United States' sugar subsidy policies, including barriers to imported sugar, have been widely and justifiably criticised for supporting a relatively small group of sugar producers at the expense of consumers, taxpayers, sugar-using industries, and the environment. A recent paper (pdf) on US sugar points to the New Zealand experience of withdrawing subsidies from farmers. I've written elsewhere about the irrationality of agricultural support. What's significant is how difficult it is to withdraw subsidies, often for reasons that are quite genuine. Take those agricultural subsidies that are paid according to volume of production (the majority). Economic theory and empirical evidence say that their main effect will be to raise the price of the least elastically supplied input, which in this case is farmland. Now many farmers will have borrowed money to pay for land inflated by the capitalised value of these subsidies. It's politically difficult to take away the subsidies and erode erode the asset base of such indebted farmers.
But it should, I believe, be done. It would at first sight seem reasonable to offer some form of compensation, or to signal in advance that the subsidies are going to end in X years. But it would be kindest of all never to get involved in subsidising farming in the first place. It would be far better to subsidise poor, deserving people, not the most effective lobbyists. The possibility that the government will withdraw of subsidies is a legitimate business risk, and those who invest in subsidy-inflated assets should be prepared to accept it.