Wednesday, August 18, 1999
Japan's Economic Planning Agency normally releases revised gross domestic product figures at the same time it announces preliminary GDP figures for the subsequent period. But 1999 is not normal. EPA's June 10 announcement that the economy surged 7.9 percent in the January-March quarter raised eyebrows, to say the least. Such a growth rate would have been amazing even in a healthy economy, but Japan's has been ailing for the better part of a decade. Many analysts had expected that GDP would be basically flat.
Apparently as a result, EPA altered its procedures and announced August 13 a revised figure without waiting until September. The expectation was that the new figure would be lower, perhaps because of the availability of current account trade figures becoming available about the same time. (For details of those figures, see JEI Report No. 32B, August 20, 1999). The current account figures, while interesting for the light they shed on currency intervention, turned out not to matter much. Instead, EPA announced that stronger capital spending than previously reported resulted in slightly higher growth, 8.1 percent annualized, than previously reported.
In short, EPA managed to pour fuel on the fire surrounding the GDP statistics. Not shy despite being burned last time out, most analysts now are arguing that the preliminary figures for the second quarter will once again show to GDP to be close to flat, or maybe even shrink a bit, when they are released in September.
The problem, as Robert Alan Feldman demonstrates in a recent contribution on the Morgan Stanley Dean Witter web site (www.ms.com) is that the confidence interval around estimates of Japanese GDP are surprisingly large. In other words, GDP is very hard to predict. Predicting that GDP growth will be negative when only days later it turns out to be positive is only one sort of embarrassing mistake that is easy to make.
The recent experience raises the question as to why the Japanese GDP number is so difficult to get right. Most analysts appear to rule out one obvious answer dishonesty with some certainty. The EPA is staffed by trusted professionals. Besides, if they were cheating they would not have churned out such grim numbers for so long.
Another explanation that the economy is just like that is somewhat more appealing. Just perhaps the numbers accurately reflect the fact that the economy really does fluctuate dramatically and behave erratically. In fact, permanent employment, year-end bonuses and lengthy national holiday periods, are just three characteristics of the Japanese economy that complicate a consistent reading across quarters.
A third explanation bad data collection should not be rejected out of hand. Although some Americans have marveled at the plethora of statistics widely available in Japan, the fact is that many Japanese economic phenomena are measured using techniques that are crude by U.S. standards, where the quality of economic data has been questioned by top economists.
This possibility raises the issue of something like the Boskin Initiative, named after Michael Boskin, a chairman of the Council of Economic Advisors in the Bush Administration and a Stanford University professor. Professor Boskin was among the first to sound the alarm about the quality of U.S. economic data. As Japan enters a period where economic growth is low year in and year out, small errors of measurement increasingly can make the difference between what is regarded as acceptable economic performance and economic inadequacy. Japan needs a Boskin Initiative.