Asymmetric Effects of Exchange Rate and Its Relationship with Foreign Investments: A Case of Indian Stock Market

Abstract

Emerging markets, including India, are witnessing an influx of foreign capital. The article investigates the role of exchange rate which influences both the net foreign institutional investments (FIIs) and the stock markets, using monthly data, from January 2008 to May 2018. The effects of real effective exchange rate are studied through non-linear ARDL co-integration. The long-run relationship is found in all the three models constructed. The results highlight the nature of FII flows in relation to exchange rate asymmetry. Real rupee depreciation has a long-run effect on their debt flows. The ‘adjustment asymmetry effect’ of exchange rate is found for equity flows in the long run. The similar effect is observed for the Nifty 50 model. Due to high volatility, even positive stock returns do not attract equity FII flows. In the short run, rupee depreciation in real terms negatively influences Nifty returns. The S&P 500 returns explain FII flows indicating information asymmetry. These outcomes serve a vital input for key stakeholders such as potential FIIs, domestic traders, regulators and policymakers.

Keywords

FIIs Real Effective Exchange Rate Stock Market Returns ARDL Asymmetry GFC

Introduction

In India, during financial liberalization, the event of lifting up of capital account controls had resulted in tremendous increase in foreign investment. On the one hand, India as a developing country required such investments for stimulating its growth prospects, improving labour productivity and helping to build foreign reserves. On the other hand, the foreign investors looked for the attractive investment destination which not only manifold their returns but also availed investment diversification benefits.

One such form of foreign capital is foreign institutional investment (FII). FIIs had impacted the emerging nations. India, on its way of development, attracted FIIs by following liberalization measures in 1991–1992 (Garg & Bodla, 2015). Especially after the Asian financial crisis, India was their most preferred destination (Samal, 1997). The boom in FIIs was experienced after the global financial crises (GFC) of 2010 and 2012 (Dhingra, Gandhi, & Bulsara, 2016). According to Bloomberg report (2018), India tops the list of FII flows among the emerging nations.

The international investment decision is not solely dependent on the growth prospects of a nation. It depends on several push and pull factors (Rai & Bhanumurthy, 2004). Furthermore, they not only invest in equity markets but also place their investments in debt markets. Moreover, FIIs prefer investing in debt market of emerging nations, with stable exchange rates, due to higher interest rates.

The impact of FIIs is not always positive. On one side, their presence ‘broaden the base’ of the existing domestic market which includes more investors, enhances liquidity, creates diversification advantages by lowering risk premium as suggested by Garg and Bodla (2015). But on the other side, FIIs invest for a short time period and thus they are called ‘hot money’. Their presence can cause volatility in the stock markets causing a ‘destabilization effect’. Since they may follow the practice of ‘positive feedback trading’ where news of positive stock returns induces a lot of foreign inflows and negative news may wipe out the money from the market within seconds creating volatility in the exchange rate of the host country. Furthermore, since India is a net importer, any exogenous shock could potentially affect its current account balances and put pressure on foreign exchanges, thereby exerting pressure on domestic currency. Thus, a risk of capital flight is always surmounted on the Indian economy, which further worsens the exchange rate position. Therefore, the presence of FIIs has been an area of concern for the policymakers, regulators, researchers and academicians. Exchange rate changes are the major determinant of FIIs flows into the country as found by Srinivasan and Kalaivani (2015). They are concerned about the value of investments in dollar terms. If rupee depreciates, the value of their investments in dollar terms reduces. The MNC’s balance sheet is also impacted by exchange rate movements as opined by Bahmani-Oskooee and Saha (2018).

In this backdrop, the present study highlights the effect of real effective exchange rate (REER) on net FIIs and Indian stock market returns through non-linear model. Specifically, it aims to examine whether the striking evidence of either real rupee appreciation or depreciation can be observed on the net FII measures considering the event of GFC. This net FII is bifurcated into debt and equity flows and on the Nifty 50 returns.

The rest of this article is structured as follows: the following section deals with the ‘Literature Review’ followed by the ‘Methodology’ that provides the details of the variables and techniques used. ‘Analysis’ provides comprehensive discussion of the results and finally this article provides ‘Conclusions and Implications’ along with suggested implications of the study.

Literature Review

There have been studies to reflect that foreign capital flows could substantially destabilize the stock market. One such eminent and early study has been on Korean stock market conducted by Choe, Kho, and Stulz (1998) for a period from 30 November 1996 to end of 1997 using monthly data which includes the period of crisis. The investors were bifurcated into foreign, institutional and individual investors. Positive feedback trading evidence was found for crisis period from the positive relation between the stock returns and equity flows for the aggregate market and for individual stocks before. There was no significant evidence to conclude that foreign participants destabilized the stock market. Another extension of this study was done by Yang (2017), who investigated whether the GFC could affect the investments of foreign investors creating a destabilization impact on Korean stock market. It was found that when impact of price is represented as a unit trade, permanent impact on prices exhibited a steep rise on foreign funds during the crisis period. But this could not prove that foreign investors can solely destabilize the stock market.

Some studies focused on determining major determinants of FII’s and stock prices. Chakrabarti (2001) observed that a high correlation was found between Indian stock market returns and FII flows. The informational asymmetry was absent between domestic and foreign participants as USA and world returns failed to explain FII flows. The nature of FII flows shifted after the Asian Crisis which showed that BSE returns were solely driving the FII flows. Furthermore, Mukherjee, Bose, and Coondoo (2002) observed that the FII flows both inward and outward was caused by domestic market returns. Returns from exchange rate variation and fundamentals influence FIIs decisions. Srinivasan and Kalaivani (2015) through ARDL technique have found exchange rate to be the driving force for FII inflows. Waqas, Hashmi, and Nazir (2015) studied the exogenous macroeconomic variables and their impact on volatility of foreign portfolio investment in Pakistan, China, Sri Lanka and India from 2000 to 2012. It was observed that the low volatility in portfolio investment is related with currency depreciation, foreign direct investment (FDI), high interest rate, high GDP growth rate and low inflation in host nations. Kim and Yi (2015) examined whether the domestic and foreign investors trading could improve the information efficiency reflected in stock prices by improving the specific information available for each firm listed on Korean Stock Exchange for the period from 1998 to 2007. In comparison with domestic trading, foreign trading induces efficiency in stock market as it improves the flow of information that gets reflected in firm-specific factors.

