Literature DB >> 34230744

Assessing the impact of COVID-19 on major industries in Japan: A dynamic conditional correlation approach.

Masayasu Kanno1.   

Abstract

This study assesses the impact of the novel coronavirus disease (COVID-19) cases on the Japanese stock market. As of October 30, 2020, the cumulative number of cases in Japan has reached over one hundred thousand. COVID-19 has significantly affected both the lifestyle and the economy in Japan. First, this study develops composite stock indices by industry sector and prefecture, taking into consideration the effects of the increase in infections on industries and firms in the core prefectures. Second, this study investigates the dynamic conditional correlations between the composite stock index returns and the increment in COVID-19 cases using dynamic conditional correlation multivariate GARCH models. Finally, it can contribute to financial research in terms of coexistence of regional business economies with COVID-19.
© 2021 Elsevier B.V. All rights reserved.

Entities:  

Keywords:  COVID-19; Composite stock index; Dynamic conditional correlation (DCC); Multivariate GARCH; Sector and regional analysis

Year:  2021        PMID： 34230744      PMCID： PMC8252700          DOI： 10.1016/j.ribaf.2021.101488

Source DB:  PubMed          Journal:  Res Int Bus Finance        ISSN： 0275-5319


Introduction

The novel coronavirus disease (COVID-19; caused by the SARS-CoV-2 virus) has caused an unprecedented pandemic. Globally, approximately 40 million people have been infected with COVID-19 as of October 17, 2020, with more than 1.1 million fatalities, according to the latest data by Johns Hopkins University (JHU, 2020). It has caused great confusion in every country. In the field of public health, infections are categorized into three types: endemic, epidemic, and pandemic, with the degree of severity increasing in that order. Finally, a pandemic has a similar meaning to an epidemic, but it refers to those infections that have the most severe effects on a global scale (Vynnycky and White, 2010). On March 11, 2020, the World Health Organization (WHO) declared COVID-19 a global pandemic (WHO, 2020). Additionally, in terms of risk management in the field of finance, the pandemic, same as the global financial crisis (2007–2009), belongs to the category of “systemic risks.” It is therefore recognized that COVID-19 should be subject to systemic risk management. In Japan, a mass outbreak of COVID-19 took place among the 3711 passengers and crew of the cruise ship, Diamond Princess, in February 2020. A total of 712 patients were confirmed positive for the virus and the threat of COVID-19 became a serious concern (NIID, 2020). As measures against COVID-19, the Japanese government has implemented trial and error style of policies up until now, although the contents of these policies have been questioned. For example, at a press conference on May 4, although former Prime Minister Shinzo Abe indicated that the anti-influenza drug Avigan would receive regulatory approval, it is yet to be approved as of the end of October 2020. Considering that vaccines have not yet been developed for past pandemic diseases, such as the Severe Acute Respiratory Syndrome (SARS; outbreak in 2002) and the Middle East Respiratory Syndrome (outbreak in 2012), the delay in regulatory approval for Avigan is expected. Examining the chain of events up until now, on April 7, 2020, former Prime Minister Abe officially issued a “declaration of a state of emergency” and expressed that “the Japanese economy is facing its greatest crisis since the end of the Second World War.” Additionally, on May 4, the declaration was extended for an additional month along with commentary that “it is necessary to prepare for a fight that extends for some period of time.” However, on May 14, the declaration was lifted in 39 prefectures, and the occasion was marked as “a day to start to return to a new everyday life.” On May 25, the declaration was wholly lifted with the assertion that “the strength of the Japanese model has been shown” (source: June 13, Nihon Keizai Shimbun). Additionally, the “Novel Coronavirus Expert Meeting” was also abolished on June 24. However, thereafter, a second wave of infection, originating from “entertainment and social activities that happen in the evening in bars and clubs” of Shinjuku Ward in Tokyo, have been rapidly increasing the number of infected persons in a wide range of age groups (Fig. 1 ). On July 10, the infections hit a record high.
Fig. 1

Cumulative cases by prefecture as of 10:00 AM, October 16, 2020. Notes: A prefecture colored in black is Tokyo. In addition, from north to south, prefectures in brown are Saitama, Kanagawa, Aichi, Osaka, and Fukuoka, and ones in pink are Hokkaido, Chiba, Kyoto, Hyogo, and Okinawa.

Cumulative cases by prefecture as of 10:00 AM, October 16, 2020. Notes: A prefecture colored in black is Tokyo. In addition, from north to south, prefectures in brown are Saitama, Kanagawa, Aichi, Osaka, and Fukuoka, and ones in pink are Hokkaido, Chiba, Kyoto, Hyogo, and Okinawa. Evaluating the policies taken by the government at the present time, when the prospects for the resolution of the COVID-19 crisis do not look positive, is not the goal of this study. Yet the effect of COVID-19 on the financial economy may be even greater than the effect of the global financial crisis. From a global perspective, for example, as of November 6, 2020, the operating profit for Toyota, which has production and sales locations worldwide, is expected to decrease by 46% for the fiscal year ending March 31, 2021. This is also true for other automobile manufacturers. The effects on the automobile industry, which includes many subsidiaries, sub-subsidiaries, and affiliates are significant. In contrast, in the domestic economy, 616 instances of COVID-19-related bankruptcies, 534 legal liquidations, and 82 suspensions of business have been identified. The rank by industry is as follows: “restaurants and dining” (88 cases), “hotels and inns” (59 cases), “apparel and general retail” (44 cases), “construction” (37 cases), “food wholesalers” (36 cases), “apparel wholesalers” (25 cases), and so on (source: Teikoku Databank Corporate, 2020 as of 4:00 PM, October 16). With the end of the COVID-19 crisis nowhere in sight, the effects are increasingly extending into the long term. Additionally, in the food wholesale industry, there are concerns regarding the increase of COVID-19 infections in distribution centers, which are controlled by avoiding the three conditions for transmission, or “Three C's,” closed spaces (closed spaces with poor ventilation), crowded places (where many people congregate), and close-contact settings (such as close conversations in which individuals are close enough to touch each other). The construction industry has been affected by the closure of job sites and delays in obtaining construction materials caused by disruption in distribution and the supply chain. The food production industry has been affected by school closures and the suspension of events, causing many firms to go bankrupt. Accordingly, this study analyzes the variations of sector stock index by prefecture as a proxy of regional firm economies. In another perspective, following the increase in COVID-19 infections, financial researchers worldwide have grown concerned about the effects of COVID-19 on the financial economy. These researchers have put together special COVID-19-related reports in academic journals, and many are calling for social contributions (Goodell, 2020). In Section 2, a literature review of existing research is conducted. In Section 3, the analysis approach and data used in the study are examined. In Section 4, the analysis results are shown and discussed, and Section 5 shows the conclusion.

