Exploring the linkages between BSE Sensex, crude oil prices and exchange rates: the financial triad in India

The relationship between crude oil prices, exchange rates and stock markets is widely studied. Recent global and domestic events have made it more relevant. India is one of the world’s largest crude oil importers. This makes the country particularly vulnerable to changes in global oil prices. The 2022 Russia–Ukraine conflict caused a surge in oil prices (Mohanty, 2023). This increased India’s import costs and led to inflationary pressures (Anand et al., 2023; Sarkar and Gupta, 2024). Similarly, the COVID-19 pandemic in 2020–21 triggered extreme stock market volatility (Li et al., 2022). It also caused exchange rate fluctuations, highlighting the interconnectedness of these factors. India’s limited domestic oil production forces a heavy reliance on imports. This makes the economy highly sensitive to crude oil price shocks and exchange rate movements. The instability in the dollar–rupee exchange rate amplifies these effects. It impacts corporate profitability, inflation and foreign investment flows. Because the stock market is a key barometer of economic health, understanding how these factors influence the BSE Sensex is essential. This study becomes timely as India navigates ongoing economic uncertainties. It highlights the importance of understanding the dynamic linkages between these financial variables.

In the financial system, BSE Sensex, crude oil and exchange rates are not just connected. They also influence each other. Crude oil prices are important for economic activities. Changes in these prices affect production costs, inflation and consumer spending (Ghosh and Kanjilal, 2016; Sahu, 2016). When oil prices rise, business expenses increase (Berument and Taşçı, 2002). This reduces profits and creates negative investor sentiment, lowering stock values. Conversely, lower oil prices can decrease the cost and improve market performance. Similar to crude oil, an exchange rate influences costs, such as crude oil, and also affects company profits and inflation. A weaker currency at home will make imports costlier (Black, 2019). Conversely, a stable and strong domestic currency would reduce the costs of imports and stabilize the stock market. Exchange rates and crude oil prices determine stock market trends. Consequently, any shock to either of these variables can induce a change in investor behavior and cause a market change. Oil prices determine the trend of the stock market (Agarwalla et al., 2021). Any shock to these variables can shift investor behavior and cause market changes.

Based on this theoretical context, numerous global studies have explored the relationship between these variables. Tian et al. (2021) observed mixed relationships, while Nwosa (2021) identified correlations in Nigeria during the pandemic. Notable Indian studies, such as those by Sahu et al. (2015), Sharma et al. (2018) and Sreenu (2022), found long-run relationships among the variables. However, most of these studies relied on daily or monthly data, leaving a gap in the exploration of annual trends. Addressing this gap, the current study will focus on analyzing annual trends between the variables. This study empirically examines the dynamic relationship among crude oil prices, exchange rates and the Indian stock market. Examining the correlation among oil prices, exchange rates and emerging stock market valuations is a significant area of study, as the growth and prosperity of emerging economies will increasingly impact the global economy. The association between crude oil prices, exchange rates and BSE Sensex has been extensively studied.

Economic and financial literature has extensively explored the relationship between crude oil prices, exchange rates and stock markets. Earlier studies primarily examined bilateral relationships, such as the impact of oil prices on stock markets or exchange rates, without fully addressing their interconnected nature. However, recent research has shifted toward a more integrated perspective, recognizing the complex interplay among these three variables and their influence on macroeconomic stability and financial markets.

Sadorsky (2001) and Boyer and Filion (2007) were among the first to document the positive correlation between oil prices and energy stock returns, particularly in Canada. Their findings suggested that stock prices of oil and gas producers tend to rise with increasing crude oil prices. Conversely, Miller and Ratti (2009) observed that global stock markets generally exhibit a negative response to oil price fluctuations, highlighting the adverse effects of oil price shocks on economic activity. Further, Anoruo (2011) provided evidence of bidirectional causality between oil prices and stock market returns in the USA, reinforcing the notion that these variables are interdependent. In emerging economies, Basher et al. (2012) analyzed the interactions between oil prices, exchange rates and stock markets using a structural vector autoregression model. Their results demonstrated that positive oil price shocks tend to depress emerging market stock prices and weaken the US dollar exchange rate in the short run. Notably, their study highlighted that stock prices in emerging markets could also drive oil prices, emphasizing the demand-side influence of rising equity valuations.

