Economic Characteristics of Inter-Industry and Inter-Regional Spread of COVID-19 Shocks

Faculty Fellow, RIETI

The exit from the COVID-19 crisis that has dramatically changed our lives since spring last year is still not in sight, even though vaccination has started. Until the risk of an explosion in infection numbers diminishes through the establishment of herd immunity, infection prevention measures that impose constrains on social and economic activities will continue to be implemented, and policy measures to compensate for the ensuing economic losses will have to be taken.

In addition to the two declarations of a state of emergency issued by the government, various measures have been taken since it began, including asking people to stay at home, recommending that businesses employ telework, and requesting shops and restaurants to temporarily close or reduce their opening hours. Based on our experience of the COVID-19 crisis in the past one year or so, it is now necessary to evaluate the respective effects of those infection prevention measures and the scale of economic losses arising from their implementation, in order to prioritize policy options.

As a result of analysis work by a governmental panel of experts and other groups, the effects of the infection prevention measures have been identified to a significant degree. As a result, the narrowly targeted measures taken under the second declaration of a state of emergency that started in January this year had positive effects compared with the broad restrictive measures that were adopted under the first declaration of a state of emergency in spring last year.

Meanwhile, data-based fact-finding has not been sufficiently conducted regarding questions such as what scale of economic losses were caused by which restrictive measures taken amid the COVID-19 crisis, in which industries the losses were concentrated, and how the losses spread across different industries and regions through inter-industry input-output relationships and inter-regional economic transactions. As a result, it cannot be denied that little more has apparently been done than to look at isolated cases of economic loss. Unfortunately, because of this situation, it is doubtful whether policy options from the economic aspect are being wisely prioritized.

From this perspective, using the inter-prefectural input-output table, we conducted an analysis as to how the impact of the decline in economic activity in spring through summer of 2020 spread across different regions and industries (for detailed information, please see Discussion Paper 21-J-010). As for what triggered the decline in economic activity, we focused our attention on two factors: (i) a change in consumption behavior caused by stay-at-home requests, temporary store closures, and reduced opening hours; and (ii) a decrease in exports resulting from the rapid shrinkage of the global economy and trade.

The analysis based on the inter-regional input-output table examined how the impacts of the changes that triggered the decline in economic activity spread through economic transactions that were conducted across different regions and industries. More specifically, we estimated how changes that occurred in individual regions and industries spread through various industries across the nation and analyzed the agglomerated data to examine the overall situation. Based on the obtained results, it is possible to conduct a factor-by-factor analysis of the change that occurred in each industry or region and use the results to forecast the intensity of impacts that may spread across industries and regions when activity in a particular industry in a particular region is restricted as an infection prevention measure.

Here, I will explain two findings that attracted our attention. The first finding concerns the cause of the greatest decline in economic activity in 2020, which occurred in May. The state of emergency declared in April continued into May. The declaration was lifted in 39 prefectures on May 14, in the Kinki region on May 21, and in Hokkaido and the Tokyo metropolitan area on May 25. Many people probably understand that the extraordinary plunge in economic activity due to a consumption slump was attributable to the fact the period of the state of emergency included the whole of the Golden Week holiday season in late April through early May, which, in other years, would be a period of greatly increased travel and consumption.

There is no doubt that the consumption slump at that time was severe, but the depressed consumption accounts for only around half of the decline in economic activity in May. In March through May last year, as global trade shrank precipitously, Japanese exports also declined rapidly. The remaining half of the steep decline in economic activity in May last year is attributable to the decrease in exports. In the manufacturing industry, which is the main driver of Japanese exports, there are deeply integrated supply chains that flow from the upstream sector of parts and raw materials supply to the downstream sector of assembly of finished goods. Through the flow of the supply chain, an impact that originated in a particular region spread to various regions.

We also separated impact of restrictive measures on the spread domestic demand into intra-regional and extra-regional effects. The intra-regional effect as defined in our analysis refers to the effect of a domestic demand impact of restrictive measures taken in a particular prefecture spreading within the same prefecture through inter-industry input-output relationships. The extra-regional effect refers to the effect of such an impact spilling over to other prefectures through inter-industry input-output relationships. It was observed that the intra-regional effect is generally stronger than the extra-regional effect.

We obtained a similar finding when we identified the impact of restriction measures taken in Tokyo and Osaka Prefecture in May last year and examined the intensity of the impact of those measures that spread to other regions. Although restrictive measures taken in Tokyo had the greatest spillover impact on its neighbor, Kanagawa Prefecture, the intensity of that impact was only a tenth of the impact on Tokyo itself. Meanwhile, restrictive measures taken in Osaka had the greatest spillover impact on its neighbor, Wakayama Prefecture, but the intensity of that impact was around a fifth of the impact on Osaka itself.

Why is the domestic demand impact of anti-COVID-19 measures taken in a particular prefecture more likely to remain within the same prefecture than to spill over to other prefectures? That is because businesses whose sales were the hardest-hit by anti-COVID-19 restrictive measures are concentrated in the services industry, particularly the dining sector. In contrast, sales of products such as home electronics have been brisk due to "nesting" demand. Manufacturing industry sectors comprise complex webs of inter-industry input-output relationships, ranging from parts supply to assembly, and are geographically spread across large areas, but in service industry sectors, inter-industry input-output relationships are weak, and the geographical reach is small. Therefore, in the case of service industry sectors that were hit particularly hard by the COVID-19 crisis, the impact of restrictive measures taken in a particular prefecture mostly remained within the same prefecture without spreading across prefectural borders through inter-industry input-output relationships.

In light of the above mentioned two facts, the most important action that can be taken to avoid massive economic losses is to make sure that the global economy does not fall into the situation of shrinking equilibrium amid the COVID-19 crisis. Even so, it is reasonable enough to decide how to strike a balance between infection prevention and social and economic activity by conducting a comparative analysis at the prefectural level. When social and economic activity in a particular region is restricted, depending on the spread of the infection in the region, it is sufficient to identify the scale of economic losses arising from the restriction, so taking measures based on this principle of region-by-region analysis is reasonable. However, the exception to this principle is tourism-related sectors, including transportation and accommodation, where consumption activity involving the movement of people occurs. It should be kept in mind that because of the nature of these sectors, the impacts of restrictive measures taken in a certain region may spread to other regions through inter-region links based on other factors than inter-industry input-output relationships.

April 8, 2021

May 6, 2021