Literature review

In finance, existing research related to pandemics is almost unknown. This is because although pandemics are a type of macro stress that severely impact the economy, pandemics that had a global effect on the financial economy like COVID-19 have not occurred in the past. Research on pandemics as infectious diseases has taken place for many years, and Vynnycky and White (2010) is the basic text for mathematical modeling. Also, Kiss et al. (2018) elucidate a mathematical model for infections based on the propagation of a virus from person to person through a complex network. Though not regarded as pandemic research, a small number of studies using infectious disease models do exist, such as Kanno (2015), in which we examine a succession of bankruptcies in the Japanese banking system following the global financial crisis by applying a Susceptible-Infected-Recovered Dead (SIRD) model. The SIRD model adds a “dead” state to a Susceptible-Infected-Recovered model, a typical mathematical model used for infectious diseases. In contrast, since March 2020, COVID-19 and financial market related articles were published in line with the COVID-19 outbreak. In terms of stock markets, Goodell and Huynh (2020) analyzed the abnormal returns of 49 industrial sectors from December 9, 2019–February 28, 2020. Shehzad et al. (2020) employed the asymmetric power GARCH model and found that COVID-19 substantially harms United States’ (US) and Japan's market returns. Mazur et al. (2021) investigated the US stock market performance during the crash of March 2020, triggered by COVID-19. Akhtaruzzaman et al. (2021) showed that dynamic conditional correlations between Chinese and G7 stock returns, financial and nonfinancial alike, increased significantly during the COVID-19 period. Zaremba et al. (2020) demonstrated that non-pharmaceutical interventions significantly increase equity market volatility. Ashraf (2020) examined the stock markets’ response to the COVID-19 pandemic using daily COVID-19 confirmed cases and deaths and stock market returns data from 64 countries over the period January 22, 2020 to April 17, 2020. Okorie and Lin (2021) investigated the fractal contagion effect of the COVID-19 pandemic on the stock markets. Also, in terms of market analysis using news analytics tool, Cepoi (2020) offered novel empirical evidence on the relationship between COVID-19 related news and stock market returns across the top six countries most affected by the pandemic and showed the COVID-19 news-related variables from the RavenPack analytics tool. Haroon and Rizvi (2020) analyzed the relationship between sentiment generated by coronavirus-related news and volatility of equity markets using the same tool. Shi and Ho (2020) examined the impact of public news sentiment on volatility states of firm-level returns using the same tool.

Analysis approach and data

This study analyzes the impact of COVID-19 on regional stock index returns as proxies of regional firm economies using the dynamic conditional correlations.

Analysis approach

Composite stock index. We compose prefectural stock index using the stock prices of firms composed of the Tokyo Stock Price Index (TOPIX). At present, TOPIX is used as a representative stock index to express economic movements in Japan as a whole. TOPIX includes all domestic common stocks listed in the First Section of the Tokyo Stock Exchange. It is assigned a market capitalization of 100 as of January 4, 1968, to which all subsequent market capitalizations are indexed (Tokyo Stock Exchange, 2020). Additionally, the 33 Tokyo Stock Exchange Sector Indices are stock indices that divide this index into 33 different industry sectors (see Table A.8). By rearranging firm stock prices applicable to a sector and a prefecture, it is possible to develop stock indices arranged both “by prefecture” and “by industry.”
Table A.8

TOPIX sector indices (33 sectors).

Sector NoSector nameShort name
1Fishery, Agriculture &ForestryFAF
2FoodsFoods
3MiningMining
4Oil and Coal ProductsOilCoal
5ConstructionConst
6Metal ProductsMetal
7Glass and Ceramics ProductsGlassCera
8Textiles and ApparelsTextAppa
9Pulp and PaperPulpPaper
10ChemicalsChemicals
11PharmaceuticalPharma
12Rubber ProductsRubber
13Transportation EquipmentTransEquip
14Iron and SteelIronSteel
15Nonferrous MetalsNonMetals
16MachineryMachinery
17Electric AppliancesEleAppli
18Precision InstrumentsPrecInstr
19Other ProductsOtherP
20Information &CommunicationInfoCom
21ServicesServices
22Electric Power and GasEPGas
23Land TransportationLandTrans
24Marine TransportationMarineTrans
25Air TransportationAirTrans
26Warehousing and Harbor TransportationWaHaTrans
27Wholesale TradeWholeTrade
28Retail TradeRetailTrade
29BanksBanks
30Securities and Commodities FuturesSecComFut
31InsuranceInsurance
32Other Financing BusinessOtherFB
33Real EstateRealEstate
Today, the majority of stock indices worldwide are weighted by market capitalization such as TOPIX, the S&P 500 Index, and the NASDAQ Composite Index of the United States. Weighting by market capitalization means that the total amount of a listed stock's market capitalization (a number that represents the firm value and is obtained by multiplying the stock price by the number of listed shares) is calculated by dividing the total market capitalization of the index at a certain point in time. This is compared to the value at a past point in time so that it expresses how much the market capitalization has increased or decreased at the time of calculation, thereby expressing the change in the price of the stock as an asset. The calculation formula of composite prefectural index for prefecture and sector at time is as follows:where and are the adjusted number of shares issued1 and the stock price for firm belonging to prefecture and sector at time , respectively. is a set for firms belonging to prefecture and sector . is a standard value for the index related to prefecture and sector at time 0. Dynamic conditional correlation. To calculate dynamic conditional correlation (DCC), we introduce the multivariate GARCH model proposed by Engle (2002). The model is a dynamic multivariate regression model, in which the conditional variances and covariances of the errors follow an autoregressive moving average structure. The DCC multivariate GARCH model uses a nonlinear combination of univariate GARCH models with time-varying cross-equation weights to model the conditional covariance matrix of the errors. In the DCC multivariate GARCH model, DCC is defined as follows:where the diagonal elements of a time-varying conditional covariance matrix of the disturbances, and , follow univariate GARCH processes, are the off-diagonal elements of the matrix, and hence follows the dynamic process. In DCC GARCH model, the conditional variance (; : the number of dependent variables) evolves according to a univariate GARCH model of the formwhere for each series , is a vector of parameters, is a vector of independent variables including a constant term, is a standardized disturbance with mean zero and variance one, the are ARCH parameters, and the are GARCH parameters. and are the number of lags for the ARCH term and the GARCH term, respectively. Using Stata 14, we estimate DCC GARCH models. To this end, for the ARCH term, for the GARCH term, suppression of the constant term in the mean equation, and the assumed distribution such as Gaussian distribution or -distribution for the errors need to be set. Additionally, in terms of the optimization algorithm in the multivariate regression model, Berndt–Hall–Hall–Hausman algorithm is adopted. Tolerance parameters are set to defaults: 1e6 in coefficients vector, 1e5 in Hessian scaled gradient; 1e7 in log-normal likelihood (Gould et al., 2010)). To reduce the calculation burden, 5–9 sectors are grouped in addition to the daily increment of COVID-19 cases as the dependent variables in a model.