Several studies have examined the linkages between crude oil prices, exchange rates and stock markets in India, given its dependence on oil imports. Sharma and Shrivastava (2024) used the Granger causality test and Johansen cointegration method to analyze macroeconomic variables over a 30-year period (1991–2020). They found that oil price fluctuations exert bidirectional short-run effects on inflation, industrial production and unemployment. The study underscored that India’s exchange rate is highly sensitive to oil price volatility, with cheaper crude oil reducing inflationary pressures and stabilizing interest rates. Kumar (2019) provided additional insights into the asymmetric nature of these relationships by using nonlinear econometric techniques, including the Hiemstra and Jones (1994) nonlinear Granger causality test and the nonlinear autoregressive distributed lag model. His study confirmed bidirectional causality between oil prices, exchange rates and stock prices in India, with asymmetric effects of oil price shocks. Positive oil price shocks had a stronger influence on exchange rates and stock prices than negative ones, emphasizing the importance of incorporating nonlinearity in financial modeling.

More recent research has refined these findings by incorporating advanced methodologies. Tiwari et al. (2024) applied a cross-quantilogram approach to assess directional predictability between energy markets, exchange rates and stock markets in emerging economies. Their results demonstrated significant causality from oil prices to financial markets, with geopolitical risks playing a crucial role in shaping these interactions. Additionally, their study highlighted that natural gas returns could serve as a short-term predictor of exchange rate movements, particularly in India, Brazil and Mexico. Agarwala et al. (2023) attempt to explain the dynamic relationship between international crude oil prices and Indian stock prices (as represented by the Bombay Stock Exchange [BSE] energy index). Consequently, they have identified a long-term association between the variables and the Granger causality of crude oil on stock prices in India.

Several studies have emphasized the evolving nature of these relationships, particularly in response to financial crises and geopolitical events. Research using wavelet coherence analysis has revealed that the strength and direction of interactions between oil prices, stock markets and exchange rates fluctuate over time, reinforcing the necessity of adopting time-varying models. For instance, studies on BRICS economies suggest that oil price volatility plays a significant role in determining exchange rate movements, especially in oil-importing nations such as India (Nath Sahu et al., 2014). Kapusuzoglu (2011) investigated the causality between oil prices and stock market indices, uncovering a persistent correlation over extended periods. Similarly, Delgado et al. (2018) examined the effects of oil price fluctuations on exchange rates and stock markets in Mexico, concluding that crude oil exerts a statistically significant negative influence on exchange rates, whereas exchange rates positively impact stock indices. A further study examined the interrelationship of oil prices, exchange rates and inflation in the MENA region, specifically in relation to the COVID-19 epidemic and the Ukraine crisis. The study revealed that the impact of oil prices varied from country to country. The impact varies based on the direction and duration of the shock (Bigerna, 2024).

Based on the literature review, the outcomes illustrated how different nations react differently because of their unique consumer perceptions and national policies. In the same context, several research studies have been done in India. In most of the studies, the authors analyzed the monthly and daily data. The analysis of annual trends was left unfinished. This became the motivation for the study. Therefore, the objective of the study is to explore the linkages between the financial triad of BSE Sensex, crude oil prices and exchange rates in India using annual data.

Annual trends help to identify long-term patterns and cycles (Padachi, 2006; Shiller, 1992). It helps to minimize the noise’s short-term volatility (Chang et al., 2018; Kuchar et al., 2024). This will help policymakers understand meaningful matters. Annual trends are critical for understanding how these policies influence variables such as stock market performance, crude oil prices or exchange rates over a longer horizon. Measuring these trends helps to capture the medium-to-long-term impact of such events, such as the Iraq War (2003), the 2008 financial crisis, the 2020 COVID-19 pandemic and the Russia–Ukraine conflict (2022), which are analyzed.

The study period (2000–01 to 2021–22) was selected to include the maximum available annual data points across all variables, ensuring consistency and comparability. Data for crude oil prices from the Ministry of Petroleum and Natural Gas of the Government of India database is only available from 2000 onwards. To maintain uniformity, the same period was applied to all variables. The data set spans 22 years and includes BSE closing data sourced from Yahoo Finance; crude oil prices from the Ministry of Petroleum and Natural Gas of the Government of India; and exchange rates from the Reserve Bank of India.

