Dynamic Interactions between Foreign Institutional Investment Flows and Stock Market Returns – The Case of India

There has been a marked increase in the magnitude of Foreign Institutional Investments (FIIs) into India since the 1990s, resulting in increased forex reserves and liquidity and a higher-valued Indian capital market. However, such investment is more volatile than other types of flows, causing disruptive effects in the form of sudden stops (for example, the crash of the Indian stock market on January 21, 2008). This study empirically examines the dynamic relationship between FIIs and Indian stock market returns. It also analyses the effects of FIIs on Indian capital market returns, using data from January, 2004 through September, 2012. The analysis employs a Cross Correlation Function (CCF) approach, a Granger Causality Test and Vector Auto Regression after dividing the data into two parts: Pre Global financial crisis and Post Global financial crisis periods. The results of the CCF suggest bi-directional causality between FIIs and Nifty returns, whereas the Granger Causality Test and the VAR analysis suggest uni-directional causality running Nifty returns to FIIs.


Introduction
India opened up its economy in the early 1990s following a major crisis led by a foreign exchange crunch that dragged the economy near default. Until 1991, India followed a restrictive policy towards Foreign Institutional Investments (FIIs) and Foreign Direct Investments (FDIs) -relying more heavily on bilat-eral and multilateral pacts with long maturities. India has continued to be attractive to FIIs since 1993, when foreign institutional investors started investing in the Indian capital market. Additionally, the government Individuals with better information and judgement are inclined to participate in this market to take advantage of such information asymmetry; the actions of participants swiftly feed this information into the market, causing changes in the prices of derivatives. Therefore, these markets indicate what is likely to happen and help improve price discovery. The empirical research carried out by Chan, Chan & Karolyi (1991), Antonios & Phil (1995), Choudhry (1997), Pericil & Koutmos (1997), Bollen (1995), Abhayankar (1998), Gulen & Mayhew (2000), Mckenzie, Brailsford & Faff (2001), Thenmozhi (2002), Shenbagaraman (2003), Hetamsaria & Swain (2003) and Mukherjee & Mishra (2006) suggest the existence of a lead-lag relationship between the derivatives market and the underlying spot market.
Attractive prospects in emerging market economies (EMEs), together with low interest rates in advanced economies, are likely to lead to continuing net capital inflows and exchange rate pressures in many emerging market economies. Along with some of the EMEs, Foreign capital is free, unpredictable and always on the lookout for higher profits. FIIs frequently move investments, and those swings can bring severe price fluctuations, resulting in increased volatility. In fact, FIIs bear significant responsibility for volatility in Indian markets. Increased investment from overseas may shift control of domestic firms to foreign hands. Foreign institutional investors play a major role in the derivatives market, as their investments, measured in rupees, greatly exceed those of domestic institutional investors Their massive buying and selling activities create problems for small retail investors, whose fortunes are driven by the actions of large FIIs. The Japanese Asset Price Bubble (1990), the East Asia Financial Crisis (1997), the Russian Financial Crises (1998)  However, portfolio flows, which move in tandem with domestic and international market sentiment, are more volatile than other types of flows. Calvo (1998) showed that a sudden stop (Dornbusch, Goldfajn & Valdes, 1995) or sudden withdrawal is followed by a large capital inflow in the form of Foreign Portfolio Investment and later Calvo (1998) proposed an analytical framework to examine the impact of a sudden and largely unexpected cut-back in foreign capital inflows to emerging economies. Calvo (2009)  decisions and actions by policy makers and small investors will be greatly facilitated. Situations of sudden stops cannot be fully avoided, but at least the adversity associated with these events can be reduced.
The paper is organized as follows: Section 2 summarizes the previous literature. The data, the sample period and the methodology used to examine dynamic interactions between stock market returns and foreign institutional investment are elaborated in section 3.
The empirical results of the study are discussed in section 4. Section 5 summarizes the findings and derives the conclusion of the study.

