This time, I would like to discuss structural issues concerning the compilation of the Japanese System of National Accounts (JSNA). There are three approaches to the measurement of gross domestic product (GDP), namely, the production approach, the expenditure approach, and the income approach. Although GDP in JSNA is based on estimates from the expenditure side, the estimates are based on the production data of goods and services. The national total of household consumption or private-sector investment based on a survey on the expenditure side often turns out to be significantly lower than the national total in the JSNA, even when adjusted to be aligned with the System of National Accounts (SNA) of the United Nations.
The Economic and Social Research Institute (ESRI) of the Cabinet Office has been trying to improve the accuracy of annual GDP estimates derived from the JSNA by utilizing the annual supply-use table (SUT), a framework for describing the relationships between the outputs and inputs of goods and services. As a result, the statistical discrepancy, which used to be as large as 2% of GDP, has been reduced to about one-tenth in the JSNA.
Such efforts are commendable, but this practice is no more than disassembling a given jigsaw puzzle and reconfiguring it with a renewed set of pieces. The key to improving the accuracy of the current GDP estimates in the JSNA lies in improving the capture of domestic production activities, not expenditures or income. The real challenge for the JSNA is to thoroughly examine whether the jigsaw puzzle actually provides a good picture of the real economy. The puzzle here refers to the Benchmark Input-Output Table (BIOT), which provides benchmark estimates for the country's production system every five years.
Here lies the challenge for the Japanese statistics system, which is highly dispersed. The ESRI makes annual estimates by extending the benchmark JSNA estimates, which are developed by processing data from the BIOT, and makes quarterly estimates by applying the growth rates calculated from smaller sample surveys to the annual estimates. The compilation of the BIOT, which provides the basis for entire estimates, is a joint work of 10 government ministries and agencies. It is unclear whether all of the agencies sufficiently capture production activities under their jurisdictions by fully recognizing the importance of revising the benchmark and with a sense of responsibility. So heavily influenced by historical values, some agencies are not forthcoming in revising the benchmark, regardless of the rapidly changing nature of global production arrangements.
Also, the conceptual difference between the BIOT and the JSNA remains unresolved. The entire JSNA, including the annual SUT, must be redesigned so as not to put a square peg in a round hole. The JSNA benchmark revision should be more coherent and carried out, not as changes in the rules of annual estimation, but as the overhaul of the structure of and empirical foundation for the benchmark SUT and IOT. In order to fully achieve the purpose of benchmark revision, it is indispensable to consolidate functions. When the importance of benchmark revision is fully recognized, it would prompt the recognition of the need to restructure the BIOT. It takes long-term efforts to further improve GDP statistics.
* Translated by RIETI.