Volatility Spillover Effects during Pre-and-Post COVID-19 Outbreak on Indian Market from the USA, China, Japan, Germany, and Australia

Round 1

Reviewer 1 Report

The paper is well-designed and well-written.

I do have a number of suggestions. I also have an important issue that has to be addressed:

1) Literature survey could be more focused. As it is, it covers unrelated areas and is not relevant to the spillover the authors are covering.

2) At the bottom of page 2, there is information on total market capitalization in India. It shows that due to Covid, it went from $2.16 T to $1.3 T, and then after the Indian government’s actions to contain the epidemic, it went up to $2.3 T.

While this is an impressive comparison, it has to be put in perspective to make sure it was due to the actions of the government. I suggest that the authors should add comparison information using a few other countries with much fewer government acts to contain the pandemic. For instance, how was it in Sweeden that the government did not do anything before, during, and after the pandemic – comparable time frame.

How did the total market capitalization in the US changes during the same period?

3) Similarly, it seems like to long-shot to title it fourth in terms of the stock market capitalization growth rate of 17.4 percent next to China, the USA, and Hong Kong. It might be difficult to explain why the Indian government's actions made such a big difference. The difference is maybe more attributable to distinct characteristics of the Indian stock exchange unrelated to pandemic-related actions.

Also, in addition to the growth rate, the authors may supply the decline rate (contraction) in market capitalization before the government actions were induced so one can compare the elasticities attributable to the pandemic-related arrangements.

4) I prefer to see a clear paragraph showing the results that EGARCH is the proper model for this. Yes, the authors do explain in several places in the paper but a definitive paragraph summarizing the results with numbers and advocating EGARCH instead of the regular GARCH model will help readers to clarify the model selection. The underlined asymmetry issue is hard to follow and clarify in the present form.

 

5) IMPORTANT: The authors highlight that the “study investigated whether specific unexpected shocks have any significant impact on the volatility on the markets and if so, whether is any spillover from one market to another” While COVID was an unexpected shock, it was a shock to ALL markets simultaneously. Spillovers are evaluated when there is a shock to ONE market and the OTHER ONE is impacted by the ONE that got the shock. As a result, the very foundation of the study is not well-defined.

The authors MUST explain why it is still considered a volatility spillover effect when Covid had a simultaneous impact on all the countries included in the study.

Author Response

We thank you for reviewing the manuscripts submitted by us. We truly appreciate the time and effort invested in helping us improve the paper. We have made a sincere effort to answer all the points raised by you. Point by point response is provided below:

Comments and Suggestions for Authors

The paper is well-designed and well-written.

I do have a number of suggestions. I also have an important issue that has to be addressed:

• Literature survey could be more focused. As it is, it covers unrelated areas and is not relevant to the spillover the authors are covering.

 

Thank you for your suggestion. The research paper is mainly focused on two aspects viz., volatility spillover and impact of covid-19 on the stock markets. Thus, the literature comprised of the papers on recent work in the area of volatility spill over and event study methodology. Few papers which fall away from these two topics have been removed from page 4. Few new additional papers on volatility spill over have been added in the literature review section on pages 3 and 4. Further, to make the review more robust, we have outlined the topics on spillover and event studies separately.

2) At the bottom of page 2, there is information on total market capitalization in India. It shows that due to Covid, it went from $2.16 T to $1.3 T, and then after the Indian government’s actions to contain the epidemic, it went up to $2.3 T.

While this is an impressive comparison, it has to be put in perspective to make sure it was due to the actions of the government. I suggest that the authors should add comparison information using a few other countries with much fewer government acts to contain the pandemic. For instance, how was it in Sweeden that the government did not do anything before, during, and after the pandemic – comparable time frame. How did the total market capitalization in the US changes during the same period?

In order to explain the reasons for increase in market capitalization in India with other countries considered in the study with government playing a very key role, we investigated stringency index developed by the University of Oxford. At the end of page 2 and beginning of page 3, two paragraphs have been introduced to compare the trend with other countries

3) Similarly, it seems like to long-shot to title it fourth in terms of the stock market capitalization growth rate of 17.4 percent next to China, the USA, and Hong Kong. It might be difficult to explain why the Indian government's actions made such a big difference. The difference is maybe more attributable to distinct characteristics of the Indian stock exchange unrelated to pandemic-related actions.

Also, in addition to the growth rate, the authors may supply the decline rate (contraction) in market capitalization before the government actions were induced so one can compare the elasticities attributable to the pandemic-related arrangements. (whether recovery is due to govt. action or attributable to other factors)

Thank you for the suggestion. Yes, it’s true that it is very difficult to separate the impact of Govt policies during pandemic and other distinct characteristics of Indian Stock Exchange. It will need another study. However, when you look at the ‘stringency index’ before and after the lockdown when the Govt policies were introduced the recovery is substantial. Thus, the recovery may be attributable to the government incentives provided to stabilize the economy. We have accordingly removed the claim.

