Examining the Effect of COVID-19 on Foreign Exchange Rate and Stock Market -- An Applied Insight into the Variable Effects of Lockdown on Indian Economy
1 Examining the Effect of COVID-19 on Foreign Exchange Rate and Stock Market – An Applied Insight into the Variable Effects of Lockdown on Indian Economy
Indrajit Banerjee , Atul Kumar , Rupam Bhattacharyya
2, * Department of Economic Studies and Policies, Central University of South Bihar Department of Biostatistics, University of Michigan *
Corresponding author. Address: 1415 Washington Heights, Ann Arbor, MI 48109, USA. Email ID: [email protected], Phone: +17348006834. Abstract
The relationship between a pandemic and the concurrent economy is quite comparable to the relation observed among health and wealth in general. Since March 25, 2020, India had been under a nation-wide lockdown announced as a response to the spread of SARS-Cov-2 and COVID-19 and has resorted to a process of ‘unlocking’ the lockdown over the past couple of months. This work attempts to examine the effect of novel coronavirus 2019 (COVID-19) and its resulting disease, the COVID-19, on the foreign exchange rates and stock market performances of India using secondary data over a span of 112 days spanning between March 11 and June 30, 2020. The study explores whether the causal relationships and directions among the growth rate of confirmed cases (GROWTHC), exchange rate (GEX) and SENSEX value (GSENSEX) are remaining the same across different pre and post-lockdown phases, attempting to capture any potential changes over time via the vector autoregressive (VAR) models. A positive correlation is found between the growth rate of confirmed cases and the growth rate of exchange rate, and a negative correlation between the growth rate of confirmed cases and the growth rate of SENSEX value. A naïve interpretation from this could be that with the rising growth rate of the number of confirmed cases, the economy took a toll, reflected by the Indian currency being depreciated while the stock exchange index suffered from a fall. However, on applying a VAR model, it is observed that an increase in the confirmed COVID-19 cases causes no significant change in the values of the exchange rate and SENSEX index. The result varies if the analysis is split across different time periods – before lockdown, the four phases of lockdown, and the first phase of unlock. To compare these periods, we had undertaken eight rounds of analyses. Nuanced and sensible interpretations of the numeric results indicate significant variability across time in terms of the relation between the variables of interest. This detailed knowledge about the varying patterns of dependence could potentially help the policy makers and investors of India in order to develop their strategies to cope up with the situation. Introduction
A 55-year-old individual from the Hubei province in China was reported to have been the first person to have contracted what is now called COVID-19, [Bryner, 2020] the disease caused by the novel coronavirus (SARS-Cov-2). [Huang et al., 2020] In India, the first case was reported on the 30 th of January 2020 in Kerala. [“COVID-19 India”, 2020] Till March 24, 2020, a total of 657 COVID-19 cases were reported among which 11 were deceased and the number of recovered cases was 6. [“COVID-19 India”, 2020] Since March 25, India had been under national lockdown, which extended till the end of May over four different phases. [Murukesh, 2020] The beginning of June saw India attempting to come out of the lockdown via what was christened ‘unlock’, a procedure that is currently at its third cycle as of August 14. [Murukesh, 2020] One of the controlling factors of any disease scenario – quality of healthy life, which includes purified water, sanitized housing, sufficient nutritious food, good health care etc., is also an essential contributor to any economy, but the affordability of the same depends on both financial stability and access to required knowledge to enhance and maintain one’s health. [Fan et al., 2018] [Farhud, 2015] [Frakt, 2018] Thus, the relationship between a pandemic and the concurrent economy has been observed to be exactly similar to the relation observed among health and wealth in general. [Bloom and Canning, 2020] [Mckee and Stuckler, 2020] [Ojong, 2020] The pandemic can hamper economic developments and trigger catastrophic events, thereby reducing the wealth of health and reducing protection against further health threats. [Bloom and Canning, 2020] A past work on the economic impact of a pandemic observed that education, health and social service, and insurance sectors suffered from a large amount of loss to GDP in UK, France, Belgium and the Netherlands due to the effects of influenza outbreaks. [Keogh-Brown et al., 2010] According to conjectures and calculations provided by the rating agency S&P, COVID-19 can slow down the growth rate of baseline GDP for the world by 0.3 percentage point (ppt), for China by 0.7 ppt; for Asia-Pacific by 0.5 ppt; and for the USA and Europe by 0.1 to 0.2 ppt. [Bloom and Canning, 2020] Another study using data from Kuala Lumpur Stock Exchange (KLSE) and exchange rate of the currency of Malaysia (MYR) found a significant effect on equity market and on exchange rate and mentioned that the outbreak of COVID-19 is creating an insecure feeling to investors in equity market. [Bakar & Rosbi, 2020] A study on the effect of COVID-19 on financial volatility index (VIX) concluded that the spread of coronavirus is increasing the financial volatility. [Albulescu, 2020] Another study to see the short-run effect of COVID-19 on the 21 leading stock markets in the world observed a sharp decline in the stock markets since the virus outbreak. [Liu et al., 2020] In any economy, both the supply and the demand are affected by the pandemic which may result in a persistent and large economic catastrophe all over the world. [Eichenbaum et al., 2020] However, according to McKenzie, there is no substantial relationship between Foreign Exchange volatility and international trade. [McKenzie, 1999] Vaguely, this means that at a standard scenario, the fluctuations in the foreign exchange rate do not affect the international trade. Bahmani-Oskooee and Saha investigated the potential link between exchange rate and stock prices using timeseries data of 24 countries and found that there may be a short-run relationship between these two variables, but the relationship doesn’t hold in a longer term. [Bahmani-Oskooee & Saha, 2018] The short-term shocks due to a pandemic can be improved in an economy by identifying and funding productive projects with better investment opportunities which is impelled by a well-governed stock market. [Nazir et al., 2010] It has also been observed that changes in monetary policy has surprisingly variable impacts across different sectors. [Prabu et al., 2020] In context of India, a disease outbreak and its impacts on the national economy is not a completely new experience. In 1918, “The Bombay Influenza” was spread from seven police sepoys in Bombay to Uttar Pradesh and Punjab through railway services by two weeks, which appeared to cause the highest number of deaths in one nation, the estimated number being 10-20 million. [Acharjee, 2020] During the Spanish flu, India experienced the lowest GDP growth rate (-10.5%) forever and an all-time high level of inflation, with a resulting supply side shock. The situation was worse compared to the first world war and the Bengal famine, as corroborated by future investigators. [Sreevatsan, 2020] Unsurprisingly enough, COVID-19 too presents a challenge before the Indian economy which can turn both ways. IMF predicted that the growth rate of GDP in India can be more than 7% in April 2021 if the nation can control the coronavirus outbreak, but the growth rate could come down to as low as 1.9% in March 2021, depending on the control of the disease and stimulating monetary and fiscal policies immediately ending up the lockdowns. [Choudhury, 2020] The stock market acts as an important ingredient of an open-market economy. In this study, we consider SENSEX value to analyse the market performance of India’s stock market because since 1989 when it had been set up, it has become the basic representative of the Bombay Stock Exchange (BSE) and has been considered as the ‘barometer’ of Indian economy. Instability of market condition results in volatility of the stock market, and the price of BSE Sensex and stock of major companies on BSE prices have previously been reported to be positively related. [Challa et al., 2018] Internationally, behaviour of experienced and smart investors has been indicated to lead to a high return. [Kaufman, 1995] Recent studies have found the market as a respondent of the potential economic outcome under the pandemic, COVID-19. [Ramelli & Wagner, 2020] Large price movements took place over the past few months because the investors started to exhibit concerns regarding the COVID-19 shock via financial ways. [Abdelnour, 2020] This study explores the impact of COVID-19 on the stock market of India and attempts to explore their relationship using Vector Autoregressive (VAR) models. [Pfaff et al., 2008] The VAR model is treated as a standard tool in determining inter-dependencies and dynamic relationships among the variables in time-series econometrics. These models are able to explain the endogenous variables solely based on their historical values, apart from the deterministic regressors. We also aim to find out whether the fluctuations in international trade of India, caused by the pandemic, is reflected in the foreign exchange rate of the country. With new incoming data, the dependence between the economic indicators and the spread of the pandemic can be tracked more accurately, which in turn can help guide policymaking to absorb the potential diminishing effects of COVID-19. Methods
Secondary data were collected for the number of infected cases from COVID-19, the stock market value and exchange rate from different sources. We have divided our work into eight Rounds considering the following eight time periods based on the decisions of government and WHO. a.
March 25 to April 14, 2020: 21 days (Lockdown 1.0), b.
April 15 to May 03, 2020: 19 days (Lockdown 2.0), c.
May 04 to May 17, 2020: 14 days (Lockdown 3.0), d.
May 18 to May 31, 2020: 14 days (Lockdown 4.0), e.
June 01 to June 30, 2020: 30 days (Unlock 1.0), f. March 25 to June 30, 2020: 98 days (Lockdown 1.0 to Unlock 1.0), g.
March 11 to April 14, 2020: 35 days (Pre-lockdown to Lockdown 1.0), h.
March 11 to June 30: 112 days (Pre-lockdown to Unlock 1.0). On 11 th March, WHO upgraded the status of COVID-19 to Pandemic from Epidemic and on that day the first death in India due to coronavirus was reported. Rounds a-e are considered to identify the relations between COVID-19 on stock market and exchange rate of India exclusively for the individual phases of lockdown and unlock. Round f serves the purpose of a holistic view over the lockdown and unlock periods combined to check whether the observed patterns in the individual rounds hold for long. Round g offers the utility of taking a closer look at the dependence between the pandemic and the economic indicators at the initial, more moderate stage of the pandemic. Finally, round h is considered for 112 days to summarize the changing patterns over the whole course of the pandemic in India till the end of June. In our analysis we use closing value of the stock market, which is significant and is a standard quantification of the market for several reasons. Investors, traders, financial institutions, regulators and other stakeholders use it as a reference point for determining performance over a specific time such as one year, a week and over a shorter time frame such as one minute or less. In fact, investors and other stakeholders base their decisions on closing stock prices. [Ellul et al., 2005] We denote the growth rates for confirmed cases, SENSEX and exchange rate as GROWTHC, GSENSEX and GEX throughout this paper.