Another set of literature focuses on volatility dynamics resulting from FIIs flows. Rai and Bhanumurthy (2004) by using monthly data from January 1994 to November 2002 examined whether risk and return in Indian stock market affects the FIIs flows. Using TGARCH model, it was found that BSE returns and US inflation exerted a positive influence, while negative relation was found with ex-ante risk of BSE and S&P and India’s inflation level. Garg and Bodla (2015) have found that daily returns on BSE were negatively impacted by FIIs introduction in the stock market. Volatility has significantly reduced in the stock market returns after FIIs introduction creating a stabilizing effect. The current set of study focuses on the GFC’s impact on FIIs and stock market. Yaha, Singh, and Rabanal (2017) using daily observations from 1999 to 2011 examined any significant evidence of abnormal responses in foreign equity flows and in stock market returns in the eve of global shocks. No evidence of abnormal responses could be traced in either of the two due to global shocks. Goh, Zam, and Sapian (2017) examined the relationship of market volatility, returns and equity fund flows of both retail and foreign investors over a period from October 2009 to December 2015 using daily data. The results conveyed that a negative association exists between the market returns and net equity flows. Volatility is negatively related to net foreign flows.

A number of studies were conducted to study the effect of exchange rate and foreign flows. Caporale, Menla Ali, and Spagnolo (2015) studied the influence of uncertainty associated with exchange rate on net bond and net equity flows of USA with respect to seven countries for a period from monthly data over a period from 1998 to 2011. By using VAR-GARCH-BEKK-in-mean model, it was concluded that uncertainty of exchange rate had a negative impact on net equity flows in Euro Area, Sweden and UK, while a positive influence was found in Australia. No impact was observed for Canada and Japan. However, the uncertainty of exchange rate adversely impacted the bond flows for every set of nations except Canada. Spillover from net equity flows to fluctuations in exchange rate was observed in Sweden and UK. Caporale, Menla Ali, Spagnolo, and Spagnolo (2017) conducted similar study with seven economies. Through Markov Switching GARCH model, it was found that equity flows into USA when each of Asian economies faces a period of high exchange rate volatility regime except Philippines. Singhania and Saini (2016) have found that negative news about the host country significantly influences the FII inflows by using the GARCH model. The stability of exchange rate and stock returns were pointed to lure FIIs. Vardhan and Sinha (2014) found that exchange rate caused the outflow of foreign capital from India. While Rashid and Fazal (2010) observed that non-linear causal relation exists between FII flows and exchange rate. Chakraborty and Kakani (2016) by employing Markov-Switching GO-GARCH model in four emerging markets confirmed the asymmetric behaviour of foreign institutional investors as compared with domestic investors in respect of market volatility characterized by inflow of bad news and high volatility regime. This behaviour confirms the positive feedback trading by the FIIs. It was concluded that bad news led higher volatility followed by lower trade volume resulting in its wider dispersion creating a greater impact on FIIs than on domestic investors. Liu, Bredin, Wang, and Yi (2014) observed that the foreign funds prefer to invest in metal and non-metals, machinery and transportation, creating diversified portfolio while considering the corporate governance indicators. On the other hand, the domestic investors prefer to distribute their investments more evenly.

The study done by Sui and Sun (2016) observed that trade flow model better explains the exchange rate transmission to BRICS stock market. Recent study done by Chang and Rajput (2018) by which asymmetric effects of different macroeconomic variables such as consumer price inflation, interest rate, industrial production and REER were studied on stock prices. Bahmani-Oskooee and Kanitpong (2017) found that the presence of asymmetry in real exchange rates on the trade balance, that is, the effect of appreciation or depreciation of the local currencies of nations have a different impact on their imports and exports. They conducted their study on seven Asian economies. Non-linear ARDL model displayed short-run asymmetric impact on Thailand, Korea, Singapore and Malaysia, while long-run asymmetric responses were supported by Japan, Korea and Indonesia. Similarly, Bahmani-Oskooee & Fariditavana (2015) while using quarterly data from 1973I to 2014II for Canada, Japan, USA and China confirmed that the effect of exchange rate on trade balance for each of the nation is non-linear while only Canada and USA have displayed linear relationship. Walid, Aloui, and Nguyen (2012) have found a bidirectional asymmetric relation between stock market and exchange rate of UK, Germany and France. Exchange rate volatility was a major determinant in formulation of portfolios by the investors. Bahmani-Oskooee and Saha (2018) have included 24 nations and examined such relationship using NARDL approach. The study revealed that changes in exchange rate have an asymmetric impact on stock prices of Canada and Malaysia. But the opposite was true in case of Germany, Brazil, Singapore and China.

Literature is available on determining the determinants of FIIs flows into India and the effect of volatility in stock prices on FIIs flow in India and vice versa through GARCH models. From the available literature, it was identified that the effect of exchange rate on FIIs flow and on stock market return was one of the most important macroeconomic determinants and only symmetric effects were studied. Bahmani-Oskooee and Saha (2015, 2016) have found that previous studies have pointed that exchange rate changes exhibit a symmetric effect on the sample countries’ stock prices. This has proved to be a major limitation of those studies as differences in expectations and beliefs of traders and various participants persist in the stock market (Dahir, Mahat, Ab Razak, & Ariffin, 2018; Lakshmi & Thenmozhi, 2018; Shin, Yu, & Greenwood-Nimmo, 2014). Due to these differences in beliefs, the effect of depreciation or appreciation might not adjust contemporaneously and produces lagged effect. Thus, the reviews suggested that there is ample scope of research in finding the asymmetric effect of exchange rate which has pro-effect on the FIIs and stock market while considering the GFC.

Methodology

The present study incorporates the net foreign institutional investments (FIIs) which were bifurcated into debt and equity investments. Another important macroeconomic variable is exchange rate. Exchange rate is one of the major determinants of FII flows and stock prices as found by Basin and Khandelwal (2014) and Srinivasan and Kalaivani (2015). Consequently, it has been considered as an endogenous variable. Furthermore, we have used REER in our study. It is the index that represents the value of rupee against the weighted average currencies of six major trading partners after adjusting for the inflation level. It reflects the competitiveness of rupee.

Table 1.

Variables Used

Variables	Source	Symbols
Monthly real effective exchange rate (base year, 2010 = 100)	fred.stlouisfed.org.	REER
Monthly Nifty 50 stock returns	Yahoofinance.com	NR
Monthly volatility of Nifty 50 returns	Calculated	NV
Monthly S&P 500 stock returns	Yahoofinance.com	SR
Monthly volatility of S&P 500 returns	Calculated	SV
Dependent variable as monthly net foreign equity investments	www.cdslindia.com	EF
Dependent variable as monthly net foreign equity investments	www.cdslindia.com	DF

Source: Authors’ compilation.