Data

Stock data. This study needs daily data enough to calculate dynamic conditional correlations. To this end, Pronexus Inc.'s eol database has been used to obtain the stock market prices and shares issued for the relevant firms. There are 2169 firms that make up TOPIX. Most of the listed firms’ headquarters are concentrated in Tokyo. Only Tokyo has firm headquarters for every industry, which makes it possible to examine the relationship between the stock return movement and the increment in COVID-19 cases (i.e., number of patients admitted to the hospital, etc.). The First Section of the Tokyo Stock Exchange is primarily composed of large firms, and firms in industries such as manufacturing do not necessarily have their employees concentrated at their headquarters. However, this is one of the few promising methods that can be used in understanding the relationship between the increase in COVID-19 cases and regional business economies. The objects of analysis in this case are industries with a large number of firms that have been driven to bankruptcy or suspension by the effects of COVID-19, specifically “hotels and inns” (included in the “service industry,” which is one of the 33 sector indices), “restaurants and dining” (included in the “retail business” sector index), “apparel and general retail” (covered by the “textile goods” and “retail industry” sector indices, respectively), as well as “transportation equipment” and the “air transport industry,” which have experienced major reductions in demand. The industries that are subject to analysis do not necessarily align perfectly with the division of the 33 TOPIX sector indices, but they do allow a general understanding to be obtained. In contrast, as an industry that is expected to contribute to resolving the increase in COVID-19 cases, “pharmaceutical firms” (the “medical products” sector index), which develop vaccines and provide testing equipment and are expected to see improved performance, are also considered. The summary statistics are shown in Table 1 . By using square-root-t method and assuming 250 business days in a year, annualizing the standard deviation of stock market returns results in a 27% volatile figure.
Table 1

Summary statistics pertaining to daily returns for the period from January 6, 2020 to October 16, 2020.

VariableObs.25%Median75%MaxMeanS.D.
Stock market returns2169−0.014−0.0000.0139.5270.0000.017
TSE Sector index returns33−0.009−0.0000.0100.113−0.0000.018

Note: Abbreviations: Obs., observations; S.D., standard deviation. 25% and 75% indicate the first quartile and the third quartile, respectively.

Summary statistics pertaining to daily returns for the period from January 6, 2020 to October 16, 2020. Note: Abbreviations: Obs., observations; S.D., standard deviation. 25% and 75% indicate the first quartile and the third quartile, respectively. COVID-19 data. In terms of COVID-19 related statistics in Japan, it became possible to collect COVID-19-related data with any degree of accuracy in a compiled form beginning from mid-March. At first, figures were publicly reported according to infection statuses reported from each prefecture based on Article 12 of the Infectious Disease Act and scrutinized by the Ministry of Health, Labor, and Welfare. Since May 8, the reporting method was changed to public reports that compile numbers independently reported by each prefecture. Thus, the data for new COVID-19 cases by prefecture since mid-March are necessary to be put together. The data are publicly available at the site of a firm of J.A.G JAPAN Corp. (J.A.G JAPAN, 2020). Fig. 2 reveals the number by day of the week pertaining to the new COVID-19 cases by prefecture. Because of the large variation in the number of the tested people by day of the week, the weekly moving average number is instead used to smooth the case curves. Additionally, to conduct correlation analysis, the increments of new COVID-19 cases are calculated corresponding to the daily stock index returns. It is impossible to calculate the daily change rates corresponding to the day without any new cases.
Fig. 2

New COVID-19 cases by prefecture. Notes: Prefecture No; 1: Hokkaido; 11: Saitama; 12: Chiba; 13: Tokyo; 14: Kanagawa; 23: Aichi; 26: Kyoto; 27: Osaka; 28: Hyogo; 40: Fukuoka.

New COVID-19 cases by prefecture. Notes: Prefecture No; 1: Hokkaido; 11: Saitama; 12: Chiba; 13: Tokyo; 14: Kanagawa; 23: Aichi; 26: Kyoto; 27: Osaka; 28: Hyogo; 40: Fukuoka. In terms of large cumulative COVID-19 cases (Fig. 1), we focus on the 10 prefectures, that is, Hokkaido, Saitama, Chiba, Tokyo, Kanagawa, Aichi, Kyoto, Osaka, Hyogo, and Fukuoka,2 and analyze the effect of COVID-19 on regional business economies. In Hokkaido, initially, infections were concentrated among Chinese tourists, but soon after several clusters, mass outbreaks of patients occurred in Sapporo City which houses the entertainment districts. The number of cases in Tokyo and the three surrounding prefectures (Saitama, Chiba, and Kanagawa) is likewise conspicuous and, needless to say, represents the center of the Japanese economy. Osaka is likewise the commercial and industrial heart of the Kansai region, and large corporations are more concentrated here than in the neighboring prefectures. Additionally, Osaka has many entertainment districts that allow for socialization and contact between people, resulting in a large number of hospitalizations and so on. Kyoto is the most prosperous city among Japanese sightseeing cities. Hyogo is also famous for Kobe City as an international city. Fukuoka has the largest commercial area in Kyushu region.

