Another important development was introduced in a paper by Robert Lucas in his 1976 paper, Econometric Policy Evaluation: A Critique. In this paper he developed a model where firms possessed only imperfect information about the price level. The assumption was that they knew the price their good fetched in their own market but not the general price level. The problem for the firm is as follows. If the price in their own market increases then they will want to increase output in response to it. If the general price level increases then they will not want to alter their output at all. The problem is that when a movement in price is observed, the firm does not know if the change represents a shift in demand in their market, a change in the general price level or a combination of the two. It is only in the next time period that the firm learns what the general price level was in the previous period. There are a number of important implications here. The first is that there will only be an output response if the inflation created by the government is unexpected. The other important implication is that agents can learn. If they observe the authorities creating inflation to manipulate demand they will not change their output in response to a price change; even if it is a change in the price in their own market which requires an output response. The implication here is that if the government wishes to control unemployment it should pursue supply side policies and try to keep the aggregate price level constant. The more general implication is that trying to manipulate a historical relationship to achieve a policy objective will not work, since the parameters of the relationship are functions of the expectations which are, in turn, affected by the attempt to achieve the policy objective.