Monthly data from January 2008 to May 2018 have been taken. The period also includes the time of GFC. The effect of such crisis was exchange rate volatility (Prakash, 2012) due to significant influence of volatile capital flows into the economy.

The variables used along with their sources are depicted in Table 1. For estimating volatility of both Nifty 50 and S&P 500 returns, realized volatility measure was adopted (Chandra & Thenmozhi, 2015). The volatility is calculated by aggregating the squared returns over a definite period of time, that is, the number of trading days in a month (here, it is assumed to be 30 days):

V O L_{t} = \sum_{t = 0}^{n} r_{i, t}^{2} .

(1)

Non-linearity can be caused due to financial or economic crisis, political instability, loss of public faith in government, structural changes observed in business cycles and poor economic fundamentals. Thus, it is important to consider non-linear effects. Thus a non-linear ARDL model is adopted to study the response of net FIIs in India and stock market returns to positive and negative REER separately.

NARDL model (Shin et al., 2014), an extension of ARDL, estimates the asymmetric effects of one regressor in the estimation of the model. It does away with the assumption of conventional co-integration models of existence of linear stationary combination of underlying non-stationary variables. This model can be applied before a pretest of order of integration through unit root test. We have used augmented Dickey–Fuller parametric test and Phillip Perron non-parametric test to check for the stationarity of the variables. Both the models have a null of unit root. The NARDL model cannot be utilized if any of the variables becomes 1 (Equation 2). Furthermore, this model can be well suited for assessing co-integration using small samples.

In our study, all the variables were either integrated in I(0) or I(1) order and we have total sample of 125 periods, and this methodology was aptly used. The model specifications are expressed as follows:

\begin{array}{l} \begin{matrix} Δ E F = C_{01} + \sum_{p = 1}^{l} δ_{p 1} Δ E F_{t - p} + \sum_{q = 1}^{l} γ_{j 1} Δ R E E R_{t - q} + \\ \sum_{r = 1}^{l} φ_{k 1} Δ N R_{t - r} + \sum_{s = 1}^{l} π_{s 1} Δ N V_{t - s} + \end{matrix} \\ \begin{matrix} \sum_{u = 1}^{l} ω_{t 1} Δ S R_{t - u} + \sum_{v = 1}^{l} α_{v 1} Δ S V_{t - v} \\ + θ_{11} E F_{t - 1} + θ_{21} R E E R_{t - 1} + \\ θ_{31} N R_{t - 1} + θ_{41} N V_{t - 1} + \\ θ_{51} S R_{t - 1} + θ_{61} S V_{t - 1} + ε_{1 t}, \end{matrix} \end{array}

(2)

\begin{array}{l} \begin{matrix} Δ D F = C_{02} + \sum_{p = 1}^{m} δ_{p 1} Δ D F_{t - p} + \sum_{q = 1}^{m} γ_{j 1} Δ R E E R_{t - q} + \\ \sum_{r = 1}^{m} φ_{k 1} Δ N R_{t - r} + \sum_{s = 1}^{m} π_{s 1} Δ N V_{t - s} + \end{matrix} \\ \begin{matrix} \sum_{u = 1}^{m} ω_{t 1} Δ S R_{t - u} + \sum_{v = 1}^{m} α_{v 1} Δ S V_{t - v} + \\ θ_{12} D F_{t - 1} + θ_{22} R E E R_{t - 1} + \\ θ_{32} N R_{t - 1} + θ_{42} N V_{t - 1} + \\ θ_{52} S R_{t - 1} + θ_{63} S V_{t - 1} + ε_{2 t}, \end{matrix} \end{array}

(3)

\begin{array}{l} \begin{matrix} Δ N R = C_{03} + \sum_{p = 1}^{n} δ_{p 1} Δ N R_{t - p} + \sum_{q = 1}^{n} γ_{j 1} Δ R E E R_{t - q} + \\ \sum_{r = 1}^{n} φ_{k 1} Δ E F_{t - r} + \sum_{s = 1}^{n} π_{s 1} Δ N V_{t - s} + \\ \sum_{u = 1}^{n} ω_{t 1} Δ S R_{t - u} + \sum_{v = 1}^{n} α_{v 1} Δ S V_{t - v} + \end{matrix} \\ \begin{matrix} θ_{13} N R_{t - 1} + θ_{23} R E E R_{t - 1} + θ_{33} E F_{t - 1} + \\ θ_{43} N V_{t - 1} + θ_{53} S R_{t - 1} + θ_{63} S V_{t - 1} + ε_{3 t}, \end{matrix} \end{array}

(4)

\begin{array}{l} \begin{matrix} Δ N R = C_{04} + \sum_{p = 1}^{0} δ_{p 1} Δ N R_{t - p} + \sum_{q = 1}^{0} γ_{j 1} Δ R E E R_{t - q} + \\ \sum_{r = 1}^{0} φ_{k 1} Δ D F_{t - r} + \sum_{s = 1}^{0} π_{s 1} Δ N V_{t - s} + \\ \sum_{u = 1}^{0} ω_{t 1} Δ S R_{t - u} + \sum_{v = 1}^{0} α_{v 1} Δ S V_{t - v} + \end{matrix} \\ \begin{matrix} θ_{14} N R_{t - 1} + θ_{24} R E E R_{t - 1} + θ_{34} D F_{t - 1} + \\ θ_{44} N V_{t - 1} + θ_{54} S R_{t - 1} + θ_{64} S V_{t - 1} + ε_{4 t}, \end{matrix} \end{array}

(5)

where l, m, n, o are optimum lags, C_0i are the constants, ε_t are the white noise error terms, θ_i are long-run coefficients and δ, γ, ϕ, π, ω, α are short-run coefficients.

These equations were modified to take into account the asymmetric effect, that is, depreciation and appreciation of REER. As proposed by Shin et al. (2014), one exogenous variable can be fragmented to form into two variables represented as sum of partial positive and negative changes:

R E E R_{i}^{p o s} = \sum_{i = 1}^{t} Δ R E E R_{i}^{p o s} = \sum_{i = 1}^{t} \max (Δ R E E R_{i}, 0)

(6)

and,

R E E R_{i}^{n e g} = \sum_{i = 1}^{t} Δ R E E R_{i}^{n e g} = \sum_{i = 1}^{t} \min (Δ R E E R_{i}, 0) .

