Search the KIT-LINK database
Search the KIT-LINK database of publications on agricultural innovation.



The Fallacy of Universal Solutions in Extension

The universal application of the T&V model of agricultural extension in more than 50 countries is one of agricultural development’s best known failures. The approach worked well in places where it was originally developed, but proved inappropriate almost everywhere else. In the September 2008 edition of the LINK News Bulletin, Rasheed Sulaiman V. and Andy Hall worry that an apparently successful extension innovation piloted in India is set to suffer a similar fate.

Developed in India, ATMA — Agricultural Technology Management Agency — was part of a large World Bank-supported project aimed at strengthening and reforming the agricultural research and extension system. The central idea of the ATMA model was that it would act as a mechanism to bring together the different agencies involved in extension in a district.

However, ATMA is currently facing numerous implementation challenges, including:
  • Insufficient support
  • Mismatch with diversity of application contexts
  • Lack of local ownership
  • Capacity and institutional constraints
ATMA seems to be going the T&V way and it is only a matter of time before we will hear people talking about it as yet another model that failed due its universal promotion. Is there an alternative to promoting turnkey, “one-size fits all” approaches in countries as vast as India — where agriculture-related poverty needs urgent solutions? Is extension still a relevant concept? Tell us what you think in the comments below.

Comments (2)Add Comment

I must agree that the ATMA model is currently floundering, but not for the reasons you suggest. The ATMA model was pilot-tested through the Innovations in Technology Dissemination (ITD) component of the National Agricultural Technology Project (NATP) from 1998-2004 and also in the Uttar Pradesh Diversified Agricultural Support Project (UP-DASP). The purpose of first pilot-testing this new decentralized, participatory, market-driven approach under NATP was to first test the model, identify and work out any problems and then, if this approach was successful, to scale it up, which the GOI decided to do under the 11th 5 year plan.

The irony of your article begins with the first section on Diseases of a Dinosaur. Yes, T&V extension did focus solely on technology transfer, but the ITD component of NATP or the ATMA model specifically addressed the needs identified in your second paragraph (market information, micro-credit, diversification, and getting farmers organized and linking them to markets). Also, in your analysis, you summarized some elements describing the ATMA model, but you failed to include any of the important impact data that was summarized in the good practice paper prepared in 2005 for the World Bank by J.P. Singh, B.E. Swanson and K.M. Singh. This paper was subsequently published (2006) in a book edited by A.W. Van den Ban and R.K. Samanta on Changing Role of Agricultural Extension in Asian Nations. Since you and Dr. Hall also wrote a chapter in this book, I would assume that you were fully aware of these findings since the purpose of this article was to focus on the success or failure of the ATMA model. Since these findings were not included, let me repeat the key points that were independently documented by the Indian Institute of Management in Lucknow (IIM-Lucknow) and summarized in both the good practice paper mentioned above and in the paper prepared by B.E. Swanson, K.M. Singh and M.N. Reddy entitled: “A Decentralized, Participatory, Market-Driven Extension System: The ATMA Model in India” that was presented at the IFPRI conference on Advancing Agriculture in Developing Countries through Knowledge and Innovation held in Addis Ababa during April 7-9, 2008 (both you and Dr. Hall also presented papers at this conference):
Farm income increased 24% in project districts vs. only 5% in non-project districts from 1999 to 2003 (Tyagi and Verma 2004);
10,000 producer groups were formed (1/3 being women’s groups) under NATP (and another 17,000 groups were organized under UP-DASP)
Substantial diversification into many different high-value crops and enterprises occurred;
Area devoted to staple food crops decreased from 55% to 47% of the land area, but yields increased, so there was no decline in staple food crop production.
250 innovations were identified, scaled up and documented by IIM-Lucknow.
Yes, there were problems during implementation. In one district, they changed the project director (PD) every 6 months, so project implementation was slow and project impacts were minimal in that district. In another district, a very effective PD retired after one year and then was not replaced for more than a year, which substantially slowed down progress in that district, and so on. In spite of these different administrative problems, this approach still had a significant, overall impact, as described above. These administrative problems are to be expected when trying to implement a complex project like this one (i.e. trying to transform a very top-down, technology-driven extension system into a bottom-up system) in 28 different districts across 7 different states across India.

Moving on to the scaling-up issues, your analysis is largely correct. You cannot transform institutions, like agricultural research and extension organizations, or universities, without considerable effort. First, in attempting to scale up this model, there were very few capacity building resources available to train the extension staff, especially at the district and sub-district level, in basic implementation procedures (i.e. how to conduct a Participatory Rural Appraisal and then develop a Strategic Research and Extension Plan; how to organize and then working with many different farmer interest groups and self-help groups in each community to determine which high-value crop, livestock or other enterprise is best suited for them, given their available resources, access to markets and agronomic conditions). How can the work behavior of extension officials and the field staff be changed unless you first train them how to work and function effectively in a more innovative extension system that has been transformed 180 degrees (bottom-up vs. top-down, and market- rather than technology-driven) and that formally engages farmers in the decision process.

