Econlib Resources
Subscribe to EconLog
XML (Full articles)RDF (Excerpts) Feedburner (Oneclick subscriptions) Subscribe by author
Bryan CaplanDavid Henderson Alberto Mingardi Scott Sumner Subscribe by email
More
FAQ
(Instructions and more options)



Bring us right back to "not growing as fast as before" vs. "not growing".
Caplan's entire post is weak. None of his criticisms directly address the point that CPI adjusted wage growth is growing slower in absolute terms and relative to the top earners. And pointing out consumer surplus is committing the basic math error of double counting, since CPI is adjusted for hedonics and other things to take into account such gains.
The huge point failure by Cowen that Caplan, Kling, and Hendersen ought to be pointing out is that wages need to be adjusted to take into account increased benefits. Basically, median worker's productivity gains have been shifted from wage comp to benefits compensation. And Caplan, Kling, and Hendersen ought to be yelling that if workers don't want all their compensation growth going into health care, they ought to demand cash comp instead of health care, and buy cheaper health insurance that has a high deductible, and is pretty much just designed to protect against catastrophe.
CPI *tries* to make hedonic adjustments. It fails. We know this because an hour's labor buys more than ever before, much more than CPI adjustments would suggest. And in what version of the world does the BLS keep track of demand prices?
"We know this because an hour's labor buys more than ever before,"
Housing? Medical care? Education? Since wages have stagnated, it is more difficult to buy more.
Steve
An example (courtesy of Steve Horwitz) of what I'm talking about is the first table at this link:
http://myslu.stlawu.edu/~shorwitz/Good/myths.htm
I believe he posted an update last year, from 1999 to 2009, but I don't see it on his page.
[miscoded link fixedEconlib Ed.]
Isn't this just the endlessly repeated critique that the CPI overstates inflation because it misses substitution effects?
Since 1985, the CPI has increased approximately 100%. Taking the standard interpretation of the CPI, I would be roughly indifferent between:
(A) $25,000 income in 1985
(B) $50,000 income in 2010
If I understand Miller, I would prefer option (B) to option (A). The percentage increase in income required to make me indifferent between 1985 and 2010 is not 100%. It is less.
Isn't this just the basic point that the CPI is similar to a Laspeyres index, and Laspeyres indices overstate inflation because they don't capture substitution effects. For example, Laspeyres indices suffer from "new good bias." While some prices have gone up, the price of any new good has declined from an infinite to a finite value, and this is not taken into account.
In equations, let p be the old price level, p' be the new price level, and let x be the old consumption bundle. Let p*x denote the dot product. Let e be the expenditure function.
e(p', u) / e(p, u) is less than (p' * x )/(p * x)