(7)

Due to such representation, asymmetric outcome can be assessed in long and short run:

\begin{array}{l} \begin{matrix} Δ E F = C_{01} + \sum_{p = 1}^{l} δ_{p 1} Δ E F_{t - p} + \\ \sum_{q = 1}^{l} (γ_{q 1}^{p o s} Δ R E E R_P_{t - q} + γ_{q 1}^{n e g} Δ R E E R_N_{t - q}) \\ \sum_{r = 1}^{l} φ_{k 1} Δ N R_{t - r} + \sum_{s = 1}^{l} π_{s 1} Δ N V_{t - s} + \\ \sum_{u = 1}^{l} ω_{t 1} Δ S R_{t - u} + \sum_{v = 1}^{l} α_{v 1} Δ S V_{t - v} + \end{matrix} \\ \begin{matrix} θ_{11} E F_{t - 1} + θ_{21}^{p o s} R E E R_P_{t - 1} + \\ θ_{21}^{n e g} R E E R_N_{t - 1} + θ_{31} N R_{t - 1} + \\ θ_{41} N V_{t - 1} + θ_{51} S R_{t - 1} + θ_{61} S V_{t - 1} + ε_{1 t} . \end{matrix} \end{array}

(8)

Similarly other equations can also be modified. $γ_{q 1}^{p o s} a n d γ_{q 1}^{n e g}$ are short-run asymmetric coefficients while $θ_{21}^{p o s} a n d θ_{21}^{n e g}$ are long-run asymmetric coefficients. These positive and negative coefficients should be assumed to be one while assessing the co-integration results.

First and foremost, the existence of co-integrating relationship is assessed using OLS estimation. The null hypothesis of $θ_{11} = θ_{21}^{p o s} = θ_{21}^{n e g} = θ_{31} = θ_{41} = θ_{51} = θ_{61} = 0$ is tested for each of the equations. The calculated F-statistics is compared with the bounds as given by Pesaran, Shin, and Smith (2001). One would reject the null hypothesis of no co-integration when the F-statistics is more than the upper bound critical value I(1). If the F-statistics value falls below the lower bound I(0), then null hypothesis cannot be rejected. Decision becomes inconclusive when F-statistics is between the values of both bounds. After co-integration is established, asymmetric effects in long run is obtained through Wald test which follows F-distribution. There has to be significant difference in the long-run asymmetric coefficients, which is tested through the following equation:

\frac{- θ_{21}^{p o s}}{θ_{11}} = \frac{- θ_{21}^{n e g}}{θ_{11}}

(9)

One has to reject this equation to establish asymmetric effect. Similarly in the short run, statistical difference has to be found in $γ_{q 1}^{p o s} a n d γ_{q 1}^{n e g}$ coefficients.

We have also employed structural break test given by Bai and Perron (1998). It identifies multiple significant break points in the data which can potentially distort the findings and also cause problem of spurious regression through minimizing residual sum of squares in OLS estimation. We allowed maximum five breaks through sequential L breaks method and 5 per cent significance level. The obtained breaks were incorporated in the regression model as dummy variables. Each dummy variable for each break was given the value 1 on and after the break date and 0 before the break date (Pesaran et al., 2001), confirmed that use of dummy variables did not affect the conclusions from co-integration.

Analysis

Descriptive Statistics

Table 2 depicts the characteristics of each of the monthly variables for the entire period of the study.

In comparing the foreign equity flows and foreign debt flows, we found that the average equity flows is more than the debt flows. The volatility measured by the standard deviation is more in equity flows which is normally expected. The third and fourth moments, that is, the skewness and kurtosis depict departure from normality for debt flows. But for equity flows, the skewness was close to 0 and kurtosis value was close to 3. For further validating the presence of normality, we used Jarque–Bera test that confirms the null hypothesis of normality in equity flows as p-value was more than 10 per cent level of significance. But the null gets rejected for debt flows.

Furthermore, while comparing the return series of exchange rate, Nifty 50, Nifty 50 volatility, S&P 500 and its volatility, we found that the REER gives a negative mean return. This reflects that on an average the rupee has appreciated with respect to US dollars in real terms. Both the Indian and US stock market return series were positive and so do their respected volatilities reinforcing the risk-return notion.

Table 2.

Descriptive Statistics

	EF	DF	REER	NR	NV	SR	SV
Mean	4,700.538	2,949.286	–0.000033	0.03159	0.01166	0.00003	0.00008
Median	4,902.7	2,115.4	0.000975	0.02971	0.00453	0.00002	0.00007
Std. deviation	10,450.03	9,404.5	0.017132	0.02034	0.02941	0.00002	0.00001
Skewness	0.15296	–0.21247	–0.469923	0.24326	2.96662	2.10819	1.41993
Kurtosis	2.54102	4.53695	3.388261	5.52326	13.18939	6.45793	3.84012
Jarque–Bera	1.58461	13.24368	5.385716	34.3933	724.099	154.8704	45.6803
Probability	0.45280	0.00133	0.067687	0.00000	0.00000	0.00000	0.00000
Observations	125	125	125	125	125	125	125

Source: Authors’ calculations.

Indian stock market exhibits high risk and high return as true for most emerging markets. It was observed that none of the return series were normally distributed except the REER which could not reject the null of normality at 1 and 5 per cent level of significance.

Correlation Matrix

Table 3 depicts the pairwise unconditional correlation of all the variables under study. Here the REER was bifurcated into positive and negative rates to get a clear idea of the nature of association among the variables. A significant correlation between equity and debt flows was observed. Both were positively associated with each other. Both increase or decrease together. Both the Nifty 50 returns and its volatility were significantly negatively correlated to negative REER return, that is, rupee depreciation. Whenever the rupee depreciates in real terms, the Nifty 50 returns became negative at 5 and 10 per cent level of significance and volatility reduces. Due to rupee depreciation, the cost of imports increases causing profits to reduce. Similar positive correlation was observed between positive REER and volatility in S&P 500 return series at 10 per cent significance level. If the US stock returns are volatile, the rupee appreciates contemporaneously. However, the negative REER was not related to either Nifty 50 or S&P 500 return series.

But it was significantly positively related to both Nifty 50 volatility and S&P 500 volatility. The Nifty 50 returns and its volatility are negatively related. The returns of Nifty 50 are negatively associated with S&P 500 returns. When returns of S&P 500 are low, investors buy Indian stocks rising the prices of Nifty 50 and thereby its returns. No significant correlation was observed between market returns and net equity flows unlike Goh et al. (2017) and Chakrabarti (2001).

However, market returns were related to net debt flows. But like Goh et al. (2017), volatility was negatively related to net equity foreign flows.