Results
Round a: Lockdown 1.0 (March 25 – April 14)
During this round, the nation was experiencing its first phase of complete lockdown due to COVID-19. The movement of the three growth rates over time for this period is summarized in Figures 1a-1c (red lines). The three variables were checked to be stationary by using unit root test (Supplementary Materials Subsection S1.1). Regression result of GEX as dependent variable on GROWTHC and GSENSEX as covariates indicates the overall model to be significant at 5% level, with the GROWTHC coefficient having a p-value of 0.08 (Supplementary Materials Subsubsection S2.1.3 and Figure 1c). One-point increase in GSENSEX causes GEX to fall by 0.058 point and the same in GROWTHC causes GEX to rise by 0.023 point (Supplementary Materials Subsubsection S2.1.3). In this period, the correlation matrix exhibits a positive correlation (0.45) between GROWTHC and GEX, but a negative correlation (-0.23) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.1). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued depreciating, whereas the stock exchange index started to fall. Also, there is a notable negative correlation (-0.40) between GSENSEX and GEX. On applying VAR, we do find a significant dependence between GSENSEX and the one-period lagged difference in GROWTHC, evident by the computed t-statistic (2.16) (Supplementary Materials Subsection 4.1). Looking at the IRF graphs (Figures 1d-e), it can be seen that during this round, due to the COVID-19 shock, GSENSEX and GEX would have taken about 5.5 periods (days) to return to their long-run growth trend.
Round b: Lockdown 2.0 (April 15 – May 03)
During this round, the nation was experiencing the second extended phase of complete lockdown due to COVID-19. The movement of the three growth rates over time for this period is summarized in Figures 2a-2c (red lines). The three variables were checked to be stationary by using unit root test (Supplementary Materials Subsection S1.2). Regression result of GEX as dependent variable on GROWTHC and GSENSEX as covariates indicates the overall model to be significant at 5% level, with the GSENSEX coefficient having a p-value less than 0.01 (Supplementary Materials Subsubsection S2.2.3 and Figure 2c). One-point increase in GSENSEX causes GEX to fall by 0.213 point and the same in GROWTHC causes GEX to rise by 0.003 point (Supplementary Materials Subsubsection S2.2.3). In this period, the correlation matrix exhibits a weaker positive correlation (0.24) between GROWTHC and GEX compared to the same from the previous round, but a stronger negative correlation (-0.29) than before between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.2). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency still continued depreciating, whereas the stock exchange index started to fall. Also, there is a notably strong negative correlation (-0.79) between GSENSEX and GEX. On applying VAR, we find no significant dependence between the growth rates (Supplementary Materials Subsection 4.2). Looking at the IRF graphs (Figures 2d-e), it can be seen that during this round, due to the COVID-19 shock, GSENSEX and GEX would have taken about 6 and 8 periods (days) respectively to return to their long-run growth trend, indicating a larger shock. Round c: Lockdown 3.0 (May 04 – May 17)
During this round, the nation was experiencing its third extended phase of complete lockdown due to COVID-19. The movement of the three growth rates over time for this period is summarized in Figures 3a-3c (red lines). The three variables were checked to be stationary by using unit root test (Supplementary Materials Subsection S1.3). All the multiple linear regression models turn out to be non-significant at an overall 5% level. Regression result of GSENSEX as dependent variable on GROWTHC and GEX as covariates shows that ne-point increase in GEX causes GSENSEX to fall by 0.284 point and the same in GROWTHC causes GSENSEX to fall by 0.858 point (Supplementary Materials Subsubsection S2.3.2). In this period, the correlation matrix exhibits a positive correlation (0.46) between GROWTHC and GEX, but a negative correlation (-0.59) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.3). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued depreciating, whereas the stock exchange index started to fall. Also, there is a notable negative correlation (-0.33) between GSENSEX and GEX. Looking at the IRF graphs from the VAR models (Figures 3d-e), it can be seen that during this round, due to the COVID-19 shock, GSENSEX and GEX would have taken about 10 periods (days) to return to their long-run growth trend.
Round d: Lockdown 4.0 (May 18 – May 31)
During this round, the nation was experiencing the fourth extended phase of complete lockdown due to COVID-19. The movement of the three growth rates over time for this period is summarized in Figures 4a-4c (red lines). The three variables were checked to be stationary by using unit root test (Supplementary Materials Subsection S1.4). All the multiple linear regression models turn out to be non-significant at an overall 5% level. Regression result of GEX as dependent variable on GROWTHC and GSENSEX as covariates indicates the overall model to be significant at 10% level, with the GROWTHC coefficient having a p-value of 0.06 (Supplementary Materials Subsubsection S2.4.3 and Figure 4c). One-point increase in GSENSEX causes GEX to rise by 0.062 point and the same in GROWTHC causes GEX to rise by 0.029 point (Supplementary Materials Subsubsection S2.4.3). In this period, the correlation matrix exhibits a positive correlation (0.50) between GROWTHC and GEX, and a very small positive correlation (<0.01) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.4). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued depreciating, whereas the stock exchange index experienced a very small upward effect. Also, the correlation between GSENSEX and GEX surprisingly becomes positive (0.32). Looking at the IRF graphs from the VAR model (Figures Round e: Unlock 1.0 (June 01 – June 30)
During this round, the nation was experiencing its first phase of unlock procedure. The movement of GSENSEX over time for this period is summarized in Figure 5a (red line). The three variables were checked for stationarity, and only GSENSEX and GEX turned out to be stationary by using unit root test (Supplementary Materials Subsection S1.5). Regression result of GSENSEX as dependent variable on GEX indicates the overall model to be non-significant at 5% level, with the GEX coefficient having a p-value greater than 0.10 (Supplementary Materials Subsubsection S2.5.1 and Figure 5a). One-point increase in GEX causes GSENSEX to fall by 1.39 points (Supplementary Materials Subsubsection S2.5.1). In this period, the correlation matrix exhibits a small positive correlation (0.12) between GROWTHC and GEX, and a larger positive correlation (0.20) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.5). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency now started improving, and the stock exchange index started to perform better in first-order growth. Also, there is a notable negative correlation (-0.30) between GSENSEX and GEX. On applying VAR, we do find a significant dependence between GSENSEX and the one-period lagged difference in GROWTHC, evident by the computed t-statistic (2.19) (Supplementary Materials Subsection 4.5). Looking at the IRF graphs (Figures 5b-c), it can be seen that during this round, due to the COVID-19 shock, GSENSEX and GEX would have taken about 8 and 4 periods (days) respectively to return to their long-run growth trend.
Round f: Lockdown 1.0 to Unlock 1.0 (March 25 – June 30)
This round was considered to observe the overall timeline of the lockdown till the end of the first unlock stage. The movement of the three growth rates over time for this period is summarized in Figures 6a-6c (red lines). The three variables were checked to be stationary by using unit root test (Supplementary Materials Subsection S1.6). Both the models with GEX and GSENSEX as responses turn out to be significant at 5% level, with both of them having extremely low (less than 0.001) p-values. Regression result of GEX as dependent variable on GROWTHC and GSENSEX as covariates indicates that one-point increase in GSENSEX causes GEX to fall by 0.078 point and the same in GROWTHC causes GEX to rise by 0.012 point (Supplementary Materials Subsubsection S2.6.3). In this period, the correlation matrix exhibits a positive correlation (0.22) between GROWTHC and GEX, but a very small negative correlation (-0.06) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.6). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued improving, whereas the stock exchange index started to experience a slightly negative growth. Also, there is a notable negative correlation (-0.39) between GSENSEX and GEX. On applying VAR, we do find a significant dependence between GSENSEX till a four-period lagged difference in GROWTHC, evident by the computed t-statistics (Supplementary Materials Subsection 4.1).
Round g: Pre-lockdown to Lockdown 1.0 (March 11 – April 14)
This round allows us to take a closer look at the relations between the growth rates of interest at the initial stage of the pandemic, till the end of the first lockdown. In this period, the correlation matrix exhibits a positive correlation (0.40) between GROWTHC and GEX, but a negative correlation (-0.13) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.7). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued depreciating, whereas the stock exchange index started to fall. Also, there is a notable negative correlation (-0.52) between GSENSEX and GEX. Looking at the IRF graphs from the VAR model (Figures 7a-b), it can be seen that during this round, due to the COVID-19 shock, GSENSEX and GEX would have taken about 3 periods (days) to return to their long-run growth trend. Round h: Pre-lockdown to Unlock 1.0 (March 11 – June 30)
This round was considered to observe the overall timeline of the pandemic till the end of the first unlock stage. The three variables were checked to be stationary by using unit root test (Supplementary Materials Subsection S1.7). In this period, the correlation matrix exhibits a positive correlation (0.28) between GROWTHC and GEX, but a negative correlation (-0.15) between GROWTHC and GSENSEX (Supplementary Materials Subsection S3.1). This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued depreciating, whereas the stock exchange index started to fall. Also, there is a notable negative correlation (-0.47) between GSENSEX and GEX. On applying VAR, we do find a significant dependence between GSENSEX and the two and four-period lagged differences in GROWTHC, evident by the computed t-statistics (Supplementary Materials Subsection 4.8).
