Is Gold a Hedge against Stock Price Risk in U.S. or Indian Markets?

Round 1

Reviewer 1 Report

The paper studies whether Gold has been a good hedge for US and Indian stock market. The study examines the relationship between Gold and stock market during different market scenarios such as extreme stock return period, or low and high volatility periods.

In Eq.(2), is there a typo ? Should it be 'bt = ..' instead of 'bt + ...'?
Line 310 refers to Eq.(1b), but Eq.(1b) is nowhere found in the paper. Could this be Eq.(2)?

In terms of conclusions, the authors state that Gold hasn't been good hedge for US or Indian stock market. However, based on results in Table 5A, Top 5% and 1% volatility periods has very low statistical significance, so it's hard to make any meaningful conclusion out of these two results. However, Top 10% volatility result has relatively higher significance, and the Top 10% volatility results show that actually in three out of the four subperiods the regression coefficient is negative, which means Gold actually played a hedging role to certain degree. Similarly in Table 5B, in Top 10% volatility periods, the same thing can be said in the US market, i.e., three out of the 4 periods Gold had negative correlation to stock market. This seems to be in contradiction to what the authors claim in the conclusion, so this should be addressed so that the conclusion is consistent with the regression results of the study.

Author Response

Comments and Suggestions for Authors

The paper studies whether Gold has been a good hedge for US and Indian stock market. The study examines the relationship between Gold and stock market during different market scenarios such as extreme stock return period, or low and high volatility periods.

In Eq.(2), is there a typo ? Should it be 'bt = ..' instead of 'bt + ...'?
Line 310 refers to Eq.(1b), but Eq.(1b) is nowhere found in the paper. Could this be Eq.(2)?

Response: 

Thank you for your review and observations.

Yes, these were typo errors and have now been corrected. Thank you very much for noticing and highlighting these.

In terms of conclusions, the authors state that Gold hasn't been good hedge for US or Indian stock market. However, based on results in Table 5A, Top 5% and 1% volatility periods has very low statistical significance, so it's hard to make any meaningful conclusion out of these two results. However, Top 10% volatility result has relatively higher significance, and the Top 10% volatility results show that actually in three out of the four sub-periods the regression coefficient is negative, which means Gold actually played a hedging role to certain degree. Similarly, in Table 5B, in Top 10% volatility periods, the same thing can be said in the US market, i.e., three out of the 4 periods Gold had negative correlation to stock market. This seems to be in contradiction to what the authors claim in the conclusion, so this should be addressed so that the conclusion is consistent with the regression results of the study.

Response:

You are right in observing that for the Top 10% volatility result, in three out of the four sub-periods, the regression coefficient is negative, which means Gold actually played a hedging role to certain degree. However, these coefficients are not statistically significant, as shown by the absence of any star against the coefficients. Based on your observation, however, we have now clarified and added the following sentences in the results section.

“There are a few cases, where the regression coefficient is negative, but not significant at even a 10% level. For example, the top 10% volatility observations, for the periods 1980-89, 1990-99, and 2010-20 for Sensex (Table 5A), and the periods 1980-89, 1990-99, and 2000-09 for S&P500 (Table 5B), are amongst such cases. However, since these values are not significant at any of our considered levels, we do not consider gold to be a hedge in any of these cases.”

Reviewer 2 Report

The paper addresses an interesting topic, but there are still issues to be addressed. First, some of the results of the paper should be better presented in the abstract. The authors should better state in the paper why U.S. has been chosen in the present analysis - there is a motivation provided for India, but the motivation for U.S. is missing.

Also, I believe that the authors should be more "mild" in presenting their paper -e.g. in row 87 it is stated that: "This is the first significant study" - how have you determined that this holds true?

Please change "literature survey" to "literature review". Survey stand for another type of analysis, which, for sure, it is not used in the paper. Please better structure the information on sub-chapters instead of using A, B, C and D in the literature review section.