To understand the nature of the data, this study first conducts a descriptive analysis. The characteristics of the variables obtained from this analysis are then used in the stationarity test (unit root test). This unit root test assists in ensuring reliable results for further econometric analysis.

A stationary test is a statistical procedure used to determine the statistical properties of the variables (mean, variance and autocorrelation) remain constant over time. In this context, the authors have chosen the Augmented Dickey–Fuller (ADF) test and Philips–Parron (PP) test (Gregory et al., 1996). Both tests are crucial for confirming stationarity.

Johansen’s cointegration test is used to ascertain the long-term equilibrium relationship among the variables. It is a vector autoregressive (VAR)-based approach suggested by Johansen (1988). In econometric terminology, two variables are considered cointegrated if they exhibit a long-term or equilibrium relationship. The concept of cointegration becomes particularly essential when the time series under analysis are non-stationary. The test checked whether the crude oil price, exchange rate and BSE Sensex are correlated in the long term.

According to Engle and Granger (1987), if there is a cointegrating vector present, then an error correction framework should be adopted. Based on that, the present study used the vector error correction model (VECM). It is a restricted form of the VAR model. It is designed to capture both short-term dynamics and long-term relationships between cointegrated variables. The VECM effectively breaks the dynamics into short-term adjustments and long-term equilibrium corrections by incorporating an error correction term (ECT). This term corrects the system from deviation in the long-run equilibrium. The model includes both differenced terms to capture short-term fluctuations and ECTs to ensure convergence to long-term stability, making it highly useful in econometric analysis.

VECM Granger causality refers to the examination of causal relationships between cointegrated time series variables within the context of the VECM. In this framework, Granger causality is tested by assessing whether past values of one variable contain useful information for predicting future values of another, after accounting for the cointegrated relationship between them. Specifically, the ECT in the VECM reflects the long-term equilibrium, and the Granger causality test can help determine whether short-term movements in one variable can influence or “cause” changes in another variable. This process involves testing the significance of lagged values of each variable to explain the movements of others, offering insight into the directionality and strength of causal links between them. It involves both the F-test and the chi-square (χ²) test.

This study applies the variance decomposition test (VDT) and impulse response function (IRF) for a more comprehensive estimation of the relationship between variables. The VDT helps determine the extent of endogeneity or exogeneity of the variables. In contrast, the IRF forecasts how a variable responds over time when an impulse is applied to the VAR system. By using both IRF and VDT, the study confirms the degree, direction and time length of the variables’ responses.

In the study, the authors transformed all three variables into logarithmic forms to better fit the statistical analysis. Specifically:

1. BSE closing data was converted to “LNBSE.”

2. The price of crude oil data was converted to “LNCRUDE_OIL.”

3. Exchange rate-related data was converted to “LNEXCHANGE_RATE.”

By applying logarithmic transformations, the authors aimed to normalize the distribution of these variables and reduce skewness, allowing for more robust statistical analysis and interpretation of results.

Based on the variables, econometric models are:

In the provided notation:

• “β₀”, “β₄” and “β₈” are the constant term of the model.

• “e_t1”, “e_t2” and “e_t3” are the error term of the model.

• “t-i” represents the lagged value of the model.

Before analysis, the authors run the descriptive statistics of the variables. The analysis will determine the nature of the data by examining its central tendency, dispersion and normality. This summary of the statistics is represented in Table 1.

From the table, it has been observed that variables show moderate to low stability throughout the study. The BSE Sensex shows the highest mean and reflects upward growth over time. Both crude oil and BSE Sensex show a high standard deviation, indicating variability in the data set. On the other hand, the exchange rate shows a moderate level of variability based on the standard deviation. The skewness values indicate a moderate asymmetry across all the variables. The BSE Sensex exhibits an extreme positive skew, while the other two variables display a slight positive skew. Kurtosis values for all variables are close to or below 3, indicating distributions that are not excessively peaked or heavy-tailed. The Jarque–Bera test confirms normality for all variables, with p-values above 0.05.

