30
- 147 - 8 Determinants of FIIs in Indian Stock Market This chapter is designed to find out the factors that determine the flow of FIIs to India. The nature and direction of causality between returns on Indian stock market and FIIs investment flows have also been examined. This chapter is divided into eight sections. Section I presents a brief background of the problem. Section II recaptures the research methodology used here in. While descriptive statistics are given in Section III, while Section IV outlines time series properties of data. Section V presents the bi-variate analysis. Section VI gives the result of the multivariate analysis while section VII describes the causality relationships and in the last section the conclusion is given. 8.1 Background The ongoing transition in the international financial markets is exhibiting several features of which capital inflows are perhaps the most important. Emerging economies have experienced massive capital inflows, which in some cases have proved to be troublesome later. The question that arises is why do the capital flows from foreign countries have brought in negative effects on the host countries? Whether such situations occurred due to incompatible macro economic and exchange rate policies or imprudent banking policy or lack of liquidity in the market. The above-mentioned question is a subject of debate at the theoretical plane and issue of future policy concern under country specific situations. Whatever be the reason, the issue boils down to the fact whether such flows are properly paced and properly sequenced such that the inflow of capital is not excessive relative to the maturity of the system in which it must be absorbed; only then the capital flows can be sustained and systemic stability ensured. As suggested by the Bhattacharya and Mukherjee (2005), this may be attained through structural and operational realignment of the domestic and financial sector variables of the economics exposed to global financial

Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

  • Upload
    others

  • View
    12

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 147 -

8 Determinants of

FIIs in Indian Stock Market

This chapter is designed to find out the factors that determine the flow of FIIs

to India. The nature and direction of causality between returns on Indian stock market

and FIIs investment flows have also been examined. This chapter is divided into eight

sections. Section I presents a brief background of the problem. Section II recaptures

the research methodology used here in. While descriptive statistics are given in

Section III, while Section IV outlines time series properties of data. Section V

presents the bi-variate analysis. Section VI gives the result of the multivariate analysis

while section VII describes the causality relationships and in the last section the

conclusion is given.

8.1 Background

The ongoing transition in the international financial markets is exhibiting

several features of which capital inflows are perhaps the most important. Emerging

economies have experienced massive capital inflows, which in some cases have

proved to be troublesome later. The question that arises is why do the capital flows

from foreign countries have brought in negative effects on the host countries?

Whether such situations occurred due to incompatible macro economic and exchange

rate policies or imprudent banking policy or lack of liquidity in the market. The

above-mentioned question is a subject of debate at the theoretical plane and issue of

future policy concern under country specific situations. Whatever be the reason, the

issue boils down to the fact whether such flows are properly paced and properly

sequenced such that the inflow of capital is not excessive relative to the maturity of

the system in which it must be absorbed; only then the capital flows can be sustained

and systemic stability ensured. As suggested by the Bhattacharya and Mukherjee

(2005), this may be attained through structural and operational realignment of the

domestic and financial sector variables of the economics exposed to global financial

Page 2: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 148 -

network. It is in the context that the interlinkage among the FII flows, stock market

return and risk, exchange rate, interest rates prevailing in investing and host country

etc. needs to be addressed. On what factors foreign institutional investments depend is

a prime issue to be sorted out in this research beside others.

The present chapter focuses on this issue in the Indian context. In fact, from

among the whole gamut of institutional reforms undertaken in India since the 1990‟s,

gradual abolishment of capital flow barriers and foreign exchange restrictions,

adoption of more flexible exchange rate arrangements deserve a special attention at

these juncture to examine the determinants of FII flows. Some researchers made

attempts to identify the determinants of FII flows to India. Most of the existing studies

found that the equity return has a significant and positive impact on the FII

(Aggarwal, 1997; Chakrabarti, 2001; and Trivedi & Nair, 2003). But given the huge

volume of investments, foreign investors could play a role of market makers and book

their profits, i.e. they can buy financial assets when the prices are increasing (Gordon

and Gupta, 2003). Hence, there is a possibility of bi-directional relationship between

FII and the equity returns. Ahmad et al (2005) make a firm level analysis of FII‟s role

in the Indian equity market. At the aggregate level, FII investments and NSE Nifty

seem to have a strong bi-directional causality.

Asha C. Parsuna (2000) finds that mainly the return in the host country stock

market attract the FIIs investments, other factors are also creating impact on the arrival

of FIIs but they are statistically insignificant. Kumar (2001) has also shown the similar

result in his study. He inferred that FII flows do not respond to short-term changes or

technical position of the market and they are more driven by fundamentals. The study

finds that there is causality from FIIs to Sensex. This is in contradiction to Rai and

Bhanumurthy (2003) results using similar data but for a larger period. A study by

Panda (2005) also shows FII investments do not affect BSE Sensex. No clear causality

is found between FII and NSE Nifty. FIIs are found to follow positive feedback

strategy and to have return clustering tendency.

Following the Asian crisis and the brust of info-tech bubble internationally in

1998-99, the net FII investments have declined by US $ 61 million. But there was

not much effect on the equity returns. Chakrabarti (2001) has marked a regime shift

in the determinants of FII after Asian crisis. The study found that in the pre-Asian

crisis period any changes in FII found to have a positive impact on the equity returns.

Page 3: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 149 -

But in the post-Asian crisis period a reverse relation was found, i.e. a change in FII

was mainly due to change in equity returns.

There are also a number of studies on exchange rate affecting stock prices

directly. Theory explained that a change in the exchange would affect a firm‟s foreign

operation and overall profits. This would, in turn, affect its stock prices. The nature of

the changes in stock prices would depend on multinational characteristics of the firm.

Conversely, a general downward movement of the stock market will motivate

investors to seek better returns elsewhere. This decreases the demand for money by

pushing interest rates down, causing further outflow of funds and hence depreciating

the currency. While the theoretical explanation was clear, empirical evidence was

mixed. Aggarwal (1981) found a significant positive correlation between the US dollar

and US stock prices while Soenen and Hennigan (1988) reported a significant negative

relationship. Soenen and Agarwal (1989) found mixed results among industrial

countries. Ma and Kao (1990) attributed the differences in results to the nature of the

countries i.e. whether the countries were export or import dominant. Morley and

Pentecost (2000), in their study on G-7 countries, argue that the reason for the lack of

strong relationship between exchange rate and stock prices may be due to the

exchange controls that were in effect in the 1980s.

Among the more recent studies Narayan and Smith (2005) in their study on

exchange rates and stock prices in South Asia found that exchange rate Granger cause

stock prices in India both in the long run and short run. Venkateshwarlu and Tiwari

(2005) analyzed bi-variate causality between stock prices and exchange rates.

Bhattacharya and Mukherjee (2006) investigate the nature of the causal relationship of

FIIs with stock return and exchange rate in India by applying co-integration and long

term Granger Causality test and find a bi-directional causality between stock return

and FIIs investments. But no causal relationship between exchange rates and net

investments by FIIs investments is found.

After going through the existing studies on the subject under reference of this

chapter, we are in a position to note some gaps in them. Firstly, the period under

reference of these research studies to relatively shorter. Secondly, the number of the

independent variables considered for examining their linkage with FIIs has remained

limited. Thirdly, the multiple regression model was applied without verifying the

properties of the times series data such as stationarity, autocorrelation etc. Lastly,

Page 4: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 150 -

majority of the authors have proceeded with only one dependent variable i.e. either

purchase or sale or net investment by the FIIs. The present study is an improvement

over the earlier studies in several ways. First, it is a comprehensive study, as it

considers as many as 24 independent variables. Second, the period under reference is

comparatively longer than the earlier studies. Third, well-established model of

examining causality (i.e. Granger Cause) has been applied. Fourth, times series

properties of the data were examined prior to the application of regression models.

Lastly, as day-to-day variation in FII flows and market return is likely to contain

relatively large random components, one would not expect high correlation between

FII flows and market return. We therefore tried to explore whether the FII flow

variables would show any stronger dependence on market return if daily variation

were filtered through moving average smoothing.