Please remove "&" from section 3 and replace it with "and". Please explain the meaning of the variables used in section 3 - some of them are explained, but not all. Please name the tables in the annex using the same style (some of them are named using numbers, while other using numbers and letters).

I found the model used in the paper to be rather simple.

The conclusions are mainly focusing on the situation in India. No words are provided for U.S. Please extend the conclusions by including such a discussion related to the behavior of the investors in the U.S.

Please use the reference style requested by the journal when citing the references in the main body of the paper.

Author Response

Comments and Suggestions for Authors

The paper addresses an interesting topic, but there are still issues to be addressed. First, some of the results of the paper should be better presented in the abstract. The authors should better state in the paper why U.S. has been chosen in the present analysis - there is a motivation provided for India, but the motivation for U.S. is missing.

Response:

Thank you for your valuable comments.

The key result that gold does not serve as a hedge or a safe haven has been already stated in the abstract. I have also added a para on the motivation for the choice of U.S. market in the introduction section.

Also, I believe that the authors should be more "mild" in presenting their paper -e.g. in row 87 it is stated that: "This is the first significant study" - how have you determined that this holds true?

Response:

We accept that your suggestion, and have modified the sentence to read as “This extends the few significant studies, known to the authors,..”. To be sure, we have not found any research paper covering both the Indian and the U.S. markets over such a long time-period of time, i.e., 40 years. Hence, we would like to mention this point in the revised format.

 

 

Please change "literature survey" to "literature review". Survey stand for another type of analysis, which, for sure, it is not used in the paper. Please better structure the information on sub-chapters instead of using A, B, C and D in the literature review section.

Response:

I have changed, as suggested. Thank you.

 

Please remove "&" from section 3 and replace it with "and". Please explain the meaning of the variables used in section 3 - some of them are explained, but not all. Please name the tables in the annex using the same style (some of them are named using numbers, while other using numbers and letters).

Response:

I have made the changes and additions, as suggested. Thank you.

Further, the numbering of the sections has been revised to correct an error in the earlier numbering. The section referred to, here, is now section 4.

 

 

 

I found the model used in the paper to be rather simple.

Response:

I agree that the model uses a simple OLS linear regression technique. We are interested in considering the relationship between the variables, on an average basis, over a long time horizon. This is why we been explicit about the focus on long term investors. In view of the same, there is some merit in employing the OLS model. This has also been used in a few previous prominent studies like Baur and McDermott (2010), as cited in the article.

Having said that, the application of quantile regression and other models in this area should also be evaluated for varied sets of portfolios. While we are not employing this method, there is scope to further study these in a separate work.

We have also now stated the limitation, in the methodology section, that we have not adopted the other models like quantile regression.

 

The conclusions are mainly focusing on the situation in India. No words are provided for U.S. Please extend the conclusions by including such a discussion related to the behavior of the investors in the U.S.

Response:

I agree. I have added a paragraph in the conclusions with a discussion on U.S. market investors.

 

Please use the reference style requested by the journal when citing the references in the main body of the paper.

Response:

Yes, the same has been followed, thanks.

Reviewer 3 Report

The authors analyse the property of gold as a safe-haven asset, both in the US and Indian markets. For this purpose, the authors use linear regression. The results show that gold returns do not show a significant negative relationship with stock returns. 

The paper is potentially publishable, but in my opinion, it needs a thorough revision. 

1. I don't understand why the authors focus on two markets separately. I would focus on a particular market or highlight the spillover effects (so I invest in the Indian market and disinvest in the US, and vice versa) between the two. I don't understand the point of treating them separately. 

2. I think the contribution needs to be better explained. Several papers are analysing the property of gold as a haven asset in the current COVID-19 environment

Akhtaruzzaman, M., Boubaker, S., Lucey, B. M., & Sensoy, A. (2021). Is gold a hedge or a safe-haven asset in the COVID–19 crisis?. Economic Modelling, 102, 105588.