Discussions
The main objective of this study is to explore the relationship among the number of infected cases, exchange rate and stock market in India and to analyze the correlation among the growth rate series of these three variables. Since October 07, 2019 to March 10, 2020, average exchange rate of USD to INR was at 71.44 but after march 11, when the COVID-19 has been declared to be Pandemic by WHO, the average exchange rate over the period considered became above 75. Overall, during the course of the pandemic till the first unlock period in India, we observed a generally positive correlation between the growth rate of infected cases and the growth rate of exchange rate. This correlation was found to be negative between the growth rate of infected cases and the growth rate of Sensex. The VAR results implied that the increase in the confirmed COVID-19 cases caused suggestive but not statistically significant changes in the values of the exchange rate and Sensex, and that due to the COVID-19 shock, the growth rate of Sensex and the growth rate of exchange rates would have taken a substantial number of days, in general, to return to their long-run growth trends. The decision of lockdown appears to be effective in controlling the spread of COVID-19 reflected by the average daily growth rate of confirmed cases and death cases. The daily average growth rate of confirmed cases was 17.804% before the lockdown which changes first to 15.684% in first phase of lockdown and then to 7.55% when the lockdown is extended. Similarly, the growth rate of deaths was 23.55% before the lockdown which first falls to 19.39% in first phase of lockdown and again to 7.27% when it is extended after April 15. These are indications that the lockdown improved the situation to some extent. During the first phase of lockdown, the growth rate of exchange rate stays lower than the daily average growth rate of exchange rate observed before lockdown. With the extension of lockdown, the daily growth rate of exchange rate becomes negative (INR being appreciated) after April 15. During Lockdown 1.0, there is a positive correlation between GROWTHC and GEX, but a negative correlation between GROWTHC and GSENSEX. This meant that with the rise in the growth rate of the number of confirmed cases, the Indian currency continued depreciating, whereas the stock exchange index started to fall. Also, there is a strong negative correlation between the exchange rate and the Sensex index, implying that with the Indian currency depreciating the stock exchange also dropped. With extension of the lockdown, clear variabilities were observed in the relationships across the variables of interest. We found a high degree of positive correlation between GROWTHC and GEX initially, which stayed positive for most of the time over the entire period considered. The negative correlation between GROWTHC and GSENSEX was almost washed out at a point and then started showing up again towards the end. Our investigations into the relations and interdependence between economic and health variables in context of the pandemic and the lockdown identify strong evidences of variability in these relationships over time. It is quite evident that these changes are extremely dynamic, and nuanced observation at the rapidly-changing scenario is required for effective policy decisions. Our framework is quite flexible and quick updating of the results is possible based on new incoming data, which is why it can potentially be useful in identifying the current needs from a decision-making perspective in the battle against COVID-19. References
Abdelnour, A. (2020, July 8).
Pricing in a pandemic: Navigating the COVID-19 crisis
Journal of Financial and Quantitative Analysis , (4), 779-801. Fan, V. Y., Bloom, D. E., Ogbuoji, O., Prettner, K., & Yamey, G. (2018). Valuing health as development: going beyond gross domestic product. Bmj , , k4371. Farhud, D. D. (2015). Impact of lifestyle on health. Iranian journal of public health , (11), 1442. Frakt, A. B. (2018). How the economy affects health. Jama, 319(12), 1187-1188. Gopinathan, R., & Durai, S. R. S. (2019). Stock market and macroeconomic variables: New evidence from India. Financial Innovation, 5 (1), 1–17. Huang, C., Wang, Y., Li, X., Ren, L., Zhao, J., Hu, Y., ... & Cheng, Z. (2020). Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China.
The Lancet , (10223), 497-506. Kaufman, G. (1995). Research in Financial Services: Banking, Financial Markets, and Systemic Risk. Greenwich, CT: JAI. Keogh-Brown, M. R., Smith, R. D., Edmunds, J. W., & Beutels, P. (2010). The macroeconomic impact of pandemic influenza: estimates from models of the United Kingdom, France, Belgium and The Netherlands. The European Journal of Health Economics, 11(6), 543-554. Liu, H., Manzoor, A., Wang, C., Zhang, L., & Manzoor, Z. (2020). The COVID-19 outbreak and affected countries stock markets response. International Journal of Environmental Research and Public Health, 17(8), 2800. McKee, M., & Stuckler, D. (2020). If the world fails to protect the economy, COVID-19 will damage health not just now but also in the future. Nature Medicine , (5), 640-642. McKenzie, M. D. (1999). The impact of exchange rate volatility on international trade flows. Journal of economic Surveys, 13 (1), 71–106. Murukesh, K. (2020, June 19). ‘Unlock 2.0’: What could happen after June 30? All you need to know Healthcare Figures
Figure 1:
Summary of results for round a: Lockdown 1.0 (March 25 – April 14).
In panels a-c, M3 and M4 respectively indicate the months of March and April. In panels d-e, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Summaries corresponding to regression of GROWTHC on GSENSEX and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel b:
Summaries corresponding to regression of GSENSEX on GROWTHC and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel c:
Summaries corresponding to regression of GEX on GSENSEX and GROWTHC where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel d:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel e:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Figure 2:
Summary of results for round b: Lockdown 2.0 (April 15 – May 03).
In panels a-c, 2020m4 and 2020m5 respectively indicate the months of April and May. In panels d-e, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Summaries corresponding to regression of GROWTHC on GSENSEX and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel b:
Summaries corresponding to regression of GSENSEX on GROWTHC and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel c:
Summaries corresponding to regression of GEX on GSENSEX and GROWTHC where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel d:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel e:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Figure 3:
Summary of results for round c: Lockdown 3.0 (May 04 – May 17).
In panels a-c, 2020m5 indicates the month of May. In panels d-e, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Summaries corresponding to regression of GROWTHC on GSENSEX and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel b:
Summaries corresponding to regression of GSENSEX on GROWTHC and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel c:
Summaries corresponding to regression of GEX on GSENSEX and GROWTHC where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel d:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel e:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Figure 4:
Summary of results for round d: Lockdown 4.0 (May 18 – May 31).
In panels a-c, 2020m5 indicates the month of May. In panels d-e, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Summaries corresponding to regression of GROWTHC on GSENSEX and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel b:
Summaries corresponding to regression of GSENSEX on GROWTHC and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel c:
Summaries corresponding to regression of GEX on GSENSEX and GROWTHC where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel d:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel e:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Figure 5:
Summary of results for round a: Unlock 1.0 (June 01 – June 30).
In panel a, M6 indicates the month of June. In panels b-c, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Summaries corresponding to regression of GSENSEX on GROWTHC and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel b:
Impulse Response Function graph of GSENSEX to D(GROWTHC) obtained from the fitted vector autoregressive model.
Panel c:
Impulse Response Function graph of GEX to D(GROWTHC) obtained from the fitted vector autoregressive model. Figure 6:
Summary of results for round f: Lockdown 1.0 to Unlock 1.0 (March 25 – June 30).
In panels a-c, M3, M4, M5, and M6 respectively indicate the months of March, April, May, and June. In panels d-e, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Summaries corresponding to regression of GROWTHC on GSENSEX and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel b:
