Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Money
in the Indian Equity Market &
An Analysis of Daily Flows during January Finance
APRIL–SEPT. 2002
1999-May 2002
✝ ✝
PARAMITA MUKHERJEE , SUCHISMITA BOSE
✜
AND DIPANKOR COONDOO
Abstract
This paper explores the relationship of foreign institutional invest-
ment (FII) flows to the Indian equity market with its possible covariates based
on a daily data-set for the period January 1999 to May 2002. The set of
possible covariates considered comprises two types of variables. The first type This paper explores
includes variables reflecting daily market return and its volatility in domestic
and international equity markets as well as measures of co-movement of the relationship of
returns in these markets (viz., relevant betas). The second type of variables, on
the other hand, are essentially macroeconomic ones like exchange rate, short- foreign institutional
term interest rate and index of industrial production (IIP)—viz., variables that
are likely to affect foreign investors’ expectation about return in Indian equity investment (FII)
market.
It may be mentioned that our analysis has been primarily motivated
flows to the Indian
by the research done in this area by Chakrabarti (2001), results of which
equity market with
appeared in a recent issue of Money & Finance. Briefly, using a monthly data-
set Chakrabarti examined the nature and causes of FII net inflow into the its possible
Indian equity market during the period May 1993 to December 1999. He
obtained some interesting results: viz., (1) the FII net inflow is correlated with covariates based on
the return in Indian equity market and the former is more likely to be the
effect than the cause of the Indian equity market return; (2) so far as invest- a daily data set.
ment in Indian equity market is concerned, foreign investors do not seem to be
at an informational disadvantage compared to domestic investors; and (3) the
Asian crisis marked a regime shift in the sense that in the post-Asian crisis
period the return in the Indian equity market turned out to be the sole driver
of the FII inflow, whereas for the pre-Asian crisis period other covariates
reflecting return in other competing markets, urge for diversification etc., were
also found to be correlated with FII net inflow.
Malaysia, Pakistan, The Philippines, Sri Lanka, Taiwan and China in Asia, Czech
Republic, Greece, Hungary, Poland Russia, Slovakia, Turkey, Egypt, Israel and
22 Jordan in Europe and the Middle East, Morocco, South Africa and Zimbabwe in
Africa and Brazil, Chile, Colombia, Mexico, Peru and Venezuela in Latin America.
reversals of market performances make foreign equity investment ICRA BULLETIN
2 This is a close approximation, since part of the FII flow data contains a
compared to debt, can be gauged from the fact that, for the financial year 2001-02,
the average monthly sales by FIIs was to the tune of Rs. 3771 crore and purchases
Rs. 3102 crore in the equity segment, while for debt sales and purchases were a
meagre Rs. 392 crore and Rs. 332 crore, respectively. 23
ICRA BULLETIN in which we have tried to identify the relevant covariates of FII flow
into and out of the Indian equity market and also to determine the
Money nature of causality between the relevant variables. Since the use of
& daily data helps in determining the nature of causality with greater
Finance precision, our results of causality between FII flows and their major
covariates like domestic equity market return may prove to be useful
APRIL–SEPT..2002
from policy viewpoint. Next, going beyond Chakrabarti’s analysis, we
have carried out an in-depth statistical analysis to find out not only the
factors that affect net FII flow, but also those which affect FII sale and
purchase decisions.
The paper is organised as follows: In Section II we discuss
Since the use of
briefly about the importance of the and factors behind portfolio flows to
daily data helps in emerging markets, the nature of such flows during the past decade,
their suggested role in causing volatility and contagion in emerging
determining the markets. In the third section, we discuss how our analysis builds upon
Chakrabarti’s; we also describe our data-set and present the main
nature of causality results of our causality and regression analyses. The last section
concludes, relating our analysis to policy implications from empirical
with greater studies in various emerging markets, and the current policies being
followed in India with respect to foreign capital.
precision, our
II. International Portfolio Flows to Emerging Markets
results of causality As is well-known, international capital flows to emerging
markets is a somewhat recent phenomenon, which began at a reason-
between FII flows
able scale in the early 90s. On the theoretical side, the case for libe-
and their major ralisation of international capital flows is built around a few basic
tenets—viz., (a) free capital movements facilitate efficient allocation of
covariates like global savings, channelling resources to countries where they will be
most productive and thereby increasing growth and welfare globally;
domestic equity (b) access to foreign capital markets enable investors to achieve a
higher degree of portfolio diversification, thus allowing them to obtain
market return may higher returns at lower risk; (c) full convertibility for capital account
transactions complement the multilateral trading system which broad-
prove to be useful ens the channels through which countries obtain trade and investment
finance on much easier terms; and (d) liberalisation improves mac-
from policy
roeconomic performance as it subjects governments to greater market
viewpoint. discipline and penalises unsound monetary and fiscal policies. On the
practical side, on the other hand, the surge in international portfolio
investment over the past decade or so has been triggered by a number
of parallel developments. First, institutionalisation of savings in the
USA and the developed world since the 1980s placed a massive and
increasing volume of funds under the management of professional
portfolio managers,5 who for tactical reasons tend to prefer a widely
5 That is, the choice of pooled funds held by pension funds, life insurance
24 companies, mutual funds and investment trusts as repositories for the majority of
diversified portfolio spread out internationally. Second, there has been ICRA BULLETIN
savings—increased the share of funds invested in securities and enhanced the role of
institutional investors compared to that of depository institutions. In the United
States, for example, the share of total financial sector assets held by institutional
investors rose from 32% in 1978 to 52% in 1993, while the share of depository
institutions fell from 57% to 34% over the same period (Federal Reserve System
[FRS], Flow of Funds).
