I'm still not sure why the substitution effect _necessarily_ means the CPI grows faster then does the cost of living. If 1) Purdue chicken costs MORE than Tyson chicken, and 2) Purdue chicken FALLS in price then the substitution should be to the HIGHER priced good. The relative price of Purdue has fallen, but the substitution is to the good with a higher absolute price. In this case, the CPI understates the rise in the cost of living. It seems that the direction of the substitution bias is an empirical matter, not a theoretical matter. This issue is whether high priced goods are more (or less) likely to experience a decline as are low priced goods. If price declines occur randomly among low/high priced goods, then the substitution effect should net out. Perhaps I'm missing something. Eric .. Eric Nilsson Department of Economics California State University San Bernardino, CA 92407 [EMAIL PROTECTED]