Correlation and Causation

Real data does not lie on a straight line.

Left-out variables.

Correlation is not a sufficient basis for econometrics  because it measures a relation between just two variables.  Most of the problems of interest in economics concern relations among more than two variables.

Example:  parents' income and SAT scores.  Could be smart parents have smart children and high incomes.  Same story for bathrooms [expensive cars]. 

Exception:  three variable model.

Exception:  supply and demand.