Scope of the study is restricted to India because it is an appropriate case for

conducting such a study as portfolio investment has become the dominant path of

foreign investment in Indian economy. India liberalized its financial market and

allowed FIIs to participate in their domestic markets in 1992. The opening up of the

market resulted in a number of positive effects. First, the stock exchanges had to

improve the quality of their trading and settlement procedures in line with the best

practices of the world. Second, the transparency and information flows improved on

account of the entry of FIIs in India. However, people are also guessing negative

effects in the form of potential destabilization because of the bulk buying and selling

activity of FIIs. The period of the second-generation economic reform i.e. March

1999 to December 2006 is taken as reference period.

8.2 Assortment of Variable and Research Methodology

Dependent Variables: The present study consider six dependent variables namely:

FIIs sales (FIIS), purchases (FIIP), net investment (FIIN) and 7 days moving averages

of each of them denoted as FIIS_MA, FIIP_MA and FIIN_MA respectively. The logic

of taking these variables is as follows: Rationally, global investors would continuously

adjust investment portfolio round the clock using available market information and

thereby tracking the returns on all possible markets. The trading behaviour of these

investors can be classified into two categories: (I) Momentum or Positive Feedback

Page 5: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 151 -

Trading and (II) Herding strategy. In the case of the Momentum Trading or Feedback

Trading, the investors have a tendency to buy and sell stocks based on their observed

return records i.e. to buy recent winners and sell recent losers. In case of Herding

strategy all investors behave in a similar manner and take decision by observing the

behaviour of other investors. To capture these behavioural patterns the investor‟s

action may be aggregated and summarized into two basic measures: (I) Sale and (II)

Purchase. Hence, we have chosen to examine the nature of FIIs flows to India in terms

of three variables: FIIs sales, FII purchases and FIIs net investment. Further, as the

time series data have been taken on the daily basis so to remove the effect of day to

day variation 7 days moving average of the above mentioned variables have also been

taken as dependent variables. Thus, in total 6 dependent variables are taken.

Explanatory Variables: The independent variables considered for this study includes:

(i) Stock return in the host country;

(ii) Stock return on the major investing country i.e. US;

(iii) Stock return in emerging countries represented by Morgan Stanley Capital

International World Index (MSCI);

(iv) One day Lag return of host country and investing country markets;

(v) Moving average of 7 days return of all index (host and Investing

countries);

(vi) Risk at host country and home country markets;

(vii) Exchange rate of US $ v/s Indian Rupee;

(viii) Federal Bank 3 months treasury bills interest rates;

(ix) Growth rate of Indian economy represented by the index of industrial

production (IIP);

(x) One day lag investment of foreign institutional investors;

(xi) Differential return between host country and investing country markets:

and

(xii) Link relation of host country and investing country market represented by

their Beta.

Page 6: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 152 -

Methodology: Before making the data analysis, the time series properties of various

data series were identified. More specifically, the stationarity problem of both

dependent and independent variables were examined. For verifying time series

properties of data, we used the Augmented Dickey-Fuller Test (ADF) for checking the

unit root of the major variables. The following form of ADF regression equation was

used:

Yt = 1 + 2t+ Yt-1 + i m

i=1 Yt-i+ t----------------------------------------(8.1)

Where t is a pure white noise error term and Yt-1 additional lagged term, included

with an idea of ensuring that the errors are uncorrelated. 1, 2, , are the coefficients

where is the first difference operator, which is equal to (p-1), estimated to test the

null hypothesis that = 0. If is equal to zero it mean there is a Unit Root which

implies non-stationarity in the series under consideration. If a series is stationary at

level, it is also integrated of zero order and a series stationary at Ist difference is

integrated of 1st order and so on.

As the use of differenced variables instead of original ones may sometimes

result in the serious loss of long run information, it is essential to keep the long run

information on the variables and to avoid the problem of spurious regression. These

two problems have to be avoided simultaneously. For this purpose possible co-

integration between the variables has to be checked. To find out the co-integration

between variables, Augmented Engle-Granger Test (AEG) test was conducted. For this

very first we had to estimate co-integration regression using variables having the same

order integration. The co-integration equation by the OLS method is as follows:

Yt = a0 + a1X1 + a2X2 + anXn + Zt----------------------------------------(8.2)

The residuals (Zt) from the co-integration regression are subject to the test

stationary by applying Augmented Dickey Fuller unit root test based on the following

equation:

(ADF) Zt = 1 + 2t+ Zt-1 + i m

i=1 Zt-i+ t-----------------------------------

(8.3)

If the Z proved stationary, it means that calculated co-integration regression is

not spurious

Page 7: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 153 -

After verifying the time series properties, we specify the proper regression

equation because if the regression equations are under or overstated then the results

may not be correct. To specify the proper regression equations data mining technique

was used. In this technique, first we used the bi-varaiate form of the OLS and find out

the respective regression coefficients, R-square values and Durbin-Watson values of

all the 24 independent variables with each of the 6 dependent variables. Only those

variables were used for the final multiple regression analysis which were significantly

associated with the particular dependent variable. The results of the bi-variate form of

OLS applied to examine the significance of various independent variables as

determinants of FII flows in India are given in Table 8.5.

After bi-variate analysis, we conducted multivariate analysis so as to study the

relationship between the various dependent variables with explanatory variables over

the period 1999-2007. For the purpose of multiple regression analysis we made

separate specification of the model for each of the dependent variables (i.e. FIIS, FIIP,

FIIN, FIIS_MA, FIIP_MA and FIIN_MA) with the appropriate independent variables

[decided as per data mining technique].

This analysis has carried out in two stages. In the first stage we took only

those variables as regressors, which are selected by data mining technique (as

specified in the equations 8.6 to 8.11). In the second stage to improve R2 and to make

the data free from the autocorrelation we introduced one more independent variable in

each model. That is FIIN_MA, FIIS_MA and FIIP_MA in case of FIIN, FIIS and

FIIP respectively while in case of other three variables it was their own one day lag

values. The results of (both the stages) multivariate form of OLS applied to examine

the significance of various independent variables as determinants of the FII

investments in India are given in Table 8.6 and 8.7.

In the last section of this chapter, the results of Granger causality test are

given. This test was used to determine causal relationship amongst the dependent and

independent variables under reference. Granger Causality Test is a bi-variate analysis

and involves estimates X(Y X) and Y(X Y) by using following pair of regressions:

Yt = 0 + n

i=1 i Xt-i + n

i=1 i Yt-i + 1t---------------------------------(8.4)

Xt = 0 + n

i=1 i Yt-i + n

i=1 i Xt-i + 2t---------------------------------(8.5)

Page 8: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 154 -

In the above equations Yt, Xt are the variables to be tested and i, i, i i are

coefficients explaining the relation of dependent variable with the lag terms of

independent variable and lag terms of dependent variable in itself. 1t, 2t are mutually

uncorrelated white noise errors. t is the time period and i is the number of lags. The

null hypothesis is i,= I = 0. If the i is statistically significant but I is not then it

means X causes Y, in the reverse case Y causes X. But if both are significant then

causality runs both ways. We have taken the 2 lags as it is prescribed by the various

researchers that 2 lags are sufficient to explain causality.

8.3 Descriptive Statistics

To begin with let us glance through the summary measures exhibited in Table

8.1 and 8.2 concerning various dependent and independent variables, respectively. It

is evident from the former table that the average daily net FII inflows in India are Rs.

98.51 crore with a standard deviation of Rs. 316.22 crore, while average daily

purchases and sales are Rs. 656.53 and Rs. 552. 29 crore respectively with standard

deviation of Rs. 703.64 in the former and Rs. 627.53 in case of the later. It shows

greater fluctuation in purchases in comparison with sales. All the three dependent

variables are positively skewed and leptokurtic in nature. Thus, the FII investments

are not normally distribution, rather they are positively skewed.

