It's truly nice to see the first string of macroeconomic interest at the Bank thinking about agriculture again, and sincere thanks to Shanta for taking the time to lay it out---but possible heresy? Let me rise to the bait. Some comments, following Shanta's numbering: 1) “Agricultural growth contributes 2 to 4 times as much to poverty alleviation as other forms or growth is an arithmetical point, not an economic one (paraphrased)” Maybe it does follow largely from the fact that 75% or the world's very poor are rural, and most of these are engaged in farming for at least part of their meagre livelihoods. The fact that agricultural spending had fallen to 4 % of ODA globally and of public expenditure in Africa at the time of World Development Report (WDR) 2008 suggests a stunning lack of arithmetical congruence at the very least. Further, much of the current discussion is about Africa, where WDR 2008 makes a compelling case that agriculture is vital for growth (unlike most of South Asia, as well as poverty alleviation (like most of South Asia). The proportional size of the sector and high transfer and transaction costs for agriculture in a good chunk of Africa are consistent with an economic view that significant numbers of rural areas in Africa, housing lots of very poor people, are demand constrained (a counter-heresy?). That is to say, there are underemployed resources because no one has any money locally to buy anything, and local products are not competitive price-wise outside the local region. When money comes into these areas (as pension payments or sales of tradable products from within to outside the local area), and it is widely distributed to consumers in the zones in question, some of these underemployed resources are brought into production to produce the things that these now slightly richer poor people consume. If these consumer items are demand-constrained nontradables like services, locally processed foods, local bricks and mats for construction, etc, this produces a multiplier effect for agricultural growth. These effects have been widely observed in Africa for both pension trucks and export crops. The latter has the advantage over the former that it is economically sustainable and does not create fiscal deficits. The entry point then is to promote the supply of (agricultural) tradables in these zones for sale outside the local area. 2) I agree absolutely with Shanta that the key is to invest in agricultural productivity growth. But following the logic in (1), it should be targeted in the first instance to tradable items (stuff that can be sold and not rot by the wayside or is not dependent on constrained local demand). Second, there is a bigger economic kick to smallholder productivity increases than for large farms if that means—as it probably does—that the income benefits of productivity increases for smallholders are more widely spread to poor consumers. Productivity increase on both large and small farms benefit all through lower food prices, but the distribution of income side is critical when demand-constrained areas are a problem. 3) The pernicious nature of the subsidy syndrome of using public good resources for private goods that Shanta alludes to is absolutely correct, and in fact given significant treatment in WDR 2008. Helping Governments use “smart subsidies”, as in Rwanda during the food price crisis say, is a different beast, provided the caveats of targeting, exit strategy, cost accounting, and transparency are followed. 4) (see above) 5) Woody Allen and public expenditures: good fun, but the serious message is that WDR 2008 calls for both “more” and “better” public expenditures. It is great to see serious prospects for “more”, but we all need to work on the “better”. Unfortunately, it is probably easier (not to say easy) to fund-raise for “more” , than to get donors to do the less glamorous job of building human capacity and institutions for “better”. 6) Democracy and “better” agricultural interventions--absolutely! --and at all levels, especially community empowerment to improve the targeting, design, delivery, and use of interventions.