One of the most persistent lessons in global development is that every program looks good as a pilot run by a really dedicated grad student. 

There are dozens of examples of promising interventions that collapsed when they tried to scale. Few are more dramatic than the Indian Community Development Program. The year is 1948. The protagonist is Albert Mayer, an American architect and urban planner. The program is an ambitious pilot to reform and modernize the rural villages of Etawah district in Uttar Pradesh. Mayer didn’t want to rely on top-down directives, as earlier British and Indian administrators had done.  Instead, his program would use a system of local advisors and village councils to determine what villagers really needed, and then do that. 

At first, it worked. New schools and new roads sprouted across Etawah. Literacy rose. Crop yields exploded. In 1952, the ICDP became a national initiative; at its height, it would reach every village in India. But the expansion failed to reproduce the success of the initial pilot. When it finally shut down in 1966, the program was — by the admission of almost everyone involved — an embarrassing failure. 

Historical accounts of community development in India tend to converge on a diagnosis for why it failed. Villages in rural India were and are deeply unequal places. Instead of serving the rural poor, the program’s resources were siphoned off to benefit local elites. Everyone else lost confidence that community development could actually deliver results, and as a result, it didn’t. 

All of this is true — yet it’s also only half of the story. The Etawah pilot was just as vulnerable to elite capture as the national program — but the Etawah pilot worked. Looking at why one succeeded and the other failed can tell us something about the structural reasons why flawed ideas can still succeed when implemented at small scales, and good ones can fall apart when they get too big. 

Albert Mayer began his career in New York City, but he was always attracted to the idea of village life. His projects — mostly large-scale apartments and housing developments in New York and across the country — married modernist urban high-rises with shared spaces meant to create a sense of community. Mayer first came to India as an Army engineer during World War II. He quickly fell in love with the country and its culture — which is to say, he immediately started working on designs for model villages. 

At first, these were just physical plans. But after doing more homework — his sources included American rural sociologists like Charles Loomis and Douglas Ensminger, missionary anthropologists like William and Charlotte Wiser, and the the village-centric philosophy of Gandhi — he came to believe that improvements to physical infrastructure would be futile without holistic social and economic reforms. “There were villages that some years before had been given brick roads which had gone to pieces for lack of maintenance, public latrines that had never been used … The single lobe, the outside gift, the unprepared general atmosphere — these are a fatal combination.” ¹ Instead of a “single-lobe approach,” Mayer wanted to strengthen village communities, giving them the tools to address the inter-linked causes of rural poverty for themselves. 

In 1945, Mayer met with Jawaharlal Nehru to discuss his proposal. Nehru was impressed.  In 1947, soon after he became Prime Minister of a newly-independent India, he made Mayer the Planning Advisor to the government of Uttar Pradesh. There, Mayer drew up the initial plans for the new city of Chandigarh, which was to feature inward-looking Barcelona-style superblocks laced together with green space. (Mayer withdrew from the project after his architect partner was killed in a plane crash, leaving the work to be completed instead by le Corbusier). After leaving the Chandigarh project, he turned his attention back to the countryside. In 1948, Mayer was given the go-ahead to launch his pilot program across 64 villages in Etawah, a district of some 500,000 people on the Western edge of the province. 

The program’s backers — who included Nehru, the Ford Foundation, and U.S. Ambassador Chester Bowles — were concerned first and foremost with agricultural productivity. India was plagued by periodic famines (and hunger, of course, could only lead to communism). So Mayer’s team distributed chemical manure and improved strains of wheat and peas. They introduced (relatively) modern farming equipment like mechanized threshers and the moldboard plow, and demonstrated new methods of livestock husbandry. They organized the construction of roads, houses, and sewers. They ran vaccination campaigns and taught adult literacy classes at their new community centers and libraries. 

