton seed
Traditional, non-commercial channels
Commercial seed companies
“Foreign seed”a
Legit. seed
Illegit. seed
“Not foreign seed”b
Ag. Ext. station
Self-saved seed
Conventional non-Bt cotton seed (control group in regression
analysis)
Illegit. seed
Legit. seed
Unappr. seed
Cotton office
Seed Prod. Base
34
Table 1. Sources of Bt-cotton seed 2001 (%)
Traditional, Non-commercial Channels
Province
Commercial
Seed companies
Extension station
Cotton office
Seed production base
Self-saved seed
Henan 55 14 7 4 20 Shandong 42 4 0 6 47 Hebei 36 0 2 27 34 Anhui 51 48 0 1 0 Jiangsu 55 45 0 0 0 Total (%) 48 23 2 7 20 Source: Farmer survey by authors.
35
Table 2. Inputs and yield of Bt cotton and non-Bt cotton in 1999-2001
Observations
Pesticide use (kg/ha)
Yield (kg/ha)
Bt seed 1374 23 3292 Commericalized Seed companies: 796 21 3347
Legitimate Jidai 349 20 3553 Illegitimate Jidai 200 23 3212 Legitimate Biocentury 66 15 3299 Illegitimate Biocenturya 126 20 3062 Unapproved varieties a 55 27 3240
Traditional, non-commercial channels: 578 27 3216
Ag. extension station a 131 49 3478
Cotton office a 66 21 2543
Seed production base 68 17 3323
Self-saved 313 20 3225
Non-bt seed 342 70 2700 Note: Data from a total of 1716 plots, 1374 Bt cotton plots and 342 non Bt cotton plots Data Source: Farm household survey by authors.
36
Table 3. Regression results for pesticide use and cotton yields in China.
Pesticide Use
(kg/ha)
Cotton Yield (Exponential damage control
function ) 1 2 Intercept 1 51.106
(5.95)*** 5.994
(0.31)***
Perception of Yield loss (%): 2 0.156
(0.03)***
Average pestic ide Price (yuan/kg)
3 -0.326 (0.04)***
Household characteristics: Age (years)
4 0.009 (0.07)
-0.034 (0.06)
Education (years) 5 -1.263 (0.28)***
0.002 (0.01)
Village leader dummy 6 0.869 (2.24)
0.079 (0.04)*
Bt cotton training dummy 7 -2.28 (1.49)
0.021 (0.03)
Farm size scale (ha) 8 -14.128 (3.40)***
Conventional inputs: Labor input (Days/ha)
9 0.031 (0.04)
Fertilizer (kg/ha) 10 0.127 (0.02)***
Other inputs (yuan/ha) 11 0.159 (0.01)***
Bt Seed source: Seed company: Legitimate Jidai
12
-39.770
(2.52)***
0.257
0.05*** I llegitimate Jidai 13 -30.567
(2.49)** 0.128
0.05*** Legitimate Biocentury 14 -41.447
(3.79)*** 0.192
0.07*** Illegitimate Biocentury 15 -33.515
(2.75)*** 0.012 0.05
Unapproved varieties 16 -38.530 (3.85)***
0.082 0.07
Traditional channels: Ag Extension Station 17 -34.683
(2.71)*** -0.004 0.05
Cotton office 18 -30.188 (3.43)***
-0.013 0.06
Self-saved 19 -33.293 (2.39)***
0.153 0.05***
Seed production base 20 -34.881 (3.77)***
0.200 0.07***
Coated seed dummy 21 -0.314 (2.46)
0.009 (0.05)
Hybrid seed dummy 22 14.123 (6.40)***
0.060 (0.04)
Year dummies T2000
23 14.004 (2.39)***
0.129 (0.04)***
T2001 24 12.008 (2.69)***
0.362 (0.04)***
Damage control paramet er estimates
C (pesticide parameter)
25 0.658
(0.37)**
C1 (Bt variety parameter) 26 3.540
(0.73)***
37
Notes: The figures in the parentheses are standard errors of estimates. ***, **, * denote significance at 1%, 5% and 10%, respectively. The model inc ludes 5 dummy variables to control for specific impacts of location (4 provincial dummies) and disaster (flood vs. normal). The estimated coefficients for these dummy variables are not included for brevity. The total observations are 1716.
