A re-examination of Granger causality between government expenditure and GDP

Based on data from different countries, a very strong correlation was noted between the growth rate of government expenditure and of nominal and real GDP. However, the direction of causality (which we specify as Granger causality) remains in dispute, especially regarding a stagnating economy such as Japan. This paper analyses the Granger causality between general government expenditure, nominal GDP and GDP deflators, using data from 38 OECD countries for the period 1980-2021. The preliminary results varied widely from country to country at different periods, but most results suggested the existence of Granger causality running from nominal GDP to government expenditure. In order to address the problem of spuriousness, which arises as a result of the statistics being recorded on an accrual basis, we tested for Granger causality by taking the lead variable (a variable at a later point in time) of government expenditure and found that the updated results often differed from the preliminary ones. To examine this finding in more detail, a Granger causality test was conducted using quarterly data from Japanese GDP statistics (1994-2021). One noteworthy outcome was that no Granger causality from nominal GDP to government expenditure was identified in Japan after 2008. When the lead variable of government expenditure was used as an explanatory variable, it became clear that the direction of Granger causality was unambiguously from government expenditure to nominal GDP, at least in the short term in Japan after 2008.