6 Whose local capital markets were proving to be a bottleneck in the
growth process.
7 Like the S&P/IFC Composite index which is used as a benchmark for
measuring equity markets returns in emerging markets as a whole or IFC’s Asia or Latin
America indices which reflects the corresponding returns in specific regional markets. 25
ICRA BULLETIN pattern: in the post-Mexican crisis period (to be more specific, during
1995-2001) the emerging equity markets, by and large, recorded much
Money smaller returns than those in the matured markets. Thus, whereas S&P
& 500 and NASDAQ recorded 14.77 and 13.62 per cent return, respec-
Finance tively, the corresponding figures for S&P/IFCI Composite, IFCI’s Asia,
EMEA (which covers emerging markets in Europe and the Middle East)
APRIL–SEPT..2002
and Latin America indices were –3.68, –9.81, 1.78 and 1.35, respec-
tively. Also, the risk associated with investment in the emerging equity
markets increased considerably during this period, whereas that for the
matured markets decreased. This was indicated by values of the Sharpe
ratio (which is the ratio of excess return of an asset over that of a risk
FII flows to the
free asset and the standard deviation of return of the asset concerned)
secondary equity recorded between –0.48 and –0.14 for the emerging markets as against
0.24 and 0.56 recorded for NASDAQ and S&P500, respectively. This
market do not have pattern, however, got reversed again during the first quarter of 2001-
2002, when the emerging markets (except those in Europe and the
any direct link with Middle East) recorded much higher return and lower risk compared to
those of S&P500 and NASDAQ.8
the level of real Let us next discuss briefly the possible effects of FII flow on the
recipient country’s economy. As such, FII flows to the secondary equity
investment in the market do not have any direct link with the level of real investment in
the economy. It is only by enhancing the efficiency and liquidity of
economy. It is only capital markets that such a flow can contribute to growth. Securities
markets in developing countries are typically both narrow and shallow.
by enhancing the
Therefore, FII participation may, a priori, induce considerable instabil-
efficiency and ity in these markets. The effect of such mobile capital flows can,
however, be quite complicated and therefore are highly controversial.
liquidity of capital In fact, country experiences differ considerably. Some studies found
clear evidence of benefits of such flow in the form of equity market
markets that such a development, capital market integration, lowering cost of capital, and
hence tend to question policy concerns regarding resource mobilisation,
flow can contribute market co-movements, contagion and volatility expressed by some
policy makers and academics to be largely unwarranted.9 The causes
to growth. of the instability and volatility of short-term portfolio capital flows to
emerging markets are often related to the way in which investment
funds are managed in order to confront uncertainty. It has been alleged
2000, analyse pooled cross-section and time series data for US equity flows to about
20 emerging countries, including India. Kim and Singal, 2000, analyse 20 countries
26 from the IFC EM Database.
that international portfolio investors seek liquidity and use ‘quick exit’ ICRA BULLETIN
the IFC) Emerging Markets Database (EMDB), which covers more than 2000 stocks
from 45 emerging markets. 27
ICRA BULLETIN thereby tracking returns in all possible markets. So far as the trading
behaviour of these investors is concerned, studies examining such
Money behaviour suggest broadly two types of trading behaviour (see Box 1
& for a brief description of these trading types). Further, for a given unit
Finance time interval (may be a time span of a few minutes, a day, a week or a
month), the investor’s actions may be aggregated and summarised into
APRIL–SEPT..2002
two basic measures—viz., sale and purchase—and a corresponding
overall measure of net sale or net purchase, as the case may be.
Following this logic, we have chosen to examine the nature of FII flow
to India in terms of three variables—viz., FII sale (henceforth denoted
by FIIS), FII purchase (henceforth denoted by FIIP) and net FII invest-
We have chosen to
ment (henceforth denoted by FIIN). It may be noted that in his analysis
use daily data, since Chakrabarti used monthly data on net FII flow as a proportion of the
size of market capitalisation (henceforth denoted by MCAP) of the
we felt that a set of previous month as the variable to be explained instead of the net FII
flow.14 As regards the data frequency, we have chosen to use daily data,
daily data should be since we felt that a set of daily data should be more appropriate for
examining the nature of causality, so far as the relationship between FII
more appropriate for flows and their major covariates are concerned. For the sake of com-
parison with Chakrabarti’s results, we have also considered FIIN as a
examining the proportion of MCAP (henceforth denoted by RATIO_FIIN). Similarly
defined ratio variables for FIIP (henceforth denoted by RATIO_FIIP)
nature of causality, and FIIS (henceforth denoted by RATIO_FIIS) have also been consi-
dered. However, since daily data on MCAP are unavailable, we have
so far as the
had to calculate a time series of daily MCAP using a method of appro-
relationship ximation.