Judging purely the numbers — though Mayer would caution as not to — the project was a success. Over the course of the pilot, the Etawah region saw an 165% increase in wheat yields. Its farmers produced 121% more potatoes and 43% more peas. ² The introduction of the improved wheat strain Punjab 591 alone brought in a return of ten times the program’s recurring annual cost. ³ In 1948, Etawah had almost no local industry, but by 1953, there were 520 brick kilns, run by cooperatives employing 42,000 workers. ⁴ Its villages were connected by brick roads and boasted new brick schoolhouses and sanitary wells. 

But developing the region was only part of Mayer’s real goal: developing the villagers themselves. He later wrote “The aim at Etawah was to make maximum progress both in improving physical productivity and in developing the people's own capacities and initiative — to give the people better land and better implements, and at the same time to alert them for the future. The pilot project aimed to stimulate the villagers not simply so that they would follow the specific better methods that it could now help them to find, but also so that they would themselves continue to seek and find a better life.” ⁵

This ideal of “village uplift” was not original: it drew heavily from the “Better Living” movement implemented by British colonial officers in Punjab in the 1920s and 30s. The most famous of these efforts was Frank Lugard Brayne’s experiment in the drought-prone district of Gurgaon. Brayne applied  “propaganda, propaganda, and more propaganda” (in the words of one colleague) in order to shake villagers out of their apathy and convince them to improve their lives. ⁶ They were instructed to be hygienic, diligent, industrious, and, if at all possible, British. (Philip Mason, a historian of the Anglo-Indian civil service, wrote that Brayne “thought he could persuade Punjab peasant women to behave like the wives of Victorian evangelical parsons in Norfolk.” ⁷ ) Where propaganda failed, Brayne used his authority as deputy commissioner of the district to requisition unpaid labor repairing roads or digging manure pits. Like many other would-be rural reformers in India before and after him, he also tried to distribute better plows, better seeds, and better livestock. None of it took. Eight years into the experiment, the Gurgaon district board was bankrupt. It had produced no discernable improvements to anything at all.

Mayer was very familiar with the results of the Gurgaon experiment. The lesson he took from it was that his development program could only succeed by actively engaging the effort of its supposed beneficiaries. Indian villagers had suffered under centuries of rural poverty and stagnation as well as the legacy of a rigid, sclerotic colonial bureaucracy. The villagers he met had no faith in the new tools and techniques being pushed on them by the government. To Mayer, this all seemed perfectly rational. What good had these modern advances ever done for them? Instead of imposing themselves, the would-be modernizer should be humble: their job was to find out what the villagers needed, and to show them, slowly and patiently, how they could get it. 

The key role in this process was that of the village level worker — in Hindi, the gram sevak. His primary job (all of the gram sevaks were men) was to talk to the villagers and elicit their “felt needs” — not what they had been taught to want, but what was genuinely most necessary. (It was assumed that the residents could agree on this point.) These needs would determine the priorities for the program workers in that village — access to roads, better seeds, clean water, or something else altogether. 

Above the gram sevaks, group level workers were responsible for coordinating development blocks of 100 villages. These group level workers in turn reported to a district planning officer, responsible for managing supply lines and interfacing with the central administration. Rounding out the administration were technicians and specialists — adult literacy teachers, agronomists, engineers, and even rural sociologists.

Two key principles animated this hierarchy. The first was the “dirty-hands” method. “Our own attitudes as government workers are most crucial,” Mayer wrote. “We must work with the people, not tell them; we must demonstrate by doing with our own hands in their own villages and fields.” ⁸ Gram sevaks ate, dressed, and spoke like the villagers they lived with. They participated in construction projects and labored in the fields. As much as possible, they were instructed to work within the bounds of village culture, adapting festivals, legends, and traditional dramatic performances in order to convince locals to adopt new practices and tools — no Victorian parson’s wives here. On one occasion, a gram sevak trained in Hindu philosophy out-debated a local Brahmin on the ethics of green manure; elsewhere, a feral antelope known as a nehil gae (blue cow) was formally reclassified as a nehil goa (blue horse), freeing villagers to drive out the bothersome animals before they destroyed the crops. ⁹

Cultural adaptation was valued, but it only went so far. By far the most important persuasive tool available to the gram sevak was visible evidence that their methods worked. Here, again, Mayer set the tone: 

“The farmer in every country, whether it be Etawah district of India, an out-of-the-way village in the United States, France, Mexico, or elsewhere, is conservative. He is conservative because he is smart … If [farmers] did not demand proof, if they did not want to see results, if they were not careful in exchanging something that they know will work for something they know nothing about, then the world would have many more famines than it has experienced to date.” 

