This paper argues against determinacy as a guide to interpret rational expectations solutions, using an example in which a determinate solution exists but differs sharply in dynamic behavior from that implied by the model considered on a sector-by-sector basis.

So-called "spurious regression" relationships are generally accompanied by clear signs of residual autocorrelation. A conscientious researcher would likely re-estimate with an autocorrelation correction. Simulations indicate that resulting test statistics are close to true values, so do not...