38
Table 4. Comparison of the benefits of Bt and conventional cotton seed in China, 1999 to 2001 (yuan/ha)
Value of yield change
Value of pesticide savings
Imputed labor cost reduction
Change in seed costs
Total benefitsb
Seed company:
Legitimate Jidai 2491 1070 660 447 3775
Illegitimate Jidai 1241 822 507 238 2333
Legitimate Biocentury 1861 1115 688 568 3096
Illegitimate Biocentury a 116 902 556 291 1283
Unapproved varieties a 795 1036 640 694 1777
Traditional channels:
Ag. extension station a 10 933 576 452 1067
Cotton office a -155 812 501 206 952
Seed production base 1900 938 579 748 2669
Self-saved 1367 896 553 -112 2927 Average impact of Bt
930
923
570
568 1856
a Coefficient in yield regression was statistically insignificant. b Benefits = value of increased yields + pesticide saving + labor cost reduction - change in seed costs
39
Table 5. Distribution of benefits from Bt cotton in China in 2001.
Sensitivity analyses
Baseline
I
II
III
IV
Assumptions: Demand Elasticity 1 1.2 0.8 1 1 Supply Elasticity 0.5 0.5 0.5 0.6 0.4 Welfare change (million yuan)a 3888 3871 3909 3913 3861 Farmers 2586 2725 2401 2440 2752 Consumers 1302 1146 1508 1473 1109 Change in profits of: Seed companies (million yuan)b 69 Biotech suppliers – Monsanto c
Biotech supplies – CAAS3 27 17
Pesticide suppliers profitsd
-96
Notes: a. Welfare changes were calculated using standard producer and consumer surplus methods described in the text. b. The benefits of the seed companies were calculated by multiplying the profit of the companies in our sample of seed companies (20 Yuan/kg for JiDai and 7.4 yuan/kg for other seed companies) times the actual quantity of Bt cotton seed sold by CAAS and Monsanto licensees. c. The benefits to the companies are the average royalties (CAAS 7.4 yuan/kg and 14 yuan per kg for Monsanto) that seed companies in our sample reported paying times the amount of legitimate seed that their affiliates sold. d. Assumes profits were 5 percent of sales decline (of Yuan 1,930 million).
40
Table 6. Simulations showing the annual benefits from reforming IPR, bio-safety management and the seed industry compared to baseline scenario in China (million yuan).
Farmer welfare
Consumer welfare
Seed
companies’ profits
Monsanto Royalties
CAAS Royalties
Baseline
2,586
1,302
69
27
17
Scenarios 1. Strengthen IPRs (PVP extended to cotton and enforced)
4,593
2,296
419
229
51
2a Biosafety enforcement: Approved Bt varieties only in approved provinces and no unapproved Bt varieties + no replacement
1,915 957 69 27 17
2b. Biosafety enforcement: Approved Bt varieties only in approved provinces and no unapproved Bt varieties + complete replacement
2,884 1,441 206 80 51
3a. Seed industry reform: Government agencies stop selling seed, their seed replaced by current mix of varieties
2,909 1,465 92 35 23
3b. Seed industry reform: No international technology – CAAS Bt varieties replace Monsanto
1,965 990 55 0 42
41
References
Alston, J., Norton G., Pardey P., 1996. Science Under Scarcity: Principles and Practice for Agricultural Research Evaluation and Priority Setting. Cornell University Press, Ithaca, NY.
Babcock, A. B., Lichtenberg E., Zilberman D., 1992. Impact of Damage Control and Quality of Output: Estimating Pest Control Effectiveness. American Journal of Agricultural Economics 72, 163-172.
Carrsco-Tauber, C., Moffitt L.J., 1992. Damage Control Econometrics : Functional Specification and Pesticide Productivity. American Journal of Agricultural Economics 72, 158-162.
Fan, S., 1999. Technological change, technical and allocative efficiency in Chinese agriculture. EPTD discussion papers 39, International Food Policy Research Institute (IFPRI).
Fang, X., Daxin C., Xu J., Xu R., Fan T., 2001. Commercial Implementation of Intellectual Property Rights of Chinese Transgenic Insect Resistant Cotton with Bt gene and Bt+CpTI genes. Journal of Agricultural Biotechnology 9 (March), 103–106
Huang, J., Rozelle S., 2003. Trade Reform, the WTO and China’s Food Economy in the Twenty-first Century. Pacific Economic Review 8 (June), 143-156.