In Etawah, this proof came in the form of a carefully staged sequence of demonstrations. A new tool or technique would first be adopted by a few volunteer farmers. Other farmers in the area would be brought to see the results. They were invited to talk with the successful cultivator about his methods. Ideally, this would take place over multiple casual visits — enough to convince skeptical neighbors that the new technique, and not some other advantage, was responsible for the results.. Finally, once all the farmers in a village were on board, the village itself could become a model community for the surrounding area. 

Mayer’s second principle was “inner democratization.” Of all his critiques of the Indian general bureaucracy — and there were many — the strongest and most persistent is that its rigid hierarchy made it impossible for information to flow upwards. The central administrators isolated themselves in “watertight compartments,” refusing any advice or input from field workers and low-level colleagues. Mayer wanted to create an environment where everyone from the district planning officer to the gram sevaks felt equally free to speak their mind — and was equally likely to be listened to. 

The central institution to enable this flow of information was the staff meeting for village level workers. At these meetings, the district and group-level officers would meet with the gram sevaks to discuss high-level plans and hear their reports on local conditions. The meeting would be chaired by one of the gram sevaks, who were to do most of the actual talking. All this served two functions: in the first place, of course, it helped higher-level administrators understand what was really going on in the villages they supervised. More importantly, though, it demonstrated to the gram sevaks that they were equal participants in the program. Morale — more specifically, “devotion, enthusiasm, initiative, the strong sense of status, validity, and recognition on the part of each and every worker of whatever rank” ¹⁰ — was critical to Mayer’s conception of the program’s success. Without it, Etawah would go the way of Gurgaon: skeptical, harassed villagers pressed into service by a delusional administration. 

The gram sevaks, as the link between the village and the upper levels of the hierarchy, played an essential role — but also a very specific one. Information about village needs or problems flowed up the chain; technical expertise flowed down. We have now gotten about as far as we can in a discussion of programs designed to counteract the natural tendency of bureaucrats to blind themselves to the specific concerns and survival strategies of peasants without briefly mentioning James C. Scott. To use his terminology: Mayer was not looking to the villagers of Etawah for metis, or practical, embodied, tacit knowledge derived from long experience. He believed in techne: chemical fertilizer, systematic cattle breeding, and Punjab 591 wheat. The villagers might know best what they needed, but they had to be taught how to get it. 

An illustrative example comes from a letter by Richard Morse, who observed the Etawah project as a fellow with the Institute of Current World Affairs. ¹¹ Morse accompanied Rudra Datt Singh, a sociologist attached to the pilot program, on a tour of the village of Muktapur. Everyone the two men spoke to, from untouchable leather-workers and street-sweepers to the high-caste thakurs and brahmins, agreed on the village’s biggest problem: intolerable flooding during the rainy season. Children lived in constant danger. Houses were regularly swept away. The village of 750 people had no success appealing to the government for help, and previous attempts to build berms in one section of the village only made the floods worse in others. The problem was only resolved — in Morse’s telling — when Singh suggested hiring a surveyor to find the best drainage outlet, after which the villagers could dig the necessary ditches themselves. 

At this, the village council objected: couldn’t they bring in digging machines? Singh explained that such machines were not available — the government could not afford them, and would not be able to until food production increased. “So they acknowledged, though somewhat tentatively at first, that the digging and moving of earth must be done by their own labor. Rudra promised that a surveyor would soon be sent, to accomplish the preliminaries.” ¹² All this is quite typical of the Etawah method: the project worker, through patient conversations with villagers from all walks of life, successfully identified a felt need. He was able to propose and implement a practical solution. And the villagers themselves had very little real influence on what that solution would be. 