We collected information which may signify the possible reasons for increase in market capitalization. Several measures were implemented by the Central Bank and the government during the first wave of the COVID-19 epidemic to boost economic activity. According to Mastertrust, the overall package which came out to Rs 20,97,053 crore, included the Rs 1.92 lakh crore stimulus from measures that were announced by the government such as the Pradhan Mantri Garib Kalyan Package worth Rs 1.7 lakh crore. Therefore, in the first wave, from March 2020 to November 2020, the Nifty index made a bottom of 7,511 and a high of 13,145.85.

Indian Benchmark stock index Sensex tanked from 41,257 points on 14^th feb 2020 to 27,590 on 3^rd April 2020 which is 33.12% fall. However, there was a sharp recovery in Sensex where the market has rebounded to 47,869 points at the end of Dec 2020. (73.50% rise).

India’s ‘Aatmanirbhar Bharat’ packages and PLI schemes for various critical sectors too have pumped fuel to the market (Source: “Indian equities running a marathon; Sensex doubled despite Covid-led shock”, Business Standard, March 25, 2022)

 Since points 2 & 3 are related, we addressed it together. One lakh is 100000 and one crore is 10000000 (10 million).

4) I prefer to see a clear paragraph showing the results that EGARCH is the proper model for this. Yes, the authors do explain in several places in the paper but a definitive paragraph summarizing the results with numbers and advocating EGARCH instead of the regular GARCH model will help readers to clarify the model selection. The underlined asymmetry issue is hard to follow and clarify in the present form.

 In page 16, we have included a paragraph advocating the results of EGARCH. 

5) IMPORTANT: The authors highlight that the “study investigated whether specific unexpected shocks have any significant impact on the volatility on the markets and if so, whether is any spillover from one market to another” While COVID was an unexpected shock, it was a shock to ALL markets simultaneously. Spillovers are evaluated when there is a shock to ONE market and the OTHER ONE is impacted by the ONE that got the shock. As a result, the very foundation of the study is not well-defined.

The authors MUST explain why it is still considered a volatility spillover effect when Covid had a simultaneous impact on all the countries included in the study.

The covid-19 was observed to spread among various countries of the world at some time intervals which can be differentiated by the intensity of cases in respective countries. Thus, the spillover can be examined from the country of origin to other countries. The spread of Covid 19 was not found to be simultaneous in nature and thus studies have considered this event for spillover effects between various countries.

The few recent studies on spillover are given below.

The recent study done by Yongmin Zhang, Jiaying Mao (June 2022), research paper titled “COVID-19′s impact on the spillover effect across the Chinese and U.S. stock markets” examined the cross-market spillover effect from China to U.S. stock market. The results revealed that there was an asymmetric transmission of return shocks to a stronger degree from the Chinese stock market to the U.S. market rather than from the opposite direction, and this asymmetric spillover effect during the COVID-19 spread period in China was three to five times stronger than during the pre-COVID-19 period and the period when COVID-19 was contained in China.  

Reviewer 2 Report

(See attached file)

Comments for author File: Comments.pdf

Author Response

We thank you for investing your valuable time in reading the manuscript and providing us suggestion to improve the paper. We have made all efforts to response to each of the points raised by you. Thanks once again for your constructive feedback. 

Reviewer – 2

A reviewer report on the paper:

Volatility Spillover effects during Pre-and-Post Covid-19 Outbreak on Indian Market from USA, China, Japan, Germany, and Australia This paper examines the transmission and volatility spillover from the five major economies to Indian stock markets during the pre and post covid-19 periods. The authors demonstrate that during the pre-covid and post-covid outbreak phases, ASX 200 and DAX are a strong influencers of Sensex performance during the pre-Covid epidemic timeframe. During the post-Covid outbreak phase, however, all foreign stock indices except DAX are strong influencer of Sensex performance. Thus, India's stock market integration with these economies varied before and after the Covid 19 outbreak. The research follows the standard EGARCH procedure to estimate the model. The contribution of this paper is marginal, and it should not be published in its current form. My comments are as follows.

1. Why market participants concern only the volatility spillover rather than the stock return volatility. In fact, the stress may be on the downside risk, not the volatility.

