Impact of Cash Transfers on Consumption during the COVID-19 Pandemic: Evidence from Japanese Special Cash Payments

Author Name HATTORI Takahiro (University of Tokyo) / KOMURA Norihiro (Kyoto University) / UNAYAMA Takashi (Faculty Fellow, RIETI)
Download / Links

This Non Technical Summary does not constitute part of the above-captioned Discussion Paper but has been prepared for the purpose of providing a bold outline of the paper, based on findings from the analysis for the paper and focusing primarily on their implications for policy. For details of the analysis, read the captioned Discussion Paper. Views expressed in this Non Technical Summary are solely those of the individual author(s), and do not necessarily represent the views of the Research Institute of Economy, Trade and Industry (RIETI).

During the COVID-19 pandemic, the Japanese government implemented a Special Cash Payment (SCP), comprising a uniform payment of 100,000 yen for each member of the resident population. Similar large-scale cash transfers were also implemented in many other countries. However, the impact of the SCP on household consumption was unclear in advance. Although it was possible that households might not increase their consumption out of fear of COVID-19, it was also possible that the SCP would significantly boost consumption due to an increasing number of households experiencing pecuniary constraints as a result of the pandemic. There was also the risk that increased household consumption would lead to the further spread of infections, depending on the type of consumption.

From this perspective, we estimate households' marginal propensity to consume (MPC) out of the SCP. We classify consumption based on the COVID-19 infection risk, estimate the MPC for each of these subcategories, and examine the ways in which households increased their consumption.

For this research, we use publicly available data from the Family Income and Expenditure Survey (FIES) published by the Statistics Bureau, Ministry of Internal Affairs and Communications of Japan. The FIES is a fundamental statistic under the Statistic Act and is regarded as the most reliable household-level expenditure data. In recent years, trends in the FIES correspond with GDP final household consumption expenditure, making it a suitable source of data for a macroanalysis of the SCP. However, the SCP is recorded in the FIES as "special income" along with other temporary sources of income. Therefore, it is necessary to take econometric concerns into account when estimating MPC.

The SCP was distributed from June to July 2020 amid dramatic changes in consumption, with the initial state of emergency declared from April to May, followed by initiatives such as the "Go To" Campaign in July. This makes it impossible to assess the effect of the SCP by simply examining household consumption time series data.

Rather, we use the differences in the timing of SCP distributions between cities to estimate the MPC. Specifically, we focus on timing differences in the start date of the acceptance of payment applications for each municipality. In the figure below (corresponding to Figure 4 in the paper), we divide the cities into two groups: those that started accepting applications for the SCP relatively early (on average before May 25; indicated by the solid line) and those that started relatively late (indicated by the broken line). We refer to these two groups as the "early group" and the "late group" respectively. The figure represents a comparison of household consumption trends between these two groups and shows how we used differences in the timing of payments to estimate the MPC.

Figure. Application Start Date and Consumption
Figure. Application Start Date and Consumption

In June, the group for which payments commenced relatively early saw a larger increase in household consumption than the late group, but this difference gradually diminished over time and was reversed in August. In other words, consumption increased earlier in cities that commenced payments earlier. This indicates the effect of the SCP in boosting consumption. In June, the early group had received of 210,000 yen as a special income, whereas the late group received only 106,000 yen; the difference in the SCP received between two groups was 104,000 yen. The difference in household consumption between the two groups in June is 9,000 yen. The MPC is the ratio of the two, 8.7% (= 9,000/104,000). In the paper, we use a more rigorous regression analysis to calculate the MPC, producing a baseline estimate of 11%.

The value of this estimate is consistent with previous studies on cash transfers in Japan. This suggests that, even during the pandemic, households responded as usual to cash transfers. Our estimate suggests that the increase in consumption in response to the SCP was mainly driven by households with liquidity constraints. It has been pointed out that the macro MPC is closely equivalent to the proportion of households with liquidity constraints, which has been estimated at approximately 13% by previous research. Our estimate of MPC, at around 10%, leads to a narrative suggesting that it was households with liquidity constraints that responded to the cash transfers.

This conclusion is highly significant for evaluating the SCP. Specifically, targeting only households with liquidity constraints could produce the same stimulus effect on consumption as a uniform benefit for the entire population. The stated purpose of the SCP was not purely to stimulate consumption. Moreover, there is no guarantee that government authorities would be able to accurately identify households with liquidity constraints when implementing a cash transfer. However, it will be difficult to justify unconditional uniform payments to the entire population when drafting cash transfer measures in the future.

Assuming that 10% of the SCP was spent for consumption, this corresponds to a boost in consumption of more than 1 trillion yen. This additional consumption may also have contributed to the spread of COVID-19. To assess this risk, we examine what goods and services were consumed by households.

By restructuring the detailed categories presented in the FIES, we are able to classify consumption according to our own criteria based on the risk of COVID-19 infection, to explore how households increased their consumption. Specifically, we classify consumption into four subcategories: spending on "face-to-face services" such as traveling and eating out, which carry a relatively higher risk; spending on "goods/services purchased at home" such as utilities and internet services, with the lowest risk; spending on "transfer payments" including gifts and remittances; and spending on other "goods/services purchased at stores." Our estimates of the MPC for each consumption subcategory are presented in the table below (corresponding to Table 3 in the paper).

Table. MPC out of SCP by Infection Risk
Table. MPC out of SCP by Infection Risk

We find that consumption of "goods/services purchased at stores" increases by about 8%, and consumption of "goods/services purchased at home" and "transfer payments" both increase by around 2%. By contrast, we do not find a statistically significant change in "face-to-face service" consumption. These results suggest that households consider the risk of infection when deciding what to consume using the SCP. Based on these changes in consumption behavior, the SCP did not significantly increase the risk of spreading COVID-19.