Unfortunately, the MOA extension leadership was unable to secure the necessary resources for these essential capacity building activities; an approach that was fully vetted under NATP and that was carefully illustrated during my presentation at the IFPRI conference in Addis Ababa in April 2008. These capacity building activities are of critical importance in bringing about institutional change and where donor support for extension is urgently needed.

In addition, a market-driven extension system needs some type of IT infrastructure to facilitate access to (and transfer of) market information, innovations, and specific solutions to specific technical and management problems as small-scale and women farmers begin taking on new high-value crops, livestock or other enterprises. Under NATP, both extension (at the district and sub-district levels, as well as the Krishi Vigyan Kendra (KVKs or farmer training centers in each district) were provided with a functional IT system, so they could access and share information within and between project districts on new high-value crops/products, innovations, markets, etc. It appears that when the ATMA model was scaled up across India, there were few if any resources available from the MOA to build this IT infrastructure in these new ATMA districts. Again, this type of investment is difficult for governments to provide out of their recurrent budgets; this is an area where donor resources can be directly invested to develop this capacity.

Your final section about Goodbye Extension: Hello Innovation was very disconcerting. Yes, the state governments are not replacing the T&V extension staff, so the actual numbers of extension workers has declined from about 100,000 approved positions to only about 60,000 extension workers who are currently on-the-job. Clearly this is a funding problem, since states have limited resources. In fact, by design under NATP, we knew that state funding for extension was a problem, so we purposely planned to phase out the Village Extension Workers (VEWs) as they retired, since most were only secondary school graduates. This model was specifically designed to engage front line extension staff at the block/mandal level (staff who had a minimum of a B.Sc. or M.Sc. degree in agriculture) who would work directly with farmer interest and self-help groups (FIGs and SHGs) and not so much with individual farmers. The goal was to reduce the cost of extension, while increasing the impact of extension on farm income and rural livelihoods. Based on the empirical evidence documented by IIM Lucknow, this approach appears to have worked very well.

In my professional opinion, it is necessary to take an institutional approach when considering how to transform an agricultural research and extension system. These systems do change over time, as the agricultural sector becomes more commercialized and as the private sector plays an increasingly important role in technology transfer. Based on NATP, there is no doubt, that extension systems can be more effective in introducing new innovations, through producer groups, into the respective farming systems of different farm households across different blocks/mandals, districts and states. Please keep in mind that IIM Lucknow identified 250 new innovations that were scaled-up under NATP alone (I don't have comparable data for UP-DASP, but I assume the numbers would be similar). However, agricultural innovations must move through some type of institutional framework. Some of this will occur through the private sector, especially to commercial farmers, but if you want to help the rural poor, then these households (with limited basic education) will need new technical and management skills/knowledge about how to produce and market these new crop, livestock, fishery or other enterprises. This was the objective of the ATMA model under NATP and the empirical evidence suggests that this approach did work.

To reiterate the need for an institutional approach, let me share with you some personal observations and experience. I was the first training officer at International Maize and Wheat Improvement Center (CIMMYT) from 1968-70 and worked closely with Dr. Norman Borlaug as he trained wheat breeders from 35 countries. He had a clear, but unwritten, institution building strategy. This CIMMYT wheat training strategy was very different from the approach being used by the International Rice Research Institute (IRRI) and conducting a comparative analysis of these two different approaches was the focus of my dissertation research in the early 1970s. The late Vern Ruttan called this research, the most successful training study ever carried out in the field of agriculture (this was back in the mid-1970s). The late Ed Schuh, who reviewed this study at a major international agricultural development conference in 1975 said that I had been extremely effective in integrating education and economic theory. This theoretical framework has been the foundation for my 40 years of institution building work in pursuit of sustainable agricultural development.

When I joined the University of Illinois faculty in 1975, I taught a post-graduate course on Education for Rural Development, which included Everett Rogers Diffusion of Innovation theory and I had at least one doctoral student who used this framework for his dissertation research. However, I largely abandoned this version of innovation theory by the early 1980s, since my focus had shifted to serving the needs of small-scale and women farmers. Diffusion of innovations worked well in North America and Europe, since larger commercial farmers were the “innovators” and “early adopters,” who in turn were followed by the medium-scale farmers (i.e. the early majority). However, it was implicitly understood that those farmers who had fewer resources (i.e. the late majority, etc.) were expected to transition out of agriculture into off-farm occupations, particularly in urban areas. However, if your goal is to increase the income and improve the livelihoods of small-scale and women farmers (when off-farm urban jobs are not readily available), then this “diffusion of innovation” strategy has significant limitations.