TABLE 8.1: SUMMARY STATISTICS OF DEPENDENT VARIABLES

(1999-2007)

(Amt. Rs. Crore)

Dependent Variables

Variable FIIN FIIP FIIS

Mean 98.51 656.53 552.29

Median 51.20 361.90 285.40

Max. 3319.60 5457.60 5427.40

Min. -2824.70 9.90 0.00

Std.Dev. 316.22 703.64 627.54

Skewness 1.31 2.01 2.32

Kurtosis 25.74 7.94 10.14

Jarque-Bera 42187.09 3244.25 5789.49

Page 9: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 155 -

(P-Value) (0.00) (0.00) (0.00)

Table 8.2 provides that, both mean and standard deviation of the daily return in

case of Indian stock markets (BSE and NSE) are found higher than those of the US

market index (S&P) as well as the emerging market index (MSCI). It means that

Indian stock market has performed better than the US and other emerging markets over

the recent years in terms of return. However the risk involved is higher in the former

than the later. While the return series are negatively skewed in case of BSE and NSE

(Indian Markets), these series are positively skewed in case of S&P and MSCI. Table

8.2 further shows that, on an average, the exchange rate of Indian currency against US

dollar is Rs.45.176. During the study period it was maximum 48.65 and minimum

39.33. As regard the kurtosis, table shows that except the FBIR and IIP series all

posses leptokurtic distribution. As per the result of Jarque-Bera statistics, each variable

under the study turned non-normal.

TABLE 8.2: SUMMARY STATISTICS OF INDEPENDENT VARIABLES

(1999-2007)

Independent Variables

Summary

Measures

BSE Sensex

Return

(%)

NSE

Return (%)

MSCI

Return

(%)

S & P

Return

(%)

FBIR US_EX IIP

Mean 0.08 0.08 0.03 0.02 3.18 45.71 187.05

Median 0.16 0.16 0.03 0.05 3.33 45.75 180.00

Max. 8.25 8.30 6.73 5.73 6.24 48.65 253.60

Min. -11.14 -12.24 -5.31 -4.92 0.80 39.33 144.40

Std.Dev. 1.58 1.56 91 1.0 1.74 1.91 28.58

Skewness -0.3868 -0.4295 0.2697 0.16 0.12 1.85 0.56

Kurtosis 6.7398 7.8178 7.6028 5.08 1.53 28.30 2.23

Jarque-Bera 1177.06 1932.88 1733.34 359.55 177.97 52777.01 151.65

(P-Value) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

8.4 Time Series Properties of Data

Table 8.3 which provides the result of ADF test indicates clearly that that all

the dependent and independent variables are stationary at level except Federal Bank

Page 10: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 156 -

Interest rate and Industrial Production Index which are stationary at Ist difference

and to make them integrated on the level we used differencing technique. The

differencing technique might result in loss of long run information which is

otherwise essential to return. To deal with the problem Augmented Engle Granger

Test was used. This test first estimate the co-integration regression using variables

having the same order integration. The co-integration was estimated using OLS

method. The stationarity of the residual of the co-integration regression was verified

by the use of ADF unit root test. The result of AEG and ADF are given in table 8.4.

It is obvious from this table that error term of the OLS regression explaining the net

FII investments (FIIN) by selected major independent variable (such as NSE and

MSCI return, risk, FBIR and IIP etc.) is significantly stationary. This implies that

the various series are co-integrated.

TABLE 8.3: RESULTS OF THE AUGMENTED DICKEY FULLER UNIT

ROOT TEST

Variable Constant, Without Trend Constant, With Trend

BSE -25.7218*, Level --25.78205*, Level

FIIN -15.8364*, Level -16.34354*, Level

FBIR -25.80197*, Ist Difference -26.01480*, Ist Difference

MSCI -25.85145*, Level -25.87004*, Level

NSE -25.6853*, Level -25.7309*, Level

S & P -27.67409*, Level -27.68075*, Level

US_EX -4.49117*, Level -4.481535*, Level

IIP -25.3460*, Ist Difference -25.35231*, Ist Difference

FIIS -5.658116* Level -9.012926* Level

FIIP -6.351345* Level -10.33795* Level

*Hypothesis rejected at 1% significance level

TABLE 8.4: RESULTS OF AUGMENTED ENGLE-GRANGER

TEST OF CO-INTEGRATION

ADF Test Statistic -21.92628 1% Critical Value* -3.4368

5% Critical Value -2.8636

Page 11: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 157 -

8.5 Bi-Variate Analysis

As mentioned in the research methodology, in order to identify the

independent variables, which deserve to be included in the multiple regression

technique, the data mining technique was used. For this, bi-variate form of OLS was

used. More specifically, only those variables are considered for the multiple

regression models, which show significant relationship with the dependent variables.

Table 8.5 consists of the result of bi-variate form of OLS.

TABLE 8.5: REGRESSION COEFFICIENT AND OTHER STATISTICS

RESULTING FROM BI-VARIATE ANALYSIS

Dependent Variables

Independent Variables

(1)

FIIN

(2)

FIIN_MA

(3)

FIIS

(4)

FIIS_MA

(5)

FIIP

(6)

FIIP_MA

(7)

BSE Return Coefficient

(P-Value)

1097.86*

(.016)

1674.421*

(.00)

196.403

(.830)

113.004

(.891)

1254.41

(.220)

1513.759

(0.099)

R2

Adjusted R2

S.D

DW

F Value

.003

.002

315.8300

1.278

5.784

.018

.018

192.314

.124

36.314

.000

.000

627.6954

309

627.6954

.000

.001

.019

.372

.019

.001

.000

703.5544

.372

1.503

.001

.001

636.5429

.15

2.723

BSE_MA Coefficient 11092.336* 9342.632* 613.900 753.28 10043.27* 9194.46*

10% Critical Value -2.5679

*MacKinnon critical values for rejection of hypothesis of a unit root.

Augmented Dickey-Fuller Test Equation

Method: Least Squares

Variable Coefficient Std. Error t-Statistic Prob.

LAG(-1) -0.921115 0.042010 -21.92628 0.0000

D(LAG(-1)) -0.190892 0.034278 -5.568990 0.0000

D(LAG(-2)) -0.141279 0.022979 -6.148240 0.0000

C -0.138156 6.233266 -0.022164 0.9823

R-squared 0.563334 Mean dependent var -0.347746

Adjusted R-squared 0.562638 S.D. dependent var 409.3216

S.E. of regression 270.6980 Akaike info criterion 14.04200

Sum squared resid 1.38E+08 Schwarz criterion 14.05376

Log likelihood -13237.61 F-statistic 809.3113

Durbin-Watson stat 2.020812 Prob(F-statistic) 0.000000

Page 12: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 158 -

(P-Value) (.00) (0.0) (.791) (.721) (00) (0.0)

R2

Adjusted R2

S.D

DW

F Value

.047

.047

307.9105

1.332

95.316

.088

.088

185.8476

.124

186.216

.000

.000

623.0570

.314

.070

.000

.000

573.5035

.009

.128

.008

.007

700.2507

.370

14.904

.008

.007

634.4510

.015

15.483

Lagged BSE Return Coeff.

(P-Value)

3840.874*

(.00)

1730.858*

(.00)

32.303

(.972)

64.522

(.938) 2650.424*

(.012)

1657.326

(.071)

R2

Adjusted R2

S.D

DW

F Value

.037

.036

310.4841

1.311

74.040

.020

.019

192.6812

.131

38.841

.000

-.001

627.8015

.309

.001

.000

-.001

573.5216

.009

.006

.003

.003

702.7936

.374

6.263

.002

.001

636.4538

.016

3.264

BSE Return Volatility Coeff.