This imbalance would matter a great deal over the next few years, when the scope of the project dramatically expanded. 

By 1952, it was clear that the Etawah pilot was a success. The next step would be to adopt its methods into a nationwide community development program. This program would be housed in the new Community Projects Administration under the leadership of Surendra Kumar Dey, an American-trained engineer turned civil servant. Instead of operating independently, like the other development ministries, the CPA was supposed to serve all of them at once. It had a small amount of funding to hire and train staff, but the resources it would be providing — seeds, agricultural equipment, or whatever else — would come from other ministries or state governments. As much as possible, labor would be supplied by the communities themselves. 

There was enormous pressure to expand: every parliamentarian wanted a community development project in their district. The Ford Foundation and the U.S. government, both of which contributed funding, wanted a solution to India’s recurrent famines that looked more attractive to its citizens than communism. Previous efforts to increase food production, like the “Grow More Food” campaign of the late 1940s, had produced disappointing results. It was crushingly apparent that the central government needed to do something. Community development was something — and it was something different from the failed, centrally-managed programs. 

Mayer had always intended to grow the program slowly. His model for national expansion was similar to the demonstration system within Etawah: achieve thorough, complete, and visible success in one district before radiating out to the impressed and curious neighbors. Politically, though, this was a non-starter. Chester Bowles, the U.S. Ambassador to India, argued that doing “a complete and perfect job to produce not only a model agricultural area along the lines of the Etawah district but to produce a model township along the lines of Nilokheri ... would result in some showplaces but the impact on India's 350 million people would be minimal.” ¹³

Everyone (except Mayer) agreed: fast growth was necessary. The first wave of CPA projects covered 16,500 villages. By 1956, its services nominally extended to one fifth of the population of India. Inevitably, as the program exploded in size, its resources were spread more thinly. The CPA’s second five year plan doubled its budget, while more than tripling the number of administrative blocks the organization was supposed to cover. 

In 1957, the Indian government appointed a committee under parliamentarian Balwantrai Mehta to evaluate the progress of the community development program. Their report was brutal: the program had failed to invoke any kind of popular enthusiasm for self-improvement. Even worse, from the central government’s perspective, it hadn’t led to any noticeable improvements in agricultural productivity. The source of these problems, they concluded, was that the already piecemeal CPA was too centralized. More power should be devolved to village councils, or panchayats. ¹⁴ Their recommendations were implemented in 1958, laying the foundation for the system of village self-government which still exists in India today. And, instead of slowing down, the community development program kept growing. By the beginning of the 1960s, it encompassed 446,000 villages containing 253.2 million people. By 1964, it had expanded to every village in India. ¹⁵

The transition to panchayat governance didn’t improve results. Nehru continued to champion the program, but it became widely unpopular across the rest of the government. While its coverage expanded, the program was hollowed out from the inside. The odd structure of the CPA didn’t help either: because funding for community development programs came from states or other ministries, the administration was constantly caught up in bureaucratic turf wars. Dey had a particularly contentious relationship with Chidambaram Subramaniam, then the Minister for Food and Agriculture. Sumbramaniam was a great centralizer who mistrusted Dey and the CPA’s bottom-up approach. Throughout the early 60s, Subramaniam directed scarce agricultural inputs like chemical fertilizers and high-yield seeds to his own preferred programs (ultimately, to much greater success — today, he is primarily known as the political father of India’s Green Revolution). 

When Nehru died in 1964, the CPA lost its last and most powerful defender. By 1966, his successors had dismantled it completely. 

