Tyler has a new mini-book out on stagnation where he accepts the idea that median income has been growing very slowly for the last 40 years relative to earlier times. One challenge to this claim is that prices are mismeasured. Tyler, of course, understands that and offers this response  (HT: Will Wilkinson ):
One might think that the CPI is skewed and there are reasonable arguments to be made in this direction. But the CPI will be most skewed to underappreciate progress when truly new goods and services  are being introduced into the marketplace or spreading to new regions. And that is (roughly) the 1870-1950 period, more than any other time. In other words, if you account for CPI bias, the slowdown in median income growth — the difference — is probably largerthan the numbers make it appear, even though in absolute terms both growth rates will be higher than measured.
It is not just a question of the number of new goods and services–it is the pace of innovation within product categories and how much each of these makes it hard to measure prices with any accuracy. So it is not just that we now have televisions that we didn’t have before. It’s the speed at which new television models are introduced and the pace of the improvement of those new models. The 1946-1973 period surely introduced some new products. But my perception is that the pace of innovation was quite slow relative to today. As I wrote on September 6, 2007:
The iPod will be six years old next month. The newly released iPod Classic with 160 GB of memory is $50 cheaper than the original iPod, holds 40 TIMES more songs and also plays color videos and displays photos. It is smaller, lighter and has a better battery. I wonder how the BLS takes account  of the quality differences when measuring the price index and inflation.
That 160 GB iPod is now $249. I assume it’s smaller and weighs less and has longer battery life. But that’s a pretty easy price change to measure. How do you deal with the existence of the nano and the iTouch? Much tougher. I have a feeling this kind of within-product innovation is much harder for the BLS to quantify and is much more prevalent in recent times than earlier ones.
To understand the challenge of measuring CPI accurately when there is a high rate of innovation, read my earlier posts here  and here . The expert on this issue is Mark Bils. Some of his papers on these issues are here .