Thank you for the comments. In the EGARCH model, we have analyzed both return spillover and volatility spillover. However, based on the objectives of the paper we have emphasized on the latter part in-depth. Moreover, the volatility spillover during crisis periods indicates possible ‘contagion’ impact that amplifies the volatility and exacerbates the stress in the financial system (Reference: “Volatility spillovers between Forex and Stock markets in India”, Sudarsana sahoo & others, RBI Occasional papers, Vol. 38, No.1&@, 2017)

 

2. In a globally integrated system, theories or findings provided by King and Wadhwani (1990), Bekaert et al. (2005) and Chen et a., 2018) should be mentioned and discussed.

The papers mentioned with few other papers in the literature have been added on page 2.

King & Wadhwani (1990) investigated stock market crisis during 1987. Bekaert et.al (2005) examined whether the Asian and Mexican crisis in the late 1980s and early 1990s caused the Contagion effect or not. Chen et al. (2018) developed a novel technical analysis-based method for stock market forecast that assist investors in making stock investment decisions.

 

 

3. Break point is achieved by minimizing Dickey-Fuller t-statistic. This statistic differs from Bai and Perron Multiple Structural Test. Explain or test for comparison.

Application of these tests depend on factors like number of breaks present in the data, null hypothesis being tested and presence of trend in the data. In this study, Dicky-Fuller break-point test is used as we look for specific structural break caused by COVID-19. Other tests like Bai and Perron  (1998) multiple breakpoint test may be appropriate when we suspect that there is more than one structural break in the data that are unable to easily explain. Explanation has been now included on page 5.

References: Bai, J., & Perron, P. (1998). Estimating and testing linear models with multiple structural changes. Econometrica, 66, pp. 47-78.

4. The last term of equation (2) is strange. ??????(???????), where is subscript. What is this meant?

However, in Table 5, it shows: δ*RES_FOREIGN MARKET INDICES. I cannot find them to be consistent. Is the residual of foreign market indies as an independent variable in the conditional variance? Why no time indicator involves? It seems to me that a lagged residual squared for the foreign stock return should be considered

The methodology on EGARCH was followed using the paper by Alok Kumar et.al (2007), which examined the volatility spillover between various markets. In the study which considers EGARCH (1, 1) model, they considered only residuals of other markets instead of squared residual, since EGARCH, by definition, ensures that volatility is positive. The time indicator has been provided in the equation. We have changed the equation number to 4.

5. The authors mention that this study examines the volatility spillover from the four top economics to Indian stock markets. However, the title and the related tables contain five economies.

Thank you. We have corrected this. We considered top four economies (based on nominal GDP) in the world and Australia as a fifth economy.

 

 

 

 

 

 

 

 

 

Reviewer 3 Report

This paper examines volatility spillover effects from 5 global stock markets to the Indian stock market. The paper is overall well written. My comments are summarized below.

1.    It is not quite rigorous to say that we are now in a post-COVID period. In fact, the pandemic is still seriously affecting all major countries. In particular, at the end of Sep-2021 (the end of investigated sample period), it was the rise of Omicron globally. You may have to more comprehensively discuss why the post-COVID period defined in this paper is appropriate.

2.    The methodology employed in this paper could be improved. Spillover effect is more widely examined using a multivariate framework in the literature. At the very least, a multivariate model may work as a robustness check to the baseline result. Also, there are other possibilities other than EGARCH to be considered, even for a univariate model.

3.    Some illustrations could be revised. It is usually not professional for academic papers to have screenshots from Eviews directly (e.g. Fig 3), especially we do not need to see ACFs of lags up to 30 (or so). Also, please keep a consistent number of digits to be reported in all tables. Usually, 3 or 4 decimal places is sufficient. In Table 5, again, it is unprofessional to use Eviews code rather than presenting the original equation in math symbols. In Table 1, since all data are daily, the label of “daily” seems redundant.

Author Response

We thank you for investing your valuable time in reading the manuscript and providing us suggestion to improve the paper. We have made all efforts to response to each of the points raised by you. Thanks once again for your constructive feedback. 

 

Reviewer 3

1. It is not quite rigorous to say that we are now in a post-COVID period. In fact, the pandemic is still seriously affecting all major countries. In particular, at the end of Sep-2021 (the end of investigated sample period), it was the rise of Omicron globally. You may have to more comprehensively discuss why the post-COVID period defined in this paper is appropriate.

Most people who develop COVID-19 fully recover by September 2021, but current evidence suggests approximately 10%-20% of people experience a variety of mid- and long-term effects after they recover from their initial illness. These mid- and long-term effects are collectively known as post COVID-19 condition or “long COVID.” (Source: Coronavirus disease (COVID-19): Post COVID-19 condition, World Health Organisation, dated 16th Dec 2021).

2. 
   The methodology employed in this paper could be improved. Spillover effect is more widely examined using a multivariate framework in the literature. At the very least, a multivariate model may work as a robustness check to the baseline result. Also, there are other possibilities other than EGARCH to be considered, even for a univariate model.