Literature Review
Although many economies liberalized during the 1980s and 1990s, several studies have documented home-bias among foreign investors (Frankel, 1991). French and Poterba (1991) and Cooper and Kaplanis (1994) showed empirically that if equity returns are negatively correlated with inflation in the home country, investors with low levels of risk aversion tend to exhibit home-bias in their equity portfolios. Information asymmetry between domestic and foreign investors has been found by Gehrig (1993), Coval and Moskowitz (1999), Brennan and Cao (1997) and Kang and Stulz (1997) to be among the main factors driving home-bias. Foreigners face 'lemons' effects, as they are poorly informed and vulnerable to being overcharged in acquiring shares of domestic firms (Gordon & Bovenberg, 1996). Categorizing investors in the Korea Stock Exchange (KSE) into domestic individual investors, domestic institutional investors and foreign investors, Choe, Kho, and Stulz (2001) find that foreign institutions are at less of a disadvantage relative to domestic institutions than they are relative to domestic individuals. Grinblatt and Keloharju (2000), in a study of the Finnish stock market, found foreign investors to be heavy momentum investors, i.e., buying past winning stocks and selling past losers. On the other hand, Finnish investors, particularly households, are contrarians -buying losers and selling winners. More generally, they find that Finnish investors in all categories are less sophisticated than foreign investors. Furthermore, in Thailand and Singapore, foreign investors are found not to be at an informational disadvantage but rather to possess superior information processing ability (Bailey, Mao, & Sirodom, 2007). Therefore, it is plausible that global institutional investors invest in acquiring information, owing to their resources, size, domain expertise, global experience and niche skills. Dvořák (2005) mediates these disagreements, finding that global investors lack local information but possess expertise.
Several researchers have shown that portfolio investment in an emerging market often gives rise to classic speculative bubbles. Foreign institutional investors pump capital into these markets, generating bubbles and increasing stock market volatility (Grabel, 1995). The process of liberalization, innovation, deregulation and globalization increases the volatility of capital markets. Foreign portfolio flows, which are unstable, act as an additional source of volatility (Claessens, Dooley & Warner 1995;Grabel, 1995), creating difficulties in the pricing of financial assets. On the constructive side, foreign portfolio flows increase the efficiency of capital markets (Clark & Berko, 1997). De Brouwer (1999) observes that the volatility of capital flows is unlikely to end: outflows were preceded by inflows, and most likely, they will be followed by inflows. The pattern of capital movements to emerging markets over the past 30 years or so has been one of ebb and flow rather than stasis. Some observers, however, believe that the builtin volatility of capital flows, as demonstrated most starkly by ''sudden stops'' (Calvo & Reinhart, 2000), ''hot money'' (Stiglitz, 1999) and even capital flight, adversely affects the economy, especially during economic downturns in countries with small "absorptive capacity" and weak investor protections (Lemmon & Lins, 2003). It is possible that openness and integration could depress growth (Ferreira & Laux, 2009). Wang & Shen (1999) observed that FIIs, due to their stabilizing and demonstration effects, positively affect local stock markets in host countries. With respect to stabilizing and demonstration effects, they argue that because FIIs in developing countries focus on stock fundamentals, their trading schemes tend to stabilize stock markets. In the long run, this strategy helps stock markets mature.
Momentum trading or the feedback trading hypothesis (Grinblatt & Keloharju, 2000) suggest that a shock to security returns leads to changes in capital inflows, causing further changes in security returns. They reported that foreign investors tend to be momentum investors, i.e., they tend to buy past winning stocks and sell past losers. Foreign institutional investors tend to exhibit return-chasing behavior, i.e., they buy when the market rises and sell when the market drops. This is destabilizing, as selling activities cause the capital market to sink further (Radelet & Sachs, 1998). Chakrabarti (2001) states that flows are highly correlated with equity returns in India and that they are more likely to be an effect rather than a cause of such Kumar (2009). The dependence of net FII flows on daily returns in the domestic equity market at a day's lag is suggestive of foreign investors' return-chasing behavior; their decisions appear to be affected by the recent history of market returns and volatility. This casts them as feedback traders (Mukherjee et al., 2002). Gordon & Gupta (2003) find a significant negative correlation between monthly flows and lagged returns and examine the determinants of FIIs in India, using a multivariate regression model. Griffin, Nardari & Stulz (2004) reveal that foreign flows are significant predictors of returns in Korea, Thailand, Taiwan and India, indicating that foreign investors buy before the market index increases. They also find that contemporaneous flows are positive and highly significant in India but fail to predict future values. The results of Ananthanarayanan, Krishnamurti & Sen (2009) are consistent with the base-broadening hypothesis; however, they do not find compelling confirmation of momentum strategies employed by foreign institutional investors and reject the claim that foreigners destabilize the market. Foreign investors have the ability to be market makers, given their voluminous investments (Babu & Prabheesh, 2008). Inoue (2008)