As regards the choice of covariates of the FII variables, to the
between FII flows extent possible we have tried to incorporate into the analysis a set of
variables that appear, a priori, to be the primary determinants of
and their major global investors’ demand/supply for/of stocks in the Indian market. To
be precise, we have considered two different sets of variables—one
covariates are relating to the Indian and other equity markets which tend to compete
for global investors’ investment and the other relating to the Indian
concerned. economy which may be relevant for investors’ expectation formation
about the Indian market. The first set of variables include daily return
in Rupee terms in the Indian market and corresponding returns in US
take place on a day to day basis (or for that matter, even over shorter time inter-
vals), use of monthly aggregates may blur the picture considerably. Further, since the
ratio of net FII flow to the total size of market capitalisation (henceforth denoted as
Mcap) is rather small for India (and the monthly variation in this ratio is even
smaller), any analysis using this ratio as the explained variable may fail to capture
the true relationship between FII flow and its various covariates identified as
important explanatory factors for this flow in the literature as well as in
28 Chakrabarti’s own analysis.
ICRA BULLETIN
BOX 1
Money
Studies examining foreign institutional investors’ behaviour suggest broadly two
types of trading behaviour—viz., momentum (M) or positive-feedback trading(PT) &
and herding(H) strategies1, which may or may not be within the bounds of rational Finance
behaviour2. Briefly, strategy M is the tendency of an investor to buy and sell stocks
based on their observed return records—i.e., to buy recent winners and sell recent APRIL–SEPT. 2002
manner.
Dollar terms in competing markets, measures of volatility of these
returns (taken as proxy of the corresponding risks), daily return implied
by day to day variations in the Rupee-Dollar exchange rate. More
specifically, this set consists of the following variables:
1. Daily return in the Indian market calculated on the basis of
day to day variations in the value of BSE Sensex (henceforth
denoted by BSE_RET);15
2. Volatility of daily return in the Indian market calculated as the
standard deviation of previous 7/ 15/ 30 days’ daily returns
based on the BSE Sensex (henceforth denoted by BSE_RETVOL);
3. Daily return in the international equity market based on the
15 BSE Sensex is compiled using a set of 30 major shares and reports from
stock markets suggest that FIIs mostly restrict their trading to the shares covered in
the BSE Sensex. 29
ICRA BULLETIN day to day variations in the value of the MSCI World Index
(henceforth denoted by MSCI_RET);16
Money 4. Daily return in the US equity market based on the day to day
& variations in the S&P500 (henceforth denoted by S&P_RET);
Finance 5. Volatility of daily return in the international equity market
calculated as the standard deviation of previous 7/ 15/ 30 days’
APRIL–SEPT..2002
daily returns based on MSCI World index (henceforth denoted
by MSCIRET_VOL);
6. Volatility of daily return in the US equity market calculated as
the standard deviation of previous 7/ 15/ 30 days’ daily returns
based on the S&P500 index (henceforth denoted by
S&PRET_VOL);
7. Extent of co-movement of daily returns in Indian and Interna-
tional equity markets as measured by the beta of returns from
BSE Sensex and MSCI World Index (henceforth denoted by
BETA_MSCI);
The second set,
8. Extent of co-movement of daily returns in Indian and the US
includes two equity markets as measured by the beta of returns from BSE
Sensex and S&P500 Index (henceforth denoted by
macroeconomic BETA_MSCI); and finally
9. Daily return from day to day variations in the Rupee—USD
variables—viz., the exchange rate (henceforth denoted by EXCH).
index of industrial The second set, on the other hand, includes two macroeconomic
variables—viz., the index of industrial production (henceforth denoted
production and the by IIP) taken as a proxy for short run real income changes and the call
money rate (henceforth denoted by CMR) taken as a proxy for short-
call money rate. term interest rate. These two variables, taken to reflect the short run
variations in the fundamentals of the Indian economy, have been used
together with the equity market-related variables to see whether or not
global investors take into account their expectations about the state of
the Indian economy. The sample period of the daily data set is January
1999—May 2002, which wholly relates to the post- Asian crisis period.
For a fuller description of these variables and the sources of data on
them, see Box 2. See also Appendix 1 some general information on FII
investment activities as well as the Indian stock market and related
charts and Appendix 2 for a description of the method used to build up
the time series of daily MCAP.
The Results
In the present analysis we have mostly used appropriate linear
regression techniques to testify various hypotheses concerning FII flows
to the Indian equity market. For the hypotheses relating to the direction
International
S&P 500 The daily series for the S&P500 stock price index; Source: website
of Standard and Poors
MSCI WI The daily series for the MSCI World Index, a weighted stock
price index for all countries1; Source: website of Morgan Stanley
Capital International Inc; www.msci.com
Others
Exch Daily exchange rate of the Indian rupee vis a vis the US dollar
CMR Daily Call money rate; Source RBI website; www.rbi.org.in
IIP Index of Industrial production reported weekly; Source: RBI
Handbook of Statistics on Indian Economy, 2001 and CMIE
Monthly Review of the Indian Economy
Estimated series
Ratio_FIIN Ratio of FIIN to previous day’s Mcap2
Ratio_FIIP Ratio of FIIP to previous day’s Mcap
Ratio_FIIS Ratio of FIIS to previous day’s Mcap
FIIN_MAk(t-1) k-day moving average value of FIIN calculated using FIIN(t-k) to
FIIN(t-1), k=7,15,30.
FIIP_MAk(t-1) k-day moving average value of FIIP calculated using FIIP(t-k) to
FIIP(t-1), k=7,15,30.
FIIS_MAk(t-1) k-day moving average value of FIIS calculated using FIIS(t-k) to
FIIS(t-1), k=7,15,30.