(P-Value)

-7197.466*

(.00)

-7280.704*

(.00)

114.349

(.956)

-968.965

(.607) -11708.2*

(.00)

-12662.7*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.025

.024

313.3609

1.293

48.918

.068

.067

188.3673

.100

138.869

.000

-.001

628.9148

.310

.003

.000

-.001

574.2313

.009

.265

.013

.013

700.3709

.376

25.504

.019

.019

631.557

013

37.368

NSE Return Coefficient

(P-Value)

972.094*

(.036)

1674.323*

(.00)

404.267

(.663)

146.562

(.861)

1275.874

(.219)

1517.470

(.103)

R2

Adjusted R2

S.D

DW

F Value

.002

.002

313.3609

1.293

43.918

.018

.017

192.8602

.124

35.191

.000

.000

627.6717

.309

.191

.000

-.001

573.5180

.009

.030

.001

.000

703.5229

.372

1.511

.001

.001

636.5544

.015

2.653

NSE_MA Coefficient

(P-Value)

11.41.690*

(.00)

9332.683*

(0.0)

728.060

(.756)

739.129

(.730) 9935.040*

(.00)

9163.094*

(0.0)

R2

Adjusted R2

S.D

DW

F Value

.046

.045

308.8388

1.338

91.829

.086

.085

186.1056

.123

180.357

.000

.000

623.0528

.314

.096

.000

.000

573.5047

.009

.120

.007

.007

700.3580

.370

14.205

.008

.007

634.5333

.015

14.978

Lagged NSE Return Coeff.

(P-Value)

3781.190*

(.00)

1786.314*

(.00)

-14.256

(.988)

99.831

(.905) 2443.094*

(.018)

1679.365

(.072)

R2

Adjusted R2

S.D

DW

F Value

.035

.034

310.7556

1.308

69.469

.020

.020

192.6176

.130

40.141

.000

-.001

627.8017

.309

.000

.000

-.001

573.5205

.009

.014

.003

.002

702.9189

.374

702.9189

.011

.010

634.3131

.013

20.405

NSE Return Volatility Coeff.

(P-Value)

-6881.910*

(.00)

-6968.443*

(.00)

2612.480

(.213)

1844.175

(.331) 8683.420*

(.00)

9464.073*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.022

.022

313.7556

1.102

1.416

.061

.061

189.0254

.099

124.571

.001

.000

628.6588

.009

1.550

.000

.000

574.1296

.009

.946

.007

.007

702.5264

.373

13.712

.011

.010

634.3131

.013

20.405

MSCI Return Coefficient

(P-Value)

-938.814

(.234)

577.221

(.234)

1522.864

(.335)

1509.085

(.291)

565.404

(.749)

1979.229

(.213)

R2

Adjusted R2

S.D

DW

F Value

.001

.000

316.1867

1.258

1.416

.001

.000

194.5401

.091

1.415

.000

.000

627.5502

.310

.932

.001

.000

573.3571

.010

1.113

.000

.000

703.8119

.370

.102

.001

.000

636.7358

.014

1.553

Lagged MSCI Return Coeff.

(P-Value)

345.173

(.662)

663.077

(.172)

1401.896

(.372)

1322.918

(.355)

2158.801

(.220)

1949.707

(.219)

R2

Adjusted R2

.000

.000

.000

.000

.000

.000

.000

.000

.001

.000

.001

.000

Page 13: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 159 -

S.D

DW

F Value

316.3679

1.259

.191

194.5172

.093

1.870

627.6710

.310

.797

573.3953

.010

.857

703.6670

.371

1.540

636.7431

.014

1.509

MSCI_MA Coefficient

(P-Value)

4119.270*

(.049)

2985.306*

(.021)

11735.28*

(.005)

10594.66*

(.005)

15343.75*

(.001)

13009.13*

(.002)

R2

Adjusted R2

S.D

DW

F Value

.002

.001

315.1231

1.276

3.891

.003

.002

194.3408

.92

5.377

.004

.004

621.7594

.316

8.039

.004

.004

572.3654

.04

7.807

.006

.005

700.7993

.368

10.812

.005

.004

635.4209

.014

9.551

MSCI Return Volatility Coeff.

(P-Value)

-9422.803*

(.00)

-9784.286*

(.00)

-61124.2*

(.00)

-60122.8*

(.00)

-72662.4*

(.00)

-72363.6*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.013

.013

315.1933

1.276

26.127

.038

.038

191.299

.367

310.716

.144

.144

581.7272

.37

32.379

.162

.162

645.2719

.450

368.198

.196

.196

571.7218

.024

468.065

.135

.135

514.369

.028

478.035

S & P 500 Return Coefficient

(P-Value)

-324.179

(.616)

745.502

(.061)

743.040

(.565)

667.639

(.569)

668.729

(.644)

1205.141

(.355)

R2

Adjusted R2

S.D

DW

F Value

.000

.000

316.2821

1.258

.251

.002

.001

194.4343

.093

3.516

.000

.000

627.6485

.310

.332

.000

.000

573.4744

.010

.324-

.000

.000

703.7913

.317

.214

.000

.000

636.8507

.013

.856

S & P_MA 500 Return Coeff.

(P-Value)

6361.99*

(.001)

4935.861*

(.00)

6692.181

(.073)

5825.854

(.088) 11424.82*

(.007)

9778.225*

(.010)

R2

Adjusted R2

S.D

DW

F Value

.001

.000

316.2972

1.262

1.055

.002

.002

194.3687

.096

4.821

.000

.000

627.7681

.309

.205

.000

.000

573.4886

.010

.229

.001

.000

703.7635

.371

.979

.000

.000

636.8364

.014

.940

Lagged S & P Return Coeff.

(P-Value)

663.920

(.305) 872.312*

(.028)

585.454

(.651)

580.571

(.633)

1433.312

(.323)

1262.233

(.332)

R2

Adjusted R2

S.D

DW

F Value

.006

.005

314.5018

1.272

11.519

.009

.009

193.6939

.095

18.318

.002

.001

622.5437

.315

3.217

.002

.001

573.0897

.010

2.915

.004

.003

701.6051

.368

7.381

.003

.002

635. 8936

.014

6.670

S & P Return Volatility Coeff.

(P-Value)

-11032.1*

(.00)

-11182.8*

(.00)

-62329.5*

(.00)

-61620.9*

(.00)

-76515.4*

(.00)

-76164.5*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.026

.026

313.1299

1.293

51.820

.071

.071

187.9698

.101

147.577

.214

.213

557.6749

.403

516.167

.250

.250

497.2324

.24

640.371

.256

.256

608.0437

.508

654.325

.310

.310

529.6443

.033

862.246

Exchange Rate Coefficient

(P-Value)

-14.699*

(.00)

-14.853*

(.00)

-81.839*

(.00)

-80.137*

(.00)

-100.593*

(.00)

-98.357*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.008

.007

315.0505

1.270

15.380

.021

.021

192.5305

.097

41.923

.062

.062

607.9289

.344

126.472

.071

.071

522.6991

.026

148.092

.075

.074

677.0991

.415

154.031

.087

.087

608.6305

.034

183.967

FBIR Coefficient

(3MTB) (P-Value)

-10.812*

(.009)

-10.402*

(.00)

67.579*

(.00)

65.643*

(.00)

57.980*

(.00)

57.155*

(.00)

R2

Adjusted R2

S.D

DW

.004

.003

315.7412

1.263

.009

.008

193.7645

.092

.035

.035

316.5424

.321

.040

.039

561.9872

.010

.021

.020

696.5312

.377

.024

.024

629.1493

.013

Page 14: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 160 -

F Value 6.874 16.899 69.884 80.002 40.305 48.392

IIP Coefficient

(P-Value)

2.144*

(.00)

2.247*

(.00)

16.928*

(.00)

16.819*

(.00)

19.550*

(.00)

19.621*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.037

.037

310.3917

1.310

75.166

.108

.108

183.7941

.102

234.110

.595

.595

399.3286

.772

2418.823

.698

.698

315.2068

.045

4458.098

.632

.632

427.2453

1.018

3279.459

.770

.770

305.4745

.071

6459.703

Lagged Investment Coefficient

(P-Value)

.445*

(.00)

.398*

(.00)

.160*

(.00)

.144*

(.00)

.515*

(.00)

.534*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.198

.197

283.6665

.225

474.911

.414

.414

149.0851

1.040

1361.752

.007

.006

626.3820

.313

12.570

.006

.006

572.3013

.081

12.137

.054

.054

685.3819

.529

108.953

.070

.069

614.9565

.106

144.236

Deferential Return 1 Coeff.