Analysis results and discussion

This section presents the empirical analysis results. Table 2 indicates summary statistics for daily stock index returns by prefecture and industry. Approximately 39% of these, or 854 firms, are headquartered in Tokyo, the second is Osaka with 206, and the third is Aichi with 107. The mean value of composite index returns nationwide is 0.052%, whereas the mean value of the index returns for 10 selected prefectures (i.e., 192 sectors) is 0.066%. This difference shows that the effect on prefectures with more cases is larger than on the others.
Table 2

Summary statistics for daily stock indices returns by prefecture and sector.

Pref1111111111111111111111111111111111111111
Sec12202122272829261012131416171819202123282933
Obs.112321711231414521211811
Med0.00−0.00−0.00−0.00−0.000.000.000.000.00−0.000.000.00−0.000.00−0.000.00−0.000.000.00−0.000.000.00−0.000.00
Max0.080.090.100.070.070.110.060.070.090.120.070.100.040.120.090.070.100.160.090.210.150.060.070.15
Mean−0.00−0.000.00−0.00−0.00−0.000.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.00−0.000.000.00−0.000.00
S.D.0.030.030.030.020.020.030.010.020.030.030.020.020.010.040.030.020.030.020.030.040.030.010.030.03
Pref121212121212121212121313131313131313131313131313
Sec2516172021232829331234567891011121314
Obs.22121227313558631913198772141611
Med−0.00−0.00−0.00−0.000.000.00−0.000.00−0.000.00−0.00−0.00−0.01−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00
Max0.130.080.090.080.100.070.080.080.070.080.060.060.090.060.070.090.090.060.100.060.060.080.090.12
Mean−0.00−0.00−0.00−0.000.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00
S.D.0.030.030.020.020.020.020.030.020.020.030.020.010.030.020.020.020.020.020.020.020.020.020.030.03
Pref131313131313131313131313131313131313131414141414
Sec1516171819202122232425262728293031323312567
Obs.145776172718161716831119781221824512332
Med−0.00−0.000.000.000.000.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.00−0.00−0.00−0.00
Max0.090.070.070.080.060.060.090.080.060.090.100.070.060.060.060.080.070.050.110.070.120.100.090.13
Mean−0.00−0.00−0.000.00−0.000.000.00−0.00−0.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.00−0.00−0.00−0.00
S.D.0.020.020.020.020.020.010.020.020.020.030.030.020.020.020.020.020.020.020.020.020.020.030.030.03
Pref141414141414141414141414141414142323232323232323
Sec810131415161718192021232627283325678101213
Obs.1481211153114425111342721113
Med−0.000.00−0.00−0.00−0.00−0.000.000.00−0.000.000.00−0.000.000.00−0.000.000.00−0.000.00−0.000.000.00−0.01−0.00
Max0.070.080.090.090.120.070.070.070.080.070.090.110.090.070.050.220.090.070.070.080.040.100.110.09
Mean−0.000.00−0.00−0.00−0.00−0.000.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.000.00−0.00−0.000.00−0.00−0.00
S.D.0.020.020.030.030.030.020.020.030.020.020.020.020.030.020.020.070.020.020.020.020.010.020.030.02
Pref232323232323232323232323232323262626262626262626
Sec141516171819202122232627282933268101113161718
Obs.2112431611223914311223113132
Med−0.000.00−0.00−0.000.000.000.00−0.00−0.00−0.00−0.00−0.000.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.000.000.00−0.00
Max0.090.130.100.080.070.130.070.060.090.090.110.090.060.090.100.090.090.060.070.110.120.070.070.13
Mean−0.00−0.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.00
S.D.0.030.020.020.020.020.030.030.020.020.020.020.020.020.020.030.030.030.020.020.030.030.030.020.02
Pref262626262626262626262727272727272727272727272727
Sec192021232627282932332567891011121314151617
Obs.2221112112714451123411713215
Med0.000.00−0.000.000.000.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.00−0.00−0.00−0.00−0.000.000.00
Max0.060.100.090.170.100.080.060.080.090.060.060.090.070.080.060.090.070.080.100.080.080.100.060.08
Mean0.000.00−0.000.00−0.000.00−0.000.000.00−0.00−0.00−0.00−0.000.00−0.00−0.000.00−0.00−0.000.00−0.00−0.000.000.00
S.D.0.020.030.030.060.030.020.020.020.030.030.020.020.020.020.020.020.020.020.030.020.020.020.020.02
Pref272727272727272727272727272828282828282828282828
Sec181920212223262728293031332456781011121314
Obs.311122743322221951121272435
Med−0.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.000.00−0.00−0.000.000.00−0.00−0.00−0.000.00−0.00−0.00−0.00
Max0.080.070.070.050.070.090.080.060.050.100.090.100.070.070.140.100.090.110.110.060.070.090.100.09
Mean−0.00−0.000.00−0.00−0.00−0.00−0.000.00−0.00−0.000.000.00−0.000.00−0.000.000.00−0.00−0.00−0.00−0.01−0.00−0.00−0.00
S.D.0.020.020.020.020.020.020.020.020.020.030.020.030.020.020.030.020.030.020.030.020.090.020.030.03
Pref282828282828282828284040404040404040404040404040
Sec15161719212326272833235717202122232728293233
Obs.285122255131223242266211
Med0.00−0.000.00−0.010.00−0.00−0.000.000.00−0.000.00−0.00−0.000.00−0.00−0.000.000.00−0.00−0.000.00−0.000.00−0.00
Max0.110.090.090.160.090.110.110.070.070.110.060.090.090.100.110.160.090.110.090.060.090.080.080.15
Mean0.00−0.000.00−0.000.00−0.00−0.00−0.000.00−0.000.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.00−0.000.00−0.000.00−0.00
S.D.0.020.020.020.040.030.020.020.040.020.030.020.020.020.020.030.030.030.020.020.020.020.020.020.04

Notes: Abbreviations: Pref, prefecture; Obs., observations; Med, median; S.D., standard deviation. Pref Number: 1: Hokkaido; 11: Saitama; 12: Chiba; 13: Tokyo; 14: Kanagawa; 23: Aichi; 26: Kyoto; 27: Osaka; 28: Hyogo; 40: Fukuoka. Refer to Table A.8 pertaining to Sector Number. Total number of observations is 1476.