Data and Methodology
The data set comprises daily closing prices of the S&P CNX Nifty of the National Stock Exchange of India Ltd. and values of different FIIs-related series, viz.  Table 1. The Chow forecast test is used to estimate two models-one employing the full set of data T and the other employing a long sub-period T 1 . The F-statistic is computed as: The CUSUM of squares test (Brown, Durbin & Evans, 1975) provides a plot of S t (expected standard error of regression) against t and a pair of 5 percent critical lines. As with the CUSUM test, movements outside the critical lines suggest parameter or variance instability.
The graph in Figure 1 indicates the presence of a structural break during the global financial crisis.
Therefore, further analysis, applied to the whole sample period (January, 2004-September, 2012  Stationarity is examined by means of an autocorrelation function (correlogram) and a unit root test.
The pioneering work on testing for unit roots in time series was performed by Dickey and Fuller (1979;1981), and later, a non-parametric test was used by Phillips and Perron (1988) to check for the presence of a unit root in time series. The Nifty is stationary after the first log difference i.e., I (1), but all FIIs-related series are I(0) and show the presence of a significant trend. Therefore, the deterministic trend is removed by regressing the series against the time trend, and the residuals thereby obtained are used as a detrended FII series.
To establish a lead-lag relationship between two time series, Nifty and FII, Cross Correlation Functions (CCF) are estimated, as CCF can help identify lags of the independent variable that might be useful in predicting the dependent variable.
In this study, the Cross Correlation Function takes the following form: where the standard deviation of each sequence is assumed to be time-independent.
Granger Causality tests are used to determine causality between two variables. In the present study, a Granger Causality test is applied to the following pair of regression equations: where m is a suitably chosen positive integer; α i , β j , γ i and δ j = 0, 1… k are parameters; t is a time or trend variable; and it is assumed that µ 1t and µ 2t are uncorrelated disturbance terms with zero mean and finite variance. The equations are estimated for each type of FII flow and Nifty returns. A vector auto-regression (VAR), as proposed by Sims (1980), is estimated to capture short-run causality between Nifty returns and FII investment. VAR is commonly used to make forecasts using systems of The VAR equations can be formulated as:

Empirical Analysis
The autocorrelation figures for daily index returns, daily trends of FIIs in the futures market and overall daily Unit root tests can be used to confirm serial correlation or autocorrelation.   Note: Lag length was selected using Schwarz Bayesian Criterion (BIC). If we compare the Akaike Information Criterion (AIC) and the Schwarz Bayesian Criterion, BIC imposes a harsher penalty for adding more variables to the model. According to BIC, there could be a significant causal relationship either at lag 1 or lag 2. Therefore, in analysis, lags up to 2 are considered in estimating the regression model. * Significant at the 5% level.

Policy Suggestions
The study offers findings that policy makers can use the Indian equity market. FIIs should be allowed greater flexibility to switch between equity and debt investments, as more balanced strategies may help stabilize movements of FIIs into and out of India (Bawa, 2012).
To address these issues, the government might seek to promote financial sector prudence, which would be an indirect effort to prevent asset bubbles in the financial markets. It can impose restrictions on bank loans, asking banks to maintain higher provisions on loans to certain sectors such as real estate or the equity market to avoid bubbles in these asset classes. Macro-economic managers can impose a ban on certain financial activities temporarily, as deemed necessary. For instance, Taiwan imposed capital controls on November 9, 2010 to curb currency appreciation. A tax on exit can be imposed on investors wishing to sell assets or withdraw money before a stipulated time. Brazil doubled its tax on foreign portfolio inflows into bonds and some other financial instruments from 2 to 4 percent in 2010 to curb currency appreciation. During the same time period, Thailand imposed a 15 percent withholding tax on capital gains and interest income from foreign investments. The Thai baht gained the most among currencies in the region except that of Japan, and the SET index of the Thai stock market soared by 30 percent in four months. Ceiling or capping inflows is a direct measure, which India is now implementing. For instance, the upper limit on FIIs in corporate bonds is fixed at $ 15 billion, and that in government bonds is $ 5 billion.
Policy makers can also implement a Tobin tax. For example, the securities transaction tax (STT) imposed by India in 2004 is a type of Tobin tax. The Taiwanese and Brazilian measures mentioned above are also examples of Tobin taxes (Kazi, 2011). There is no universally appropriate solution to sudden stops. Based on circumstances, a country can adopt one or a combination of policy measure(s).

Conclusion
This study has examined the Lead-lag relationships be- To investigate the causal relationship between FII flows and market returns, a cross correlation approach and a combination of Granger causality and Vector Auto Regression approaches have been adopted. The results provide insights into the behavior of FIIs that would be useful to the formulation of utilitarian policies. Under the cross-correlation approach, bidirectional causality is found between FII flows and Nifty returns. In Phase 2, strong unidirectional causality is found to run from FIIs to Nifty returns, confirming the behavior of FIIs as feedback traders. Since January, 2008 FII outflows have been persistent, giving rise to episodic sudden reversals. A Granger Causality Test suggests that FII flows are significantly affected by returns in the equity market; however, the latter is not significantly influenced by variations in these flows.
A Vector Auto Regression indicates that variations in the time series are strongly explained by its own lags.
However, variations in Futures Buy and Futures Sell are also explained by Nifty along with their own lags.
The reason for this result is that FIIs in the Indian market extrapolate trends in stock price changes. Thus, after some price decrease, they anticipate a further dip in stock prices and hence sell shares. Such actions, when taken by a large number of investors, suggest that stock prices will continue to decline in the future. Therefore, investors' expectations lead them to sell their shares following a decrease in index prices, leading to negative feedback trading behavior among FIIs. Thus, the turbulent effects of FIIs cannot be disregarded.