One problem in affecting aggregate measures of the economy, such as employment or demand, through policy is that the natural rates of these measures represent an equivocated basket of heterogeneous information.
For example, aggregate demand really represents the sum of the demand of different kinds of goods. One resulting problem is that if we say, stimulate demand, we may stimulate the wrong kind of demand. Or, if we create a job program to boost employment, we may foster the wrong kind of employment.
In short, it’s not enough to calculate the natural rate of unemployment, the natural price level, or the natural levels of supply or demand, we must also calculate the kinds. Otherwise we may be encouraging the ideal level of aggregate activity, but from an entirely wrong heterogeneous microfoundation.
A microfoundationally malformed macroeconomy will rapidly collapse if we are lucky, then rebound, rendering the activity a full waste if not an outright harm. If we are unlucky, the microfoundaiton will stick around for a while, creating massive opportunity costs and a stagnating economy over time.
What is the natural price level of grain? Of computers? What is the natural employment rate of petroleum engineers? What is the natural income level of teachers? And so on. If these values are not all calculated then government stimulus is likely to result in some degree of malconsumption or other distortion.
Two good examples would be the “green technology” focus of Obama’s stimulus, which was a huge mix of opportunity cost and outright cronyish waste, and the housing market collapse which caused the need for the stimulus. The Great Recession has been traced back to malconsumption in the housing market which was stimulated through government policy.