Summaries corresponding to regression of GSENSEX on GROWTHC and GEX where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel c:
Summaries corresponding to regression of GEX on GSENSEX and GROWTHC where the blue line, the red line and the green line represent residual, actual and fitted values respectively.
Panel d:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel e:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Figure 7:
Summary of results for round g: Pre-lockdown to Lockdown 1.0 (March 11 – April 14).
In panels a-b, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel b:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Figure 8:
Summary of results for round h: Pre-lockdown to Unlock 1.0 (March 11 – June 30).
In panels a-b, 1 in the x-axis indicates the beginning of the period considered.
Panel a:
Impulse Response Function graph of GSENSEX to GROWTHC obtained from the fitted vector autoregressive model.
Panel b:
Impulse Response Function graph of GEX to GROWTHC obtained from the fitted vector autoregressive model. Supplementary Document for Examining the Effect of COVID-19 on Foreign Exchange Rate and Stock Market – An Applied Insight into the Variable Effects of Lockdown on Indian Economy
Indrajit Banerjee , Atul Kumar , Rupam Bhattacharyya
2, * Department of Economic Studies and Policies, Central University of South Bihar Department of Biostatistics, University of Michigan *
Corresponding author. Address: 1415 Washington Heights, Ann Arbor, MI 48109, USA. Email ID: [email protected], Phone: +17348006834. Supplementary Materials
S1. Results of unit-root tests across different rounds
S1.1. Round a: Lockdown 1.0 (March 25 – April 14)
S1.1.1. Unit-root test for GROWTHC
Null Hypothesis: GROWTHC has a unit root Exogenous: Constant Lag Length: 0 (Automatic - based on SIC, maxlag=4) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -4.566172 0.0020 Test critical values: 1% level -3.808546 5% level -3.020686 10% level -2.650413 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC) Method: Least Squares Date: 07/08/20 Time: 12:22 Sample (adjusted): 3/26/2020 4/14/2020 Included observations: 20 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GROWTHC(-1) -1.072300 0.234836 -4.566172 0.0002 C 0.169459 0.044689 3.791935 0.0013 R-squared 0.536679 Mean dependent var -0.006341 Adjusted R-squared 0.510939 S.D. dependent var 0.145101 S.E. of regression 0.101473 Akaike info criterion -1.643401 Sum squared resid 0.185343 Schwarz criterion -1.543828 Log likelihood 18.43401 Hannan-Quinn criter. -1.623963 F-statistic 20.84993 Durbin-Watson stat 1.945374 Prob(F-statistic) 0.000239
S1.1.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=4) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -5.051439 0.0000 Test critical values: 1% level -2.685718 5% level -1.959071 10% level -1.607456 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 12:23 Sample (adjusted): 3/26/2020 4/14/2020 Included observations: 20 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -1.028991 0.203703 -5.051439 0.0001 R-squared 0.570748 Mean dependent var -0.003490 Adjusted R-squared 0.570748 S.D. dependent var 0.047269 S.E. of regression 0.030970 Akaike info criterion -4.062912 Sum squared resid 0.018223 Schwarz criterion -4.013126 Log likelihood 41.62912 Hannan-Quinn criter. -4.053193 Durbin-Watson stat 2.228737
S1.1.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=4) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -5.370109 0.0000 Test critical values: 1% level -2.685718 5% level -1.959071 10% level -1.607456 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 12:24 Sample (adjusted): 3/26/2020 4/14/2020 Included observations: 20 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -1.206194 0.224613 -5.370109 0.0000 R-squared 0.602810 Mean dependent var 6.10E-05 Adjusted R-squared 0.602810 S.D. dependent var 0.009668 S.E. of regression 0.006093 Akaike info criterion -7.314523 Sum squared resid 0.000705 Schwarz criterion -7.264737 Log likelihood 74.14523 Hannan-Quinn criter. -7.304805 Durbin-Watson stat 1.544222
S1.2. Round b: Lockdown 2.0 (April 15 – May 03)
S1.2.1. Unit-root test for GROWTHC
Null Hypothesis: GROWTHC has a unit root Exogenous: Constant Lag Length: 0 (Automatic - based on SIC, maxlag=3) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -3.650822 0.0152 Test critical values: 1% level -3.857386 5% level -3.040391 10% level -2.660551 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 18 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC) Method: Least Squares Date: 07/08/20 Time: 12:26 Sample (adjusted): 4/16/2020 5/03/2020 Included observations: 18 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GROWTHC(-1) -0.908795 0.248929 -3.650822 0.0022 C 0.064823 0.018265 3.549110 0.0027 R-squared 0.454456 Mean dependent var -0.000112 Adjusted R-squared 0.420359 S.D. dependent var 0.023140 S.E. of regression 0.017617 Akaike info criterion -5.135438 Sum squared resid 0.004966 Schwarz criterion -5.036508 Log likelihood 48.21894 Hannan-Quinn criter. -5.121797 F-statistic 13.32850 Durbin-Watson stat 1.991684 Prob(F-statistic) 0.002155
S1.2.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=3) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -3.846746 0.0007 Test critical values: 1% level -2.699769 5% level -1.961409 10% level -1.606610 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 18 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 12:27 Sample (adjusted): 4/16/2020 5/03/2020 Included observations: 18 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -0.920242 0.239226 -3.846746 0.0013 R-squared 0.465027 Mean dependent var 0.000562 Adjusted R-squared 0.465027 S.D. dependent var 0.022954 S.E. of regression 0.016789 Akaike info criterion -5.282262 Sum squared resid 0.004792 Schwarz criterion -5.232797 Log likelihood 48.54036 Hannan-Quinn criter. -5.275441 Durbin-Watson stat 1.931701
S1.2.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=3) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -3.652402 0.0011 Test critical values: 1% level -2.699769 5% level -1.961409 10% level -1.606610 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 18 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 12:27 Sample (adjusted): 4/16/2020 5/03/2020 Included observations: 18 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -0.784460 0.214779 -3.652402 0.0020 R-squared 0.435893 Mean dependent var -0.000419 Adjusted R-squared 0.435893 S.D. dependent var 0.005236 S.E. of regression 0.003932 Akaike info criterion -8.185261 Sum squared resid 0.000263 Schwarz criterion -8.135796 Log likelihood 74.66735 Hannan-Quinn criter. -8.178440 Durbin-Watson stat 1.957083
S1.3. Round c: Lockdown 3.0 (May 04 – May 17)
S1.3.1. Unit-root test for GROWTHC
Null Hypothesis: GROWTHC has a unit root Exogenous: Constant Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -4.185030 0.0081 Test critical values: 1% level -4.057910 5% level -3.119910 10% level -2.701103 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 13 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC) Method: Least Squares Date: 07/08/20 Time: 12:29 Sample (adjusted): 5/05/2020 5/17/2020 Included observations: 13 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GROWTHC(-1) -0.723950 0.172986 -4.185030 0.0015 C 0.040648 0.010599 3.835088 0.0028 R-squared 0.614231 Mean dependent var -0.002831 Adjusted R-squared 0.579161 S.D. dependent var 0.011666 S.E. of regression 0.007568 Akaike info criterion -6.789090 Sum squared resid 0.000630 Schwarz criterion -6.702174 Log likelihood 46.12908 Hannan-Quinn criter. -6.806955 F-statistic 17.51447 Durbin-Watson stat 2.233425 Prob(F-statistic) 0.001524 S1.3.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -6.894809 0.0000 Test critical values: 1% level -2.754993 5% level -1.970978 10% level -1.603693 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 13 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 12:30 Sample (adjusted): 5/05/2020 5/17/2020 Included observations: 13 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -1.064886 0.154448 -6.894809 0.0000 R-squared 0.790394 Mean dependent var 0.004568 Adjusted R-squared 0.790394 S.D. dependent var 0.023782 S.E. of regression 0.010888 Akaike info criterion -6.128461 Sum squared resid 0.001423 Schwarz criterion -6.085004 Log likelihood 40.83500 Hannan-Quinn criter. -6.137394 Durbin-Watson stat 2.930780
S1.3.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -5.042100 0.0001 Test critical values: 1% level -2.754993 5% level -1.970978 10% level -1.603693 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 13 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 12:30 Sample (adjusted): 5/05/2020 5/17/2020 Included observations: 13 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -1.187481 0.235513 -5.042100 0.0003 R-squared 0.675744 Mean dependent var -0.000756 Adjusted R-squared 0.675744 S.D. dependent var 0.007430 S.E. of regression 0.004231 Akaike info criterion -8.018989 Sum squared resid 0.000215 Schwarz criterion -7.975532 Log likelihood 53.12343 Hannan-Quinn criter. -8.027922 Durbin-Watson stat 2.193739
