Multilateral agencies and economists with much influence have been urging laissez-faire in agriculture. While success with the rich countries has been minimal despite the commitments under the WTO, many poor countries with much agricultural potential in the long run have been coaxed to adopt near free trade in agriculture with disastrous results especially for the poor in these economies. There are fundamental problems in achieving even global (leave aside optimum for any particular country) optimality through world trade in agriculture given the immovability of land. Additionally the fact that poor countries start their transformation process with much of their population engaged in agriculture imposes special requirements upon agriculture. Incomes have to rise in agriculture to overcome poverty and to constitute rising domestic demand for modern manufactures and therefore the infant industry argument holds with additional force. We bring together the historical experience of agricultural development, the relationship between economic development and agriculture, trade in agriculture, the role of state action especially in the late industrialisation context. The differences between land endowed and land poor countries are recognised in their analyses. We develop a perspective on the comparative advantage of nations in agriculture and the evolution of the same. The metrics of agriculture and trade, arising out of the dynamics of the share of agriculture in GDP, the dependence of agriculture on land endowments, the biological limits to consumption of agricultural products, underlie a dynamic structural model of the revealed comparative advantage which is developed and tested using panel data from about 100 countries. The nature of agricultural products on several dimensions-its long lead in production, its perishability in some cases, its storability in others, but above all the grouping of many agricultural products into low price and income elasticity of demand-is used. The purpose is to draw insights that can usefully inform the content of state intervention, and trade policy especially from the point of view of a country like India which is likely to lose its comparative advantage in many agricultural products as incomes rise. The comparative advantage of countries in agriculture is most usefully characterized as rising of the arable land endowments per person and declining as the per capita income rises relative to the worlds "average" per capita income. A structural model on the lines above is estimated empirically. The Model is also dynamic since the rise in per capita incomes at a faster rate in transforming countries can be used as data to predict with a high degree of reliability that they would see a decline in their competitiveness. Similarly countries with low arable land per person would see a rapid fall in their competitiveness. Yet land abundance in poor countries does not automatically result in high competitiveness. [The abundance of easily mined other natural resources like fuels acting through the balance of payments could lower greatly the revealed competitiveness of agriculture]. To realize the same, much land has to be brought under the plough and enhanced, a task where the role of the state is important. Irrigation development as also the use of machinery on land enhances the competitiveness of agriculture. And the former is dependent much upon the ability of the state to put together public irrigation and support private irrigation. Even more importantly the investments in storage, market support, transportation, information provision, demonstration of new technologies and extension all of which are required at the beginning of the agricultural transformation require active intervention of the state. The problem for the poor countries with land abundance is compounded by the large distortion of international prices resulting from subsidization by rich countries as they face declining competitiveness in agriculture due to very high incomes. The coaxing of land rich poor countries in this situation to embrace laissez faire policies by the multilateral agencies is shameful and nothing short of suicide for these countries. The costs of subsidization in the rich countries are very small and the political benefits very large, so a roll back of subsidization is least likely. Agriculture is the first industry where surpluses can arise to stoke development as such. The historical evidence that no country of substantial size has been able to industrialize without a prior or simultaneous agricultural revolution has to be noted. And the infant industry argument is valid for agriculture as much as for industry. Both these further condemn the laissez faire position. Protection of agriculture is therefore the least distortionary way for the "large" land-poor poor countries as they advance to protect their employment. Protection alone without active support of the state to overcome the significant market distortions in agriculture and its inputs may not be enough. Protection in land scarce economies ought to be scaled down only as such economies are able to absorb labour shed by an advancing agriculture in other segments of the economy. Functionality also demands that the role of the state in agriculture and subsidization recognizes not only the market failures arising out of the public good nature of many inputs, but also the perversities that low price and income elasticities, when combined with the long "lead" can bring to the functioning of markets. Similarly the structure of the value chain from production to final consumption in distant lands-especially the fact that the aggregators and processors in the value chain would be able to capture rents-creates the basis for a crucial role for the state in trading, stocking and processing. Shortages and variations in output again create the need for buffer stocking. Successful late agricultural transformations have been built upon the state playing these roles. The state's role in processing while crucial has not generally been successfully realised, the complexity of the tasks being a basic bottleneck. Laissez faire policies in agriculture when without reference to the stage of development, and state failure to compensate for the market perversities underlie the disaster that agriculture has been for poor countries with much agricultural potential.