<$BlogRSDURL$>

2004-06-04

Should Eurostat adjust Europe's economic outputs by Purchasing Power Parity? 

Yesterday Eurostat published its first 'GDP per capita nowcast' for the EU-25. It runs under the headline "GDP per capita in new Member States ranges from 42% of EU25 average in Latvia to 83% in Cyprus".

For the non-economist that I am, there are a few points in this that look puzzling.

The first mistake one can make is to think of these data points as a measure of average population wealth, or average income. Such a mistaken interpretation is somehow suggested by Eurostat's adjustment of its data to purchasing power parity (PPP). PPP-adjustment scales the raw figures to reflect the different cost of a representative basket of goods and services in each country. But of course, GDP is not wealth or income, but rather something close to the total turnover in a national economy. And so the PPP-adjustment is not done to show how much the average consumer can buy, but only to get a measure of the size of each country's economy in terms of its own price levels. The advantage of this is that it irons out short-term swings in currency exchange rates, which can lead to huge swings in unadjusted GDP/head ratios in multi-year time series. Here are some good illustrations of this difference. So far so good.

Next, all honest sources admit that the econometrics that leads to PPP-adjusted data is complicated, leads to time delays, and can never be as precise as reporting raw GDP/head figures. But, as Walter Stanners explains,
Be all that as it may, Keynes’ observation applies. It is better to be approximately right than to be precisely wrong, and it is abundantly clear that cross-country comparisons using precise market currency conversion rates are too volatile ever to be right.
But my concern is a different one. We are supposed to use PPP-adjusted data for comparing different countries. So, for example, to compare the economy of Ireland to the one of Austria and to the one of Lithuania. The Eurostat figures have Ireland at 131 units, Austria at 121 units, and Lithuania at 46 units. However, while Austrian GDP/head in 2003 was 27849 euro, Lithuanian GDP/head was 4684 euro (source). What people do with comparisons is to relate the compared items to each other in various ways. And therefore, inspite of all the warnings, when people like myself read these figures reported in the daily press (which loves such stories), they have thoughts like: the average Irish company/individual has X times more spending power than its Austrian equivalent, so when it comes to Austria it can buy X times more than at home. Eurostat wants me to think in PPP for the first half of the sentence, and in unadjusted GDP for the second half. But companies invest, people travel or migrate - and it's exactly at those times that they think about cross-national economic differences. PPP-adjustement is misleading here, especially where economic differences between compared countries get larger and PPP diverges further from GDP/head. The Lithuanian thinking about economic opportunities in Austria would have a ratio of 5.9 in mind, and not the PPP-adjusted 2.6 suggested by Eurostat.

So PPP-adjustment is complicated, imprecise, slow, and misleading for most of the audience of such reports. Is this not a case for, at least, reporting unadjusted GDP/head along with it?


0 comments:
Post a Comment

Back to Main Page
This page is powered by Blogger. Isn't yours?