(BSE- S & P500) (P-Value)

895.189*

(.020)

928.172*

(.00)

-124.578

(.872)

-155.725

(.823)

655.134

(.449)

651.232

(.400)

R2

Adjusted R2

S.D

DW

F Value

.003

.002

315.8606

1.274

5.409

.008

.007

193.8328

.108

15.529

.000

-.001

627.6987

.309

.026

.000

.000

573.5157

.009

.050

.000

.000

703.5250

.370

.574

.000

.000

636.8752

.013

.708

Deferential Return 2 Coeff.

(BSE- MSCI) (P-Value)

1066.985*

(.007)

11117.062*

(.00)

-233.511

(.768)

-292.879

(.683)

806.782

(.564)

644.309

(.419)

R2

Adjusted R2

S.D

DW

F Value

.001

.000

318.0102

1.262

1.168

.011

.010

193.5464

.113

21.208

.000

.000

327.6888

.309

.087

.000

.000

573.4978

.010

.167

.000

.000

703.6792

.370

.823

.000

.000

636.8841

.013

.654

Beta_MSCI Coefficient

(P-Value)

1.210*

(.00)

1.359*

(.00)

1.495*

(.00)

1.461*

(.00)

1.670*

(.00)

1.679*

(.00)

R2

Adjusted R2

S.D

DW

F Value

.007

.007

317.3311

1.268

13.452

.023

.022

193.4266

.094

44.819

.275

.275

536.3765

.432

716.036

.308

.308

478.5412

.034

846.715

.273

.273

601.8127

.525

710.096

.330

.330

522.4681

.041

938.255

Beta_S&P 500 Coefficient

(P-Value)

2.996

(.280)

1.764

(.300)

9.414

(.087) 1.325*

(.008)

2.529*

(.00)

1.576*

(.005)

R2

Adjusted R2

S.D

DW

F Value

.001

.000

318.0102

1.262

1.168

.001

.000

195.1193

.093

1.076

.002

.001

628.8434

.319

2.941

.004

.003

574.355

.018

7.012

.009

.008

703.6537

.372

17.004

.004

.004

637.7951

.021

8.041

Note: 1) Figures in brackets are P values that show the significance of value of t- test.

2) * Significant at a level of 5 percent

3) DW is the Durbin-Watson Value

This table presents the values of regression coefficients, coefficients of

determination (R2) and adjusted R

2 values, Standard deviation and Durbin-Watson

value. The coefficient of determination (R2) indicates the percentage of the total

variance in the dependent variable, explained by the independent variables. It is

noteworthy from the table that the values of the coefficients of determination (R2)

resulting from the various regression equations are very low irrespective of the

Page 15: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 161 -

nature of the variable. The highest explanatory power amongst the various

independent variables, works out in case of lagged investment by FIIs (41.4%),

followed with wide difference, in down side, by Industrial Production Index (10.8%),

7 days moving average of BSE Sensex return (8.8%), moving average of NSE Nifty

return (8.6%), and volatility of BSE return (6.8%). Though the above-mentioned

values of R2

are obtained when moving average of net investments by FIIs (i.e.

FIIN_MA) is taken as the dependent variables for fitting the OLS equations, the

aforesaid phenomenon holds good in case of all of the dependent variables with a

few exceptions. Thus, the results make indication of the dependency of FII flows

primarily on past investments by FIIs, industrial growth in the host country, stock

market return and volatility.

Table 8.5 further shows that the regression coefficients concerning BSE

Return, BSE_MA, Lagged BSE Return, NSE Return, NSE_MA, Lagged NSE Return

and Differential Return 1 and 2 are positive and statistically significant at 5 percent

level of significance, insofar as the dependent variable FIIN and its seven days

moving average (FIIN_MA) are concerned. The next revelation of the table is that

both BSE_MA as well as NSE_MA have obtained positive and statistically

significant regression coefficient in case of the regression equations having FIIN,

FIIN_MA, FIIP and FIIP_MA as the dependent variables. Similarly, lagged BSE

Return and Lagged NSE Return have established positive relation with three

dependent variables namely FIIN, FIIN_MA and FIIP. Thus net inflows as well as

gross purchases of securities are dependent on both current and past information

regarding stock returns in Indian capital market.

Regarding the relationship between the return in the foreign stock markets and

FIIs flows to India, it may be seen from the table under reference that the return has a

negative relationship with FIIN, but positive relationship with rest of the dependent

variables. However, the regression coefficients pertaining to the two variables MSCI

return and S&P return are found insignificant irrespective of the dependent variable. It

is also noteworthy that seven days moving average series of the MSCI and S&P 500

is found having positive and significant values of regression coefficient in case of four

dependent variables FIIN, FIIN_MA, FIIP and FIIP_MA.

The volatility in return of both domestic as well as foreign stock markets

affects negatively the FII flows to Indian stock market. To be precise, volatility in

Page 16: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 162 -

return of BSE, NSE, MSCI and S&P 500 have obtained negative values of regression

coefficient against majority of the dependent variables. Moreover, these coefficients

are significant at 1 percent level. The effect of Beta, measures of co- movement of

domestic and foreign market in both cases (S&P and MSCI) have obtained positive

values. While the coefficient of BETA_MSCI turned significant at 1 percent against

each of the model specifications, BETA_S&P is found significant in case of only three

dependent variables namely, FIIS_MA, FIIP and FIIP_MA.

Regarding the effect of the various macro-economic factors taken as

determinants of FIIs, table shows that Exchange Rate of Indian Rupee with US Dollar

has a significant negative relationship with each and every dependent variable

representing foreign portfolio investment in India. In contrast, Federal Bank Interest

Rate has significant positive relationship with four of the dependent variables namely,

FIIS, FIIS_MA, FIIP and FIIP_MA. The Federal Interest Rate has a significant

negative relationship with FIIN and FIIN_MA. As expected, FII flows to India are

found having positive relationship with growth of Indian economy represented by

industrial production index.

Based on the above mentioned bi-variate analysis we established the following

multiple regression models:

FIIN = (Constant, BSE, BSE_MA, L_BSE, R_BSE, NSE, NSE_MA, L_NSE,

R_NSE, MSCI_MA, R_MSCI, S&P_MA, R_S&P, US_EX, FBIR, IIP, L_FIIN,

D_RET1, D_RET2, BETA_MSCI, Error term-------(8.6)

FIIN_MA = (Constant, BSE, BSE_MA, L_BSE, R_BSE, NSE, NSE_MA, L_NSE,

R_NSE, MSCI_MA, R_MSCI, S&P_MA, L_S&P, R_S&P, US_EX, FBIR, IIP,

L_FIIN, D_RET1,D_RET2, BETA_MSCI, Error term-------(8.7)

FIIS = (Constant, MSCI_MA, R_MSCI, R_S&P, US_EX, FBIR, IIP, L_FIIN,

BETA_MSCI, Error term-------(8.8)

FIIS_MA = (Constant, MSCI_MA, R_MSCI, R_S&P, US_EX, FBIR, IIP, L_FIIN,

BETA_MSCI, BETA_S&P Error term-------(8.9)

FIIP = (Constant, BSE_MA, L_BSE, R_BSE, NSE_MA, L_NSE, R_NSE,

MSCI_MA, R_MSCI, S&P_MA, R_S&P, US_EX, FBIR, IIP, L_FIIN, BETA_MSCI,

BETA_S&P Error term-------(8.10)

Page 17: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 163 -

FIIP_MA = (Constant, BSE_MA, R_BSE, NSE_MA, R_NSE, MSCI_MA, R_MSCI,

S&P_MA, R_S&P, US_EX, FBIR, IIP, L_FIIN, BETA_MSCI, BETA_S&P Error term----

(8.11)

The description of various equations is available in Chapter III, (Research

Methodology).

