Do Wild Fluctuations in Quarterly Inventory Investment Data Matter?: A Study of Japanese GDP Statistics, 1994-2010



For more than half a century, inventory investment has attracted wide attention as a major cause of short-term macroeconomic fluctuations, and the mechanisms involved have been the focus of many major studies. Yet microeconomists and business people familiar with corporate behavior have frequently expressed misgivings about the enterprise. Using Japanese quarterly GDP inventory investment statistics both by commodity and by category, 1994~2010, I investigate the nature of quarterly inventory statistics and the inventory investment behavior, and draw two conclusions. First, statisticians estimate the quarterly statistics under severe time constraints, and their resulting figures incorporate seasonal variations which dominate the quarterly fluctuations. This fluctuation mostly disappears in annual data. Secondly, when I examine the inventory variation after the Lehman Shock in the autumn of 2008, I find neither a notable increase in inventory stock nor a long-run stock adjustment process. Given the size of this unforeseen exogenous shock, most observers expected a large inventory stock accumulation to follow. That the accumulation did not follow suggests that the focus on inventory variation may be misplaced. For inventory investment data estimation, Japan is an ideal OECD country, with generous statistics availability. The conclusions of this research, drawn from the quarterly GDP inventory statistics, will stimulate the interest both in the study of inventory data in other countries focusing on its estimation process and source statistics, and in the great variety of inventory investment. At the same time, the conclusions pose a grave implication not only for re-evaluation of the literature in inventory investment variations but also for other research topics in macroeconomics like monetary transmission mechanisms including "financial accelerator" theory.