Summary statistics for daily stock indices returns by prefecture and sector. Notes: Abbreviations: Pref, prefecture; Obs., observations; Med, median; S.D., standard deviation. Pref Number: 1: Hokkaido; 11: Saitama; 12: Chiba; 13: Tokyo; 14: Kanagawa; 23: Aichi; 26: Kyoto; 27: Osaka; 28: Hyogo; 40: Fukuoka. Refer to Table A.8 pertaining to Sector Number. Total number of observations is 1476. Table 3, Table 4, Table 5, Table 6, Table 7 denote the optimization results of DCC multivariate models. Sector names in the Tables are referred to “Short name” column in Table A.8. The brute-force optimization trials are conducted in terms of for the ARCH term, for the GARCH term, suppressing the constant term in the mean equation, and the assumed distribution for the errors (i.e., Gaussian distribution or t-distribution) in Eq. (3). In case of high volatile increment in cases, -distribution is apt to be fitted and GARCH models are not necessarily optimized (i.e., ). For example, the models for Tokyo are GARCH(2,1) (i.e., ) are obtained, whereas the models for Osaka are ARCH(2) (i.e., ) are obtained.
Table 3

Estimation results of DCC multivariate GARCH models (1/5).

Hokkaido (8 sectors)
Saitama (16 sectors)
Chiba (10 sectors)
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|
FAFFoodsPrecInstrFoodsServices
cons0.0010.196archarcharcharch
FAFL1.0.427***0.001L1.0.304***0.009L1.0.0580.495L1.0.0710.447
archL2.0.0230.633L2.0.020.782cons0.001***0cons0.001***0
L1.0.509***0.002cons0***0cons0***0ConstLandTrans
L2.0.1610.276MetalOtherParcharch
cons0***0archarchL1.0.120.334L1.0.1960.297
FoodsL1.0.0520.52L1.0.2630.193cons0.001***0cons0.001***0
cons0.0010.574L2.0.0540.348L2.0.459*0.051MachineryRetailT
Foodscons0.001***0cons0***0archarch
archChemicalsInfoComL1.0.0590.585L1.0.797***0
L1.0.1080.38archarchcons0.001***0cons0***0
L2.0.1810.142L1.0.2670.104L1.0.0340.715EleAppliBanks
cons0.001***0L2.0.232*0.061L2.0.0850.456archarch
InfoComcons0***0cons0.001***0L1.0.0530.454L1.0.2820.159
cons0.005**0.044RubberServicescons0.001***0cons0.001***0
InfoComarcharchInfoComRealEstate
archL1.0.669***0.002L1.0.0630.438archarch
L1.0.0230.766L2.0.140.103L2.0.0720.351L1.0.0320.401L1.0.1130.418
L2.0.1160.344cons0***0cons0.001***0cons0.001***0cons0.001***0
cons0.001***0TransEquipLandTransCaIncrCaIncr
Servicesarcharcharcharch
cons0.003.L1.0.263*0.06L1.0.347**0.015L1.0.707*0.069L1.0.4280.118
ServicesL2.0.1510.278L2.0.0370.652cons1.943***0.003cons2.288***0
archcons0.000***0cons0.001***0
L1.0.043.IronSteelRetailT
L2.0.087.archarch
cons0***0L1.0.369**0.028L1.0.0930.443
EPGasL2.0.0150.784L2.0.1450.239
cons0.0010.225cons0***0cons0***0
EPGasMachineryBanks
archarcharch
L1.0.2390.105L1.0.12*0.086L1.0.020.701
L2.0.27*0.091L2.0.0080.908L2.0.365**0.023
cons0***0.004cons0***0cons0***0
WholeTEleAppliRealEstate
cons0.002*0.069archarch
WholeTL1.0.0340.652L1.0.047***0
archL2.0.0620.158L2.0.1080.227
L1.0.672***0cons0.001***0cons0.001***0
L2.0.0050.335CaIncrCaIncr
cons0***0archarch
RetailTL1.0.2060.14L1.0.1870.172
cons0.002**0.038L2.0.120.494L2.0.1030.505
RetailTcons2.754***0cons2.911***0
arch
L1.0.0520.504
L2.0.1270.114
cons0***0
Banks
cons0.0020.119
Banks
arch
L1.0.403**0.026
L2.0.0470.733
cons0***0
CaIncr
cons0.10.416
CaIncr
arch
L1.0.3240.134
L2.0.465**0.024
cons0.955***0.003

***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr: cases increment; cons: constant.

Table 4

Estimation results of DCC multivariate GARCH models (2/5).

Tokyo (33 sectors)
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|
FAFPulpPaperEleAppliAirTransCaIncr
archarcharcharcharch
L1.0.1110.356L1.0.1060.293L1.0.0220.666L1.0.1230.303L1.0.358.
L2.0.0330.837L2.0.0410.54L2.0.0370.501L2.0.040.721L2.0.203***0
garchgarchgarchgarchgarch
L1.0.0580.851L1.1.53***0L1.0.0840.787L1.0.662***0.008L1.1.272.
cons0.001***0cons00.442cons0.001***0cons0.003***0cons4.586.
FoodsChemicalsPrecInstrWhHaTrans
archarcharcharch
L1.0.0270.718L1.0.1070.111L1.0.0270.497L1.0.318*0.094
L2.0.2740.107L2.0.0560.318L2.0.0210.626L2.0.1520.219
garchgarchgarchgarch
L1.0.486*0.074L1.1.213***0.01L1.0.331.L1.0.633**0.03
cons00.204cons00.479cons0***0cons00.16
MiningPharmaOtherPWholTrade
archarcharcharch
L1.0.0890.294L1.0.1780.496L1.0.0130.947L1.0.271**0.031
L2.0.0030.962L2.0.1680.414L2.0.0290.836L2.0.277*0.066
garchgarchgarchgarch
L1.0.787***0L1.0.2370.467L1.0.8060.108L1.0.1220.717
cons0.002***0cons0*0.091cons00.466cons0***0.01
OilCoalRubberInfoComRetailT
archarcharcharch
L1.0.077**0.034L1.0.1510.113L1.0.0250.442L1.0.0680.446
L2.0.114***0L2.0.0230.723L2.0.1***0L2.0.2060.178
garchgarchgarchgarch
L1.0.2580.18L1.0.350.421L1.0.751***0L1.0.3750.218
cons0***0cons00.116cons0***0cons00.126
ConstTransEquipServicesBanks
archarcharcharch
L1.0.498**0.024L1.0.0560.268L1.0.303**0.05L1.0.2140.187
L2.0.0650.491L2.0.0930.323L2.0.0790.46L2.0.0150.904
garchgarchgarchgarch
L1.0.363*0.063L1.0.47*0.073L1.0.0710.838L1.0.3260.278
cons0***0.008cons0.001***0cons0.001***0cons0*0.063
MetalIronSteelEPGasSecComFut
archarcharcharch
L1.0.404**0.021L1.0.188**0.045L1.0.988***0.004L1.0.1260.315
L2.0.0290.845L2.0.0040.977L2.0.115***0L2.0.2450.304
garchgarchgarchgarch
L1.0.2130.406L1.0.0750.851L1.0.0720.163L1.0.5890.251
cons0**0.012cons0.001**0.046cons0***0cons00.489
GlassCeraNonMetalsLandTransInsurance
archarcharcharch
L1.0.26**0.026L1.0.1210.152L1.0.1860.344L1.0.0120.926
L2.0.0660.52L2.0.0250.696L2.0.823**0.02L2.0.0180.905
garchgarchgarchgarch
L1.0.1910.592L1.0.764***0L1.0.0970.39L1.0.4360.34
cons0**0.011cons0.001***0cons0***0.001cons0.001**0.013
TextAppaMachineryMarineTransOtherFB
archarcharcharch
L1.0.1050.413L1.0.0790.109L1.0.0110.928L1.0.1570.162
L2.0.0770.541L2.0.287***0.001L2.0.2780.263L2.0.262**0.023
garchgarchgarchgarch
L1.0.0150.955L1.0.5740.22L1.0.4260.322L1.0.562***0
cons0.001***0cons0.001**0.011cons0.0010.185cons0.001***0
CaIncrCaIncrCaIncrRealEstate
archarcharcharch
L1.0.173.L1.0.148.L1.0.6850.42L1.0.199*0.092
L2.0.113***0L2.0.104***0L2.0.170.682L2.0.0530.839
garchgarchgarchgarch
L1.1.045.L1.1.037.L1.0.760.164L1.0.6590.37
cons0.522.cons0.325.cons31.185.cons00.654