S1.4. Round d: Lockdown 4.0 (May 18 – May 31)
S1.4.1. Unit-root test for GROWTHC
Null Hypothesis: GROWTHC has a unit root Exogenous: Constant, Linear Trend Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -3.858449 0.0478 Test critical values: 1% level -4.886426 5% level -3.828975 10% level -3.362984 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 13 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC) Method: Least Squares Date: 07/08/20 Time: 12:32 Sample (adjusted): 5/19/2020 5/31/2020 Included observations: 13 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GROWTHC(-1) -1.069484 0.277180 -3.858449 0.0032 C 0.061259 0.015776 3.882963 0.0030 @TREND("5/18/2020") -0.001017 0.000381 -2.667005 0.0236 R-squared 0.602591 Mean dependent var -6.36E-06 Adjusted R-squared 0.523110 S.D. dependent var 0.006069 S.E. of regression 0.004191 Akaike info criterion -7.912626 Sum squared resid 0.000176 Schwarz criterion -7.782253 Log likelihood 54.43207 Hannan-Quinn criter. -7.939423 F-statistic 7.581512 Durbin-Watson stat 1.356884 Prob(F-statistic) 0.009913
S1.4.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -3.601657 0.0017 Test critical values: 1% level -2.754993 5% level -1.970978 10% level -1.603693 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 13 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 12:32 Sample (adjusted): 5/19/2020 5/31/2020 Included observations: 13 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -0.797588 0.221450 -3.601657 0.0036 R-squared 0.508009 Mean dependent var 0.002644 Adjusted R-squared 0.508009 S.D. dependent var 0.017825 S.E. of regression 0.012503 Akaike info criterion -5.851947 Sum squared resid 0.001876 Schwarz criterion -5.808489 Log likelihood 39.03765 Hannan-Quinn criter. -5.860879 Durbin-Watson stat 1.283765
S1.4.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=2) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -5.392857 0.0001 Test critical values: 1% level -2.754993 5% level -1.970978 10% level -1.603693 *MacKinnon (1996) one-sided p-values. Warning: Probabilities and critical values calculated for 20 observations and may not be accurate for a sample size of 13 Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 12:33 Sample (adjusted): 5/19/2020 5/31/2020 Included observations: 13 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -1.386764 0.257148 -5.392857 0.0002 R-squared 0.707436 Mean dependent var 0.000200 Adjusted R-squared 0.707436 S.D. dependent var 0.005171 S.E. of regression 0.002797 Akaike info criterion -8.846919 Sum squared resid 9.39E-05 Schwarz criterion -8.803461 Log likelihood 58.50497 Hannan-Quinn criter. -8.855852 Durbin-Watson stat 2.166973
S1.5. Round e: Unlock 1.0 (June 01 – June 30)
S1.5.1. Unit-root test for GROWTHC
Null Hypothesis: D(GROWTHC) has a unit root Exogenous: Constant, Linear Trend Lag Length: 0 (Automatic - based on SIC, maxlag=7) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -6.131553 0.0001 Test critical values: 1% level -4.323979 5% level -3.580623 10% level -3.225334 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC,2) Method: Least Squares Date: 07/08/20 Time: 12:39 Sample (adjusted): 6/03/2020 6/30/2020 Included observations: 28 after adjustments Variable Coefficient Std. Error t-Statistic Prob. D(GROWTHC(-1)) -1.170694 0.190929 -6.131553 0.0000 C -0.000570 0.001281 -0.444976 0.6602 @TREND("6/01/2020") 7.36E-06 7.34E-05 0.100304 0.9209 R-squared 0.601475 Mean dependent var -0.000166 Adjusted R-squared 0.569593 S.D. dependent var 0.004775 S.E. of regression 0.003132 Akaike info criterion -8.593100 Sum squared resid 0.000245 Schwarz criterion -8.450364 Log likelihood 123.3034 Hannan-Quinn criter. -8.549464 F-statistic 18.86563 Durbin-Watson stat 2.161623 Prob(F-statistic) 0.000010
S1.5.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=7) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -6.063706 0.0000 Test critical values: 1% level -2.647120 5% level -1.952910 10% level -1.610011 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 12:45 Sample (adjusted): 6/02/2020 6/30/2020 Included observations: 29 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -1.020402 0.168280 -6.063706 0.0000 R-squared 0.565833 Mean dependent var -0.000980 Adjusted R-squared 0.565833 S.D. dependent var 0.015219 S.E. of regression 0.010028 Akaike info criterion -6.333000 Sum squared resid 0.002816 Schwarz criterion -6.285852 Log likelihood 92.82850 Hannan-Quinn criter. -6.318234 Durbin-Watson stat 2.210104
S1.5.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=7) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -4.720960 0.0000 Test critical values: 1% level -2.647120 5% level -1.952910 10% level -1.610011 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 12:46 Sample (adjusted): 6/02/2020 6/30/2020 Included observations: 29 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -0.881777 0.186779 -4.720960 0.0001 R-squared 0.443046 Mean dependent var 5.06E-05 Adjusted R-squared 0.443046 S.D. dependent var 0.003081 S.E. of regression 0.002300 Akaike info criterion -9.278205 Sum squared resid 0.000148 Schwarz criterion -9.231057 Log likelihood 135.5340 Hannan-Quinn criter. -9.263439 Durbin-Watson stat 1.899165
S1.6. Round f: Lockdown 1.0 to Unlock 1.0 (March 25 – June 30)
S1.6.1. Unit-root test for GROWTHC
Null Hypothesis: GROWTHC has a unit root Exogenous: None Lag Length: 11 (Automatic - based on SIC, maxlag=11) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -2.577656 0.0104 Test critical values: 1% level -2.592129 5% level -1.944619 10% level -1.614288 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC) Method: Least Squares Date: 07/08/20 Time: 15:58 Sample (adjusted): 4/06/2020 6/30/2020 Included observations: 86 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GROWTHC(-1) -0.075050 0.029115 -2.577656 0.0119 D(GROWTHC(-1)) -0.743955 0.101624 -7.320633 0.0000 D(GROWTHC(-2)) -0.411056 0.122011 -3.369021 0.0012 D(GROWTHC(-3)) -0.366975 0.112700 -3.256203 0.0017 D(GROWTHC(-4)) 0.066533 0.119279 0.557798 0.5787 D(GROWTHC(-5)) 0.383748 0.107652 3.564708 0.0006 D(GROWTHC(-6)) 0.211301 0.104451 2.022957 0.0467 D(GROWTHC(-7)) 0.127963 0.084095 1.521662 0.1324 D(GROWTHC(-8)) 0.068154 0.061528 1.107686 0.2716 D(GROWTHC(-9)) -0.106908 0.055595 -1.922985 0.0583 D(GROWTHC(-10)) -0.158045 0.048610 -3.251271 0.0017 D(GROWTHC(-11)) -0.124442 0.036883 -3.373981 0.0012 R-squared 0.858176 Mean dependent var -0.001527 Adjusted R-squared 0.837094 S.D. dependent var 0.033244 S.E. of regression 0.013418 Akaike info criterion -5.655668 Sum squared resid 0.013323 Schwarz criterion -5.313201 Log likelihood 255.1937 Hannan-Quinn criter. -5.517841 Durbin-Watson stat 2.149824
S1.6.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=11) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -10.46675 0.0000 Test critical values: 1% level -2.589020 5% level -1.944175 10% level -1.614554 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 15:59 Sample (adjusted): 3/26/2020 6/30/2020 Included observations: 97 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -1.000680 0.095606 -10.46675 0.0000 R-squared 0.532641 Mean dependent var -0.000733 Adjusted R-squared 0.532641 S.D. dependent var 0.027881 S.E. of regression 0.019061 Akaike info criterion -5.072111 Sum squared resid 0.034878 Schwarz criterion -5.045568 Log likelihood 246.9974 Hannan-Quinn criter. -5.061378 Durbin-Watson stat 2.127587 S1.6.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=11) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -10.79482 0.0000 Test critical values: 1% level -2.589020 5% level -1.944175 10% level -1.614554 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 15:59 Sample (adjusted): 3/26/2020 6/30/2020 Included observations: 97 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -1.096586 0.101584 -10.79482 0.0000 R-squared 0.548295 Mean dependent var 5.58E-06 Adjusted R-squared 0.548295 S.D. dependent var 0.006167 S.E. of regression 0.004145 Akaike info criterion -8.123748 Sum squared resid 0.001649 Schwarz criterion -8.097205 Log likelihood 395.0018 Hannan-Quinn criter. -8.113016 Durbin-Watson stat 1.795259 S1.7. Round h: Pre-lockdown to Unlock 1.0 (March 11 – June 30)
S1.7.1. Unit-root test for GROWTHC
Null Hypothesis: D(GROWTHC) has a unit root Exogenous: None Lag Length: 3 (Automatic - based on SIC, maxlag=12) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -11.07113 0.0000 Test critical values: 1% level -2.586753 5% level -1.943853 10% level -1.614749 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GROWTHC,2) Method: Least Squares Date: 07/08/20 Time: 16:04 Sample (adjusted): 3/16/2020 6/30/2020 Included observations: 107 after adjustments Variable Coefficient Std. Error t-Statistic Prob. D(GROWTHC(-1)) -3.225016 0.291300 -11.07113 0.0000 D(GROWTHC(-1),2) 1.444382 0.240247 6.012078 0.0000 D(GROWTHC(-2),2) 0.961579 0.169819 5.662363 0.0000 D(GROWTHC(-3),2) 0.436950 0.086341 5.060775 0.0000 R-squared 0.833401 Mean dependent var 0.001263 Adjusted R-squared 0.828548 S.D. dependent var 0.128502 S.E. of regression 0.053209 Akaike info criterion -2.992530 Sum squared resid 0.291608 Schwarz criterion -2.892611 Log likelihood 164.1004 Hannan-Quinn criter. -2.952024 Durbin-Watson stat 1.951236 S1.7.2. Unit-root test for GSENSEX
