SNA Revision: Has the picture of the Japanese economy changed? Jun Saito, Senior Research Fellow Japan Center for Economic Research January 11, 2017 Japanese SNA revised in December 2016 Japanese system of national accounts (SNA) underwent a major revision last December. While the stock tables are still to be published (expected in January 19), most of the flow tables are already available to the public. I would like to take this opportunity to see if the new figures change the picture of the Japanese economy. Reasons for the Revision The latest revision of the SNA is a major one in the sense that three independent revisions are made simultaneously at this timing. First is the finalization of the preliminary figures that have been published quarterly in the past year. The revision of this nature takes place every December, when the new information provided by the annual statistics such as the Census of Manufactures and central and local governments budgets are incorporated. They are used to replace the GDP estimated from the demand side by GDP estimated from the supply side by applying the commodity-flow methodology. Second is the benchmark revision that takes place every five years. Such rebasing takes place to incorporate the information included in the input-output tables and other major statistics that are published every five years. At this timing, the base year for the deflators is also changed to the year of the latest input-output table: The base year has been changed this time from CY to CY. Although the revision from this aspect usually takes place every five years, it has taken six years since the last revision was made. That is because it took six years to compile the latest input-output table for, which tried to incorporate information that is collected through the newly launched economic census of. Third is revision related to the adaptation of the national account system to the latest SNA manual published by the United Nation; the SNA. It includes changes such
as the change in the treatment of research and development (R&D) activities. Spending on R&D is now treated as a part of capital formation instead of treating it as a part of intermediate input. It leads, for instance, to considerable change in the level of private business investment. Changes in GDP growth rates The latest SNA revision has led to changes in the GDP growth rates in both nominal and real terms. The changes in the nominal GDP growth rates (new growth rate minus old growth rates) in the 21 years since FY range between -0.4 percentage points (for FY) and +1.0 percentage points (for FY). The average changes in percentage points for the whole period is +0.2 percent, but the changes in the last six years is constantly positive. In particular, for the last three years, the upward revision has been relatively large at an average of +0.7 percent. Similarly, the changes in the real GDP growth rates (new growth rate minus old growth rates) in the 21 years since FY range between -0.7 percentage points (for FY) and +0.8 percent (for FY). The average of the changes in percentage points is +0.1 percent, but the changes in the last three years are positive with a relatively large upward revision of +0.5 percent on average (Table 1). (Percent) 5.0 3.0 1.0-1.0 Table 1: Revisions to Real GDP Growth Rate -3.0-5.0-7.0 Difference due to the revision Real GDP ( SNA, Benchmark year=) Real GDP (1993 SNA, Benchmark year=) As for the rate of change of the GDP deflator, there is only a minor upward revision by an average of +0.1 percentage points for the whole period. It is significantly limited compared to the past revisions. It owes to the introduction of the chain-linked index to - 2 -
the deflators: It removed the shortcoming of the simple Paasche type fixed-based index that often led to major upward revisions of deflators. The magnitude of the revision of GDP deflators for the last three years is also about the same at +0.1 percentage points on average. Therefore, the relatively large upward revision of the real GDP growth rate in the last three years can mostly be attributed to the relatively large upward revision of nominal GDP growth rates rather than the changes in GDP deflators. Changes in GDP levels The changes in the GDP growth rates obviously reflect the changes in the level of nominal (and real GDP after adjusting for the price changes). As Table 2 shows, the level of nominal GDP has been revised upwards by a magnitude of 6.8 trillion yen in FY (1.4 percent of the old nominal GDP) to 31.6 trillion yen in FY (6.3 percent of the old nominal GDP). The revision is particularly large in the last three years, FY to FY, where the upward revision amounted to 25.0 trillion yen, 28.3 trillion yen, and 31.6 trillion yen, respectively (on average, it is equivalent to 5.8 percent of the old GDP). 540 520 500 480 460 Table 2: Revisions to Nominal GDP 440 420 400 Nominal GDP ( SNA, Benchmark year=) Nominal GDP (1993 SNA, Benchmark year=) According to the information provided by the organization in charge of compiling the SNA data in Japan, the Department of National Accounts of the Economic and Social Research Institute of the Cabinet Office, the upward revision is largely due to its adaptation to the SNA, as Table 3 shows. However, as the same Table also shows, the increase in the upward revision in the three years to FY is due to revisions - 3 -
other than the adaptation. It should come from the impact of the changes in the industrial structure that was revealed in the latest input-output table for. Table 3: Factors behind the Revision of Nominal GDP 35 30 25 Due to adaptation to SNA Due to other revisions 20 15 10 5 0-5 -10 Changes in the picture of the Japanese economy One of the interesting features that the new SNA series reveals is that the speed of the recovery from the global financial and economic crisis now seems more rapid than we thought. With regards the profile of the developments in real GDP, it is not so different from before. As Table 4 shows, while the level of the real GDPs has fallen, the profile maintains that the real GDP recovered the pre-crisis level in FY. Table 4: Revisions to Real GDP 550 530 510 490 470 450 430 410 390 370 Real GDP ( SNA, Benchmark year=) Real GDP (1993 SNA, Benchmark year=) 350-4 -
As for the nominal GDP, as the preceding Table 2 shows, the figure for FY in the old series was 500.6 trillion yen, still 5.7 percent lower than the pre-crisis peak of 513.0 trillion yen at FY. However, after the latest revision, the nominal GDP for FY is now 532.2 trillion yen, which is slightly above the FY level of 531.0 trillion yen. Although it has taken more time than real GDP, nominal GDP also seems to have recovered the loss incurred by the global financial and economic crisis. The changes in other main indicators are more limited. The picture, therefore, is not that different from before the latest revision. I will touch upon three of them. First is the labor share of the value added (share of compensation of employees in the national income). It has been revised downwards, except for the period between FY and FY. However, it still maintains the declining trend that can be observed since FY: It is 67.8 percent in FY, the lowest level since FY. Second is the household savings rate. It has been revised upwards, though the magnitude of the revision is limited for the years after FY. It still retain the profile of showing a rebound in FY: the new series shows that, after reaching -1.3 percent in FY, it has recovered somewhat to 0.7 percent in GY. Third is the net lending of households. It has been revised upwards for most of the period. However, for the years after FY, the figures are revised downwards. As a result, the fall in net lending since FY is steeper than before. In fact, net lending for FY now shows that it is -2.5 trillion yen, the first negative figure for the series. Historical data need to be extended As we have seen, the changes brought about by the latest SNA revision does not call for a significant change of the picture of the Japanese economy. The latest revision, however, reveals some new features that economic analyses need to address. It is hoped that economic analysis using the new SNA series will quickly be undertaken and published. However, for the analysis to be meaningful, data series need to be long enough. Although the new SNA goes back 21 years to FY, much longer than the past revisions, it is hoped that it goes back further: hopefully to cover the bubble period of late-1980s and the early years of the lost decade (1990s). - 5 -