All the variables are transformed to their natural log form. It normalizes the scale of the variables, reduces skewness, stabilizes variance and makes the data more suitable for econometric analysis. These adjustments enhance the reliability and interpretability of results.

Following the descriptive statistics, researchers move on to a stationarity test, which is vital for time series analysis to ensure the data’s statistical properties remain stable over time. The study used the ADF and PP tests for the stationary check. The ADF test extends the basic Dickey–Fuller test by including lagged differences of the series to account for higher-order autocorrelation, making it more robust in identifying unit roots. Similarly, the PP test also examines the presence of a unit root but adjusts for serial correlation and heteroskedasticity in the error terms without adding lagged difference terms. In a stationary test, the hypothesis is formulated to determine whether a time series is stationary or not. The null hypothesis (H0) states that the time series contains a unit root, meaning that it is not stationary and exhibits a trend or other non-stationary behavior.

Table 2 shows the result of the unit root test of the data. All the variables have p-values greater than 0.05 in both tests. Therefore, the authors fail to reject the null hypothesis, indicating that the data has a unit root and is not stationary. However, in contrast, all the variables are stationary and exhibit no unit root at the first difference, as evidenced by p-values less than 0.05 for all variables.

As mentioned earlier, the VAR model required an appropriate lag length. Table 3 shows the results of the optimal lag length. Including more lags can enhance the model’s explanatory power but may also introduce multicollinearity. The authors selected Lag 3 for further analysis based on the Akaike information criterion (AIC), final prediction error (FPE) and Hannan–Quinn information criterion (HQ).

As the variables are stationary at I(1), the authors go for the Johansen co-integration test. It is a powerful statistical method used to determine the presence and number of cointegrating relationships between multiple non-stationary time series variables. The test involves two statistics: the trace statistic and the maximum eigenvalue statistic, both of which assess the rank of the cointegration matrix. The trace statistic tests the null hypothesis of no cointegration against the alternative of at least one cointegrating vector, while the maximum eigenvalue statistic tests the null of r cointegrating vectors against the alternative of r + 1 vectors. By identifying cointegrating relationships, the Johansen test helps in understanding long-term equilibrium relationships among economic variables, providing a robust foundation for further analysis such as VECM, which can capture both short-term dynamics and long-term equilibria.

Table 4 shows the results of the Johansen cointegration test, where the authors found that the trace statistic and maximum eigenvalue statistic are 52.89 and 33.22, respectively, when there is no cointegrating vector in the equation (r = 0). The p-value for both statistics is less than 0.05, leading the authors to reject the null hypothesis at the 5% level of significance. Similarly, the trace statistic and maximum eigenvalue statistic are 19.67 and 19.02, respectively, with p-values less than 0.05 when the null hypothesis suggests there is only one cointegrating vector present in the model (r = 1). Therefore, the authors conclude that there is more than one cointegrating vector present in the model. This conclusion is further confirmed as the p-value is greater than 0.05 when the null hypothesis suggests that there are two cointegrating vectors in the equation (r = 2). Thus, the authors fail to reject the null hypothesis and conclude that there are two cointegrating vectors present in the model.

The VECM is an econometric model that integrates short-term dynamics with long-run equilibrium relationships between time series variables.

In the VECM, long-run dynamics refer to the equilibrium relationships among the cointegrated variables that persist over time. These dynamics are captured through the cointegrating vectors, which represent the long-term equilibrium conditions that the variables tend to revert to, even though they may deviate in the short term. Here, the long-run cointegration equations are as follows:

Based on the cointegration equations, the exchange rate has a positive and significant impact on the Sensex. This indicates that an increase in the exchange rate is associated with an increase in the Sensex value. Conversely, the exchange rate has a negative and significant impact on crude oil prices, suggesting that an increase in the exchange rate leads to a decrease in crude oil prices.

• The error correction term

The ECT in the context of the given Table 5 represents the coefficients of the cointegrating equations (ECT(γ₁) and ECT(γ₂)) in the error correction model. These terms indicate how deviations from the long-run equilibrium relationship (cointegrating vectors) impact the short-term dynamics of the dependent variable. Here, in the case of ΔLNBSE, the speed of adjustment, or the ECT, is negative and significant. The table indicates that 36.53% for ECT(γ₁) and 68.28% for ECT(γ₂) represent the speed of adjustment. A larger magnitude implies a faster adjustment back to equilibrium. In this case, ECT(γ₂) has a larger magnitude, suggesting that deviations from the equilibrium defined by ECT(γ₂) are corrected more quickly than those defined by ECT(γ₁).

