No. 22 — June 9, 2000

 

Weekly Review

JAPAN'S GDP DATA PROVOKE INTERNATIONAL CONTROVERSY
--- by Douglas Ostrom

What did the late Prime Minister Keizo Obuchi have in common with The New York Times, the Organization for Economic Cooperation and Development and any number of Tokyo-based economists? Like them, he was skeptical about the accuracy and the usefulness of Japan's broadest measure of economic activity — the gross domestic product — in general and the fourth-quarter 1999 figures in particular.

The controversy began to unfold after the Economic Planning Agency's mid-March announcement that GDP had contracted an annualized 5.5 percent on a price-adjusted basis in the October-December period from the previous quarter. This report conflicted with other data that led EPA to assert only days later that the economy had reached a state of "self-sustaining recovery" (see JEI Report No. 11B, March 17, 2000). American analysts, in contrast, saw in the new GDP figures evidence of a recession since the economy had shrunk in the July-September quarter.

At a mid-March cabinet meeting, Mr. Obuchi directed EPA and the Management and Coordination Agency, another of the principal government organizations involved in the collection of economic data, to investigate the discrepancies. Analysts speculated that Mr. Obuchi was looking for a way to head off talk about a recession before nationwide elections then expected to take place in the fall. An EPA official noted at the time that a prime minister's involvement in such a technical matter was unprecedented.

Mr. Obuchi's incapacitating stroke in early April and his death in mid-May had the effect of advancing the lower house elections to June 25 while doing nothing to end the data controversy. Revisions to the fourth-quarter GDP statistics, which EPA released May 11, put the economy's contraction at 5.6 percent, or slightly more than originally estimated. The New York Times reported May 24, however, that the agency had opted out of making additional changes that would have enlarged the decline even more.

At issue is capital spending by financial institutions in the October-December period. In the GDP statistics released in mid-March, EPA had relied on a forecast for this activity. Ordinarily, it would have used survey data for the revised GDP numbers, but the latter information had seemed so out of line that the agency decided to recycle the original projection.

The obvious speculation was that EPA had yielded to political pressure to report a more favorable GDP number ahead of the expected lower house elections, although most analysts, foreign and domestic, argued that the agency and its leaders were above such shenanigans. These experts noted that, intuition notwithstanding, economic forecasts not infrequently provide numbers that are more reliable than preliminary survey data. One EPA insider, however, couched his denial of political interference in a less-than-reassuring manner, admitting that EPA sometimes does bow to political influence in its economic pronouncements.

Perhaps more persuasive than the denials was the reality that EPA would have little motive to make last year's October-December figures look better if, by doing so, it robbed growth from the first quarter of 2000. The January-March GDP statistics are scheduled for release June 9, just a little more than two weeks before the actual lower house elections.

In an extraordinary response to a story originally broken by a foreign reporter, EPA officials announced May 25 that further investigation had shown the less optimistic survey figure on the financial sector's investment spending to be appropriate. They said that the agency would release fourth-quarter GDP data revised for a second time along with the numbers for the January-March period. Although EPA revises a given set of quarterly GDP statistics repeatedly over the years, two revisions within a one-month span are highly unusual, if not unprecedented.

EPA may have reversed itself because it also was feeling heat from another source — the OECD. In its semiannual world economic outlook, which presumably was prepared before publication of The New York Times article and released in late May, the Paris-based organization took aim at the quality of Japan's economic data, noting that seemingly related statistics often are inconsistent. In yet another unusual move, in early June, EPA published on its Web site four pages of "comments" that appeared to constitute a rebuttal of the OECD criticism.

Beyond the back and forth on a seemingly arcane subject are several larger issues. The flap over capital spending by the financial sector concerns something bigger than the relatively small revision to the GDP data that would result from plugging in the survey information. Given the general acknowledgement that Japanese financial institutions have fallen behind their foreign counterparts in the deployment of new technology and the widespread view that information technology is the magic bullet that will solve the problems of Japan's economy, a decline in business investment, if confirmed, is anything but good news to those hoping for a sustained recovery.

Moreover, with the world's second-largest economy on what is at best a rocky growth path, timely access to robust data is particularly important. This information will help to determine, for example, whether another stimulus package might be in order. Three decades ago, when Japan's annual growth rate was close to double digits, a percentage-point error in measurement hardly mattered. In today's environment, such a mistake could mean the difference between contraction and expansion.

A third issue concerns national pride. EPA authorities are proud not only of their integrity but also of the accuracy of the agency's information, particularly vis-a-vis counterpart organizations in other countries. As such, the suggestion of Brian Rose, a veteran economist at the Tokyo office of Warburg Dillon Read LLC, is particularly galling to EPA. The coauthor of a 1994 book on Japanese economic statistics, Mr. Rose told the Financial Times in March that "[t]hey really need to start from scratch again — and they could do with outside help from countries such as the [United States]."

Yet another issue is the underlying economic reality. EPA explained in its response to critics that Japan's economic data vary because the economy fluctuates to an unusual degree. Although this argument sounds like an excuse, it cannot be quickly dismissed. Such institutional factors as semiannual bonuses and cultural proclivities like the tendency of Japanese workers to take vacations at the same time do produce what is, by U.S. standards, unusual volatility in the economy.

EPA also is on firm ground when it implies that alternative economic data are not necessarily better than its GDP statistics and, in fact, may be worse. For example, the Ministry of International Trade and Industry's industrial production series, which sometimes is used instead of GDP because it comes out sooner, covers mainly manufacturing and double-counts some economic activities within that sector. Moreover, the index's volatility is extraordinary. Despite recent improvements in the methodology used for seasonal adjustment, which have smoothed out some variability, factory output jumped an unbelievable 32.9 percent on an annualized basis between February and March 2000 before plunging 4.7 percent in April, the most recent month for which information is available.

Finally, the brouhaha over Japan's economic data concerns in part the technical expertise of government professionals. Graduates of elite universities, they tend to be generalists and are rotated from position to position within their agencies. Few appear to spend their careers focusing on a handful of economic statistics, as their counterparts in the Department of Commerce's Bureau of Economic Analysis, for example, have been known to do.

While the Japanese system arguably produces a high average level of competence, it may yield few world-class experts. Moreover, in many instances, it appears to have fostered an extraordinarily conservative approach to data collection — one that includes few means through which changes in practices can be implemented. What the growing ranks of EPA's critics, domestic and foreign alike, are saying is that these procedures no longer are good enough.

The views expressed in this report are those of the author
and do not necessarily represent those of the Japan Economic Institute

Issue Index aaaa 2000 Archive Index aaaa Subscriber Area aaaa Home