Playing with wages and price indices

Graph

Doing two things, experimenting with economic data; and looking at actual, price index adjusted wages; based on a post in Cafe Hayek

Bear with me, this is an experimental post. But I am looking at the cost of a blue collar worker in 2005 dollars, vs the CPU-U, a price index of a consumer basket of goods in 2005 dollars, and PCE, a price index based on actual consumer purchases. The actual cost of labor went up with the PCE, but not as fast as the CPU-U.

It looks like the labor cost index rose with CPI-U from 1999 to 2004, then rose with PCE.

CPI-U is a goods sellers index, it tells the merchandizer what things are selling for. PCE is a buyers index. PCE should track closer to actual purchases, but CPU-U has to equalize over time, but generally requires manual updates to the basket.

This seems to confirm some CATO work and this study. I am missing any benefit adjustments to wages here.

Update:
I added the unemployment rate so we can see slack vs tight labor supplies, and again, under tight labor we follow CPI, under loose labor we follow PCE.

The debate is always about whether some change in the relative wage distribution justifies some change in standard of living.