BSE_Ret Daily returns on the BSE Sensex, calculated as the excess of the
logarithm of the index value on a date over the logarithm of the
index value on the previous day.
BSERET_MAk(t-2) k-day moving average value of return on BSE calculated using
BSE_Rte(t-k) to BSE_RET(t-2), k=7,15,30.
BSE_RetVol k(t-2) k-day moving average volatility of, returns on BSE Sensex or the
standard deviation of BSE_Ret(t-k) to BSE_Ret(t-2), k=7,15,30.
that FII activities of causation between FII flows and some of their covariates, we have
used the technique of pair-wise Granger Causality test. The results are
exert a strong
presented below.
demonstration effect
A. Direction of Causation between FII flows and Return in the
and thus drive the Indian Stock Market
For any type of investor, domestic or foreign, market return
domestic stock is generally the prime driver of equity investments. However, when
it comes to the case of foreign investment in a thin equity market like
market. that of India, there is a prevalent feeling that FII activities exert
a strong demonstration effect and thus drive the domestic stock
market.
In other words, some believe that the day to day FII trading in
Indian market, rather than being influenced by the market return,
induces the daily market return to be what it is. As a starting point we
examine the nature of pair-wise causality between daily measures of FII
investment and corresponding BSE_RET separately for the three FII
variables (viz., FIIP, FIIS and FIIN). The results are given in Table 1. It
may be mentioned here that Chakrabarti (2001) also examined the
32 nature of causality between FIIN and BSE_RET mostly on the basis of
monthly data on these variables.17 However, given the facts that ICRA BULLETIN
TABLE 1
Pairwise Granger Causality Tests between BSE Return and FII
Lags —> 2 3 4 5 6 7
PANEL 1
BSE Return does not
Granger Cause FIIN 25.98* 17.33* 15.95* 13.53* 11.72* 10.21*
FIIN does not Granger
That equity market
Cause BSE Return 1.90 1.41 2.87 2.02 1.45 1.42
responses are
PANEL 2
BSE Return does not
Granger Cause FIIP 4.32+ 4.42* 5.67* 5.75* 4.70* 4.18*
typically extremely
FIIP does not Granger
Cause BSE Return 1.44 0.94 4.39* 3.33 3.19* 2.99* quick and of very
PANEL 3
short run nature,
BSE Return does not
Granger Cause FIIS 8.69* 6.18* 5.55* 4.84* 4.34* 3.69*
FIIS does not Granger
examination of
Cause BSE Return 7.37* 4.94* 3.62 2.99 3.03 2.69
causality based on a
PANEL 4
BSE_RET does not Granger
Cause RATIOFIIN 28.91* 19.09* 17.27* 14.11* 12.17* 10.49*
monthly data set
RATIOFIIN does not
Granger Cause BSE_RET 2.89 2.24 3.41 2.41 1.84 1.71 may fail to capture
PANEL 5
the exact nature of
BSE Return does not Granger
Cause RATIOFIIP 3.04+ 2.80+ 4.25* 3.97* 3.35* 2.83*
RATIOFIIP does not Granger
causality.
Cause BSE Return 0.96 0.60 4.02* 3.07 2.96 2.67
PANEL 6
BSE Return does not
Granger Cause RATIOFIIS 18.16* 11.37* 8.50* 6.86* 5.76* 4.88*
RATIOFIIS does not
Granger Cause BSE Return 6.22* 4.20 2.98 2.74 2.80 2.31
*’ Denotes rejection at 1% level of significance
+’ Denotes rejection at 5% level of significance
17 He has done the causality test taking the daily data for a limited period
of one year (1999, in the post-Asian crisis period), most of his analysis of causality
for the pre-crisis period was based on monthly data (1993 to 1999). 33
ICRA BULLETIN BSE_RET to FIIN and not the other way. This is true for FIIS and FIIP
and their ratios also, barring a few cases where no definite conclusion
Money can be drawn. For these variables, causality, if found statistically
& significant, is unidirectional running from BSE_RET to FII flows. 18 This
Finance lends further credence to the supposition that FII flows to India are
mostly in response to contemporaneous returns in the Indian stock
APRIL–SEPT..2002
markets (in the post Asian crisis period) rather than FII inflows and
outflows being the cause of returns in the national markets.