Unit Root Tests

It is necessary to check the stationarity of the time series in order to avoid possibility of spurious regression. Table 4 reflects the results of both augmented Dickey–Fuller and Phillip–Perron tests. The null of both tests is that variables have a unit root.

Both the test results were in convergence. All the variables at level rejected the null of a unit root at 5 per cent level of significance except the positive and negative real effective exchange rate. However, at first difference the null of unit root was rejected for both the positive and negative real effective exchange rate. Thus, it was concluded that all the variables are either integrated in I(0) or I(1) order at 1 per cent level of significance.

Structural Break Test

It may happen that due to the presence of structural break in the data, our co-integration results become unreliable if not accounted for. Its incorporation in the study gives valuable insights and makes the study more accurate. Thus before employing the NARDL co-integration technique, we check for the presence of structural break in the data set via running the original NARDL model for each of the dependent variable through structural break test (Bai & Perron, 1998).

The test results are depicted in Table 5.

The findings reflect that with equity flows as the dependent variable, no break date was ascertained. This signalled that foreign equity flows did not affect the NARDL equation. But with debt flows, three break dates effected the original equation (2). The probable reasons could be:

Table 3.

Unconditional Correlation Matrix

	EF	DF	REER_N	REER_P	NR	NV	SR	SV
EF	1
	–
DF	0.3316*	1
	(0.0002)	–
REER_P	0.00984	0.10582	1
	(0.913)	(0.2402)	–
REER_N	–0.02330	–0.06983	–0.9771*	1
	(0.796)	(0.439)	(0.00)	–
NR	–0.02132	–0.181**	–0.187**	0.12516	1
	(0.813)	(0.043)	(0.036)	(0.164)	–
NV	–0.3614*	–0.11458	–0.4548*	0.465*	–0.296*	1
	(0.00)	(0.203)	(0.00)	(0.00)	(0.0008)	–
SR	–0.1016	–0.05952	–0.0679	0.05835	–0.17***	0.0158	1
	(0.259)	(0.5096)	(0.452)	(0.5181)	(0.057)	(0.86)	–
SV	–0.11888	–0.06832	–0.156	0.17***	0.0212	0.0825	–0.083	1
	(0.187)	(0.449)	(0.1012)	(0.057)	(0.8144)	(0.36)	(0.353)	–

Source: Authors’ calculations.

Notes: Parentheses reflect p-values of correlation coefficients. Significance at 1%, 5% and 10% levels are reflected as *, ** and ***, respectively.

Table 4.

Unit Root Tests

Variables	At Level				At First Difference
	ADF		PP		ADF		PP		Order of Integration
	T-statistics	P-value	T-statistics	P-value	T-statistics	P-value	T-statistics	P-value	Order of Integration
EF	–7.6354	0.00*	–7.61909	0.00*					I(0)
DF	–5.749	0.00*	–5.74544	0.00*					I(0)
REER_P	–1.1206	0.7062	–1.22562	0.6619	–10.9351	0.00*	–10.9659	0.00*	I(1)
REER_N	–2.0999	0.2452	–2.02179	0.2773	–10.3248	0.00*	–10.356	0.00*	I(1)
NR	–3.981	0.002*	–6.87076	0.00*					I(0)
NV	–5.5227	0.00*	–5.33811	0.00*					I(0)
SR	–12.78	0.00*	–12.7935	0.00*					I(0)
SV	–14.193	0.00*	–14.1319	0.00*					I(0)

Source: Authors’ calculations.

Note: * Signifies the significance at 1% level.

Table 5.

Bai–Perron Multiple Breakpoint Test

Dependent Variable	Break Test	Scaled F-statistics	Critical Value	Break Dates
EF	0 v/s 1	17.25877	27.03	0
DF	0 v/s 1	31.94223	27.03	2016; 10 DM1
	1 v/s 2	41.90693	29.24	2015; 03 DM3
	2 v/s 3	35.57285	30.45	2013; 06 DM2
	3 v/s 4	19.29288	31.45
NR (equity)	0 v/s 1	70.56786	27.03	2009; 11 DM4
	1 v/s 2	18.34912	29.24
NR (debt)	0 v/s 1	76.30789	27.03	2009; 12 DM5
	1 v/s 2	15.68422	29.24

Source: Authors’ calculations.

In June 2013, FII flows witnessed a bearish phase particularly in the debt flows reaching its all-time low. Around 33,000 crores INR were taken out of the country. Since FIIs were jittery about credibility of Indian economy, it witnessed steep increase in exchange rate volatility in June 2013 reaching 14.75 per cent from 4.47 per cent in May 2013 particularly due to quantitative easing programme adopted by USA (Prakash, 2012).

In March 2015, the period pertains to the commodity slump and net exporting countries experiencing slowdown. The downfall in prices of commodities was sharp during this period. However, India benefited through improvement in current account deficit position, growth of forex reserves, strengthening credibility of economy which brought influx of FII investment into the country.

In October 2016, lack of government consensus in various internal issues such as land acquisition, bankruptcy laws, unexpected election results along with US interest rate hike expectations by policymakers. All such issues resulted in outflow of foreign debt funds.

The Nifty 50 returns impacted the equation including equity flows with one break date. The period 2008–2009 was a period of GFC. November 2009 was marked as a period of recovery for the Indian economy with significant positive returns. The rupee–dollar exchange rate became stable and rupee appreciated during 2009–2010. A lagged effect could be observed for the equation including debt flows, where the break date is December 2009.

Non-linear ARDL Bounds Test with Structural Breaks

Before studying the asymmetric impact of the exchange rate on FIIs and on Nifty 50, it is necessary to check whether any co-integrating relationship exists between the variables. Table 6 shows the bounds tests estimation for assessing co-integration among the three set of dependent variables. The unrestricted non-linear ARDL model has been chosen for estimation.

Table 6.

Panel A: NARDL Bounds Testing Results

NARDL Equation	F-statistics	K	Selected Model	Conclusion
F(EF\|REER_P, REER_N, NR, NV, SR, SV)	18.22418	5	ARDL(1, 1, 2, 0, 0, 0, 0)	Co-integration
F(DF\|REER_P, REER_N, NR, NV, SR, SV, DM1, DM2, DM3)	3.274013	8	ARDL(1, 1, 0, 1, 1, 4, 0, 1, 1, 0)	Co-integration
F(NR\|REER_P, REER_N, EF, NV, SR, SV, DM4)	6.533816	6	ARDL(2, 0, 1, 0, 4, 1, 0, 3)	Co-integration
F(NR\|REER_P, REER_N, DF, NV, SR, SV, DM5)	5.616489	6	ARDL(2, 0, 1, 0, 4, 1, 0, 4)	Co-integration

Source: Authors’ calculations.