Null Hypothesis: GSENSEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=12) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -10.70777 0.0000 Test critical values: 1% level -2.585962 5% level -1.943741 10% level -1.614818 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GSENSEX) Method: Least Squares Date: 07/08/20 Time: 16:05 Sample (adjusted): 3/12/2020 6/30/2020 Included observations: 111 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GSENSEX(-1) -1.020718 0.095325 -10.70777 0.0000 R-squared 0.510363 Mean dependent var -2.76E-05 Adjusted R-squared 0.510363 S.D. dependent var 0.038513 S.E. of regression 0.026949 Akaike info criterion -4.380768 Sum squared resid 0.079888 Schwarz criterion -4.356358 Log likelihood 244.1326 Hannan-Quinn criter. -4.370865 Durbin-Watson stat 1.912996 S1.7.3. Unit-root test for GEX
Null Hypothesis: GEX has a unit root Exogenous: None Lag Length: 0 (Automatic - based on SIC, maxlag=12) t-Statistic Prob.* Augmented Dickey-Fuller test statistic -11.77495 0.0000 Test critical values: 1% level -2.585962 5% level -1.943741 10% level -1.614818 *MacKinnon (1996) one-sided p-values. Augmented Dickey-Fuller Test Equation Dependent Variable: D(GEX) Method: Least Squares Date: 07/08/20 Time: 16:05 Sample (adjusted): 3/12/2020 6/30/2020 Included observations: 111 after adjustments Variable Coefficient Std. Error t-Statistic Prob. GEX(-1) -1.110241 0.094288 -11.77495 0.0000 R-squared 0.557590 Mean dependent var 4.40E-05 Adjusted R-squared 0.557590 S.D. dependent var 0.006579 S.E. of regression 0.004376 Akaike info criterion -8.016300 Sum squared resid 0.002107 Schwarz criterion -7.991890 Log likelihood 445.9047 Hannan-Quinn criter. -8.006398 Durbin-Watson stat 1.924101 S2. Results of multiple linear regression models across different rounds
S2.1. Round a: Lockdown 1.0 (March 25 – April 14)
S2.1.1. MLRM for response GROWTHC
Dependent Variable: GROWTHC Method: Least Squares Date: 07/08/20 Time: 16:07 Sample: 3/25/2020 4/14/2020 Included observations: 21 Variable Coefficient Std. Error t-Statistic Prob. C 0.163634 0.020572 7.954115 0.0000 GSENSEX -0.176696 0.677014 -0.260993 0.7971 GEX 6.817109 3.703791 1.840576 0.0822* R-squared 0.203451 Mean dependent var 0.160851 Adjusted R-squared 0.114945 S.D. dependent var 0.097658 S.E. of regression 0.091874 Akaike info criterion -1.805235 Sum squared resid 0.151935 Schwarz criterion -1.656017 Log likelihood 21.95497 Hannan-Quinn criter. -1.772851 F-statistic 2.298734 Durbin-Watson stat 1.837904 Prob(F-statistic) 0.129096
S2.1.2. MLRM for response GSENSEX
Dependent Variable: GSENSEX Method: Least Squares Date: 07/08/20 Time: 16:08 Sample: 3/25/2020 4/14/2020 Included observations: 21 Variable Coefficient Std. Error t-Statistic Prob. C 0.010188 0.014999 0.679268 0.5056 GROWTHC -0.021336 0.081750 -0.260993 0.7971 GEX -2.060027 1.316234 -1.565092 0.1350 R-squared 0.166906 Mean dependent var 0.007212 Adjusted R-squared 0.074340 S.D. dependent var 0.033183 S.E. of regression 0.031926 Akaike info criterion -3.919257 Sum squared resid 0.018346 Schwarz criterion -3.770040 Log likelihood 44.15220 Hannan-Quinn criter. -3.886873 F-statistic 1.803097 Durbin-Watson stat 1.958781 Prob(F-statistic) 0.193307
S2.1.3. MLRM for response GEX
Dependent Variable: GEX Method: Least Squares Date: 07/08/20 Time: 16:26 Sample: 3/25/2020 4/14/2020 Included observations: 21 Variable Coefficient Std. Error t-Statistic Prob. C -0.003539 0.002412 -1.467526 0.1595 GROWTHC 0.023235 0.012624 1.840576 0.0822* GSENSEX -0.058147 0.037152 -1.565092 0.1350 R-squared 0.296211 Mean dependent var -0.000221 Adjusted R-squared 0.218012 S.D. dependent var 0.006065 S.E. of regression 0.005364 Akaike info criterion -7.486765 Sum squared resid 0.000518 Schwarz criterion -7.337548 Log likelihood 81.61104 Hannan-Quinn criter. -7.454381 F-statistic 3.787919 Durbin-Watson stat 2.467532 Prob(F-statistic) 0.042363*
S2.2. Round b: Lockdown 2.0 (April 15 – May 03)
S2.2.1. MLRM for response GROWTHC
Dependent Variable: GROWTHC Method: Least Squares Date: 07/08/20 Time: 16:39 Sample: 4/15/2020 5/03/2020 Included observations: 19 Variable Coefficient Std. Error t-Statistic Prob. C 0.072971 0.004116 17.72677 0.0000 GSENSEX -0.291959 0.413779 -0.705592 0.4906 GEX 0.100851 1.530551 0.065892 0.9483 R-squared 0.087252 Mean dependent var 0.071411 Adjusted R-squared -0.026841 S.D. dependent var 0.016682 S.E. of regression 0.016905 Akaike info criterion -5.178533 Sum squared resid 0.004572 Schwarz criterion -5.029411 Log likelihood 52.19607 Hannan-Quinn criter. -5.153296 F-statistic 0.764745 Durbin-Watson stat 1.741838 Prob(F-statistic) 0.481732
S2.2.2. MLRM for response GSENSEX
Dependent Variable: GSENSEX Method: Least Squares Date: 07/08/20 Time: 16:39 Sample: 4/15/2020 5/03/2020 Included observations: 19 Variable Coefficient Std. Error t-Statistic Prob. C 0.010307 0.010825 0.952161 0.3552 GROWTHC -0.103361 0.146489 -0.705592 0.4906 GEX -2.822533 0.575886 -4.901204 0.0002 R-squared 0.635001 Mean dependent var 0.005080 Adjusted R-squared 0.589376 S.D. dependent var 0.015696 S.E. of regression 0.010058 Akaike info criterion -6.216918 Sum squared resid 0.001619 Schwarz criterion -6.067796 Log likelihood 62.06072 Hannan-Quinn criter. -6.191680 F-statistic 13.91786 Durbin-Watson stat 1.615159 Prob(F-statistic) 0.000315*
S2.2.3. MLRM for response GEX
Dependent Variable: GEX Method: Least Squares Date: 07/08/20 Time: 16:40 Sample: 4/15/2020 5/03/2020 Included observations: 19 Variable Coefficient Std. Error t-Statistic Prob. C 0.000125 0.003054 0.040914 0.9679 GROWTHC 0.002690 0.040824 0.065892 0.9483 GSENSEX -0.212652 0.043388 -4.901204 0.0002* R-squared 0.623746 Mean dependent var -0.000763 Adjusted R-squared 0.576714 S.D. dependent var 0.004243 S.E. of regression 0.002761 Akaike info criterion -8.802649 Sum squared resid 0.000122 Schwarz criterion -8.653527 Log likelihood 86.62517 Hannan-Quinn criter. -8.777412 F-statistic 13.26221 Durbin-Watson stat 1.237533 Prob(F-statistic) 0.000402*
S2.3. Round c: Lockdown 3.0 (May 04 – May 17)
S2.3.1. MLRM for response GROWTHC
Dependent Variable: GROWTHC Method: Least Squares Date: 07/08/20 Time: 16:42 Sample: 5/04/2020 5/17/2020 Included observations: 14 Variable Coefficient Std. Error t-Statistic Prob. C 0.057291 0.002811 20.37766 0.0000 GSENSEX -0.324580 0.157552 -2.060143 0.0638* GEX 0.729374 0.600518 1.214575 0.2500 R-squared 0.429432 Mean dependent var 0.059747 Adjusted R-squared 0.325692 S.D. dependent var 0.012190 S.E. of regression 0.010010 Akaike info criterion -6.183114 Sum squared resid 0.001102 Schwarz criterion -6.046174 Log likelihood 46.28180 Hannan-Quinn criter. -6.195791 F-statistic 4.139519 Durbin-Watson stat 1.294806 Prob(F-statistic) 0.045676
S2.3.2. MLRM for response GSENSEX
Dependent Variable: GSENSEX Method: Least Squares Date: 07/08/20 Time: 16:42 Sample: 5/04/2020 5/17/2020 Included observations: 14 Variable Coefficient Std. Error t-Statistic Prob. C 0.045904 0.024857 1.846707 0.0918 GROWTHC -0.857766 0.416362 -2.060143 0.0638* GEX -0.283725 1.036099 -0.273839 0.7893 R-squared 0.357295 Mean dependent var -0.005595 Adjusted R-squared 0.240440 S.D. dependent var 0.018671 S.E. of regression 0.016272 Akaike info criterion -5.211315 Sum squared resid 0.002913 Schwarz criterion -5.074375 Log likelihood 39.47921 Hannan-Quinn criter. -5.223992 F-statistic 3.057587 Durbin-Watson stat 2.276306 Prob(F-statistic) 0.087915
S2.3.3. MLRM for response GEX
Dependent Variable: GEX Method: Least Squares Date: 07/08/20 Time: 16:43 Sample: 5/04/2020 5/17/2020 Included observations: 14 Variable Coefficient Std. Error t-Statistic Prob. C -0.008942 0.007798 -1.146666 0.2759 GROWTHC 0.162125 0.133483 1.214575 0.2500 GSENSEX -0.023864 0.087148 -0.273839 0.7893 R-squared 0.214641 Mean dependent var 0.000878 Adjusted R-squared 0.071848 S.D. dependent var 0.004898 S.E. of regression 0.004719 Akaike info criterion -7.686930 Sum squared resid 0.000245 Schwarz criterion -7.549989 Log likelihood 56.80851 Hannan-Quinn criter. -7.699607 F-statistic 1.503164 Durbin-Watson stat 2.639018 Prob(F-statistic) 0.264775
S2.4. Round d: Lockdown 4.0 (May 18 – May 31)
S2.4.1. MLRM for response GROWTHC
Dependent Variable: GROWTHC Method: Least Squares Date: 07/08/20 Time: 17:00 Sample: 5/18/2020 5/31/2020 Included observations: 14 Variable Coefficient Std. Error t-Statistic Prob. C 0.050839 0.001318 38.56308 0.0000 GSENSEX -0.058791 0.091211 -0.644559 0.5324 GEX 0.964298 0.464557 2.075739 0.0622* R-squared 0.281463 Mean dependent var 0.050459 Adjusted R-squared 0.150820 S.D. dependent var 0.005187 S.E. of regression 0.004780 Akaike info criterion -7.661285 Sum squared resid 0.000251 Schwarz criterion -7.524344 Log likelihood 56.62900 Hannan-Quinn criter. -7.673961 F-statistic 2.154444 Durbin-Watson stat 1.198755 Prob(F-statistic) 0.162357