TABLE 2.2
Regression of FII_MA on BSE Return_MA
Regressands ——>
TABLE 3A
Stage 1 Regression results of FII on all relevant variables
REGRESSION A: Stage 1
Regression Coefficients FIIN Ratio FIIN_MA FIIP Ratio FIIP_MA FIIS Ratio FIIS_MA
FIIN 15(t-1) FIIP 15(t-1) FIIS 15(t-1)
Constant -29.06 -0.001 -9.84 137.05 0.01 136.47 166.10 0.01 146.30
(0.14) (0.12) (0.58) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0)
Lagged BSE Return 1285.21 0.04 -15.80 408.05 0.01 -437.76 -877.39 -0.03 -422.01
(0.0) (0.0) (0.87) (0.18) (0.18) (0.016) (0.0005) (0.0) (0.0095)
BSE Return 15-day
MA(t-2) 776.92 0.02 3426.23 1455.58 -0.01 433.96 677.97 -0.03 -2993.03
(0.52) (0.59) (0.0004) (0.44) (0.78) (0.79) (0.71) (0.41) (0.05)
BSE Return 15-day
Volatility(t-2) 587.96 0.02 2134.64 3164.84 0.03 4783.72 2577.51 0.01 2649.34
(0.49) (0.48) (0.004) (0.03) (0.35) (0.0003) (0.04) (0.67) (0.0099)
Lagged MSCI Return 393.554 -0.002 -260.48 -1237.87 -0.05 -922.41 -1630.71 -0.05 -662.13
(0.73) (0.96) (0.53) (0.33) (0.14) (0.28) (0.15) (0.07) (0.299)
MSCI Return 15-day
MA(t-2) -16601.52 -0.46 -2105.64 -16763.99 -0.65 -10479.46 -160.46 -0.19 -8372.86
(0.017) (0.03) (0.66) (0.19) (0.018) (0.35) (0.99) (0.39) (0.36)
MSCI Return 15-day
Volatility(t-2) -7446.42 -0.18 -592.82 -7372.09 -0.01 -2060.30 72.42 0.17 -1470.89
(0.07) (0.14) (0.86) (0.297) (0.94) (0.73) (0.989) (0.15) (0.76)
Lagged S&P Return -150.21 0.01 45.69 621.42 0.03 403.22 771.12 0.02 357.56
(0.85) (0.69) (0.86) (0.51) (0.18) (0.51) (0.33) (0.19) (0.46)
S&P Return 15-day
MA(t-2) 17196.60 0.48 2571.63 16217.83 0.60 10508.50 -981.58 0.11 7935.17
(0.009) (0.0096) (0.57) (0.19) (0.02) (0.35) (0.92) (0.57) (0.38)
S&P Return 15-day
Volatility(t-2) 9631.50 0.27 970.09 7840.28 0.11 1623.88 -1790.15 -0.16 656.41
(0.008) (0.02) (0.72) (0.20) (0.49) (0.75) (0.68) (0.12) (0.87)
Beta_MSCI 15-day (t-2) 7.53 0.00 2.06 -5.45 0.0001 -15.80 -12.98 0.00 -17.86
(0.697) (0.66) (0.87) (0.87) (0.94) (0.58) (0.68) (0.76) (0.44)
Beta (S&P) 15-day (t-2) 11.03 0.00 14.88 15.42 0.001 48.63 4.39 0.0003 33.74
(0.69) (0.78) (0.47) (0.70) (0.61) (0.13) (0.91) (0.77) (0.23)
R-squared 0.09 0.10 0.15 0.05 0.05 0.18 0.05 0.07 0.18
Adjusted R-squared 0.08 0.08 0.14 0.04 0.03 0.17 0.04 0.06 0.17
S.E. of regression 115.99 0.003 49.25 138.57 0.00 84.32 110.59 0.003 70.36
Durbin-Watson stat 1.73 1.70 0.06 1.19 1.48 0.05 0.99 1.24 0.04
moving average period of 7, 15 and 30 days, respectively. This was done essentially
36 to see whether or not the regression results would be sensitive to the choice of the
period of moving average.
BSE_RET along with the set of non-domestic return determinants listed ICRA BULLETIN
above. This analysis was done in several stages.21 Here let us first
summarise the qualitative results obtained on the nature of relationship
Money
of different types of FII flows with the variables that we identified to be &
possible covariates of these flows. (We present here only some repre- Finance
sentative results in Tables 3A and 3B.)
APRIL–SEPT. 2002
TABLE 3B
Stage 2 Regression Results of FII on all relevant variables
REGRESSION A: Stage 2
Regression Co-efficients FIIN Ratio FIIN_MA FIIP Ratio FIIP_MA FIIS Ratio FIIS_MA
FIIN 15(t-1) FIIP 15(t-1) FIIS 15(t-1)
Constant -17.21 -0.001 -0.28 69.84 0.005 2.40 57.25 0.003 1.15
(0.27) (0.19) (0.79) (0.0) (0.0) (0.07) (0.0) (0.0) (0.23)
Lagged BSE Return 1252.26 0.04 -9.43 -694.50 -0.03 -14.78
(0.0) (0.0) (0.65) (0.0007) (0.0) (0.37)
BSE Return 15-day MA(t-2) 159.07 187.86
(0.13) (0.003)
BSE Return 15-day Volatility(t-2) 74.92 1132.59 -31.30 324.36 21.04
(0.25) (0.04) (0.67) (0.48) (0.67)
Lagged MSCI Return -0.01
(0.18)
MSCI Return 15-day MA(t-2) -14793.27 -0.37 -0.50
(0.02) (0.04) (0.01)
MSCI Return 15-day Volatility(t-2) -5046.68
(0.13)
S&P Return 15-day MA(t-2) 14899.95 0.39 0.43
(0.01) (0.02) (0.03)
S&P Return 15-day Volatility(t-2) 7411.93 0.12
(0.01) (0.002)