He, Xie, Tetsuya Takiguchi, Tadahiro Nakajima, and Shigeyuki Hamori. "Spillover effects between energies, gold, and stock: the United States versus China." Energy & Environment 31, no. 8 (2020): 1416-1447.

Uddin, G. S., Hernandez, J. A., Shahzad, S. J. H., & Kang, S. H. (2020). Characteristics of spillovers between the US stock market and precious metals and oil. Resources Policy, 66, 101601.

Ghazali, M. F., Lean, H. H., & Bahari, Z. (2020). Does gold investment offer protection against stock market losses? Evidence from five countries. The Singapore Economic Review, 65(02), 275-301.

3. The authors use a linear regression model. However, with time series, another approach would be welcome. In my opinion, the relationship with other econometric models should be tested, especially quantile regression models. For example:  

Shahzad, S. J. H., Mensi, W., Hammoudeh, S., Sohail, A., & Al-Yahyaee, K. H. (2019). Does gold act as a hedge against different nuances of inflation? Evidence from Quantile-on-Quantile and causality-in-quantiles approaches. Resources Policy, 62, 602-615.

Chkili, W. (2017). Is gold a hedge or safe haven for Islamic stock market movements? A Markov switching approach. Journal of Multinational Financial Management, 42, 152-163.

Shahbaz, M., Tahir, M. I., Ali, I., & Rehman, I. U. (2014). Is gold investment a hedge against inflation in Pakistan? A co-integration and causality analysis in the presence of structural breaks. The North American Journal of Economics and Finance, 28, 190-205.

Conlon, T., Lucey, B. M., & Uddin, G. S. (2018). Is gold a hedge against inflation? A wavelet time-scale perspective. Review of Quantitative Finance and Accounting, 51(2), 317-345.

4. I would like the authors to expand the results section by adding policy implications from an economic and investor perspective. 

Author Response

Comments and Suggestions for Authors

The authors analyse the property of gold as a safe-haven asset, both in the US and Indian markets. For this purpose, the authors use linear regression. The results show that gold returns do not show a significant negative relationship with stock returns. 

The paper is potentially publishable, but in my opinion, it needs a thorough revision. 

1. I don't understand why the authors focus on two markets separately. I would focus on a particular market or highlight the spillover effects (so I invest in the Indian market and disinvest in the US, and vice versa) between the two. I don't understand the point of treating them separately. 

Response:

Thank you for your valuable comments.

The paper addresses the question of hedging with gold, faced by several investors who may invest in the U.S. and the Indian stock markets separately. As rightly highlighted by you, for the investors investing in both the markets together, a portfolio approach would be required. That would require the study of the covariance between the two markets in varying periods and situations. While it is an important area, our objective here is restricted to determining the relationship between stock and gold for a single market at a time. There is scope to take this study further, in another paper, in a multi variate context. This has also been now mentioned in the conclusions section.

 

2. I think the contribution needs to be better explained. Several papers are analysing the property of gold as a haven asset in the current COVID-19 environment

Akhtaruzzaman, M., Boubaker, S., Lucey, B. M., & Sensoy, A. (2021). Is gold a hedge or a safe-haven asset in the COVID–19 crisis?. Economic Modelling, 102, 105588.

He, Xie, Tetsuya Takiguchi, Tadahiro Nakajima, and Shigeyuki Hamori. "Spillover effects between energies, gold, and stock: the United States versus China." Energy & Environment 31, no. 8 (2020): 1416-1447.

Uddin, G. S., Hernandez, J. A., Shahzad, S. J. H., & Kang, S. H. (2020). Characteristics of spillovers between the US stock market and precious metals and oil. Resources Policy, 66, 101601.

Ghazali, M. F., Lean, H. H., & Bahari, Z. (2020). Does gold investment offer protection against stock market losses? Evidence from five countries. The Singapore Economic Review, 65(02), 275-301.

