Yes, another arguement for using data that more closely tracks actual household life. The further you get away from real life the more chances there are for mistakes or mischief (GDP takes National Accounts and imputes it to real life; PPP takes GDP and imputes it to a US GDP basket and in a particularly inaccurate way) .
I was amazed to see that the Soviet Union had superb (and largely truthful!) household surveys produced by Goskomstat and used for some key living standard decisions, like calculating the "normale" (the base unit then used for adjusting minimum wages, pensions, benefits). Even more marginal Republics (e.g. Moldova) had accurate household surveys that survived the initial transition and functioned when they were needed most - in times of upheaval. I was therefore equally dismayed (as were most CIS and foreign statisticians) in the early '90s when the World Bank insisted on reforming/dismantling the system throughout the CIS and ensuring that the statistical series were broken before they could track the impact of proposed economic reforms.
Admittedly Russia today is an extreme case with such large amounts of the economy being moneterized but 'off the books'. But, worldwide, since the advent of the Bretton Woods' structural reform programs they have been declared a quick success by the simple expedient of having large amounts of funds move 'on the books' as measures such as hard currency liberalization or tax "reform" simply legalized what was kept in the black. Of course in Russia today one also has the expedient of simply changing the official "estimates" of unofficial activity.
To be fair, this exchange with Chris is a criticism of the abuse of national accounts in general and not just the World Bank's version of PPP. But the Bank's PPP lends itself (through the inherent design in its math) to a new level of recklessly inaccurate and ideologically driven calculations. We began with the example of the Swedish paper (France and Germany are poorer than all but 4 US states; etc, etc) - but I have seen countless similar silly use of PPP numbers to drive ideology - and the numbers almost always look different (and more intuitive) without PPP.
Paul