Fluctuations in the Number of Stores by Industry During the COVID-19 Pandemic Based on Japanese Phone Book Entries

Economic activities can be affected by various exogenous events, including health crises such as the worldwide outbreak of the novel coronavirus disease (COVID-19) [1‐4]. The first case of COVID-19 in Japan was reported in January 2020, followed by several large outbreaks with clusters of the infection. As of April 2022, the cumulative number of cases of COVID-19 in Japan has exceeded 7.6 million. In its efforts to combat the spread of the virus, the Japanese government has declared a state of emergency four times: from April to May 2020, January to March 2021, April to June 2021, and July to September 2021. In each case, the people were asked to refrain from going out after 8 p.m. and to limit their social interactions. Unsurprisingly, these measures have had a serious impact on economic activities throughout the country.

It is estimated that 700 Japanese firms with total indebtedness greater than 10 million yen went bankrupt due to the pandemic in 2020, followed by 1700 more firms in 2021, according to Teikoku Databank, Ltd [5]. Measures against such widespread company bankruptcies are required for economic stability [6]. Currently, precise information on bankruptcies in Japan is mostly limited to the company information provided by domestic research firms such as Teikoku Databank and Tokyo Shoko Research [7]. Little if any detailed information regarding the changes in the number of stores/facilities in Japan during the COVID-19 pandemic appears to be available

In this study, we used telephone directory data containing not only phone numbers but also information on the location and industrial category of each store/facility in order to estimate the changes in the number of stores/facilities by industry before and after the COVID-19 outbreak. Using such data, we were able to count the number of records for stores/facilities in a certain prefecture or in a certain industrial category. We approximated the number of stores/facilities using the number of records in the data and calculated the value at each of 12 time points. By evaluating the temporal change in the number or change rate of the number, we were able to identify areas where the number of stores/facilities in various industrial categories decreased significantly after the outbreak.

The rest of this paper is organized as follows. In Sect. 2, we explain the data and the method. We show the results on the change in the number of stores/facilities in Sect. 3, followed by a discussion in Sect. 4.

To conduct our study, we analyzed the NTT Hello Page phone book provided by Eins Co., Ltd. (https://teldata.jp/). This dataset is updated every few months. For each store or facility, the telephone directory data record includes the store or facility’s telephone number, name, zip code, prefecture, city, address, and industrial category. Multiple industrial categories are sometimes assigned to a single store or facility. Thus, the dataset provides an approximation of the number of existing stores and facilities both temporally and geographically. Each industry code is represented by an alphabetical character, and each industrial category is further classified into detailed industrial categories, each of which has a specific code called a detailed industry code, denoted by a three-digit number. For example, the detailed industry code A100 denotes a “dental clinic” in the medical industrial category.

In the directory data analyzed, there were 25 industry codes and 332 detailed industry codes, as shown in Table 1. We analyzed the data at 12 time points, from September 2018 to May 2021. Each time point t is listed in Table 2. The number of stores at time t is denoted as \(N_t\). Here, each time point t at which the information is aggregated in the data is not evenly distributed, whereas the data are provided regularly. Approximately five million total records are represented at each time point (Table 2). We can observe a trend before the outbreak of COVID-19 in which the number of stores/facilities recorded in the data generally decreases. However, it should be noted that although such a decrease in the total number of records can be observed, this does not necessarily mean an increase in the number of stores that have closed, since the telephone directory dataset provides an incomplete picture of the number of existing stores/facilities. For example, it does not include stores/facilities that are not registered in the telephone directory data or those without a landline telephone. However, if there is a significant difference in the number of phone book entries before and after the COVID-19 outbreak, it can be expected that this difference reflects, at least to some extent, the impact of the outbreak.

We investigated the change in the number of stores/facilities recorded in the Japanese telephone directory before and after the outbreak of COVID-19 in Japan and found a significant decrease in the number of leisure/amusement stores/facilities such as karaoke venues and travel agencies after the outbreak. Such declines were prominent in Ibaraki, Osaka, and Hyogo prefectures, whereas the decline was less apparent in Tochigi. We also found that the number of restaurants and travel agencies significantly decreased after the outbreak in urban areas such as Osaka.