The results differ in the case of ΔLNEXCHANGE_RATE and ΔLNCRUDE_OIL. For ΔLNCRUDE_OIL, the ECT (γ₁) is insignificant, and the coefficient for ECT (γ₂), which ideally should lie between 0 and −1, exceeds −1. This suggests that the term is insignificant. In the case of LNEXCHANGE_RATE, both ECT coefficients are significant, but ECT (γ₂) is positive, indicating that 35.25% of the deviation from the long-run equilibrium. This positive ECT suggests instability rather than a return to equilibrium. Therefore, the authors conclude that there is no short-run relationship between these variables, as the expected correction mechanism does not function as theorized.

• Explanatory variables

In Table 5, the first lag of ΔLNBSE (ΔLNBSE(−1)) has a significant negative impact on the current period’s change in ΔLNBSE, where a 1% increase in ΔLNBSE(−1) leads to approximately a 0.55% decrease in ΔLNBSE in the current period. Conversely, the second lag of ΔLNBSE (ΔLNBSE(−2)) has a positive and statistically significant impact on the current period’s change in ΔLNBSE, with a 1% increase in ΔLNBSE(−2) resulting in approximately a 0.44% increase in ΔLNBSE in the current period. In contrast, ΔLNBSE(−3) has no impact on ΔLNBSE, as it is not statistically significant.

Similarly, the coefficients indicate that the first and second lags of ΔLNCRUDE_OIL have a positive and significant impact on ΔLNBSE, with a 1% increase in ΔLNCRUDE_OIL(−1) and ΔLNCRUDE_OIL(−2) leading to approximately a 0.99% and 0.46% increase in ΔLNBSE, respectively. ΔLNCRUDE_OIL(−3) is not significant, indicating no effect on ΔLNBSE in the short run.

On the other hand, the first and second lags of ΔLNEXCHANGE_RATE have a positive and significant impact on the dependent variable, with a 1% increase in the first and second lags of ΔLNEXCHANGE_RATE leading to approximately a 3.6% and 4.9% increase in ΔLNBSE, respectively. Although the third lag of ΔLNEXCHANGE_RATE and the constant term are statistically insignificant.

Therefore, the authors can conclude that ΔLNBSE is dependent on itself and other variables as well. With an R-squared value of 0.93, it can be said that the variables well explain the dependent variable.

Proposed model:

In the short run, ΔLNCRUDE_OIL is not explained by its own or any other variables. On the other hand, ΔLNEXCHANGE_RATE is slightly explained by ΔLNCRUDE_OIL but is mostly explained by its own lagged values. Because the ECT terms are negative for both variables, their short-term relationship does not function as it should.

In Table 6, the results of the Granger causality test are presented. In Model 1, the authors found that the p-value is lower than 0.05. Therefore, they reject the null hypothesis and conclude that both independent variables, ΔLNCRUDE_OIL and ΔLNEXCHANGE_RATE, Granger-cause ΔLNBSE. In Model 2, where the dependent variable is ΔLNCRUDE_OIL, the authors found that neither of the independent variables causes the dependent variable. In contrast, in Model 3, the authors found that ΔLNBSE does not Granger-cause ΔLNEXCHANGE_RATE. However, they found that ΔLNCRUDE_OIL Granger-causes ΔLNEXCHANGE_RATE, as they rejected the null hypothesis based on the p-value.

Hypothesis of the VEC Granger causality/block exogeneity Wald test:

The study further explores the variance decomposition function under the VECM model. Table 7 underscores the complexity and interdependence of the variables within the VECM model, illustrating how the forecast error variances evolve over different periods for LNBSE, LNCRUDE_OIL and LNEXCHANGE_RATE. Results highlight that none of the variables shows exogeneity in terms of the percentage of forecast error variance explained in both the short run and the long run.