Other Regressors Lag1 Lag1 Lag1 Lag1 Lag1 Lag1 Lag1 Lag1 Lag1
FIIN Ratio FIIN_ FIIP Ratio FIIP_ FIIS Ratio FIIS_
FIIN MA15 FIIP MA15 FIIS MA15
(Lagged Variables) 0.17 0.19 1.18 0.296 0.29 1.23 0.36 0.32 1.27
Lag2 Lag2 Lag2 Lag2 Lag2 Lag2
FIIN_ FIIP FIIP_ FIIS Ratio FIIS_
MA15 MA15 FIIS MA15
-0.21 0.302 -0.24 0.31 0.22 -0.10
Lag3
FIIS_
MA15
-0.18
R-squared 0.11 0.12 0.96 0.27 0.10 0.99 0.37 0.26 0.99
Adjusted R-squared 0.11 0.11 0.96 0.27 0.10 0.99 0.37 0.25 0.99
S.E. of regression 114.26 0.00 10.56 120.77 0.004 10.30 89.83 0.003 7.80
Durbin-Watson stat 2.10 2.12 2.04 2.11 2.13 2.11 2.08 2.06 2.05
Note: 1. Figures in brackets are the p-values of coefficients
2. P-values for all the lagged dependent variables are less than or equal to zero.
3. Since the package we used doesn’t provide Durbin h-stats, we report the DW here.
21
First, we have checked for stationarity of all the variables concerned.
Then, we have specified the regressions keeping in view to the order of integration of
the variables concerned. In all cases the regressions have been done in two stages,
detecting residual auto-correlation at the first stage, we have run the regressions
again. In stage 2, we included only the significant variables at stage 1 and the
lagged dependent variable as the regressors. 37
ICRA BULLETIN C1. FII Flows and BSE Return
As regards the relation between daily FII net inflow and BSE
Money return, a definite positive association with the previous day’s return was
& observed in all the regressions estimated. Similarly, for daily FII sale
Finance the association with the previous day’s return was always found to be
negative and significant. However, no definite association between
APRIL–SEPT..2002
daily FII purchase and previous day’s market return was noticed. One
might surmise that daily FII activities might be based, not just on
previous day’s return information, but on the history of return over the
previous few days. Thus, for example, daily FII sale may be thought to
depend, not just on previous day’s return but also on the average and
the standard deviation of returns observed during the previous 7 or 15
days (the standard deviation being taken as a measure of volatility of
returns). Ceteris paribus, FII sale/purchase should be positively/nega-
tively related to average return and negatively/positively related to
volatility. We, however, failed to observe any definite association with
One might surmise
average return or return volatility of any of the daily FII flow variable.
that daily FII Very similar results were obtained when the FII variables were used in
ratio form. But use of moving average variables—i.e., FIIN_MA,
activities might be FIIP_MA and FIIS_MA—showed definite positive association of FII net
inflow, FII purchase and FII sale with return and volatility, but not with
based, not just on average return. While a positive association of return volatility with FII
sale seems intuitively justifiable, its negative association with the other
previous day’s two FII flows appears somewhat unrealistic.
return information, C2. FII Flows and Return and Volatility in Competing markets
As already mentioned, in our regression strategy, we included
but on the history of
the return and volatility variables relating to the S&P500 index and
return over the MSCI index (viz., MSCI_RET, S&P_RET, MSCI_RETVOL and
S&P_RETVOL) in the set of possible regressors, as indicators of return
previous few days. and risk involved in investing in the US and the World equity markets,
other than the Indian one (see Box 3 in this context). The hypothesis of
portfolio return maximisation suggests reallocation of investment in
favour of markets in which risk is minimised (as measured by Sharpe
ratio—i.e., excess return per unit of volatility) and this in turn means,
ceteris paribus, a positive/negative partial association with competing
market returns/volatility of FII net inflow and FII purchase and a nega-
tive/positive association with FII sale. In our exercise, we did get result
supporting the hypothesis of return maximisation only in some cases,
not unequivocally. This is in conformity with the result that Chakrabarti
(2001) obtained for the post-Asian crisis sub-period using monthly data.
We also found in most of the cases that S&P variables and
MSCI variables are simultaneously significant. We checked the correla-
tion between these variables which, turned out to be quite high
(Table 4) implying thereby the possibility of multicollinearity, given the
extent of integration of the world and US markets. We therefore ran
38 separate regressions for S&P variables and MSCI variables. The results
ICRA BULLETIN
BOX 3
Money
We have also tried to see to what extent the performance of matured equity markets
affect that of the Indian equity market and therefore we examined the possible inter- &
dependence of BSE and S&P returns. To see whether or not S&P return exerts a Finance
significant causal effect on the BSE return we carry out a Granger causality test for the
APRIL–SEPT. 2002
same the results are as below.
There is clear evidence of S&P return causing BSE return unilaterally. This was to be
expected and perhaps suggestive of the fact that a fluctuation in the US (matured)
equity market would get transmitted to the Indian market.
(Table 5) showed that while only the MSCI return volatility but not the
average return itself was significant, in more than one cases both S&P
return and its volatility were significant. See, however, Table 3A which
suggests that whereas FII net inflow is affected by return and volatility
of both MSCI and S&P, the corresponding purchase and sale flows are
mostly not affected by these.
