Intermediate targets are the economic factors that (1) stand in a cause-effect relationship to the ultimate policy goals and (2) the government can affect, but not directly control.
According to the Tinbergen model the economic experts are not supposed to tell politicians what goals to pursue. Dialogue about preferences-that is, trying to influence preferences for income distribution and taxes and the like-would not be instrumentally rational. Preferences for X-amount of military expenditure and Y-amount of social welfare expenditure are assumed by the expert to preexist and be imperturbable, like one's taste for a cup of vanilla and spoonful of cookie crumbs. Instead, Tinbergen said, much in the spirit of Lionel Robbins, a contemporary at the London School of Economics famous for the "means and ends" definition of economics, that economists should play the role of The Expert, advising politicians on how to achieve ultimate goals, whatever the goals happen to be. Where do the goals come from? Politicians elicit them from constituents, Tinbergen said, and declared them to be. The economist-expert, for his or her part, can't just order 3 percent unemployment or 2 percent inflation. He lacks the means to act so directly - an important truth to learn in an economics course. But Tinbergen took it further, and claimed that the economist should not even try to be in the goal-making game. Not every economist agrees with him, though most academic economists do.
Still, in Tinbergen's way of seeing things, the economist-expert can still do a lot. The science of economics tells the politician that unemployment, inflation, economic growth, and other economic concerns are caused by numerous factors, only some of which the government can control. The trick is to find those factors that both influence the ultimate goals and can be influenced by the government: the intermediate targets.
Suppose the policy goal is growth in GDP per capita (economic growth). To accomplish it you the economist would first resolve whether the policy goal should focus on the aggregate supply curve, the aggregate demand curve, or both. If you focus on aggregate supply you may choose, for example, "after-tax corporate profits" as your intermediate target. You choose profits because your economic theory predicts that higher profits entice corporations to expand their supply and because you believe the government can increase those profits -- by lowering corporate taxes (these are your warrants). Similarly, to focus on aggregate demand, you might target "new housing starts." And you choose new housing starts because your theory predicts that new housing is an indicator of new economic confidence or consumption possibilities, something the government can affect through for example lower interest rates and "first home owner" subsidies. So the two requirements for an intermediate target-influence and ability to control-are met.
To target aggregate demand you might, to take one more example, choose interest rates as your intermediate target on the theory that lower interest rates will increase spending. Again, your theory holds that lower interest rates will stimulate aggregate demand. The target will influence the goal. The other factor-the "control" factor-is met if the government can bring down interest rates by changing monetary policy (no certain thing).