Is it possible that the economic policy debate in Japan is being distorted by bad data?
A recent article by Robin Harding in the Financial Times (“Japan revamps GDP numbers over accuracy fears,” September 30)
suggests that growth in Japan since the Global Financial Crisis may be significantly higher than previously believed. Perhaps most notably, an experimental data series produced the the Bank of Japan suggests that the Japanese economy expanded by 2.4 percent in 2014, following the rise in the country’s consumption tax from 5 to 8 percent, rather than contracting as indicated by official data.
If this were true, several implications would emerge. First, Abenomics may be working better than any of us imagined. This would at least be broadly consistent with the fact that the labor market is extremely tight, with labor shortages emerging in a number of sectors, and that tax revenues and corporate profits have been sharply rising.
Second, growth in Japan may be considerably above potential (although estimates of potential would also need to be re-estimated for consistency with actual new growth estimates). If that was in fact the case, then the view expressed by many that the first increase in the consumption tax in 2014 wold need to be reconsidered, as would the wisdom of failing to implement the planned second hike last year and the planned large fiscal stimulus recently proposed by the Abe administration.
At the same time, signing on to these experimental data would only raise additional questions about the failure of prices to rise substantially toward the BoJ goal of 2 percent. Why would massive monetary accommodation contribute to strong growth but not to significantly higher inflation? While the recent yen appreciation and depressed energy prices play an important role here, even core inflation is currently barely above zero.
Of course, there is no reason to assume that the BoJ estimates are the right ones. But there are reasons to believe that the official estimates are problematic, including a falling response rate and some apparent inconsistencies with other data. The BoJ estimates are based on tax return data which are more comprehensive but, unfortunately, available only with a one-year lag.