Table 6.

Panel B: Critical Values with Unrestricted Constant and No Trend

	K = 5		K = 8		K = 6
	I(0)	I(1)	I(0)	I(1)	I(0)	I(1)
AT 1%	3.41	4.68	2.79	4.10	3.15	4.43
AT 5%	2.62	3.79	2.22	3.39	2.45	3.61
AT 10%	2.26	3.35	1.95	3.06	2.12	3.23

Source: Authors’ compilation.

Optimal lags have been chosen through Akaike Information Criteria and maximum lags were taken as 4 due to limited data points. The modified equation (7) was run.

The results depict that for all the NARDL equations, the calculated F-statistics was more than the critical upper bound, that is, I(1), implying the existence of long-run equilibrium relationship among each NARDL equation at 1 per cent significance level except for the debt flows model. For example, when the equity flows was dependent variable, F-statistics was 18.22418 which was more than the upper bound critical value of 4.43 at 1 per cent significance level. But for the debt flows model, F-value of 3.274 is greater than only the upper bounds at 10 per cent significance level.

Since long-run co-integration was established, it was possible to find out long- and short-run dynamics of each of variables with dependent variables, and thus to find out the exchange rate asymmetric effect on each of three dependent variables.

Table 7 presents the long-run effects and Table 8 presents the short-run effects. Table 7 also includes the error correction terms (ECT) which should be negative and less than 1. ECT denotes the speed of adjustment towards equilibrium level. It also presents several diagnostic tests for each of the NARDL model. LM test checks the null of serial correlation on residuals. R² checks whether the model fits the data. For stability of the model, CUSUM and CUSUM square tests were used.

Table 7.

Long-run NARDL Coefficient

IDV\ DV	EF	DF	NR (Equity)	NR (Debt)
REER_P	–30,923.6 (0.1583)	79,361.08	–0.048033	–0.015675
	–30,923.6 (0.1583)	(0.0667)**	(0.197)	(0.112)
REER_N	–485,998 (0.0238)**	98,058.33	–0.080202	–0.036689
	–485,998 (0.0238)**	(0.0257)**	(0.0728)***	(0.1365)
NR	–189,590 (0.0022)*	21,947.65	–	–
	–189,590 (0.0022)*	(0.8475)
NV	–246,465.8 (0.00)*	57,167.17	–1.135244	–0.105996
	–246,465.8 (0.00)*	(0.5052)	(0.00)*	(0.472)
EF	–	–	–0.00000076	–
	–		(0.0031)*
DF	–	–	–	–6.01E-07
	–			(0.0278)*
SR	–115,000,000 (0.0344)**	91,427,277	–83.00575	–20.00406
	–115,000,000 (0.0344)**	(0.6621)	(0.6243)	(0.912)
SV	–205,000,000 (0.0546)**	–228,000,000	-43.8267	25.87445
	–205,000,000 (0.0546)**	(0.166)	(0.8379)	(0.9089)
DM1	–	3,993.44	–	–
	–	(0.3113)
DM2	–	6,758.233	–	–
	–	(0.1947)
DM3	–	–9,159.941	–	–
	–	(0.0223)**
DM4	–	–	–0.014186	–
	–		(0.0098)*
DM5	–	–	–	–0.033925
	–			(0.0036)*
ECT	–0.765294 (0.00)*	–0.435575	–0.55032	–0.53095
	–0.765294 (0.00)*	(0.00)*	(0.00)*	(0.00)*
Diagnostic Checks
LM	0.8643	0.7479	0.0432	0.2406
CUSUM	S	S	S	S
CUSUMSQ	S	S	S	S
R²	0.378249	0.484337	0.680832	0.665943
WALD-Long	0.34821 (0.556)	47.83228	2.620392	3.668296
	0.34821 (0.556)	(0.00)*	(0.1086)	(0.058)***
WALD-Short	4.010729 (0.0208)*	2.684875	9.283	7.7269
	4.010729 (0.0208)*	(0.073)***	(0.005)*	(0.008)*

Source: Authors’ calculations.

Notes: Parentheses reflect p-values. Significance at 1%, 5% and 10% levels are reflected as *, ** and ***, respectively. S symbolizes that the model is stable.

Table 8.

Short-run NARDL Coefficients

IDV\ DV	EF		DF		NR (Equity)		NR (Debt)
IDV\ DV	Coefficients	P-value	Coefficients	P-value	Coefficients	P-value	Coefficients	P-value
D(NR(-1))	–	–	–	–	–0.382666	0.00*	–0.383361	0.00*
D(REER_P)	–165,346	0.03**	–76,314.85	0.1883
D(REER_N)	208,764.7	0.025**	62,074.6	0.2294	–0.260227	0.043**	–0.339702	0.0137**
D(REER_N(-1))	200,157.7	0.02**	–	–
D(NR)	–	–	–90,961.7	0.008*
D(NV)	–	–	–50,448.85	0.08***	–0.369128	0.00*	–0.329575	0.0001*
D(NV(-3))	–	–	–	–	0.129853	0.037**	0.119347	0.063**
D(SR)	–	–	18,814,081	0.00*	–114.2841	0.005*	–197.3318	0.00*
D(SR(-1))	–	–	14,862,430	0.00*	–	–	–	–
D(SR(-2))			–28,571,682	0.00*
D(SV)			–	–	–	–	–271.3484	0.0003**
D(DM1)	–	–	–8,694.201	0.2176	–	–	–	–
D(DM2)	–	–	–15,768.91	0.043**	–	–	–	–
D(DM4)	–	–	–	–	–0.022423	0.07***	–	–
D(DM4(-2))	–	–	–	–	0.034905	0.008*	–	–
D(DM5(-3))	–	–	–	–	–	–	0.003817	0.005*

Source: Authors’ calculations.

Note: Significance at 1%, 5% and 10% levels are reflected as *, ** and ***, respectively.

The Wald test results were also presented for assessing the presence of long- and short-run asymmetry. The test results show that long-run asymmetric impact of exchange rate was found only on the debt flows model at 1 per cent significance level with more pronounced effect of rupee depreciation. It is due to the fact that in order to manage the rupee depreciation, borrowings increase both at corporate and government levels. At that time, debt flows increase.

However, such effect was absent for both the foreign equity models, indicating that exchange rate has a symmetric effect on equity foreign net investments. Asymmetry was not found in Nifty 50 returns. However, the short-run Wald test results indicate the presence of asymmetric effect of REER on equity flow model at 5 per cent significance level as well as on both Nifty 50 returns models at 1 per cent significance level, where the rupee depreciation has a more pronounced effect than real rupee appreciation.