Instead, the approach implemented through the ATMA model, was to first identify innovative farmers who were already successfully producing and marketing different high-value crops, livestock and other products. If feasible (i.e. sufficient market demand), the strategy was to scale-up those innovations to appropriate groups of small-scale and women farmers in different communities, depending on their available resources and access to markets. A key part of this process was to conduct PRAs in each district in first identifying and then seeking the help and leadership of these innovative farmers. If these innovations could be scaled-up, then they became part of the strategic research and extension plan (SREP) for that particular district.

Seeking the cooperation of these innovative farmers is a delicate process, but most are willing to cooperate. One way of doing so was to call them “farmer professors” and let them become community leaders for these new enterprises. In short, if you want these innovators to share their knowledge and to take a leadership role for these producer groups in local communities, then give them the credit and respect they deserve. Once this process got started, then different types of farmers and farm women got organized into different groups, with KVKs providing the needed training for these groups and with both the extension system and the KVKs looking to the larger ICAR research system for technical backstopping and help in solving specific problems.

The bottom line of this approach in disseminating agricultural innovations, is that there must be a market for the products being produced and those enterprises must fit the resource base (especially, land and labor), as well as the agro-ecological conditions of small-scale and women farmers within these different communities. As you saw from the Patna slide (developed by K.M. Singh) during the IFPRI conference presentation in Addis Ababa, more than 30 different high-value crops/products or enterprises were pursued within this one district alone, to diversify risk and to fit these innovations within the different farming systems, agro-ecological zones and access to markets, (see attached paper). As demonstrated in this one district and the different mix of crops, livestock and other enterprises that emerged in the other 27 NATP districts, many of these innovations are very location specific.

Returning to the extension or innovation issue, there must be some type of institutional capacity to introduce, educate, organize and provide technical and management support to the small-scale and women farmers who want to pursue new opportunities (for them, innovations). The research system, including KVKs at the district level, does not have sufficient capacity to train and serve all of the small-farm households in India (and other countries), so who can deliver these services? The commercial farm sector routinely develops a working relationship with input supply dealers and this linkage will strengthen over time. However, there are still 220 million rural poor people in India (<$1.25/day) and most of them will not be helped by the private sector. NGOs can assist, but most of their field staff members lack the necessary technical and management expertise needed to introduce appropriate innovations within different communities. Also, they need continuing resources to operate; therefore, once public or donor resources are phased out, then these NGOs disappear or move on to other projects or tasks.
Addressing the problems of the rural poor, especially in helping them increase household income and to improve their livelihoods, will take considerable time and resources. As empirically documented through the NATP, it is possible to make significant progress in achieving these goals, but to do will require specific types of investments to transform the existing extension system (e.g. capacity building, IT, policy support, etc.) into one that is more decentralized, participatory and market-driven. Without such investments, you are correct, the ATMA model will fail.

You indicated in your article that the ATMA model was struggling to cope with the wide diversity of Indian agriculture. In fact, the pilot project states and districts were specifically selected to deal with this diversity issue. Let me explain why: T&V Extension worked well in increasing the staple food crop production in irrigated areas with the introduction of high yielding wheat and rice varieties. However, this approach was not very effective in rainfed areas, since there was little in terms of new or appropriate technologies to disseminate to farmers in these areas. Therefore, we wanted to see if this ATMA model would work equally well in very diverse states and districts. If you carefully examine the diversity of crop innovations that were introduced across the different states and districts, you will see that this approach did work equally well across most districts in India, which is why IIM-Lucknow identified 250 innovations. Many of these innovations could be scaled up across much of India as the number of middle class consumers (and their incomes) continue to increase. In short, as the markets for different high-value crops, livestock, aquaculture and other products (e.g. silk) continues to increase, there will be many new opportunities for small-scale and women farmers to increase their household incomes.

One final point that I would like to mention in closing. Before helping to develop this ATMA model for pilot-testing in India, I prepared another World Bank project in China, starting in 1990 (Agricultural Support Services Project), which was also designed to strengthen crop and livestock extension systems. The crops extension part of this project, which I prepared, used a very similar approach (decentralized, market-driven extension system) in about 55 counties (similar to districts) and over 700 townships (similar to blocks/mandals), but the actual structure, investments, etc. needed to implement this model was very different, given conditions in China. Unfortunately, impact data was not collected on this project; therefore, it is not possible to carry out a comparative analysis, but it is my professional opinion that the final impacts of this ASSP were more significant.