***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant.

Table 5

Estimation results of DCC multivariate GARCH models (3/5).

Kanagawa (21 sectors)
Aichi (23 sectors)
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|
FAFTransEquipInfoComFoodsIronSteelEPGas
cons0***0archarcharcharcharch
FoodsL1.0.266**0.021L1.0.030.424L1.0.2980.159L1.0.0840.406L1.0.484**0.029
cons0.001***0L2.0.071***0L2.0.067***0L2.0.0360.783L2.0.0330.692L2.0.1420.481
Constcons0.001***0cons0***0garchcons0.002***0.001cons0***0.002
cons0.001***0IronSteelServicesL1.0.454*0.085NonMetalsLandTrans
Metalarcharchcons0.001***0archarch
cons0.001***0L1.0.416***0.001L1.0.303**0.026ConstL1.0.662**0.044L1.0.1050.549
GlassCeraL2.0.0530.597L2.0.291*0.06archL2.0.2070.407L2.0.441*0.077
cons0.001***0cons0.001***0cons0***0L1.0.2540.21cons0***0cons0***0.003
TextAppaNonMetalsLandTransL2.0.0590.601MachineryWhHaTrans
cons0.001***0archarchgarcharcharch
ChemicalsL1.0.252*0.062L1.0.175**0.028L1.0.709***0L1.0.1820.215L1.0.1610.386
cons0***0L2.0.0070.853L2.0.362***0.003cons0.001***0L2.0.078***0L2.0.322*0.077
CaIncrcons0.001***0cons0***0Metalcons0.001***0cons0***0
cons10.709***0MachineryWhHaTransarchEleAppliWholeT
archarchL1.0.2970.221archarch
L1.0.0870.262L1.0.085***0L2.0.1350.358L1.0.148***0.001L1.0.399***0.003
L2.0.0540.321L2.0.1440.314garchL2.0.020.857L2.0.0370.534
cons0***0cons0.001***0L1.0.685***0cons0.001***0cons0***0
EleAppliWholeTcons0.001***0PrecInstrRetailT
archarchGlassCeraarcharch
L1.0.295*0.096L1.0.0250.709archL1.0.0920.699L1.0.1790.133
L2.0.0090.897L2.0.229**0.045L1.0.0150.832L2.0.0310.846L2.0.0080.918
cons0***0cons0***0L2.0.117*0.097cons0.001***0cons0***0
PrecInstrRetailTgarchOtherPBanks
archarchL1.0.240.345archarch
L1.0.0160.633L1.0.1680.113cons0.001***0L1.1.125**0.034L1.0.1410.14
L2.0.309**0.018L2.0.1620.2TextAppaL2.0.0250.743L2.0.230.147
cons0***0cons0***0archcons0.001***0cons0***0
OtherPRealEstateL1.0.449*0.073InfoComRealEstate
archarchL2.0.2870.146archarch
L1.0.19**0.039L1.0.23**0.037garchL1.0.0430.755L1.0.0950.264
L2.0.1080.41L2.0.274**0.029L1.0.722*0.065L2.0.2990.195L2.0.0520.514
cons0***0cons0.003***0cons00.414cons0.001***0cons0.001***0
CaIncrCaIncrChemicalsServicesCaIncr
archarcharcharcharch
L1.0.217*0.054L1.0.272**0.042L1.0.2080.422L1.0.1070.272L1.0.596***0
L2.0.198**0.039L2.0.17*0.066L2.0.2430.303L2.0.1050.504L2.0.418***0.005
cons5.503***0cons5.476***0garchcons0.001***0cons0.599***0
L1.0.416**0.044CaIncr
cons0**0.041arch
RubberL1.1.513***0.001
archL2.1.042**0.015
L1.0.0980.573cons0.721*0.098
L2.0.0570.649
garch
L1.0.638***0.001
cons0.004***0
TransEquip
arch
L1.0.0210.814
L2.0.0410.382
garch
L1.0.2780.39
cons0.001***0
CaIncr
arch
L1.1.164***0.006
L2.1.046**0.013
garch
L1.0.0890.345
cons0.4290.119

***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant.

Table 6

Estimation results of DCC multivariate GARCH models (4/5).