Let’s turn to the typical account of what went wrong. While Mayer and Dey tried to encourage democratic participation, they failed to account for entrenched issues of caste, class, and gender. Gram sevaks were supposed to be “natural leaders” in their communities — that is, high-caste, land-owning men. More than half of gram sevaks were Brahmins. Only 1% were actually agricultural laborers. ¹⁶ In practice, “voluntary participation” in local construction projects looked more like forced labor, mostly by poor, low-caste people. S.C. Dube, an Indian observer, writing in 1957, found that the panchayats “could manage to collect a reasonable number of people for this work but a great majority of participants had very little understanding of its purpose … a very considerable proportion of the participants viewed it as a form of coercion and submitted to it in a spirit of obedience to government directives.” ¹⁷

As historial Daniel Immerwahr puts it: “Community development’s bias toward the rural elite was not just a problem for the distribution of its benefits. It was a problem for the accomplishment of any of its goals.” ¹⁸ The majority of farmers in India did not own land. If they managed to produce any agricultural surplus, they would simply have to pay it back to their landlords in rent. This left them with very little incentive to adopt modern, higher-yield techniques.

This matches the results of more modern community development programs. In a review of the literature from 2013, Ghazala Mansuri and Vijayendra Rao found that people who participate in community development decision-making are generally wealthier, more politically connected, and of higher social status than those who don’t, and that the (relatively) well-off tend to benefit more than the very poor. This elite capture is more severe in communities which are already poor, largely illiterate, and have extreme existing race, gender, or caste inequalities — all true of village India. ¹⁹

As we’ve seen, this explanation is correct, but incomplete. Elites tend to benefit disproportionately from community development — but these programs typically do succeed at delivering tangible public goods. ²⁰ It would be easy to say that the program grew too quickly, but this answer also isn’t especially useful. India in the 1950s had a national-scale food crisis that needed a national-scale solution. In hindsight, we know it was possible to convince large swaths of the Indian farming population to adopt modern industrial agriculture because it happened in the late 1960s. So why didn’t community development do the job? What did Etawah have that the CPA couldn’t replicate? 

In the first place: Albert Mayer was a perfectionist — systematic, detail-oriented, and obsessed with measurement and self-correction. He was constantly traveling from village to village, sleeping in huts, observing and modulating and adjusting as he went. His colleagues noticed, not always approvingly. Horace Holmes, the American consultant in extension at Etawah, wrote that “Mayer brought something to this thing [the pilot project] in terms of methodical planning and of target making and of follow-up and a doggedness that, at times, was disconcerting, but certainly was an important input. He also had the nerve that it took to break into anybody's office. He was also a good salesman, as shown by the fact he got the thing set up, and was able to do it at the highest governmental level.” ²¹ (This last point was also critical: Mayer had Nehru’s personal support and didn’t have to worry about fending off other bureaucrats.) V.C. Sharma, an Indian administrator attached to the project, put it more bluntly: “The whole thing was so much under the general supervision of Albert Mayer that the rest of us didn't count.”

Mayer was especially concerned with the composition of the village meeting and the staff meeting. These were the key institutions of the project, which he believed depended on the free flow of information from villagers to gram sevaks, and from gram sevaks to upper-level administrators. And he knew how it was most likely to go wrong. He instructed his staff that 

“We must be careful to attract people from all levels, and if we do not find them attending, to ask them to attend. At such meetings we should encourage people to speak their minds. Too often, a very few people do all the talking, and we hear little of what others are really thinking and feeling. We should be particularly watchful for people who are saying nothing, and should try to bring out their opinions. We must also be watchful that they are not quietly penalized later for having expressed themselves … One of our great weaknesses is failure to observe who is not talking, to be satisfied when our friends talk and say what we like to hear.” 

There is a great deal missing from this memo — for example, any explicit recognition of who might not be talking, or why. Mayer was not inclined to think in terms of structural inequality —  he was more likely to label the different caste interests in a village as “factions” who needed to be brought into harmony. The diagnosis was wrong, but the procedural cure was basically correct. It’s common for modern community development programs to have participation quotas for women, young people, or oppressed castes or ethnic minorities. The mandate to include “people from all levels” (or really, men from all levels) is a more limited, less systematized gesture in the same direction — but informal enough that it completely vanished when the people in charge didn’t care to enforce it. 