If you look at the changes in the production of high-value crop and livestock products in China, no other country in the world has ever experienced such major increases in the production of these products. Under the ASSP, 12,000 producer groups were organized and linked to markets, but country-wide over 100,000 groups were organized (by 2001). However, conditions in China during the 1990s were quite different. For example, many of the younger generation of farmers are better educated (many have 9 years of vocational agricultural education) and most are very entrepreneurial. Also, during this period, many of their wives worked in township industries and their salaries were frequently reinvested in the farm itself. First, a 2-wheel tractor was purchased, since the average farmer could only work about 125 days/year on their one-half acre farm and they used these small tractors to generate additional, off-farm income hauling of building materials. Next, these off-farm financial resources were used to diversify into high-value crops and livestock enterprises, using more fertilizers and other production inputs (e.g. plastic greenhouses).

In short, the structure of the agricultural sector was very different in China, but the same basic principles were used in helping to reorganize the extension system, which worked equally well in China. Also, since we designed a very different means of financing the extension system at the county and township levels (given the lack of input supply firms, this approach could only be done in China), I have been told that China still has about 370,000 crop extension workers and a comparable number in livestock extension. Look at these productivity and income levels and think about the impact that this extension system is having in comparison with India, which now has only about 60,000 extension workers for a comparable number of rural people (about 780 million in 2005) to be served.

I hope that this message clarifies some of the points covered in your paper about the relationship between extension institutions and innovations being up-scaled as carried out through the ATMA model under both NATP and UP-DASP. In my professional opinion, India will not be able to address the needs of small-scale and women farmers without strengthening its extension system so that small-scale and women farmers can learn how to successfully carry out these labor intensive, high-value crop, livestock and other enterprises that reflect the growing consumer demand in India and in worldwide export markets. As demonstrated by the ATMA model that was pilot-tested under NATP, by getting farmers organized into groups, providing them with the necessary training and backstopping, and then linking these groups to markets is an effective extension approach of increasing farm household income and improving livelihoods. In short, if you say goodbye extension, then you will also say goodbye to innovations for small-scale farm households.

Thanks for giving me this opportunity to comment on this article.

Burton E. Swanson
Professor Emeritus of Rural Development
University of Illinois at Urbana-Champaign
a guest
January 06, 2009

The “Fallacy” of universal solutions implies that sui generis solutions could be the answer. And I firmly believe so. But why there are no sui generis solutions? Simply because solutions that do not bring money- whatever be the strings attached- along with it will not find the daylight. So who is having the money ? It is the WB or the so called international experts who assume that they are the god-sent solution providers. And there is no wonder that their solutions end up in the fallacy you are deriding.

ATMA or TV are no exceptions. So too will be the fate of solutions you dream. Of course innovation systems provide a new tool to diagnose the disease. But stethoscope is not the remedy. (Sometimes it also works! Like homeopathy-yes the poor needs a patient listening and chronicling of problems on a diachronic fashion).

It is easy to yell iconoclastic cries. Yes it is better to have new institutional structures which could act as learning organizations. But it is difficult too. Instead of seriously caring to reinvent ( or even to provide a decent burial to !) the existing structures ( for eg, departments of agriculture) we go for new solutions( mostly consultant given-the one who borrows your watch to tell you the time” sort of fellows) that after sometime prove no different. This is the fate of any institution. Instead of characterizing them as incurables why don’t you use the so called advance knowledge we have in social sciences or policy epistemology to prescribe remedies to the ills of the existing organizations ( mind you the department of agriculture is supposed to have qualified extension professionals!)or to the ills of the linking mechanisms? “A number of diagnostic studies are available. Number of solutions have been suggested. But no one seems to have heeded” this could be the typical refrain. “We have the remedy, but no takers” . What does it mean? It means a lot. And we need to restart from this point. With the humility to allow self criticism.

As you rightly imply the best thing an extension system(or extension professionals – but are there any one ?) should understand is that there are no universal solutions. It underlines the need for contextual flexibility. How best you can provide a space for it in a given institutional configuration ( in a given period of time) is the question we need to address. It is here that you can apply your tool of innovations system approach. Yes, as a better diagnostic tool. Let the solutions be an emergent property.

If such solutions demand new structures no one can resist them. But someone somewhere has to do it. But that needs tremendous amount of grounded energy which “fly by night” extension thinkers can not afford. Perhaps we need to invoke the vintage Gandhi “ let every farmer be a scientist and every village a research laboratory”. Any takers?

Thank you very much for the excellent provocation. I look forward to the next issue.

Yours sincerely

(Originally published at
a guest
January 06, 2009

Write comment


Latest entries


New globalalliances: the end Of development assistance?
Designed by vSplash