Kyoto (19 sectors)
Osaka (27 sectors)
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|
FoodsPrecInstrFoodsTransEquipEPGas
archarcharcharcharch
L1.0.427***0.001L1.0.304***0.009L1.0.1190.144L1.0.19*0.056L1.0.429**0.037
L2.0.0230.633L2.0.020.782L2.0.165**0.043L2.0.466**0.02L2.0.301**0.027
cons0***0cons0***0cons0***0cons0***0cons0***0.001
MetalOtherPConstIronSteelLandTrans
archarcharcharcharch
L1.0.0520.520L1.0.2630.193L1.0.0850.366L1.0.1970.222L1.0.292**0.028
L2.0.0540.348L2.0.459*0.051L2.0.0340.613L2.0.0730.6L2.0.62***0.001
cons0.001***0cons0***0cons0***0cons0.001***0cons0***0
ChemicalsInfoComMetalNonMetalsWhHaTrans
archarcharcharcharch
L1.0.2670.104L1.0.0340.715L1.0.2570.191L1.0.0850.325L1.0.214*0.053
L2.0.232*0.061L2.0.0850.456L2.0.10.213L2.0.0120.835L2.0.044**0.041
cons0***0cons0.001***0cons0***0cons0.001***0cons0***0
RubberServicesGlassCeraMachineryWholeT
archarcharcharcharch
L1.0.669***0.002L1.0.0630.438L1.0.507***0.005L1.0.2550.113L1.0.38**0.012
L2.0.140.103L2.0.0720.351L2.0.0120.859L2.0.1820.272L2.0.1820.131
cons0***0cons0.001***0cons0***0cons0***0.005cons0***0
TransEquipLandTransTextAppaEleAppliRetailT
archarcharcharcharch
L1.0.263*0.06L1.0.347**0.015L1.0.158*0.09L1.0.0420.631L1.0.0520.578
L2.0.1510.278L2.0.0370.652L2.0.080.342L2.0.1160.523L2.0.1380.198
cons0***0cons0.001***0cons0***0cons0***0cons0***0
IronSteelRetailTPulpPaperPrecInstrBanks
archarcharcharcharch
L1.0.369**0.028L1.0.0930.443L1.0.1280.286L1.0.257**0.018L1.0.2080.233
L2.0.0150.784L2.0.1450.239L2.0.1230.369L2.0.0660.282L2.0.126**0.019
cons0***0cons0***0cons0***0cons0***0cons0.001***0
MachineryBanksChemicalsOtherPSecComFut
archarcharcharcharch
L1.0.12*0.086L1.0.020.701L1.0.544***0.006L1.0.0720.386L1.0.061***0
L2.0.0080.908L2.0.365**0.023L2.0.1120.36L2.0.0210.785L2.0.010.759
cons0***0cons0***0cons0***0cons0***0cons0.001***0
EleAppliRealEstatePharmaInfoComInsurance
archarcharcharcharch
L1.0.0340.652L1.0.047***0L1.0.1470.244L1.0.1050.516L1.0.1560.264
L2.0.0620.158L2.0.1080.227L2.0.283**0.049L2.0.0510.682L2.0.1890.141
cons0.001***0cons0.001***0cons0***0cons0***0cons0.001***0
CaIncrCaIncrRubberServicesRealEstate
archarcharcharcharch
L1.0.2060.14L1.0.1870.172L1.0.1070.494L1.0.0660.573L1.0.0010.987
L2.0.120.494L2.0.1030.505L2.0.315*0.079L2.0.0420.502L2.0.1770.102
cons2.754***0cons2.911***0cons0.001***0cons0***0cons0***0
CaIncrCaIncrCaIncr
archarcharch
L1.0.615***0.005L1.0.736***0.005L1.0.789***0.006
L2.0.859***0.006L2.0.809***0.009L2.0.922***0.003
cons0.949*0.093cons0.7550.163cons0.5190.159

***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant.

Table 7

Estimation results of DCC multivariate GARCH models (5/5).

Hyogo (21 sectors)
Fukuoka (14 sectors)
t(3), Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
t(3), Pr > chi2 = 0.000
Gaussian, Pr > chi2 = 0.000
Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|Coef.P>|z|
FoodsPharmaOtherPFoodsPrecInstr
archarcharcharcharch
L1.0.2290.11L1.0.1340.7L1.0.1280.458L1.0.83**0.012L1.0.304***0.009
L2.0.264*0.082L2.0.183***0L2.0.0470.79L2.0.0530.66L2.0.020.782
cons0***0cons0.001***0cons0.001***0cons0.001***0cons0***0
OilCoalRubberServicesMetalOtherProd
archarcharcharcharch
L1.0.5130.103L1.0.1560.44L1.0.2030.203L1.0.0250.858L1.0.2630.193
L2.0.0310.892L2.0.1110.636L2.0.0130.909L2.0.0660.301L2.0.459*0.051
cons0.001***0cons0.001***0cons0.001***0cons0.002***0cons0***0
ConstTransEquipLandTransChemicalsInfoCom
archarcharcharcharch
L1.0.563**0.045L1.00.999L1.0.2490.172L1.0.587*0.057L1.0.0340.715
L2.1.123***0.005L2.0.0130.957L2.0.2160.257L2.0.0410.826L2.0.0850.456
cons0**0.029cons0.002***0cons0***0cons0***0cons0.001***0
MetalIronSteelWhHaTransRubberServices
archarcharcharcharch
L1.0.0110.94L1.0.304*0.087L1.0.1270.276L1.0.887**0.024L1.0.0630.438
L2.0.3190.101L2.0.157***0L2.0.192**0.05L2.0.2450.148L2.0.0720.351
cons0.001***0cons0.001***0cons0***0cons0.001***0cons0.001***0
GlassCeraNonMetalsWholeTradeTransEquipLandTrans
archarcharcharcharch
L1.0.449*0.074L1.0.061***0L1.1.712***0L1.0.3640.159L1.0.347**0.015
L2.0.477*0.1L2.0.0250.735L2.0.0010.249L2.0.0260.897L2.0.0370.652
cons0.001***0.001cons0.001***0cons0***0cons0***0cons0.001***0
TextAppaMachineryRetailTradeIronSteelRetailT
archarcharcharcharch
L1.0.0360.829L1.0.0910.61L1.0.120.347L1.0.666*0.051L1.0.0930.443
L2.0.2330.332L2.0.1690.27L2.0.0350.724L2.0.0360.639L2.0.1450.239
cons0.002***0cons0.001***0cons0***0cons0.001***0cons0***0
ChemicalsEleAppliRealEstateMachineryBanks
archarcharcharcharch
L1.0.0140.901L1.0.060.678L1.0.572***0.008L1.0.2080.155L1.0.020.701
L2.0.1230.39L2.0.010.934L2.0.0050.956L2.0.0460.767L2.0.365**0.023
cons0.001***0cons0.001***0cons0.001***0cons0.001***0cons0***0
CaIncrCaIncrCaIncrEleAppliRealEstate
archarcharcharcharch
L1.0.7370.118L1.1.017**0.049L1.0.1840.165L1.0.0110.898L1.0.047***0
L2.0.5570.311L2.0.4270.432L2.0.1210.434L2.0.214***0L2.0.1080.227
cons4.46**0.033cons3.893*0.052cons2.784***0cons0.001***0cons0.001***0
CaIncrCaIncr
archarch
L1.1.203*0.057L1.0.1870.172
L2.1.1050.185L2.0.1030.505
cons2.6830.196cons2.911***0