Our approach enabled us to capture the influence of the COVID-19 outbreak on stores/facilities that could be considered closed spaces, crowded places, or close-contact settings, the so-called three Cs referred to as conditions to be avoided in the government policy for reducing the spread of the virus. We showed that the increase rate \(N_{12}/N_6\) of the number of facilities in category U (education and culture) after the outbreak is lower than \(N_6/N_1\), the rate before the outbreak, in urban areas like Tokyo and Osaka. The previous studies have revealed the effect of school closures on suppressing the spread of viral diseases [8, 9]. In Japan, school closures were widely implemented as a measure to prevent the spread of COVID-19 [10]. Industrial category U includes the subcategories of “after school activities and lessons.” Therefore, a possible reason for the low values of \(N_{12}/N_6\) in U is that facilities for these activities and lessons were closed as well as schools, which had likely made their management more difficult. Moreover, to control the concentration of tourists at potentially crowded sightseeing spots, the government’s Go To Travel project, originally designed to promote tourism, was cancelled nationwide from the end of December 2020 to early January 2021. Consistent with the suspension of the project, we found that the ratio \(N_t/N_1\) of T671 (travel agencies) decreased in May 2021 in Tokyo and Osaka. Finally, our results showed a significant decrease in the number of karaoke facilities after the outbreak in many prefectures, as the Japanese people generally believe that singing in indoor spaces may increase the risk of infection [11]. In sum, our results would seem to confirm the impact of the COVID-19 outbreak on stores/facilities associated with the three Cs.

Our results also appear to confirm the effect of the subsidies made available for the management of stores/facilities during the pandemic. We show a relatively small temporal decline in the number of stores/facilities in October 2020, particularly for the categories of restaurants, sports and leisure, and travel. This likely reflects the impact of the small business sustainability subsidy issued by the Japan Chamber of Commerce and Industry (JCCI) in August 2020, which was seen as a way of preventing or reducing store bankruptcies and closures.

Our results also captured the influence of the emergency declarations on the management of stores/facilities in Japan. In our detailed examination of temporal changes in the number of stores/facilities \(N_t/N_{t-1}\) for S645 (“karaoke, bowling, game centers, discos, etc.”) and T671 (“travel agency”), we observed sharp decreases in the values after March 2021 in Ibaraki, Osaka, Hyogo, and Tokyo prefectures. These results can be partially attributed to a series of government emergency declarations, including an emergency declaration issued in January 2021 in Tokyo, Saitama, and other neighboring prefectures of Ibaraki, and another issued in April 2021 for six cities/towns in Ibaraki reducing the open hours of stores and requesting residents to refrain from going out. A state of emergency was also declared in Osaka and Hyogo in mid-January 2021 and extended through early March. The decreases in \(N_t/N_{t-1}\) in our results would seem to reflect the impact of these repeated emergency declarations on stores/facilities and their role in restricting the flow of customers.

Of course, our study is not without limitations. The most significant limitation stems from the fact that the telephone directories that served as our principal data source are unable to precisely show the number of bankruptcies over the period studied. Nevertheless, many of our results based on changes in the number of stores/facilities recorded in the directories can be interpreted as reflecting the impact of COVID-19-related events on stores/facilities management. Our analysis did not include stores/facilities that were not registered in the telephone directory data or those without a landline telephone. In seeking to reduce the bias in our data, we based our analysis on the ratios of the number of stores/facilities recorded in the data rather than on the absolute values. Using these ratios, we evaluated whether the number of stores/facilities recorded after the outbreak was significantly lower than the number recorded before the outbreak, under conditions in which both records should be edited with similar biases. We also confirmed the frequency of updates of the data by prefecture. Acknowledging that any analysis requires a consideration of the various biases specific to the data, we nevertheless suggest that the impact of the COVID-19 pandemic can be captured by the type of phone book analysis employed in our study.

The impact of the COVID-19 pandemic on the economy and society has been explored extensively from various industrial perspectives. In this study, we examined changes in the number of stores/facilities registered in the Japanese telephone directory over a period extending from before to after the outbreak of the pandemic. Specifically, we examined the change in the number of stores/facilities by prefecture, industry, and detailed industry. Many of the results were consistent with previous findings regarding the impact of COVID-19 on various industries. Similar to our attempt in this study, it seems possible to evaluate the impact of an epidemic using an approximation based on the number of records found in phone book data.

Future studies may be able to determine the distinctive characteristics of stores/facilities that experience difficulties in their management during a pandemic by analyzing in greater detail information such as the location of closed and newly opened stores/facilities. The results of such an evaluation can then be compared with those from a previous study [12] that visualized stay-at-home levels by prefecture and municipality in order to assess the characteristics of regional pandemic responses.