Yes, we agree many other univariate models and multivariate models were widely employed in assessing spillover effect. The methodology on EGARCH was followed using the paper by Alok Kumar et.al (2007), which examined the volatility spillover between various markets which was appropriate for analysing for the current context.

3. 
   Some illustrations could be revised. It is usually not professional for academic papers to have screenshots from Eviews directly (e.g. Fig 3), especially we do not need to see ACFs of lags up to 30 (or so). Also, please keep a consistent number of digits to be reported in all tables. Usually, 3 or 4 decimal places is sufficient. In Table 5, again, it is unprofessional to use Eviews code rather than presenting the original equation in math symbols. In Table 1, since all data are daily, the label of “daily” seems redundant.

Thank you. We have removed the output from eviews and mentioned in the results in sentences. The codes of Eviews have been removed and only the diagnostic check results are emphasized. .  

Round 2

Reviewer 2 Report

The authors made some changes but not all the comments I made. The authors tried to avoid the question of downside risk spillover, which may be more important than the volatility spillover. The very reason is that the volatility is based on the premise that return fluctuations are symmetrical.  If I am an investor, don’t I worry the downside risk spillover? 

There are some minor errors need to be addressed:

·     In equation (4), what is meant by (error -1) in the third and fourth terms?  Should -1 be part of subscript?

·         In the same equation, why “??????(???????)” contains no time subscript?

·         In Table 3, SSE Composite R, Prob. = 0.0624**, is this P-value significant at the 5% level?

·       (Adj.) R-squared should be kept in two decimal points as in conventional approach.

·         No robustness test is provided.  

Author Response

Thank you very much for helping us improve our paper. Your inputs have been of great help in furthering our research. Appreciate the time that you have invested in providing us suggestions and ideas for future research.

The authors made some changes but not all the comments I made. The authors tried to avoid the question of downside risk spillover, which may be more important than the volatility spillover. The very reason is that the volatility is based on the premise that return fluctuations are symmetrical.  If I am an investor, don’t I worry the downside risk spillover? 

Thank you very much for allowing us to investigate the paper from the angle of the downside risk spillover among the various markets. We examined several studies conducted in the past and during the period of Covid-19 pandemic. The point raised is important to examine the consequences of the pandemic at the micro and macro levels of the countries along with its impact on the investors’ portfolio decisions. Various value-at-risk models have been used to examine the downside risk spillovers in the studies considered during 2021 and 2022. We have included the references for these studies on pages 2, 3 and 5 as well.  However, our study primarily differentiates itself from the other studies by capturing information transmission and volatility spillover from the five major economies to Indian stock markets specifically during the pre and post covid-19 periods. Although including downside risk spillover is important to the investors, we feel that it will go beyond the objective of our paper. With your permission, we would like to consider the issue of downside risk for our future research papers. Study of downside risk has been added as part of future research on page 13 as well.

There are some minor errors need to be addressed:

•  In equation (4), what is meant by (error -1) in the third and fourth terms?  Should -1 be part of subscript?

Thank you very much for keen observation into this detail. In Equation (4), the (error-1) has been corrected and made part of the subscript.

• In the same equation, why “????(???????)” contains no time subscript?

     

       With regard to this question, we had considered several papers and references. Most of the previous studies considered on EGARCH(1,1) have very less emphasis on spillover effects. Thus, in order to examine the volatility spillover, we considered methodology from the paper cited below (equation 5, page. 348) to be more appropriate. In this paper, the equation for residuals is without time subscript. By default, though it implies to include  _{????(???????)(t)} in the equation(4), we restricted ourselves to the model of the study.

Mishra, A.K., Swain, N. and Malhotra, D.K. (2007) ‘Volatility Spillover between Stock and Foreign Exchange Markets: Indian Evidence’, International Journal of Business, Vol. 12, No. 3, pp. 343-358.

• In Table 3, SSE Composite R, Prob. = 0.0624**, is this P-value significant at the 5% level?

Thank you for your suggestion. The correction has been done as the p-value is significant at the 10% level.

•      (Adj.) R-squared should be kept in two decimal points as in conventional approach.

The stated correction has been made.

• No robustness test is provided.  

      The points mentioned on robustness test has been well considered. The diagnostic tests for heteroscedasticity and serial correlation were conducted and the results are shown in Table 6. We found no significant issues with respect to heteroscedasticity and serial correlation.

Reviewer 3 Report

I thank the authors for revising the paper according to my comments.

Author Response

I thank the authors for revising the paper according to my comments.

Thank you very much for helping us improve our paper. Your inputs have been of great help in furthering our research. Appreciate your time.