Furthermore, we observe that the ‘adjustment asymmetry’, as specified by Bahmani-Oskooee and Kanitpong (2017), does take place in case of short-run equity flows model, where the equity FIIs takes time to adjust to rupee depreciation effect because of the belief that adjustments are required at times to correct for the overvaluation of rupee in real terms with respect to its trading partners and of the expectations that if the country’s foreign exchange reserve is sufficient then the adjustment process will be smooth. Due to this, in short run, real rupee depreciation does not cause outflows.

But such effect is found to be reversed in the long run adversely affecting the foreign equity flows. In the Nifty returns case, the rupee depreciation plays a significant and negative role in effecting Nifty 50 returns and causing an asymmetric effect. The stock market investors believe that due to real depreciation, the corporate profitability will be reduced due to higher cost of imports and exports will take time to adjust in correcting the current account deficit (CAD) due to J-curve effect. These conclusions help the firm manager who can be cautious of the time lag effect on the currency depreciation observed. Furthermore, the vital inputs can be drawn for the regulators as such conclusions directly impact the trade balance.

Thus, a negative influence of rupee depreciation is observed in Nifty 50 returns for both long and short run. The Nifty 50 returns and volatility influence equity and debt flows negatively in the long run and short run, respectively. If the stock market displays persistence volatility, equity investments are adversely affected. Also, if stock market returns are negative, then it induces foreign investors to make debt investments. Such kind of conclusions are vital for domestic traders who can have fair understanding of exchange rate movements and its resulting consequences on Indian stock market in long and short term. This will enable them to better time the market.

Figure 1.

Source: Authors’ compilation.

However, positive stock market returns does not necessarily positively influence foreign equity flows in the long run due to other political factors in an economy, or increase in proportion of debt relative to equity investments. In short run, volatility in Nifty 50 returns causes foreign debt investment to shrink due to short-term negative perception about the stock market and the economy.

The debt flows were influenced only by event of quantitative easing programme by the USA fed through dummy variable 2 in the short run and the event of commodity slump in 2015 through dummy variable 3. In the long run, the S&P 500 return and its volatility could influence the Nifty 50 returns negatively. While in the short run, lags of S&P 500 return negatively influenced foreign debt investments in India.

Such conclusion is vital for the international investors who are concerned about the US market returns and volatility. In long run, the foreign investment causes a negative influence on the Nifty 50 returns. In short run, the dummy variable concerning the event of recovery from financial crisis is significant.

The ECT for all four equations was significant and negative. Equity flows model adjusts the disequilibrium caused by shocks in the previous month at a speed of 76 per cent, while the debt flows model retains its long-run equilibrium position at 43 per cent. The Nifty 50 returns model adjusts at 55 per cent when equity flows model is considered and 53 per cent when bond flows model is considered.

The diagnostic model shows that all four models are stable. The models with equity flows and debt flows as dependent variable have no serial correlation in residuals, but the Nifty 50 model with EF has such correlation. The R² value is good for all the models except the one with dependent variable as equity flows. Figure 1 shows the graphical presentation of stability diagnostics with CUSUM (cumulative sum) and CUSUM of squares tests. The two graphs show the stability of the modelling based on equity flows and debt flows and Indian stock market as dependent variables, respectively. The red lines represent the critical bounds (upper and lower bounds) at 5 per cent significance level. If the blue line representing the coefficients of estimated ARDL models crosses these bounds, then the estimated models are not stable. But from the graphs of Figure 1, it is reflected that all the estimated coefficients of four models are stable. This finding also serves as a robustness measure for the ARDL modelling.

Conclusions and Implications

This study examined the effect of changes in rupee on India’s net FII flow and on Indian stock market. It incorporated several control variables such as returns of Nifty 50 and S&P 500 and their volatility to examine such relation. Discussion on preliminary statistics suggests that equity flows was found to be highly volatile. Both debt and equity flows were normally distributed signalling that they were driven by same macroeconomic fundamentals.

The co-integrating relation among variables was found using the NARDL model, while adjusting for the structural breaks in the time series. The results depict that foreign institutional debt flows and Nifty 50 returns were affected by exogenous events, of which GFC was the vital one. Such effects were incorporated into the co-integration model as dummy variables. The important NARDL findings reported an existence of long-run co-integrating relation for the three models suggested.

Exchange rate asymmetry was present in debt flows model in the long run. While short-run effects of asymmetry are present in all models except debt flows. In the long run, appreciation and depreciation of currency in real terms positively influence net debt FIIs. But the negative relationship of real rupee depreciation was found for the equity flows. Returns from exchange rate influences FIIs decisions as found by Mukherjee, Bose, and Coondoo (2002) and Srinivasan and Kalaivani (2015). Positive feedback trading was absent. S&P 500 returns explain FII flows indicating informational asymmetry unlike found by Chakrabarti (2001). In short run, real rupee depreciation negatively influenced Indian stock market returns.

The conclusions drawn from the study serves as a vital implication for international investors, domestic traders and regulators and policymakers. First, this study concludes that foreign flows causes volatility in the Indian stock market and provides insights on nature of foreign flows with respect to their behaviour with rupee appreciation or depreciation is useful for the domestic traders by helping them better to time the market. Second, it expressly finds that despite positive Indian stock returns, foreign investors remained reluctant to invest. Furthermore, the role of US stock market returns was significant in explaining FII flows. This is useful for the potential international investors who perceive India as an attractive investment destination. Third, the role of ‘adjustment asymmetry’ and the resulting J-curve effect phenomenon potentially serve as vital inputs for business managers at the firm level. Finally, it is also vital for regulators at the aggregate level who can identify the key areas like the effect on the ‘trade balance’ on the first place when significant real rupee changes are made. Important considerations can be taken by the policymakers from this study for opening up of the Indian debt market.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

References

Bahmani-Oskooee

, & Fariditavana

(2015). Nonlinear ARDL approach, asymmetric effects and the J-curve. Journal of Economic Studies, 44(3), 519–530.

Bahmani-Oskooee

, & Kanitpong

(2017). Do exchange rate changes have symmetric or asymmetric effects on the trade balances of Asian countries? Applied Economics, 49(46), 4668–4678.

Bahmani-Oskooee

, & Saha

(2015). On the relation between stock prices and exchange rates: A review article. Journal of Economic Studies, 42(4), 707–732. Retrieved from https://doi.org/10.1108/JES-03-2015-004

Bahmani-Oskooee

, & Saha

(2016). Do exchange rate changes have symmetric or asymmetric effects on stock prices. Global Finance Journal, 31, 57–72.

