Macroeconomic Stories and Patterns by Ed Leamer
While Leamer's analyses are US-centric, he approaches macro's problems with the pragmatic poise of a statistician. This seems best read as a counterpart to Gordon or Mankiw's intro to macro texts.
Correlation is in the data. Causation is in the mind of the observer.
In a normal consumer recession, weakness in spending on homes and autos creates pent-up demand, and the low interest rates in the aftermath of a recession move sales forward in time, capturing sales that had not been made during the recession. But there were no lost sales of homes and autos during the 2001 Comeuppance. So where do you suppose the sales in 2003 - 2005 came from? They came from the future. Dr. Greenspan was moving sales backward in time, not forward. That created economic strength in 2003 - 2005, but inevitable weakness later when there were fewer potential home and auto buyers. To put this another way, monetary policy needs to be conceived explicitly as an intertemporal control problem, with the decisions made today affecting the options later on.