Brad de Long slams as ‘fast food journalism’ a piece in the Financial Times by John Kay, on the topic of US and European retail productivity. His main complaint is that Kay implies that productivity statistics for retailing take no account of differences in the quality of service between friendly local retail markets and gigantic US-style megamarts. As Brad correctly says, the Bureau of Labor Statistics in the US takes a lot of trouble over this kind of thing.
Brad goes on to say, or at least imply, that Kay’s preference for the friendly local market over the Carrefour supermarket is snobbish and takes no account of the benefits to low-income workers from low-margin stores like WalMart, which are denied them by European regulation. This seems a bit unfair – as I read Kay, he’s saying that the regulations are superfluous and that the local markets can survive the competition.
Turning to the more general issue of productivity comparisons, Brad’s post raises quite a few points. First, while the BLS no doubt does its best, it is still true that retail productivity is hard to measure, and that claims of large gains must be taken with a grain of salt. For example, in principle, the BLS should take account of the increased travel time required for shoppers to go to edge-of-town Walmarts, but I don’t think they do so.
Second, as I pointed out in another recent post on this topic, the quality adjustments undertaken by the BLS do not take account, even in principle, of the negative externalities arising from people driving long distances, externalities that include traffic congestion, the loss of green space and thousands of extra road deaths every year.
Third, the BLS generally does a lot more of this ‘hedonic adjustment’ than do European statistical agencies (or, as far as I can tell, the Australian Bureau of Statistics). I think the BLS estimates are probably more accurate on balance. Regardless of which estimates are better, the inconsistency means that comparisons of US and European GDP and productivity growth are systematically biased in favor of the US, by about 0.5 percentage points per year.
Finally, the really big question is whether the differences in US and European consumption patterns, working hours and so on are driven by different tastes, different relative prices and regulatory constraints or some complex combination of the two. I’m still planning a big post on this question Real Soon Now.