S2.4.2. MLRM for response GSENSEX Dependent Variable: GSENSEX Method: Least Squares Date: 07/08/20 Time: 17:01 Sample: 5/18/2020 5/31/2020 Included observations: 14 Variable Coefficient Std. Error t-Statistic Prob. C 0.034774 0.048810 0.712439 0.4910 GROWTHC -0.619044 0.960414 -0.644559 0.5324 GEX 2.152433 1.655699 1.300015 0.2202 R-squared 0.133190 Mean dependent var 0.003097 Adjusted R-squared -0.024412 S.D. dependent var 0.015325 S.E. of regression 0.015511 Akaike info criterion -5.307098 Sum squared resid 0.002647 Schwarz criterion -5.170157 Log likelihood 40.14969 Hannan-Quinn criter. -5.319775 F-statistic 0.845101 Durbin-Watson stat 1.316734 Prob(F-statistic) 0.455599
S2.4.3. MLRM for response GEX
Dependent Variable: GEX Method: Least Squares Date: 07/08/20 Time: 17:01 Sample: 5/18/2020 5/31/2020 Included observations: 14 Variable Coefficient Std. Error t-Statistic Prob. C -0.015124 0.007131 -2.120914 0.0575 GROWTHC 0.291874 0.140612 2.075739 0.0622* GSENSEX 0.061873 0.047594 1.300015 0.2202 R-squared 0.353633 Mean dependent var -0.000205 Adjusted R-squared 0.236111 S.D. dependent var 0.003009 S.E. of regression 0.002630 Akaike info criterion -8.856362 Sum squared resid 7.61E-05 Schwarz criterion -8.719421 Log likelihood 64.99453 Hannan-Quinn criter. -8.869038 F-statistic 3.009094 Durbin-Watson stat 2.524826 Prob(F-statistic) 0.090706*
S2.5. Round e: Unlock 1.0 (June 01 – June 30)
S2.5.1. MLRM for response GSENSEX
Dependent Variable: GSENSEX Method: Least Squares Date: 07/08/20 Time: 17:03 Sample: 6/01/2020 6/30/2020 Included observations: 30 Variable Coefficient Std. Error t-Statistic Prob. C 0.002500 0.001911 1.308231 0.2014 GEX -1.392128 0.849926 -1.637940 0.1126 R-squared 0.087438 Mean dependent var 0.002527 Adjusted R-squared 0.054847 S.D. dependent var 0.010766 S.E. of regression 0.010467 Akaike info criterion -6.216930 Sum squared resid 0.003067 Schwarz criterion -6.123517 Log likelihood 95.25395 Hannan-Quinn criter. -6.187046 F-statistic 2.682849 Durbin-Watson stat 2.126028 Prob(F-statistic) 0.112625
S2.6. Round f: Lockdown 1.0 to Unlock 1.0 (March 25 – June 30) S2.6.1. MLRM for response GROWTHC
Dependent Variable: GROWTHC Method: Least Squares Date: 07/08/20 Time: 17:05 Sample: 3/25/2020 6/30/2020 Included observations: 98 Variable Coefficient Std. Error t-Statistic Prob. C 0.075868 0.006517 11.64243 0.0000 GSENSEX 0.079016 0.351351 0.224892 0.8225 GEX 3.556297 1.698950 2.093233 0.0390* R-squared 0.048075 Mean dependent var 0.075727 Adjusted R-squared 0.028034 S.D. dependent var 0.064690 S.E. of regression 0.063777 Akaike info criterion -2.636708 Sum squared resid 0.386416 Schwarz criterion -2.557576 Log likelihood 132.1987 Hannan-Quinn criter. -2.604701 F-statistic 2.398873 Durbin-Watson stat 0.935450 Prob(F-statistic) 0.096302*
S2.6.2. MLRM for response GSENSEX
Dependent Variable: GSENSEX Method: Least Squares Date: 07/08/20 Time: 17:06 Sample: 3/25/2020 6/30/2020 Included observations: 98 Variable Coefficient Std. Error t-Statistic Prob. C 0.002236 0.002955 0.756779 0.4511 GROWTHC 0.006734 0.029944 0.224892 0.8225 GEX -1.913943 0.467740 -4.091900 0.0001* R-squared 0.153385 Mean dependent var 0.002947 Adjusted R-squared 0.135561 S.D. dependent var 0.020025 S.E. of regression 0.018619 Akaike info criterion -5.099176 Sum squared resid 0.032932 Schwarz criterion -5.020044 Log likelihood 252.8596 Hannan-Quinn criter. -5.067169 F-statistic 8.605763 Durbin-Watson stat 2.130143 Prob(F-statistic) 0.000367*
S2.6.3. MLRM for response GEX
Dependent Variable: GEX Method: Least Squares Date: 07/08/20 Time: 17:06 Sample: 3/25/2020 6/30/2020 Included observations: 98 Variable Coefficient Std. Error t-Statistic Prob. C -0.000813 0.000594 -1.370114 0.1739 GROWTHC 0.012397 0.005923 2.093233 0.0390* GSENSEX -0.078289 0.019133 -4.091900 0.0001* R-squared 0.190280 Mean dependent var -0.000105 Adjusted R-squared 0.173233 S.D. dependent var 0.004141 S.E. of regression 0.003766 Akaike info criterion -8.295696 Sum squared resid 0.001347 Schwarz criterion -8.216564 Log likelihood 409.4891 Hannan-Quinn criter. -8.263689 F-statistic 11.16226 Durbin-Watson stat 2.357087 Prob(F-statistic) 0.000044* S3. Correlation summaries across different rounds
S3.1. Round a: Lockdown 1.0 (March 25 – April 14)
GROWTHC GSENSEX GEX GROWTHC 1 -0.2313756707880113 0.4477010580872346 GSENSEX -0.2313756707880113 1 -0.4046639177145718 GEX 0.4477010580872346 -0.4046639177145718 1
S3.2. Round b: Lockdown 2.0 (April 15 – May 03)
GROWTHC GSENSEX GEX GROWTHC 1 -0.2949655590059144 0.2425924212055322 GSENSEX -0.2949655590059144 1 -0.7897109760059276 GEX 0.2425924212055322 -0.7897109760059276 1
S3.3. Round c: Lockdown 3.0 (May 04 – May 17)
GROWTHC GSENSEX GEX GROWTHC 1 -0.594065752176441 0.4574787175868636 GSENSEX -0.594065752176441 1 -0.3306317126371629 GEX 0.4574787175868636 -0.3306317126371629 1
S3.4. Round d: Lockdown 4.0 (May 18 – May 31)
GROWTHC GSENSEX GEX GROWTHC 1 0.003590958809302893 0.5043063458499434 GSENSEX 0.003590958809302893 1 0.3169402889983672 GEX 0.5043063458499434 0.3169402889983672 1 S3.5. Round e: Unlock 1.0 (June 01 – June 30)
GROWTHC GSENSEX GEX GROWTHC 1 0.2009233369750571 0.1263611033248855 GSENSEX 0.2009233369750571 1 -0.2956992777823893 GEX 0.1263611033248855 -0.2956992777823893 1
S3.6. Round f: Lockdown 1.0 to Unlock 1.0 (March 25 – June 30)
GROWTHC GSENSEX GEX GROWTHC 1 -0.064572941470039 0.2181006766284255 GSENSEX -0.064572941470039 1 -0.3910676790329568 GEX 0.2181006766284255 -0.3910676790329568 1
S3.7. Round g: Pre-lockdown to Lockdown 1.0 (March 11 – April 14)
GROWTHC GSENSEX GEX GROWTHC 1 -0.1284512832334094 0.4014368887815389 GSENSEX -0.1284512832334094 1 -0.5201629315803736 GEX 0.4014368887815389 -0.5201629315803736 1
S3.8. Round h: Pre-lockdown to Unlock 1.0 (March 11 – June 30)
GROWTHC GSENSEX GEX GROWTHC 1 -0.1531029244245676 0.2808517762927862 GSENSEX -0.1531029244245676 1 -0.4680084500326706 GEX 0.2808517762927862 -0.4680084500326706 1
S4. Summaries of vector autoregression models across different rounds
S4.1. Round a: Lockdown 1.0 (March 25 – April 14) Vector Autoregression Estimates Date: 07/10/20 Time: 18:06 Sample (adjusted): 3/26/2020 4/14/2020 Included observations: 20 after adjustments Standard errors in ( ) & t-statistics in [ ] GROWTHC GSENSEX GEX GROWTHC(-1) -0.096864 0.165784 -0.017008 (0.28060) (0.07670) (0.01665) [-0.34520] [ 2.16137] [-1.02145] GSENSEX(-1) 0.088108 -0.031461 -0.027988 (0.79413) (0.21708) (0.04712) [ 0.11095] [-0.14493] [-0.59396] GEX(-1) 1.105948 -1.213721 -0.145464 (4.74500) (1.29705) (0.28156) [ 0.23308] [-0.93576] [-0.51664] C 0.173128 -0.023198 0.002741 (0.05305) (0.01450) (0.00315) [ 3.26364] [-1.59977] [ 0.87093] R-squared 0.008675 0.229388 0.116142 Adj. R-squared -0.177198 0.084898 -0.049581 Sum sq. resids 0.184703 0.013801 0.000650 S.E. equation 0.107443 0.029370 0.006375 F-statistic 0.046673 1.587573 0.700820 Log likelihood 18.46862 44.40862 74.95878 Akaike AIC -1.446862 -4.040862 -7.095878 Schwarz SC -1.247715 -3.841716 -6.896731 Mean dependent 0.157606 0.004083 -0.000218 S.D. dependent 0.099027 0.030702 0.006223 Determinant resid covariance (dof adj.) 2.54E-10 Determinant resid covariance 1.30E-10 Log likelihood 142.4779 Akaike information criterion -13.04779 Schwarz criterion -12.45035 Number of coefficients 12
S4.2. Round b: Lockdown 2.0 (April 15 – May 03)
Vector Autoregression Estimates Date: 07/11/20 Time: 06:50 Sample (adjusted): 4/16/2020 5/03/2020 Included observations: 18 after adjustments Standard errors in ( ) & t-statistics in [ ] GROWTHC GSENSEX GEX GROWTHC(-1) 0.108998 0.071118 0.024792 (0.26222) (0.25508) (0.06073) [ 0.41567] [ 0.27880] [ 0.40821] GSENSEX(-1) 0.532489 -0.038780 0.054235 (0.44174) (0.42971) (0.10231) [ 1.20543] [-0.09025] [ 0.53009] GEX(-1) 2.109313 -0.096562 0.302748 (1.60531) (1.56160) (0.37181) [ 1.31396] [-0.06184] [ 0.81426] C 0.062396 0.000973 -0.003043 (0.01968) (0.01915) (0.00456) [ 3.17020] [ 0.05082] [-0.66745] R-squared 0.122113 0.007481 0.062185 Adj. R-squared -0.066005 -0.205202 -0.138775 Sum sq. resids 0.004396 0.004160 0.000236 S.E. equation 0.017720 0.017238 0.004104 F-statistic 0.649130 0.035173 0.309440 Log likelihood 49.31588 49.81285 75.64446 Akaike AIC -5.035098 -5.090317 -7.960496 Schwarz SC -4.837238 -4.892456 -7.762636 Mean dependent 0.071340 0.005924 -0.001224 S.D. dependent 0.017163 0.015702 0.003846 Determinant resid covariance (dof adj.) 4.34E-13 Determinant resid covariance 2.04E-13 Log likelihood 186.3632 Akaike information criterion -19.37369 Schwarz criterion -18.78011 Number of coefficients 12
S4.3. Round c: Lockdown 3.0 (May 04 – May 17)
Vector Autoregression Estimates Date: 07/11/20 Time: 06:53 Sample (adjusted): 5/05/2020 5/17/2020 Included observations: 13 after adjustments Standard errors in ( ) & t-statistics in [ ] GROWTHC GSENSEX GEX GROWTHC(-1) 0.320293 -0.030908 -0.011748 (0.24960) (0.20481) (0.14084) [ 1.28322] [-0.15090] [-0.08341] GSENSEX(-1) 0.065734 -0.267138 -0.068037 (0.15345) (0.12592) (0.08659) [ 0.42837] [-2.12151] [-0.78576] GEX(-1) 0.083271 -1.834478 -0.273525 (0.52877) (0.43389) (0.29836) [ 0.15748] [-4.22794] [-0.91675] C 0.038309 0.000524 0.000744 (0.01454) (0.01193) (0.00820) [ 2.63527] [ 0.04397] [ 0.09069] R-squared 0.205515 0.706862 0.130332 Adj. R-squared -0.059313 0.609149 -0.159558 Sum sq. resids 0.000616 0.000415 0.000196 S.E. equation 0.008276 0.006791 0.004670 F-statistic 0.776031 7.234076 0.449591 Log likelihood 46.27093 48.84176 53.71006 Akaike AIC -6.503221 -6.898732 -7.647702 Schwarz SC -6.329390 -6.724901 -7.473871 Mean dependent 0.057227 -0.001457 0.000190 S.D. dependent 0.008041 0.010863 0.004337 Determinant resid covariance (dof adj.) 4.22E-14 Determinant resid covariance 1.40E-14 Log likelihood 152.0070 Akaike information criterion -21.53954 Schwarz criterion -21.01805 Number of coefficients 12