TABLE 4
Correlation
BSE_RETLAG1 1.00 -0.03 0.01 0.01 -0.02 0.14 -0.02 0.13 0.07 0.01 0.14
BSE_RETVOL15(t-2) 1.00 -0.32 -0.06 0.06 0.05 0.50 -0.05 0.05 0.50 -0.02
BSERET_MA15(t-2) 1.00 -0.05 -0.16 -0.03 -0.31 0.42 -0.03 -0.30 0.36
BETA_MSCI15(t-2) 1.00 0.86 -0.02 0.07 -0.15 -0.02 -0.06 -0.16
BETA_SP15(t-2) 1.00 -0.02 0.11 -0.24 -0.01 -0.01 -0.23
MSCI_RETLAG1 1.00 0.09 0.04 0.89 0.11 0.05
MSCI_RETVOL15(t-2) 1.00 -0.12 0.08 0.88 -0.10
MSCIRET_MA15(t-2) 1.00 0.01 -0.08 0.95
SP_RETLAG1 1.00 0.10 -0.01
SP_RETVOL15(t-2) 1.00 -0.07
SPRET_MA15(t-2) 1.00
39
ICRA BULLETIN C3. FII Flows and Betas of the Indian Equity Market
Following the logic of hedging and diversification of portfolio,
Money one would expect a significant negative (positive) partial association of
& the betas for Indian market with respect to S&P500 and MSCI indices
Finance (i.e., BETA_S&P and BETA_MSCI) with FII net inflow and sale (FII
purchase), in case the hypothesis that FII flow to India essentially serve
APRIL–SEPT..2002
as a means of diversification is true. Our results failed to find support
for this hypothesis as in most cases these variables turned out to be
statistically non-significant.22
TABLE 5
Regression of FII on MSCI and S&P separetely
22
Initially both the beta variables were included together in the regression
equations on the presumption that they would be independent, as the concerned
markets were very different. However, the sample correlation of these variables was
found to be very high, implying thereby a high degree of integration of the markets.
We next tried to see the effects of these variables on FII flows separately. Even in
40 that case the effects of these variables were found statistically non-significant.
and CMR as indicators of the real economic activity in India and tried ICRA BULLETIN
REGRESSION A
Stage 1
Coefficients FIIN_MA FIIP_MA FIIS_MA
15(t-2) 15(t-2) 15(t-2)
Constant -237.05 -558.58 -321.47
(0.01) (0.0) (0.0)
Lagged BSE Return -32.25 -503.87 -471.69
(0.78) (0.02) (0.00)
BSE Return Weekly MA(t-2) 1172.53 -2078.30 -3251.30 Some evidence of
(0.07) (0.04) (0.0)
IIP 1.60 4.20 2.60 positive association
(0.0) (0.0) (0.0)
CMR 1.52 12.00 10.48 of both IIP and CMR
(0.42) (0.0) (0.0)
& REGRESSION B
Stage 1
Finance Coefficients FIIN_MA FIIP_MA FIIS_MA
APRIL–SEPT..2002 15(t-2) 15(t-2) 15(t-2)
Constant -263.9 -711.3 -447.3
(0.01) (0.0) (0.0)
Lagged BSE Return -20.2 -461.2 -441.1
(0.87) (0.03) (0.005)
BSE Return Weekly MA(t-2) 1199.1 -1964.1 -3163.6
(0.07) (0.07) (0.0)
IIP Weekly MA(t-2) 1.7 4.8 3.1
(0.0) (0.0) (0.0)
Lagged CMR -0.1 4.9 5.0
(0.95) (0.15) (0.06)
CMR Weekly MA(t-2) 3.5 14.5 11.0
Such regression (0.34) (0.01) (0.007)
R-squared 0.10 0.26 0.26
results would have Adjusted R-squared 0.09 0.25 0.26
S.E. of regression 50.45 80.11 66.58
economic Durbin-Watson stat 0.05 0.04 0.05
Stage 2
explanation in terms
Coefficients FIIN_MA FIIP_MA FIIS_MA
15(t-2) 15(t-2) 15(t-2)
some kind of
Constant -4.13 -3.22 -6.30
(0.57) (0.75) (0.38)
dynamic adjustment Lagged BSE Return -9.82 -24.71
(0.64) (0.16)
mechanism being BSE Return Weekly MA(t-2) 219.08 238.76 27.57
(0.0) (0.0) (0.43)
involved in the IIP Weekly MA(t-2) 0.03 0.02 0.04
(0.48) (0.76) (0.37)
Lagged CMR 0.02
determination of (0.95)
CMR Weekly MA(t-2) 0.24 0.32
current daily value a (0.41) (0.22)
Other Regressors LAG1FII LAG1FII LAG1FII
given FII flow. N_15MA P_15MA S_15MA
(Lagged Variables) 1.16 1.20 1.36
LAG2FII LAG2FII LAG2FII
N_15MA P_15MA S_15MA
-0.19 -0.21 -0.37
R-squared 0.96 0.99 0.99
Adjusted R-squared 0.96 0.99 0.99
S.E. of regression 10.46 10.15 7.99
Durbin-Watson stat 2.02 2.08 2.15
Note: 1. Figures in brackets are the p-values of coefficients
2. P-values for all the lagged dependent variables are less than or equal to
zero.
3. Since the package we used doesn’t provide Durbin h-stats, we report the
42 DW here.
evidence of positive association of both IIP and CMR with FIIN was ICRA BULLETIN
23 But one cannot possibly avoid this auto-correlation, specially when the
domestic investor
flows between 14 countries, during1989-96 find that asset flows depend on market
size in both source and destination country as well as trading costs, in which both
information and the transaction technology play a role. Portes and Rey, 2000.
Garibaldi et al (2002) analysing capital flows to 25 European transition economies
showed that FPI was volatile and concentrated in a handful of countries (notably
Russia). Regressions showed that the presence of a financial market infrastructure
and property rights indicator were the only significant explanatory variables for FPI.