For LNBSE, in the short run, LNEXCHANGE_RATE explains 53% of the variance in LNBSE by period two. However, this control decreases over time. Conversely, LNCRUDE_OIL has a minimal effect on LNBSE in both the short and long run.

When LNCRUDE_OIL is the dependent variable, the percentage of forecast error variance is mostly explained by LNCRUDE_OIL itself. However, the explanatory power of LNBSE and LNEXCHANGE_RATE increases over time. For LNEXCHANGE_RATE, a greater percentage of forecast error variance is explained by LNCRUDE_OIL, except in the first period. In the long run, the forecast error variance for LNEXCHANGE_RATE is more significantly explained by LNBSE rather than LNEXCHANGE_RATE itself.

Figure 1 presents the IRFs derived from the VECM. These functions illustrate how a shock to one variable impacts other variables in the system over time. Each panel represents the response of one variable to a shock in another variable (or to itself).

The responses of each variable to its own shocks exhibit typical decay patterns, starting with strong initial reactions that gradually diminish over time. The BSE Sensex generally shows a positive initial response to shocks in crude oil prices and exchange rates, followed by a decline. Crude oil prices tend to fall after shocks in the BSE Sensex and exchange rates. The exchange rate initially responds positively to shocks in both the BSE Sensex and crude oil prices but then declines. These patterns highlight the dynamic relationships between these variables in the Indian financial context, demonstrating that shocks to one variable can have significant but varying impacts on the other variables over time.

All the R-squared values are less than the Durbin–Watson statistics, which indicates that the regression is not spurious (Bhaumik, 2015). Another test conducted by the authors is the inverse roots of the AR characteristics polynomial, where they found that all the eigenvalues are within the unit circle, indicating that the VECM model is stable.

In the diagnostics test, the authors conducted several assessments to ensure the model’s appropriateness. One of the initial tests performed was the Wald coefficient test, which evaluates the significance of individual coefficients and helps identify which independent variables are crucial in explaining the variation observed in the dependent variable. With the test statistic value being less than 0.05, the authors rejected the null hypothesis, concluding that not all coefficients are equal to zero. This indicates that the model is statistically significant and valid.

Following the Wald test, the authors proceeded with the autocorrelation LM test. This test aims to determine whether there is any autocorrelation or serial correlation present in the model’s residuals. Autocorrelation in the residuals suggests that the model may be incomplete or not adequately capturing the temporal dependencies in the data. Because the p-value associated with the autocorrelation LM test is greater than 0.05, it indicates no significant autocorrelation in the residuals. This suggests that the model adequately captures the temporal dependencies in the data, enhancing its reliability.

Proceeding to the third test, the researchers conducted a normality test using the Cholesky decomposition of covariance (Lütkepohl) orthogonalization method. If the residuals follow a normal distribution, it indicates that the model is suitable for the data and the estimates derived from it are reliable. Conversely, significant deviations from normality may suggest potential issues with the model’s validity, necessitating further investigation or adjustments. Because the p-value associated with the residuals, considered jointly, is greater than 0.05, the authors fail to reject the null hypothesis, implying that all the residuals are normally distributed.

In this study, researchers investigate the relationship between the BSE Sensex, crude oil prices and exchange rates. Cointegration tests revealed a long-run relationship among the variables, highlighting India’s dependency on crude oil imports, which account for around 80% of its energy needs (Dalei and Gupta, 2020).The results align with studies by Boyer and Filion (2007), Nath Sahu et al. (2014) and Sahu et al. (2015). However, Sharma et al. (2018) did not find any long-run dynamics between these variables.

The VECM revealed that in the short run, crude oil prices and exchange rates are primarily influenced by their own past values, while the BSE Sensex is influenced by both its own past values and changes in crude oil prices and exchange rates. The short-term independence of crude oil prices and exchange rates from the stock market could be because of their reliance on global factors such as global industrial production or global monetary policies (Avdjiev et al., 2019; Dash, 2012; Ha et al., 2020; Maurya, 2017; Ratti and Vespignani, 2016). Conversely, the BSE Sensex reacts to crude oil price and exchange rate changes as investors price in their impact on corporate earnings, inflation and macroeconomic stability (Keswani et al., 2024; Kumar et al., 2021; Shang and Hamori, 2021; Singhal and Ghosh, 2016). These findings are consistent with findings by Aggarwal and Manish (2020) and Shirodkar (2017). However, these results differ from studies by Sahu et al. (2015) and Sariannidis et al. (2010), which found no impact of crude oil or exchange rates on stock market indices.

