Fiscal multiplier in the Russo-Japanese War: A business cycle accounting perspective

In this paper, we use business cycle accounting, introduced by Chari et al. (2007, Econometrica 75 (3), 781-836) to estimate the fiscal multiplier in Japan during the Russo-Japanese War, 1904-1905. Business cycle accounting divides factors that affect the economic variables (real GNP, consumption, investment, and labor supply) into four wedges: efficiency, labor, investment, and government consumption wedges. These wedges exactly replicate allocation in the economy. Allowing for the spillover effect on other variables and wedges, this method allows us to evaluate the effect of government spending. Moreover, we do not need to specify the frictions in the model. Since the battlefields were in Korea and China, the war caused little damage to Japan’s physical capital or labor supply. On the other hand, because the military spending was huge, the Japanese government issued a large number of foreign bonds. Additionally, before the war people were not convinced that Japan and Russia would battle, and during 1904 people did not expect that the war would last for nearly two years. Given these conditions, we can use this event as a natural experiment. We define the fiscal multiplier as the fraction of an increase from the steady state in real GNP to that in government spending. Using business cycle accounting, we calculated the value of the fiscal multiplier as 0.86. We also estimated the military spending multiplier at the value of 0.81. Both figures were less than unity, indicating that the multipliers are relatively ineffective. These results are consistent with the previous literature, which estimates the multiplier using econometric models such as structural vector autoregressive (VAR) models.