Bahmani-Oskooee

, & Saha

(2018). On the relation between exchange rates and stock prices: A non-linear ARDL approach and asymmetry analysis. Journal of Economics and Finance, 42(1), 112–137.

Bai

, & Perron

(1998). Estimating and testing linear models with multiple structural. Econometrica, 66(1), 47–78.

Basin

, & Khandelwal

(2014). Relationship between foreign institutional investment, exchange rate and foreign exchange reserves: The case of India Using ARDL bounds testing approach. International Journal of Financial Management, 4(2), 33–39.

Caporale

G. M.

, Menla Ali

, Spagnolo

, & Spagnolo

(2017). International portfolio flows and exchange rate volatility in emerging Asian markets. Journal of International Money and Finance, 76, 1–15.

Caporale

G. M.

, Menla Ali

, & Spagnolo

(2015). Exchange rate uncertainty and international portfolio flows: A multivariate GARCH-in-mean approach. Journal of International Money and Finance, 54, 70–92.

10.

Chakrabarti

(2001). FII flows to India: Nature and causes. Money and Finance, 2(7), 61–81.

11.

Chakraborty

, & Kakani

R. K.

(2016). Institutional investment, equity volume and volatility spillover: Causalities and asymmetries. Journal of International Financial Markets, Institutions and Money, 44, 1–20.

12.

Chandra

, & Thenmozhi

(2015). On asymmetric relationship of India volatility index (India VIX) with stock market return and risk management. Decision, 42(1), 33–55.

13.

Chang

, & Rajput

2018, Do the changes in macroeconomic variables have a symmetric or asymmetric effect on stock prices? Evidence from Pakistan. South Asian Journal of Business Studies, 7(3), 312–331.

14.

Choe

, Kho

B.-C.

, & Stulz

R. M.

(1998). Do foreign investors destabilize stock markets? The Korean experience in 1997. Journal of Financial Economics, 54, 227–264.

15.

Dahir

, Mahat

, Ab Razak

, & Ariffin

(2018). Revisiting the dynamic relationship between exchange rates and stock prices in BRICS countries: A wavelet analysis. Borsa Istanbul Review, 18(2), 101–113.

16.

Dhingra

V. S.

, Gandhi

, & Bulsara

H. P.

(2016). Foreign institutional investments in India: An empirical analysis of dynamic interactions with stock market return and volatility. IIMB Management Review, 28(4), 212–224.

17.

Garg

, & Bodla

B. S.

(2015). Impact of the foreign institutional investments on stock market: Evidence from India. Indian Economic Review, 46(2), 303–322.

18.

Goh

Y. M.

, Zam

, & Sapian

(2017). Return, volatility and fund flows linkages: Malaysian evidence. Management & Marketing Journal, 15(2), 59–69.

19.

Kim

J. B.

, & Yi

C. H.

(2015). Foreign versus domestic institutional investors in emerging markets: Who contributes more to firm-specific information flow? China Journal of Accounting Research, 8(1), 1–23.

20.

Lakshmi

, & Thenmozhi

(2018). Impact of foreign institutional investor trades in Indian equity and debt market: A three-dimensional analysis, DECISION, 45(3), 225–233.

21.

Liu

, Bredin

, Wang

, & Yi

(2014). Domestic and foreign institutional investors’ behavior in China. European Journal of Finance, 20(7–9), 728–751.

22.

Mukherjee

, Bose

, & Coondoo

(2002, May). Foreign Institutional Investment in Indian Equity Market: An Analysis of Daily Flows during January 1999–May 2002. Money and Finance, 2(9–10), 21–51.

23.

Pesaran

M. H.

, Shin

, & Smith

R. J.

(2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16(3), 289–326.

24.

Prakash

(2012). Major episodes of volatility in the Indian foreign exchange market in the last two decades. Reserve Bank of India Occasional Papers, 33(1–2), 162–199.

25.

Rai

, & Bhanumurthy

N. R.

(2004, December). Determinants of foreign institutional investment in India: The role of return, risk and inflation. The Developing Economies, 42(4), 479–493.

26.

Rashid

, & Fazal

(2010). Capital inflows, inflation and exchange rate volatility: An investigation for linear and nonlinear causal linkages (PIDE Working Paper No. 63). Islamabad: Pakistan Institute of Development Economics.

27.

Samal

K. C.

(1997). Emerging equity market in India: Role of foreign institutional investors. Economics & Political Weekly, 32(42), 2729–2732.

28.

Shin

, Yu

, & Greenwood-Nimmo

(2014). Modelling asymmetric cointegration and dynamic multipliers in a nonlinear ARDL framework. In Sickles

Robin C.

& Horrace

William C.

(Eds.), Festschrift in honor of Peter Schmidt: Econometric methods and applications (pp. 281–314). New York, NY: Springer.

29.

Singhania

, & Saini

(2016). FII, stock exchange return and the leverage effect: Evidence from India. Journal of Wealth Management, 19(1), 103–119.

30.

Srinivasan

, & Kalaivani

(2015). Determinants of foreign institutional investment in India: An empirical analysis. Global Business Review, 16(3), 364–376.

31.

Sui

, & Sun

(2016). Spillover effects between exchange rates and stock prices: Evidence from BRICS around the recent global financial crisis. Research in International Business and Finance, 26, 459–471.

32.

Vardhan

, & Sinha

(2014). Influence of foreign institutional investments (FIIs) on the Indian stock market (Munich Personal RePEc Archive). Retrieved from http://mpra.ub.uni-muenchen.de/53611/

33.

Walid

, Aloui

, & Nguyen

(2012). Asymmetric effects and long memory in dynamic volatility relationships between stock returns and exchange rates. Journal of International Financial Markets, Institutions and Money, 22, 738–757.

34.

Waqas

, Hashmi

S. H.

, & Nazir

M. I.

(2015). Macroeconomic factors and foreign portfolio investment volatility: A case of South Asian countries. Future Business Journal, 1(1–2), 65–74.

35.

Yaha

, Singh

, & Rabanal

J. P.

(2017). How do extreme global shocks affect foreign portfolio investment? An event study for India. Emerging Markets Finance and Trade, 53(8), 1923–1938.

36.

Yang

C. W.

(2017). Do foreign investors destabilize stock markets? A reexamination of Korea in 2008. Asia-Pacific Journal of Financial Studies, 46(5), 734–759.