8.6 Results of Multiple Regression Model

It is clearly shown by Table 8.6 that the coefficients of determination (R2), and

adjusted value of R2

in case of stage I of the model vary from 0.24 in case of FIIN to

as high as 0.52 when the moving average of the purchase is taken as dependent

variable. It means the finally selected explanatory variables explain 24 to 52 percent

of the variation in various dependent variables. Hence, if taken jointly, these variables

are the important determinants of foreign portfolio investment on the Indian bourses.

What is interesting to note is the R2 and adjusted R

2 obtained at stage 2. Both of the

model (Table 8.7) indicates the extent of variation in dependent variables have got

substantially higher values in comparison to those at stage I of the model. At the stage

II, these indicators have obtained largest value (i.e. 0.988) in the case of the sixth

specification of the model and lowest value is 0.44 in case of first model. The values

of Durbin-Watson statistics pointed out absence of autocorrelation in various models

as they are close to 2.

Regarding the relationship between the returns offered by the Indian stock

market and FII flows, we might observe from the table that these variables are

positively associated. The regression coefficient of both NSE and BSE return are

having positive relationship with net investment by FIIs in India. The relationship is

judged significant at 10 percent level of significance. Similar to the study carried out

by Bose, Suchismita (2002) we found that Lagged BSE return as well as moving

average of NSE return are having positive relationship with net foreign portfolio

investments. And the same is found significant at 5 percent level. It implies that the

net foreign investment is significantly affected by the current and past information

regarding the Indian market return. Higher the return higher would be the net FII

inflows to India. Lagged BSE return is also positively correlated with gross purchases

by foreign institutional investors. In contrast 7 days moving average of BSE‟s return

Page 18: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 164 -

(BSE_MA) was seen negatively associated with gross purchase and its moving

average by FIIs at Indian bourses.

TABLE 8.6: REGRESSION COEFFICIENTS AND OTHER STATISTICS

RESULTS FROM MULTIVARIATE ANALYSIS

(First Stage)

Dependent

Variables

Independent

Variables(1)

FIIN

(2)

FIIN_MA

(3)

FIIS

(4)

FIIS_MA

(5)

FIIP

(6)

FIIP_MA

(7)

Constant 435.54*

(.014)

437.86*

(.000)

1504.10*

(.000)

2916.40*

(.000)

1687.09*

(.000)

1659.76*

(.000)

BSE Return 2542.41**

(.071)

1555.29*

(.000)

BSE_MA -4218.1**

(.073)

-3412.7**

(.057)

Lagged BSE

Return

2943.61*

(.000)

1636.881**

(.062)

BSE Return

Volatility

-3817.21*

(.000)

-51430.83*

(.000)

-52215.1*

(.000)

NSE Return 2451.33**

(086)

NSE_MA 2754.59*

(033)

3246.14*

(.000)

Lagged NSE

Return

670.909*

(.003)

NSE Return

Volatility

-3237.27*

(.001)

50472.83*

(.000)

50551.34*

(.000)

MSCI Return

Lagged MSCI

Return

MSCI_MA 9481.82*

(.007)

8280.548*

(.005)

MSCI Return

Volatility

3370.57*

(.040)

14575.2*

(.018)

24541.41*

(.000)

23427.81*

(.000)

24262.24*

(.000)

Page 19: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 165 -

S & P 500

Return

S & P_MA 500

Return

3023.461*

(.009)

14863.183*

(.000)

14003.21*

(.000)

Lagged S & P

500 Return

568.96**

(.052)

S & P 500

Return

Volatility

-4088.16*

(.009)

-5878.8*

(.000)

48564.5*

(.000)

-42127.3*

(.000)

-58974.68*

(.000)

-59783.7*

(.000)

Exchange Rate

(Rs V/s US $)

-6.396**

(.099)

-6.498*

(.002)

17.075*

(.016)

-47.107*

(.000)

-18.955*

(.017)

-18.178*

(.006)

Federal Bank

Interest Rate

(3MTB)

-88.366*

(.000)

Industrial

Production

Process

8.165*

(.007)

-10.740**

(.064)

Lagged

Investment by

FIIs

0218*

(.000)

353*

(.000)

-.09438*

(.003)

.310*

(.000)

.307*

(.000)

Deferential

Return 1

(BSE- S &

P500)

-680.98*

(.020)

Deferential

Return 2

(BSE- MSCI)

Beta_MSCI .000345*

(.034)

.0113*

(.000)

.01652*

(.000)

.01312*

(.000)

.01320*

(.000)

Beta_S&P 500 .000076**

(.059)

.0000183*

(.000)

.0000852*

(.028)

R-Squared .246 .495 .355 .442 .435 .522

Adjusted R-

Squared

.239 .490 .353 .440 .430 .519

S.E.of

Regression

277.0898 139.8983 503.3723 431.0069 533.2597 443.5306

Page 20: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 166 -

Durbin-Watson

Stat.

2.228 1.195 1.496 1.153 1.770 1.193

*Significant at 5%

** Significant as 10 %

Figure in parenthesis are p-value of the regression coefficients.

Regarding the relationship between the return in the foreign stock markets and

FII flows to India, it may be seen from the table under reference that the return in

emerging markets based on MSCI and S&P 500 indicators were excluded by

backward elimination alternative of stepwise model. The above also holds true about

lagged MSCI returns. The reason of their exclusion may be the multiple co- linearity

brought in by these variables. However, 7 days moving average of MSCI‟s return and

S&P 500 returns was included in the final regressors. The former (i.e. MSCI_MA)

amongs the above-motioned variables is positively associated with gross sales and

moving average of sales. The coefficients are found significant at 1 percent level of

significance. Moving average of S&P 500 return (S&P_MA) is having positive and

significant relationship with FIIN, FIIP and FIIP_MA. Hence, grouped information

about foreign return affects FIIs. Coondoo, Dipankor (2002) find more or less similar

results about the investing country market return and its moving average relation with

FIIN, FIIP and FIIS and their moving averages.

The volatility of stock returns at host country market is having found its

negative affect on 7 days moving average of net investments, gross purchases and

moving average of gross purchases. The above relationship is found significant at one

percent level. Interestingly the same holds true in case of S&P 500 return volatility.

However, MSCI return volatility is found positively associated with each of the

dependent variables except net investment by FIIs. It means more the volatility in the

emerging markets more will be the flows to foreign funds to India.

The analysis is further provided that the purchases, sales, moving average of

purchases and sales do have positive relation with Beta of MSCI as well as Beta of

S&P 500 return series. While the coefficient of BETA_MSCI turned significant at 5

percent against all the model speciation except the FIIN, BETA_S&P is found

significant only in case of only two dependent variables namely FIIP and FIIP_MA.

Page 21: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 167 -

The positive association of beta values with purchases by the foreign investors implies

that FIIs want to reap the benefits of portfolio diversification by switching over the

markets when risk profile of the markets get changed. It is found that the differential

return of BSE and S&P are found positively related with FIIN_MA. While past

information regarding the FIIs Investment is found having statistically significant

positive relationship with all the dependent variable except FIIS.

Regarding the macro-economic factors taken as determinants of FIIs, table

shows that exchange rate of Indian rupee with US dollar has a significant negative

relationship with each and every dependent variable except the FIIS. This is quite

contrast to the study given by Mukherjee, Paramita (2002), who found no significant

relation of exchange rate with any of the six dependent variables. The FIIS showed

positive association. In contrast the Federal Bank interest rate (FBIR) has significant

negative relation with FIIS_MA. As expected, FII flows to India are found

significantly related with growth of Indian economy represented by Index of

Industrial Production (IIP).