***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant.

Estimation results of DCC multivariate GARCH models (1/5). ***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr: cases increment; cons: constant. Estimation results of DCC multivariate GARCH models (2/5). ***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant. Estimation results of DCC multivariate GARCH models (3/5). ***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant. Estimation results of DCC multivariate GARCH models (4/5). ***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant. Estimation results of DCC multivariate GARCH models (5/5). ***, **, * represent statistical significance at 1%, 5%, and 10% levels, respectively. L1 and L2 are a first-order lag and a second-order lag, respectively. Sector name refers to “Short name” in Table A.8. CaIncr, cases increment; cons, constant. In terms of robustness check, the header in Table 3, Table 4, Table 5, Table 6, Table 7 reports the assumed distribution for the errors and the Wald test for the goodness of fit of a model against the null hypothesis that all the coefficients in the mean equations are zero, where the null hypothesis is rejected at the 5% level. Additionally, Table 3, Table 4, Table 5, Table 6, Table 7 show the statistical significance for the coefficients on the variables of the models. Fig. 3 indicates the DCCs between the daily return of prefectural stock index and the daily increment of COVID-19 cases in 10 prefectures. As shown in these panels, the DCCs were negative in almost all sectors and prefectures for the period and the variations were volatile in many sectors and prefectures. Hence, prefecture industries with positive DCCs and some noticeable sectors with negative DCCs are examined as follows:
Fig. 3

Dynamic conditional correlation between prefectural stock index and the increment of COVID-19 cases.

Dynamic conditional correlation between prefectural stock index and the increment of COVID-19 cases. In Hokkaido, only retail trade industry had slightly positive DCCs from August, 2020 to September, 2020. Regarding the retail industry in Hokkaido, large-scale supermarkets and furniture stores such as Nitori Holdings and Aeon Hokkaido were able to operate even under the declaration of a state of emergency. These firms actually had more customers than usual and their stock prices increased. In Chiba, the DCCs for foods and retail trade industries were positive in a consistent manner. This results from the government's declaration of a state of emergency which thereby increases time spent at home. In Saitama, the DCCs for transportation industry as well as information and communication industry were positive. The former is due to the increasing products sale by mail-order. The latter resulted from the increase for remote work and students’ staying at home. In Tokyo, the DCCs for information and communication industry as well as electric power and gas industry were positive because of remote work and staying at home. Precision instrument industry's increase reflected the increasing demands in home electrical appliances by remote work and medical equipment by COVID-19 cases. Regarding the pharmaceutical industry in Tokyo, contrary to expectations, the correlation coefficient was negative. The number of COVID-19 cases reached a peak during Golden week (i.e., from the end of April to the beginning of May) and thereafter decreased (Fig. 3), owing to new business styles such as remote work, staggered working hours, online meeting, and suspensions of various types of shops and restaurants. In Kanagawa, the DCCs for the information and communication industry were positive. In Aichi, the DCCs for foods and warehousing and harbor transportation industries were positive. In contrast, the DCCs for transportation equipment and rubber products industries as well as iron and steel and machinery industries were negative. Because Toyota motors is located in Aichi, its performance reflects the related industries. In Kyoto, foods, iron and steel, land transportation, and information and communication had positive DCCs; especially, land transportation, including SG holdings with a home delivery subsidiary: Sagawa Express, which had a high business demand. In Osaka, electric appliance, transportation equipment, and information and communication industries had positive DCCs, reflecting the increased demand by remote work and staying at home. By contrast, for the textile production industry, the DCCs were consistently negative, because the number of COVID-19 cases peaked in the middle of the declaration of a state of emergency and hereafter decreased sharply (Fig. 3). However, firms such as Shikibo and Kurabo Industries were fortunate to significantly benefit from the unique increase in demand for masks. Shikibo published test results showing that textile materials treated with their anti-viral processing, called “Flutect,” were effective against COVID-19, thereby gathering the interest of individual investors. Kurabo Industries has responded to government requests not only for masks but also for the production of materials such as medical gowns. In Hyogo, pharmaceutical industry's positive correlation coefficient was as expected. Additionally, foods, nonferrous metals, electric appliances, services, wholesale trade, retail trade is mentioned as industries of positive DCCs. In Fukuoka, foods, iron and steel, land transportation, and information and communication industries had positive DCCs.

Conclusion

This study contributed to analyzing the dynamic conditional correlations of COVID-19 cases on the Japanese stock market. First, we developed stock index by prefecture and sector using the data for all domestic common stocks listed in the First Section of the Tokyo Stock Exchange. This stock index represents the regional economic circumstances by sector. Second, investigating the dynamic correlation between the stock index returns as a proxy of performance of firms with their headquarters in major prefectures and the daily increment of COVID-19 cases using DCC multivariate GARCH models, contributes to financial research related to pandemics by allowing the visualization of the effects of COVID-19 on regional firms’ economies. The financial related data for this study are limited to those of the daily stock market, because it hasn’t been long since the outbreak of COVID-19. Thus, for future studies, the use of corporate financial data, macroeconomic data, and other market data such as credit default swaps is recommended.
  2 in total

1.  The sum of all SCARES COVID-19 sentiment and asset return.

Authors:  Md Tanvir Hasan
Journal:  Q Rev Econ Finance       Date:  2022-08-18

2.  Risk contagion of COVID-19 in Japanese firms: A network approach.

Authors:  Masayasu Kanno
Journal:  Res Int Bus Finance       Date:  2021-07-07
  2 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.