Response:

You are right that several papers have recently covered the topic in the context of Covid. We have, on the other hand, gone much further, to cover a long time-period of forty years. The two sets of papers would help address two very useful, yet differentiated, questions for the investors. Our time horizon (1980 onwards) is also long enough to include multiple periods of calmness as well as high volatility, including the first phase of Covid. This has been also mentioned as the contribution of our paper to the existing literature.

 

3. The authors use a linear regression model. However, with time series, another approach would be welcome. In my opinion, the relationship with other econometric models should be tested, especially quantile regression models. For example:  

Shahzad, S. J. H., Mensi, W., Hammoudeh, S., Sohail, A., & Al-Yahyaee, K. H. (2019). Does gold act as a hedge against different nuances of inflation? Evidence from Quantile-on-Quantile and causality-in-quantiles approaches. Resources Policy, 62, 602-615.

Chkili, W. (2017). Is gold a hedge or safe haven for Islamic stock market movements? A Markov switching approach. Journal of Multinational Financial Management, 42, 152-163.

Shahbaz, M., Tahir, M. I., Ali, I., & Rehman, I. U. (2014). Is gold investment a hedge against inflation in Pakistan? A co-integration and causality analysis in the presence of structural breaks. The North American Journal of Economics and Finance, 28, 190-205.

Conlon, T., Lucey, B. M., & Uddin, G. S. (2018). Is gold a hedge against inflation? A wavelet time-scale perspective. Review of Quantitative Finance and Accounting, 51(2), 317-345.

Response:

The model uses a simple OLS linear regression model. We are interested in considering the relationship between the variables, on an average basis, over a long time horizon. This is why we been explicit about the focus on long term investors.  In view of the same, there is some merit in employing the OLS model. This has also been employed in a few previous prominent studies like Baur and McDermott (2010), and Hood & Malik (2013) as cited in the article.

Having said the above, we agree that the application of quantile regression in this area may also be evaluated. A comparative work would be quite useful, though we believe that justice would be done by elaborately studying it as a separate paper, and elaborating on their respective utilities.

We have now also explicitly qualified our methodology and stated the limitation as follows:

“It may be mentioned here that the OLS model does have its limitation of not capturing the non-normal behaviour of extreme observations. Alternate models like dynamic conditional model (Chkili, 2016), time varying dynamic model (Beckman et al, 2013) have also been used in other studies. We believe that the limitation is mitigated to a large extent for a long term investor, who is interested in the average relationship between gold and stocks over a long time-horizon.”

 

4. I would like the authors to expand the results section by adding policy implications from an economic and investor perspective. 

Response:

This is a very good suggestion and I have significantly expanded on the policy implications in the conclusion section. Thank you.

Reviewer 4 Report

• This paper makes interesting contributions to the literature.
• The topic is presented in a comprehensive manner;
• The methods are adequately described;
• The manuscript includes relevant references;
• The conclusions of the paper are comprehensive and appropriate;
• The study is interesting and could be published.

Author Response

Response:

Thank you very much for your comments. Sincerely appreciate.

Round 2

Reviewer 1 Report

The author addressed some of the comments from previous version

Author Response

Thanks for your your review and comments.

Reviewer 2 Report

I have no further comments. Thank you!

Author Response

Thanks for your review and comments, appreciate them.

Reviewer 3 Report

The authors have taken onboard almost all of my suggestions. I appreciated it. However, I would appreciate it if the authors would add a section on the robustness of the analysis, with other models for example (as I suggested in my previous review). I think a robustness analysis is crucial in these empirical papers. 

Author Response

Thank you for your suggestion. I agree that doing the relevant robustness tests is important. The test for heteroscedasticity would be relevant with the OLS regression model, in the context that we have used. Hence, we have done the Breusch-Pagan test for heteroscedasticity. A sub-section on robustness test has been added at the end of the methodology section, explaining the same.

In the results section, we have stated that the B-P tests have not rejected the hypothesis of homoscedasticity.