S4.4. Round d: Lockdown 4.0 (May 18 – May 31)
Vector Autoregression Estimates Date: 07/11/20 Time: 10:55 Sample (adjusted): 5/19/2020 5/31/2020 Included observations: 13 after adjustments Standard errors in ( ) & t-statistics in [ ] GROWTHC GSENSEX GEX GROWTHC(-1) 0.448112 -0.452878 0.039041 (0.32142) (0.77135) (0.20524) [ 1.39415] [-0.58713] [ 0.19022] GSENSEX(-1) -0.128920 0.169235 -0.013566 (0.09816) (0.23555) (0.06268) [-1.31342] [ 0.71846] [-0.21644] GEX(-1) -0.301169 -0.684895 -0.401475 (0.58049) (1.39304) (0.37066) [-0.51882] [-0.49165] [-1.08313] C 0.028297 0.028191 -0.002041 (0.01647) (0.03953) (0.01052) [ 1.71796] [ 0.71320] [-0.19401] R-squared 0.341703 0.151239 0.169498 Adj. R-squared 0.122271 -0.131682 -0.107336 Sum sq. resids 0.000227 0.001309 9.26E-05 S.E. equation 0.005025 0.012058 0.003208 F-statistic 1.557215 0.534563 0.612274 Log likelihood 52.75850 41.37858 58.58994 Akaike AIC -7.501307 -5.750550 -8.398453 Schwarz SC -7.327477 -5.576720 -8.224622 Mean dependent 0.050619 0.005979 -2.08E-05 S.D. dependent 0.005363 0.011335 0.003049 Determinant resid covariance (dof adj.) 2.51E-14 Determinant resid covariance 8.33E-15 Log likelihood 155.3880 Akaike information criterion -22.05969 Schwarz criterion -21.53820 Number of coefficients 12
S4.5. Round e: Unlock 1.0 (June 01 – June 30)
Vector Autoregression Estimates Date: 07/11/20 Time: 10:59 Sample (adjusted): 6/03/2020 6/30/2020 Included observations: 28 after adjustments Standard errors in ( ) & t-statistics in [ ] D(GROWTHC) GSENSEX GEX D(GROWTHC(-1)) -0.188619 1.230814 0.095264 (0.18958) (0.56261) (0.13823) [-0.99494] [ 2.18767] [ 0.68918] GSENSEX(-1) 0.085598 -0.289173 -0.006063 (0.06150) (0.18252) (0.04484) [ 1.39180] [-1.58435] [-0.13521] GEX(-1) -0.023454 -0.865482 0.052088 (0.26410) (0.78378) (0.19257) [-0.08880] [-1.10424] [ 0.27049] C -0.000613 0.002012 0.000239 (0.00059) (0.00176) (0.00043) [-1.03575] [ 1.14554] [ 0.55476] R-squared 0.113041 0.218625 0.026743 Adj. R-squared 0.002171 0.120953 -0.094914 Sum sq. resids 0.000225 0.001979 0.000119 S.E. equation 0.003060 0.009080 0.002231 F-statistic 1.019584 2.238361 0.219821 Log likelihood 124.5317 94.07367 133.3770 Akaike AIC -8.609405 -6.433833 -9.241217 Schwarz SC -8.419090 -6.243518 -9.050902 Mean dependent -0.000414 0.001179 0.000206 S.D. dependent 0.003063 0.009685 0.002132 Determinant resid covariance (dof adj.) 2.86E-15 Determinant resid covariance 1.80E-15 Log likelihood 356.0975 Akaike information criterion -24.57839 Schwarz criterion -24.00745 Number of coefficients 12 S4.6. Round f: Lockdown 1.0 to Unlock 1.0 (March 25 – June 30) Vector Autoregression Estimates Date: 07/11/20 Time: 11:05 Sample (adjusted): 3/30/2020 6/30/2020 Included observations: 93 after adjustments Standard errors in ( ) & t-statistics in [ ] GROWTHC GSENSEX GEX GROWTHC(-1) 0.106327 0.093560 -0.010703 (0.09433) (0.04075) (0.00934) [ 1.12719] [ 2.29600] [-1.14606] GROWTHC(-2) 0.125962 -0.158900 0.014734 (0.09668) (0.04176) (0.00957) [ 1.30288] [-3.80467] [ 1.53936] GROWTHC(-3) 0.048033 0.113527 0.002565 (0.09708) (0.04194) (0.00961) [ 0.49477] [ 2.70703] [ 0.26688] GROWTHC(-4) 0.079726 -0.091596 -0.005485 (0.10133) (0.04377) (0.01003) [ 0.78679] [-2.09249] [-0.54675] GROWTHC(-5) 0.547137 0.069865 0.005569 (0.09649) (0.04168) (0.00955) [ 5.67021] [ 1.67605] [ 0.58300] GSENSEX(-1) 0.290082 0.072721 -0.005520 (0.27116) (0.11714) (0.02684) [ 1.06980] [ 0.62082] [-0.20561] GSENSEX(-2) -0.574185 0.037102 -0.025701 (0.26446) (0.11424) (0.02618) [-2.17117] [ 0.32476] [-0.98164] GSENSEX(-3) -0.587609 0.130703 -0.020135 (0.25823) (0.11155) (0.02557) [-2.27554] [ 1.17168] [-0.78758] GSENSEX(-4) 0.212495 -0.255499 0.002964 (0.26171) (0.11306) (0.02591) [ 0.81194] [-2.25991] [ 0.11439] GSENSEX(-5) 0.219097 -0.180139 0.005189 (0.23543) (0.10170) (0.02331) [ 0.93063] [-1.77124] [ 0.22264] GEX(-1) 2.132429 -1.240362 -0.011273 (1.25553) (0.54237) (0.12430) [ 1.69844] [-2.28691] [-0.09070] GEX(-2) -2.875951 1.054550 -0.219043 (1.28579) (0.55545) (0.12730) [-2.23672] [ 1.89856] [-1.72074] GEX(-3) -1.244016 -0.692982 -0.253952 (1.32514) (0.57245) (0.13119) [-0.93878] [-1.21056] [-1.93574] GEX(-4) 0.724500 -0.064064 -0.015019 (1.23977) (0.53557) (0.12274) [ 0.58438] [-0.11962] [-0.12236] GEX(-5) 1.058369 -0.628109 -0.051972 (1.16425) (0.50294) (0.11526) [ 0.90906] [-1.24887] [-0.45090] C 0.003972 0.000420 -0.000394 (0.00754) (0.00326) (0.00075) [ 0.52669] [ 0.12878] [-0.52713] R-squared 0.686938 0.348797 0.105779 Adj. R-squared 0.625952 0.221939 -0.068420 Sum sq. resids 0.111975 0.020896 0.001098 S.E. equation 0.038134 0.016474 0.003775 F-statistic 11.26386 2.749513 0.607230 Log likelihood 180.6156 258.6754 395.6888 Akaike AIC -3.540120 -5.218825 -8.165350 Schwarz SC -3.104404 -4.783109 -7.729634 Mean dependent 0.072244 0.001871 1.56E-05 S.D. dependent 0.062352 0.018676 0.003652 Determinant resid covariance (dof adj.) 4.14E-12 Determinant resid covariance 2.35E-12 Log likelihood 849.2679 Akaike information criterion -17.23157 Schwarz criterion -15.92442 Number of coefficients 48 VAR Residual Serial Correlation LM Tests Date: 07/11/20 Time: 11:06 Sample: 3/25/2020 6/30/2020 Included observations: 93 Null hypothesis: No serial correlation at lag h Lag LRE* stat df Prob. Rao F-stat df Prob. 1 24.33359 9 0.0038 2.839592 (9, 175.4) 0.0038 2 12.85928 9 0.1691 1.452433 (9, 175.4) 0.1692 3 12.91615 9 0.1664 1.459091 (9, 175.4) 0.1666 4 20.56982 9 0.0147 2.374738 (9, 175.4) 0.0147 5 6.490126 9 0.6900 0.719998 (9, 175.4) 0.6901 6 8.963914 9 0.4406 1.001384 (9, 175.4) 0.4408 Null hypothesis: No serial correlation at lags 1 to h Lag LRE* stat df Prob. Rao F-stat df Prob. 1 24.33359 9 0.0038 2.839592 (9, 175.4) 0.0038 2 34.00352 18 0.0126 1.975189 (18, 195.6) 0.0127 3 45.77812 27 0.0134 1.783477 (27, 193.4) 0.0137 4 58.60878 36 0.0100 1.728888 (36, 186.9) 0.0105 5 77.10891 45 0.0020 1.865565 (45, 179.0) 0.0022 6 108.0107 54 0.0000 2.313120 (54, 170.7) 0.0000 *Edgeworth expansion corrected likelihood ratio statistic. S4.7. Round g: Pre-lockdown to Lockdown 1.0 (March 11 – April 14)
Vector Autoregression Estimates Date: 07/11/20 Time: 11:24 Sample (adjusted): 3/12/2020 4/14/2020 Included observations: 34 after adjustments Standard errors in ( ) & t-statistics in [ ] GROWTHC GSENSEX GEX GROWTHC(-1) -0.059037 0.109057 -0.007271 (0.19580) (0.09091) (0.01243) [-0.30152] [ 1.19964] [-0.58475] GSENSEX(-1) 0.578975 0.014340 -0.016787 (0.44759) (0.20781) (0.02842) [ 1.29354] [ 0.06901] [-0.59059] GEX(-1) 2.687893 0.382948 -0.224175 (3.50390) (1.62685) (0.22252) [ 0.76712] [ 0.23539] [-1.00745] C 0.179608 -0.022226 0.002212 (0.03635) (0.01688) (0.00231) [ 4.94171] [-1.31707] [ 0.95820] R-squared 0.053721 0.065679 0.067779 Adj. R-squared -0.040907 -0.027753 -0.025443 Sum sq. resids 0.272500 0.058743 0.001099 S.E. equation 0.095306 0.044250 0.006052 F-statistic 0.567705 0.702956 0.727067 Log likelihood 33.80622 59.89208 127.5317 Akaike AIC -1.753307 -3.287770 -7.266570 Schwarz SC -1.573735 -3.108198 -7.086998 Mean dependent 0.169509 -0.003490 0.000876 S.D. dependent 0.093415 0.043649 0.005977 Determinant resid covariance (dof adj.) 3.88E-10 Determinant resid covariance 2.66E-10 Log likelihood 230.0444 Akaike information criterion -12.82614 Schwarz criterion -12.28743 Number of coefficients 12
S4.8. Round h: Pre-lockdown to Unlock 1.0 (March 11 – June 30)
Vector Autoregression Estimates Date: 07/11/20 Time: 11:27 Sample (adjusted): 3/16/2020 6/30/2020 Included observations: 107 after adjustments Standard errors in ( ) & t-statistics in [ ] D(GROWTHC) GSENSEX GEX D(GROWTHC(-1)) -0.809543 0.065822 -0.006828 (0.09653) (0.04221) (0.00759) [-8.38671] [ 1.55946] [-0.89974] D(GROWTHC(-2)) -0.476650 -0.140864 0.014838 (0.11388) (0.04980) (0.00895) [-4.18549] [-2.82879] [ 1.65736] D(GROWTHC(-3)) -0.524470 -0.002937 0.011770 (0.11008) (0.04813) (0.00865) [-4.76437] [-0.06102] [ 1.36005] D(GROWTHC(-4)) -0.379081 -0.108689 0.005617 (0.09964) (0.04357) (0.00783) [-3.80451] [-2.49463] [ 0.71701] GSENSEX(-1) 0.075213 0.248085 -0.042861 (0.24439) (0.10686) (0.01921) [ 0.30776] [ 2.32152] [-2.23080] GSENSEX(-2) -0.342825 -0.142794 -0.026046 (0.24158) (0.10563) (0.01899) [-1.41912] [-1.35179] [-1.37141] GSENSEX(-3) -0.483460 -0.090150 0.015680 (0.24169) (0.10568) (0.01900) [-2.00037] [-0.85304] [ 0.82523] GSENSEX(-4) -0.373565 -0.026276 -0.024239 (0.23232) (0.10159) (0.01826) [-1.60795] [-0.25865] [-1.32712] GEX(-1) 0.698684 -0.543759 -0.104594 (1.45825) (0.63764) (0.11464) [ 0.47912] [-0.85276] [-0.91234] GEX(-2) -3.011890 1.663799 -0.215196 (1.42298) (0.62222) (0.11187) [-2.11661] [ 2.67397] [-1.92362] GEX(-3) 0.325406 -1.465898 -0.09059871