Huang, J., Rozelle S., Pray C. E., Wang Q., 2002a. Plant Biotechnology in China. Science 295, 674-677.
Huang, J., Hu R., Rozelle S., Qiao F., Pray C. E., 2002b. Transgenic Varieties and Productivity of Smallholder Cotton Farmers in China. Australian Journal of Agricultural and Resource Economics 46, 367-387.
Huang, J., Hu R., Rozelle S., 1999. China’s Seed Industry Toward 21st Century. Problems of China's Agricultural Economics 2, 14-21.
Jia, S., 2004. Transgenic Cotton. Science Press. Beijing/New York.
Keeley, J., 2003. The Biotech Nevelopmental State? Investigating the Chinese Gene Revolution. IDS Working Paper 207.
Li, E., Yan X., 2005. Seed Companies and Industry Development in China. Seed World 2005 (2), 1-4.
Lichtenberg, E., Zilberman D., 1986. The Econometrics of Damage Control: Why Specification Matters. American Journal of Agricultural Economics 68, 261-273.
Louwaavs, N.P., Tripp R., Eaton D., Henson-Appollonio V., Hu R., Mendoza M., Muhhuku F., Pal S., Wekundah J., 2005. Impacts of Strengthened Intellectual Property Rights Regimes on the Plant Breeding Industry in Developing Countries. Report commissioned by the World Bank. Wageningen, The Netherlands.
MOA [Ministry of Agriculture], 2005. The Office of Agricultural Transgenic Biosafety Management. Collection of Laws and Regulations on Agricultural Transgenic Biosafety Management, China’s Agricultural Press, Beijing.
42
MOA [Ministry of Agriculture], 1999. The Application and Approval on Agricultural Biological Genetic Modified Organisms and Its Products Safety. Administrative Office on Agricultural Biological Genetic Engineering 4, 35-37.
Moschini, G., Harvey L., 1997. Intellectual Property Rights and the Welfare Effects of Agricultural R&D. American Journal of Agricultural Economics 79, 1229-42.
Pray, C. E., Fuglie K. O., 2001. Private Investments in Agricultural Research and International Technology Transfer in Asia. ERS Agricultural Economics Report No. 805. November (http://www.ers.usda.gov/publications/aer805/).
Pray, C., Ma D., Huang J., Qiao F., 2001. Impact of Bt Cotton in China. World Development 29, 813-825.
Pray, C., Huang J., Hu R., Rozelle R., 2002. Five Years of Bt Cotton in China: the Benefits Continue. The Plant Journal 31(4): 423-430.
Putterman, L., Chiacu A., 1994. Elasticities and Factor Weights for Agricultural Growth Accounting: A Look at the Data for China. China Economic Review 5, 191-204.
Qian, K., 1999. To share IPR of plant germplasm together with public and private sector and farmers: Analysis of social welfare and incentives to innovation. Report to the Drafting Groups of the National People’s Congress for China Seed Law, Beijing. Mimeo.
Saha, A., Shumway C.R., Havenner A., 1997. The Economics and Econometrics of Damage Control. American Journal of Agricultural Economics 79, 773-785.
The State Council. 1997. Regulations on Plant Varietal Protection of the People’s Republic of China. Decree # 213, the State Council. March 20, 1997. Beijing.
Wu, K.M., Guo Y.Y., 2005. The Evolution of Cotton Pest Management Practices in China. Annual Rev. Entomology. 50, 31-52.
Yang, P., Iles M., Yan S., Jolliffe F., 2005a. Impacts of Transgenic Bt Cotton and Integrated Pest Management Education on Smallholder Cotton Farmers. International Journal of Pest Management 51, 231-244.
Yang, P., Iles M., Yan S. Jolliffe F., 2005b. Farmers’ Knowledge, Perceptions and Practices in Transgenic Bt Cotton in Small Producer Systems in Northern China. Crop Protection. 24, 229-239.
Zhang, X., Fan, S., 1999. Estimating crop-specific production technologies in Chinese agriculture. EPTD discussion papers 50, International Food Policy Research Institute (IFPRI).