As stated previously, the multiple regression analysis was also conducted is

stage two where in the moving average of the basic variables was an additional

independent variable and for seven days moving average their one day lag was taken

as additional variable. The results of second stage model are given in Table 8.7.

It is obvious from the table that now the value R2 has risen considerably in

case of each of the six models under appreciation. While R2is found the lowest

(0.447) in case of FIIN, it was the highest in case of 0.988. Hence, the explanatory

power of the models have increased significantly. Similarly to stage one results,

lagged BSE Return, NSE‟s Return, FIIN_MA turned significance in case of model

one (i.e. FIIN as dependent variable). The regression coefficient of BSE‟s return,

NSE_MA, lagged FIIN_MA turned as significant. Hence, these are the determinants

of FIIN_MA. While FIIS is dependent only on FIIS_MA. MSCI‟s volatility and its

beta are found determinants of FIIS_MA. FIIP is caused by lagged BSE return,

Beta_S&P 500and FIIP_MA.

Page 22: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 168 -

TABLE 8.7: REGRESSION COEFFICIENTS AND OTHER STATISTICS

RESULTS FROM MULTIVARIATE ANALYSIS

(Second Stage)

Dependent

Variables

Independent

Variables(1)

FIIN

(2)

FIIN_MA

(3)

FIIS

(4)

FIIS_MA

(5)

FIIP

(6)

FIIP_MA

(7)

Constant -51.229

(.740)

-19.387

(.624)

16.55

(.913)

52.15

(.214)

8.433

(.966)

26.774

(.586)

BSE Return 361.268*

(.015)

BSE_MA 916.475*

(.001)

Lagged BSE

Return

2289.45*

(.000)

1011.434*

(.031)

NSE Return 2699.77*

(.025)

NSE_MA 1786.489*

(.000)

MSCI Return

Volatility

1265.818**

(.062)

S & P 500

Return

Volatility

-1451.304*

(.016)

Deferential

Return 1

(BSE- S &

P500)

-311.649*

(.009)

Beta_MSCI .000249*

(.024)

.000193*

(.048)

Beta_S&P 500 .000000971*

(.000)

Page 23: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 169 -

FIIN_MA 1.034*

(.000)

FIIS_MA 1.000*

(.000)

FIIP_MA 1.007*

(.000)

1 Lag

FIIN_MA

.927*

(.000)

1 Lag

FIIS_MA

.991*

(.000)

1 Lag

FIIP_MA

.985*

(.000)

R-Squared .447 .915 .852 .991 .832 .988

Adjusted R-

Squared

.444 .914 .852 .991 .832 .988

S.E.of

Regression

236.1591 57.2738 239.7555 55.6346 288.8921 71.0380

Durbin-Watson 2.263 1.659 2.167 1.710 2.162 2.122

*Significant at 5%

** Significant as 10 %

Figure in parenthesis are p-value of the regression coefficients

The first stage of daily FII flows indicated that these were stationary in nature

(i.e., contained no significant time trend but were auto-correlated). Except the first

equation this auto-correlation got reflected in all the regression equation estimated to

find out statistically significant covariates of various of the measures of FII flows.

This means none of the covariates-be it related to equity market performance or to the

performance of Indian economy could explain singly or jointly observed auto-

correlation of the FII flows. Use of lagged values of the concerned FII flow variable

as a regressor, however, removed the auto-correlation altogether. But inclusion of the

lagged value of FII flow variable in most of the cases caused the erstwhile significant

determinants to turn non-significant. As a result, the only statistically satisfactory

regression results turned out to be the ones having market return and lagged value of

the concerned FII flow. Such regression results would have economic explanation in

terms some kind of dynamic adjustment mechanism being involved in the

determination of current daily values a given FII flow. In other words, these results

Page 24: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

- 170 -

may be taken to mean that for individual FII flow there is a desired level determined

solely by BSE return or some variant of it and the actual value constantly tries to

reach this desired level.

8.7 Causal Relationship

After identifying the determinants of FIIs purchases, sales and net

investments and the moving average of their values, we made an attempt to

investigate whether there is any casual relationship between various dependent and

independent variables under reference of the study. For this objective Granger

Casualty Test was used. Table 8.8 shows the result of this test. The table indicates

that the hypothesis „FIIN does not Granger causes BSE returns‟ is accepted. In

contrast, its corresponding hypothesis „BSE return does not Granger causes FIIN‟

gets rejected at 5 percent level of significance. It refers that net foreign portfolio

investment are influenced by the stock market return of the host country. But the

vice-versa is not found true. The above result confirms to that found by Chakrabarti

(2001). Lagged return (one day lag) is also found causing net foreign investment

significantly. Thus the host country return is the strongest force that attracts foreign

institutional investors to India. The above result is also supported by the study

carried out by the Mukherjee, Paramita et al. (2002).

The table further reveals that seven days moving average of the NSE return

showed bi-directional causality with FIIN which implies that net investment is

attracted by current and past information about the return. Risk in the host country is

also caused by the foreign institutional investments. Further, risk in the investing

country (US Market) affects the quantum of foreign investment in host country.

Index of industrial production has a bi-directional causality with net foreign

institutional investments. It implies that the industrial growth of a country attracts the

foreign investments and later on these investments lead to further development in the

host country economy. Similar to the study of Bhattacharya and Mukherjee (2005),

this work found no causal relation between exchange rate and foreign portfolio

investments.

Page 25: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

171

TABLE 8.8: RESULTS OF GRANGER CAUSALITY TEST (with FIIN)

H0 / Lags 2 3 4 5 6 7

BSE does not Granger Cause FIIN

FIIN does not Granger Cause BSE

47.1021*

2.2314

31.5781*

2.46645

23.4463 *

1.45525

18.8483*

1.38726

15.8335*

1.074

13.5569*

1.26521

L_BSE does not Granger Cause FIIN

FIIN does not Granger Cause L_BSE

10.9746*

2.71584

8.58891*

2.86813*

6.44354*

1.83620

5.18584*

1.87128

4.32176*

1.67609

3.77838*

1.33846

NSE does not Granger Cause FIIN

FIIN does not Granger Cause NSE

45.8000*

2.83612

30.5540*

1.86114

22.9999*

0.92290

18.4615*

0.95504

15.4936*

0.80272

13.2155*

0.89740

NSE_MA does not Granger Cause FIIN

FIIN does not Granger Cause NSE_MA

22.6492*

5.78518*

16.0593*

6.16215*

12.2312*

5.79970*

12.4918*

5.58126*

10.6547*

4.69646*

9.59090*

4.97615*

R_NSE does not Granger Cause FIIN

FIIN does not Granger Cause R_NSE

6.17590*

7.63659*

2.36998

5.28720*

1.43693

4.13413*

1.16284

3.58659*

1.02523

3.21135*

0.86375

2.72261*

R_S&P does not Granger Cause FIIN

FIIN does not Granger Cause R_S&P

7.84588*

1.33067

3.41252*

2.54626

2.19745

2.85686*

1.86850

2.29138*

1.78131

1.93221*

1.38975

1.67961

S&P_MA does not Granger Cause FIIN

FIIN does not Granger Cause S&P_MA

1.67575*

2.55236*

1.69683*

2.69358*

1.76736*

2.15457*

1.39057*

1.70162*

1.73627*

1.44859*

1.62171*

1.42282*

US_EX does not Granger Cause FIIN

FIIN does not Granger Cause US_EX

1.92866

1.54418

0.88674

1.40103

0.63133

1.54481

0.94387

1.67121

0.83979

1.83839

0.71730

2.01256

IIP does not Granger Cause FIIN

FIIN does not Granger Cause IIP

12.5275*

3.76028*

6.01352*

3.27252*

4.21989*

2.88887*

3.17777*

2.60935*

2.81033*

2.27366*

2.13818*

2.15229*

L_FIIN does not Granger Cause FIIN

FIIN does not Granger Cause L_FIIN

21.9383*

6675.75*

2.53136*

5147.56*

3.48017*

4433.48*

1.58851*

3596.49*

3.49175*

3156.90*

4.58444*

3067.16*

F-Values are given in table * H0 Rejected at 5% Level of Significance

Page 26: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

172

After describing the results of Granger Casualty test with reference to the net

investment by the FIIs, let us now present the results of this test in case of gross sales.

