New research offers real hope that India can elevate the worst of the nation's persisting rural poverty in a way that is sustainable, affordable, and politically feasible. Poverty in India is overwhelmingly rural. The worst-off segment of the rural population is the completely landless, conservatively estimated at 11 per cent of rural families or 65 million people. Such landlessness is a far better indicator of rural poverty than either caste or illiteracy. Yet in most Indian states there is little or no political will to address the problem directly through traditional land reform. Indeed, with a few notable exceptions (including West Bengal and Kerala), traditional land reform in India has met strenuous opposition and has not been adequately implemented. But new research just completed in India, coupled with fresh review of experience in a series of other countries, suggests a highly practical alternative to the traditional land reform approach. The traditional approach has rested on the assumption that the landless will need at least 1 to 3 acres of land to serve as a new, primary (if not sole) source of livelihood. An average of 1 to 3 acres for anything close to all of India's landless households requires far more land than the states or the Central government can afford to purchase at market prices. Hence the 'traditional' approach typically involves attempted near-confiscation of the needed land, resulting in strong opposition and resourceful evasion by the 'above-ceiling' landowners and virtually no actual implementation. The alternative approach arises out of an important realization that if the completely landless received ownership of a small homestead plot - one sufficiently large to have space for a garden, trees, and animals adjacent to the house - they could supplement their current income and nutrition from activities and resources on the homestead plot. They will then have a greatly enhanced opportunity to lift themselves out of poverty. Such a homestead-cum-garden plot needs to be only a fraction of an acre. Distributing such plots to all of the 13 million completely landless rural households would require less than one-third of one per cent of India's arable land, transforming all existing assumptions about the affordability and feasibility of 'land reform'. Recent research conducted in West Bengal and Karnataka, together with a review of experience in other countries, indicates what could be accomplished by a programme to distribute homestead-cum-garden plots and at what modest cost. The research included interviews of 246 rural households that owned homestead plots but little or no agricultural land. The results show that even very small plots typically allowed an agricultural labourer family to achieve at least some significant non-housing benefits such as increased income, status, food, consumption, and access to credit. An analysis of homestead plot size shows that the economic uses and benefits increase substantially as the plot size expands beyond that needed for the house itself. Specifically, those who received homestead plots from 1,800 to 6,500 square feet in size (about 4 to 15 cents or 0.04 to 0.15 acre) were realizing substantial benefits. They had planted an average of seven trees yielding food or income (and nearly all had planted vegetables or other crops). Two, they had kept an average of nearly three large animals on the plot, generally cows or buffaloes. Three, they cited an average of roughly two positive impacts on their lives, most prominently increased status in the village, increased income, and increased access to credit. There was also a resulting increase in the family's capital assets. In that same-size group, these poor families had substantially increased the market value of their homestead plots by planting trees and making other improvements (including cattle sheds, storage, fencing, wells and so on). These plots of 0.04 to 0.15 acre in size had given a previously landless household not only income, status and credit access, but also a place to put 'sweat equity'and other resources that substantially increased the value of household assets. The research results from West Bengal and Karnataka parallel the findings on benefits from similar small homestead plots from a variety of settings around the world, ranging from Indonesia to the Caribbean and the former Soviet Union. The research also points to the potential of transforming existing rural housing or house site schemes into programmes that provide real economic and social opportunity for landless households - the poorest of India's poor. The existing housing schemes, both state and Central, focus on the house alone and not the house sites. Therein lies a significant limitation. The research pointed to three problems with the existing housing schemes that substantially limit their potential: One, the plots provided are often too small, as little as one cent (435 sq ft), and almost always less than 3 cents (1,307 sq ft). Two, existing schemes either require the beneficiary to already have land (and the government then provides funding for a house), or they distribute homestead plots from the existing available government land. But those who already have land are not the poorest of the rural poor, and available government land - if it exists at all - is likely to be common land needed for other purposes, or land of very poor quality or location, unsuitable for homestead-and-garden use. Nearly all of the substantial resources spent on rural housing schemes go into house construction for the rural poor, not into land acquisition. Yet, experience in India and elsewhere shows that once a landless household owns a plot (even a very small plot) of land, they find the means to construct their own housing, starting with the very simple and improving it over time. This last point is especially important on the question of costs and benefits. In recent years, the Central government alone has been spending about Rs 1,700 crore per year on rural housing schemes, nearly all of it earmarked for housing construction costs on very small 1 to 3 cent plots. Recent research in India and experience elsewhere suggests that the government could be providing far more benefits and to larger numbers of families if more of these resources were devoted to acquiring larger house sites and less on the house itself. In fact, the Central government could provide 10 cent plots (1/10 acre, or about 45x100 ft) to all 13 million landless households in India over 10 years for a fraction of what it currently spends on rural housing. The average market price for enough dryland to provide an adequate homestead-cum-garden plot of 10 cents is about Rs 3,500 in a typical Indian setting. Minimal infrastructure costs to support colonies of 10 cent plots (simple roads and electricity line, and a hand pump well) might double the cost to Rs 7,000 per beneficiary. This is still less than the Rs 20,000 per beneficiary currently spent on rural housing schemes. An India-wide programme to provide 1/10 acre plots to all of the country's 13 million landless agricultural households would require some 13 lakh acres (only 1/3 of one per cent of the nation's farmland). Government purchase of all such land at market prices (through mandatory acquisition if necessary) would cost roughly Rs 4,300 crore or Rs 430 crore per year spread over 10 years. Government provision of minimal infrastructure to these new colonies of 10 cent plots (through new or existing schemes) might double the cost to Rs 860 crore per year, half of what the Central government alone now spends on the largest rural housing schemes. Poverty in India is inextricably linked to landlessness. Government efforts to alleviate rural poverty cannot afford to ignore this connection. This new evidence points to feasible option for addressing poverty by giving completely landless rural families ownership of 45x100 ft homestead plot. The Central and state governments should consider pilot-testing the concept. Roy Prosterman is professor at the University of Washington School of Law and president of the Seattle-based Rural Development Institute Tim Hanstad is executive director of RDI and currently heads their office in Bangalore