This concern for procedural equality mattered even more within the bureaucracy. It’s hard to say how much gram sevaks at Etawah actually listened to the concerns of low-caste, landless farmers (we don’t have their written retrospectives). But contemporary reports suggest that gram sevaks, at the lowest tier of the project’s administration, really did participate in high-level decisions. S.K. Jain, an academic, observed that “village workers constantly bring field experience to headquarters and are also able to participate in programme planning.” Harish Chandra Seth, a development officer, later recalled that “There was great flexibility during this period. Decisions were democratic and everyone felt he had a stake in the project. This atmosphere pervaded to the lowest worker.” 

Of course, we shouldn’t conclude too much from these individual observations — but we can contrast them to the situation when the program expanded under the CPA. Here, we see a familiar pattern: as planning became more centralized, village level workers had less authority to make changes based on local conditions. Since the programs run through the CPA were funded by the states or the central ministries, there was increasing pressure to achieve results that looked good on paper. The introduction of the panchayat system nominally made community development more democratic, but in reality, communities had less and less influence on how the program allocated its resources. Block and district level officers were judged by how many wells were dug, schools built, or miles of road laid — whether or not the villages needed them. 

The CPA’s own Project Evaluation Organization reported that “the performance of block officers would be judged primarily by the amount of money they were able to spend.” Another study, in Mysore, found that “While 92 percent of VLWs [village level workers, or gram sevaks] expect BDOs [block development officers] to guide them, only 45 percent of the BDOs felt this was their role.” Instead, they gave orders and expected to be obeyed. In turn, the gram sevaks started paying more attention to metrics like number of demonstrations conducted than the needs of their clients. The practices which had proved effective Etawah were carried out by rote. Gerald Sussman, an American academic, interviewed several gram sevaks in 1965 — the dying days of the CPA. His observations were quite different from Seth’s or Jain’s: “I asked them, ‘If your program was to provide demonstrations to the farmers on the use of chemical fertilizers, would you still do it if you found the farmers already using them?’ To my surprise, all three replied ‘yes.’”

At all levels, the CPA struggled to attract dedicated, effective staff. This was an area where Mayer and his colleagues were extremely selective. Their criteria included good grades, but they were more interested in intangibles — confidence, dedication, an attitude of service and a willingness to work with their hands. U.L. Goswami, Dey’s deputy at the CPA, was impressed by the zeal of the Indian Administrative Service workers drawn to Etawah: “They felt they were really doing something.” It was simply impossible to maintain the same standards as the program ballooned. But the CPA also failed to retain the staff they did have. These early administrators had developed a reputation for excellence — which meant they were poached by the central ministries. Dey complained about this “centrifugal pull,” which made it impossible for him to keep good people. This was exacerbated by a structural flaw in the program: a lack of options for career development and long-term support for staff. Mayer may have felt that a job done well was its own reward, but the administrative elite didn’t always agree. 

All this left the CPA with inadequate staff, limited resources, and an unworkably broad mandate. They were supposed to solve rural poverty, fix famine, and reform ancient and entrenched cultural practices, all while fighting the rest of the government. In contrast, the Etawah pilot had two main advantages. The first, as we’ve seen, is that a combination of Mayer’s leadership, Nehru’s support, and the more receptive political environment of the late 1940s protected them from interference. The second was a functional — and highly specific — program culture. On paper, the pilot program was still trying to do too many things at once. It was still subject to elite capture. If it produced results, it was because everyone involved in actually implementing the program believed in its goals and methods and pushed themselves to work against the natural incentives to please rich landowners or produce results that looked good on paper. 

The practical lesson of Etawah is that serious structural and ideological flaws can be overcome by competent, dedicated administrators. If you’ve got a promising pilot project and want it to expand, pay attention to the intangibles — and temper your expectations. This was Albert Mayer’s conclusion when he looked back at the collapse of the community development program: “I struggled to make clear what were the components of success, what are the appropriate limits to rate of expansion; but it looked so easy to apply the formulas and use the multiplication table. And of course the pressures were terrific — everywhere the people and their representatives wanted such projects ... So perhaps in Nehru's place, I might have done the same.”