These results are further validated by the Granger causality test. Crude oil prices act as a leading indicator for exchange rates because rising oil prices increase import costs, leading to currency depreciation (Soundarapandiyan and Ganesh, 2017). Similarly, the influence of crude oil prices and exchange rates on the BSE Sensex reflects the cascading effects of inflationary pressures and investor sentiment on the stock market (Dua and Goel, 2021).

The VDT reveals that in the short run, exchange rates significantly influence the BSE Sensex, though this influence wanes over time. Conversely, crude oil prices exhibit a minimal impact on the BSE Sensex across both short- and long-term horizons. When considering crude oil prices as the dependent variable, it is primarily self-explanatory in the short term, but the influence of both the BSE Sensex and exchange rates amplifies over time. Interestingly, for exchange rates, crude oil prices predominantly explain the forecast error variance, except in the initial period. In the long run, the BSE Sensex emerges as a more substantial explanatory variable for exchange rate variances compared to the exchange rates themselves. Additionally, the IRFs illustrate dynamic responses where the BSE Sensex initially reacts positively to shocks from both crude oil prices and exchange rates before declining, while crude oil prices and exchange rates show distinctive decay patterns following shocks from the other variables.

The short-term dominance of exchange rates reflects the sensitivity of foreign portfolio investments and trade to currency movements (Jacob et al., 2022; Makoni, 2020). The initial positive response to shocks in crude oil prices may reflect optimism around lower production costs in oil-importing sectors, but sustained increases in oil prices eventually dampen investor sentiment because of inflationary pressures (Rafiq et al., 2016; Varghese, 2017).

In this study, the authors extensively present evidence on the relationships among crude oil prices, exchange rates and Sensex prices in India. They found a long-term relationship exists between these variables. Specifically, they observe that crude oil prices and exchange rates significantly affect the BSE Sensex. It was further confirmed that none of the variables show endogeneity in terms of the impact of shocks from other variables on their own values. Further analysis reveals that the forecast error variance of the BSE Sensex is well explained by crude oil prices and exchange rates. Similarly, the exchange rate’s forecast error variance is influenced by two other variables, while crude oil prices are affected primarily by BSE Sensex prices. The authors also found, from the IRF, that shocks in one variable have a persistent and prolonged effect on the other variables.

These findings highlight the significant impact of global economic factors on India’s stock market. As a major crude oil importer, India is highly vulnerable to fluctuations in global oil prices, which directly affect its trade balance, inflation and economic stability. Similarly, exchange rate movements influence import costs, corporate earnings and investor sentiment. This intricate relationship underscores the importance of strategic monitoring and intervention to minimize the adverse effects of global economic volatility on India’s financial markets.

The implications of these results are crucial for policymakers, investors and financial analysts. Policymakers can use these insights to address external vulnerabilities. For example, they can implement strategies to stabilize the economy from crude oil price shocks or adopt fiscal and monetary policies to stabilize exchange rates. These proactive measures would strengthen economic resilience and protect financial markets from external shocks. For investors, the study provides valuable guidance for anticipating market trends and optimizing portfolio strategies. By hedging against oil price fluctuations or leveraging exchange rate dynamics, investors can enhance their decision-making and returns. Financial analysts can also benefit by using these findings to refine market forecasts and risk assessments, helping stakeholders navigate uncertainties with greater precision.

Additionally, the study emphasizes the importance of diversifying energy sources and adopting policies to build resilience against global economic shocks. By tackling these challenges, policymakers, investors and analysts can contribute to a more stable and sustainable economic environment in India.

While the study offers significant insights, it is not without limitations. First, the focus on India limits the generalizability of the findings to other economies with different macroeconomic and financial conditions. Therefore, incorporating them into future research could provide new perspectives. Furthermore, the relationship between the variables may not apply to every sector, necessitating further sector-specific analysis.

Funding acknowledgements: No funding received during this work.

Author contribution: All authors contributed equally.

Conflict of interest: The authors have no conflict to disclose.