Table 8.9 reveals that sale of securities by FIIs has turned as a cause of seven days

moving average of return of BSE as well as NSE. FIIS shows a bi-directional

relationship with interest rate of three months Treasury bills issued by Federal Bank

(FBIR) and Beta of MSCI with BSE, which implies the co-movement in both markets.

Sale has also been found as a cause of index of industrial production.

TABLE 8.9: RESULTS OF GRANGER CAUSALITY TEST (with FIIS)

H0 Value of F-Static at lag 2

BSE_MA does not Granger Cause FIIS

FIIS does not Granger Cause BSE_MA

0.16210

4.18233*

NSE_MA does not Granger Cause FIIS

FIIS does not Granger Cause NSE_MA

0.97739

3.28154*

R_NSE does not Granger Cause FIIS

FIIS does not Granger Cause R_NSE

1.70421

1.24028

US_EX does not Granger Cause FIIS

FIIS does not Granger Cause US_EX

2.84845

1.52264

FBIR does not Granger Cause FIIS

FIIS does not Granger Cause FBIR

4.03831*

5.32049*

IIP does not Granger Cause FIIS

FIIS does not Granger Cause IIP

67.1341*

1.26211

L_FIIN does not Granger Cause FIIS

FIIS does not Granger Cause L_FIIN

0.75498

0.69146

BETA_MSCI does not Granger Cause FIIS

FIIS does not Granger Cause BETA_MSCI

13.5005*

4.96754*

* Rejected at a significance level of 5 percent

Besides net investments and gross sales the bi-variate Granger Causality Test was

also applied to the gross purchase done by FIIs in the Indian stock market. The results of

the same are shown in table 8.10. It is clearly shown by the table that purchases done by

the foreign institutional investors in Indian stock market are having cause and effect

Page 27: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

173

relationship with Federal Bank Interest Rate (FBIR), Index of Industrial Production (IIP),

one day lagged investment made by the FIIs in Indian market and Beta of MSCI with

Indian Stock Market (BSE). But gross purchases by FIIs were found a cause of Beta of

S&P and exchange rate of Indian rupess v/s US dollar.

TABLE 8.10: RESULTS OF GRANGER CAUSALITY TEST (with FIIS)

H0 Value of F-Static at

lag 2

L_BSE does not Granger Cause FIIP

FIIP does not Granger Cause L_BSE

2.76832

0.96979

BSE_MA does not Granger Cause FIIP

FIIP does not Granger Cause BSE_MA

1.76902

1.02979

NSE_MA does not Granger Cause FIIP

FIIP does not Granger Cause NSE_MA

1.77779

0.48288

R_NSE does not Granger Cause FIIP

FIIP does not Granger Cause R_NSE

2.19904

1.43212

FBIR does not Granger Cause FIIP

FIIP does not Granger Cause FBIR

3.97360*

9.80895*

IIP does not Granger Cause FIIP

FIIP does not Granger Cause IIP

110.924*

4.49611*

L_FIIN does not Granger Cause FIIP

FIIP does not Granger Cause L_FIIN

52.5724*

359.707*

BETA_MSCI does not Granger Cause FIIP

FIIP does not Granger Cause BETA_MSCI

20.2460*

4.57112*

BETA_SP does not Granger Cause FIIP

FIIP does not Granger Cause BETA_SP

4.71766*

1.84593

US_EX does not Granger Cause FIIP

FIIP does not Granger Cause US_EX

3.31516*

2.36270

* Rejected at a significance level of 5 percent

Page 28: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

174

8.8 Conclusion

The results suggest that though FII flows to India are significantly affected by

return in the domestic equity market, the later is not influenced by variation in these

flows. It is also found that apart from the return in domestic market there are other

covariates of such flows. While the dependence of the net FII flows on daily return in the

domestic equity market- at a day‟s lag, to be more specific – is suggestive of foreign

investors return chasing behavior. Their decision seems to get affected also by the recent

history of the market return and its volatility in international and domestic market as well.

We also found that the set of factors affecting FII Sales and purchases were not the same.

It appeared that some factors would affect purchase or sale decision of foreign investors,

but not the corresponding net FII flows. For example, while FII purchase and sale to the

Indian stock market appeared to be sensitive to the volatility of the emerging market

returns (both purchase and sale responding positively to the risk in emerging markets),

but corresponding net inflows shows no association with that. One more interesting fact

we have found out that the foreign institutional investors reaped the benefits of portfolio

diversification by investing in Indian market as the betas of the BSE Sensex with respect

to the MSCI world and S&P 500 indices turned positively associated with FIIs purchase.

8.9 References:

Agarwal, R.N. (1997), “Foreign Portfolio Investment in Some Developing

Countries: A study of Determinants and Macroeconomic Impact”, Indian

Economic Review, 32, No. 2, pp. 217-229.

Aggarwal, R. (1981), “Exchange Rates and Stock Prices: A Study of The US

Capital Markets Under Floating Exchange Rates”, Akron Business and Economic

Review, Vol No. 12, pp 7-12.

Ahmed, K.M., Ashif, S. and Ahmed, A. (2005), “An Empirical Investigation of

FII‟s role in the Indian Equity Market”, ICFAI Journal of Applied Finance 11(8),

pp. 21-33.

Page 29: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

175

Ananthanarayanan, Sandhya, Krishnamurti, Chandrasekher and Sen, Nilanjan

(2003), “Foreign Institutional Investors and Security Returns: Evidence from

Indian Stock Exchanges”.

Chakrabarti, Rajesh (2001), “FII Flows to India: Nature and Causes”, Money and

Finance ICRA Bulletin 2, No. 7.

Gorden, Jmes and Gupta, Poonam (2003), “Portfolio Flows into India: Do

Domestic Fundamentals Matters?” IMF Working Paper Number WP/03/02.

Mukherjeee, Parmita and Bose, Suchismita and Coondoo, Dipankar (2002),

“Foreign Institutional Investment in the Indian Equity Market: An Analysis of

Daily Flows during January 1999 to May 2002”, Money and Finance ICRA

Bulletin, April –September, pp. 21-51.

Prasuna, C. Asha (2000), “Determinents of Foreign Institutional Investment in

India”, Finance India, January, pp. 411-4221.

Rai. K. and Bhanumurthy, N.R. (2003), “Determinents of Foreign Institutional

Investments in India: The Role of Return, Risk and Inflation”, JEL Classification:

E44, G15, G11.

Kumar, S.S.S. (2000), “Foreign Institutional Investment: Stabilizing or

Destabilizing?”, Abhigyan, pp. 23-27.

Soenen, L.A. and Aggarwal, R. (1989), “Financial Prices as Determinants of

Changes in Currency Values”, Paper Presented at the 25th

Annual Meeting of

Eastern Finance Association, Philadelphia.

Soenen, L.A. and Hennigar, E.S.(1989), “ An Analysis of Exchange Rates and

Stock Price: The US Experience Between 1980 and 1986”, Akron Business and

Economic Review, Vol No. 19, pp 71-76.

Venkateshwarlu, M and Tiwari, R. (2005), “ Causality Between Stock Prices and

Exchanges Rates: Evidence for India”.

Page 30: Determinants of FIIs in Indian Stock Marketietd.inflibnet.ac.in/bitstream/10603/4121/16/16_chapter 8.pdf · Conversely, a general downward movement of the stock market will motivate

176