Claessens et al (2002) analysing data from 77 countries, find that factors such as
shareholder protection and the quality of local legal systems which make it and
easier for investors to buy shares and firms to list in public markets play a prime role
in determining the degree of integration with international capital markets.
26 Liljeblom and Löflund (2000), using company specific data on degree of
foreign ownership, in the Finnish market, which recently abolished capital controls.
27 Classens et al (2002).
28 The present stance on capital controls of Indian policy makers is in
place; policy makers in India have justifiably treated capital account liberalisation
as a process and not an event and reiterated that capital account liberalisation and
reform of the financial system should move in tandem (Rangarajan and Prasad,
1999; Rangarajan, 2000; Jalan, 2002).
29 To take care of malpractice which discourage stock market participation
by a majority of savers in our country (see NSE, 2001, in this context). A survey by
SEBI and NCAER showed that alleged malpractice like insider trading and low
confidence in brokers/sub-brokers, company management/auditors were the main
causes behind lack of domestic savers’ confidence in the equity markets. Further
regulatory authorities would need to look into alleged restrictive practices by FIIs
like price rigging as suggested by Samal (1997). 45
ICRA BULLETIN References
Bae, Kee-Hong, Kalok Chan and Angela Ng (2002), “Investability and Return
Money Volatility in Emerging Equity Markets”, Presented to the International
47
ICRA BULLETIN Appendix 1: Some Information Relating to FII Operations
in India
Money • FIIs in India include Asset Management Companies, Pension
& Funds, Mutual Funds, Investment Trusts as Nominee Compa-
Finance nies, Incorporated/Institutional Portfolio Managers or their
Power of Attorney holders, University Funds, Endowment
APRIL–SEPT..2002
Foundations, Charitable Trusts and Charitable Societies.
• SEBI acts as the nodal point in the entire process of FII registra-
tion. RBI approval under FEMA enables an FII to buy/sell
securities on Stock Exchanges and open foreign currency and
Indian Rupee accounts with a designated bank branch. Invest-
The presence of ment by FIIs in India is regulated under SEBI (FII) Regulations ,
1995 and Regulation 5(2) of FEMA.
Sectoral Cap/ • FIIs are required to allocate their investment between equity and
debt instruments in the ratio of 70:30. (However, it is also possi-
Statutory Ceiling ble for an FII to declare itself a 100 per cent debt FII in which
case it can make its entire investment in debt instruments).
means that foreign
• FIIs can buy/sell securities on Stock Exchanges. They can also
investment from all invest in listed and unlisted securities outside Stock Exchanges,
where the price has been approved by RBI.
sources cannot • No individual FII/sub-account can acquire more than 10 per
cent of the paid-up capital of an Indian company. All FIIs and
exceed a specified their sub-accounts taken together cannot acquire more than 24
per cent of the paid-up capital of an Indian Company. Indian
level. A Company to Companies can raise the above-mentioned 24 per cent ceiling
to the Sectoral Cap/Statutory Ceiling as applicable by passing
which no sectoral a resolution by its Board of Directors followed by passing a
Special Resolution to that effect by its General Body. The
cap/statutory ceiling presence of Sectoral Cap/Statutory Ceiling means that foreign
investment from all sources cannot exceed a specified level. A
is applicable can Company to which no sectoral cap/statutory ceiling is applica-
ble can raise the limit of permissible FII investment to 100 per
raise the limit of cent of the paid-up capital. A Company to which a 49 per cent
Sectoral Cap is applicable can raise the limit of permissible FII
permissible FII investment to 49 per cent and if there is an existing foreign
direct investment of 15 per cent, possible FII investment can
investment to 100%
only be up to 34 per cent.
• No permission from RBI is needed so long as the FIIs purchase
of the paid-up
and sell on recognised stock exchange. However, all non-stock
capital. exchange sales/purchases require RBI permission.
• FIIs can avail of the Forward Cover Facility from the Autho-
rised Dealer subject to certain conditions.
• High Net Worth Individuals /foreign corporates can invest
through SEBI Registered FIIs subject to a sub-limit of 5 per cent
each, within the aggregated limit of 24 per cent.
• FIIs can trade in Exchange Traded Derivative Contracts.
31-May-02
1-Jan-99
29-Mar-00
16-Mar-01
4-Mar-02
28-Sep-99
21-Sep-00
7-Sep-01
29-Dec-99
18-Dec-00
5-Jul-99
27-Jun-00
13-Jun-01
6-Dec-01
6-Apr-99
-5000 -0.3
CHART A2.1
Net FII and the Sensex, for 3 sub periods of the sample.
100 4500
0
4000
-100
-200 3500
-300 3000
14-May-99
1-Jan-99
21-Mar-00
5-Mar-99
18-Jan-00
16-Nov-99
9-Sep-99
16-Dec-99
14-Jun-99
11-Aug-99
13-Jul-99
20-Apr-00
9-Apr-99
17-Feb-00
3-Feb-99
12-Oct-99
49
50
ICRA BULLETIN
Money
APRIL–SEPT..2002
&
Finance
FII net FII net
-150
-100
1000
1200
-600
-400
-200
100
150
200
250
300
350
-50
200
400
600
800
50
0
0
2-May-01 2-May-00
28-May-01 25-May-00
21-Jun-01 19-Jun-00
17-Jul-01
FIIN
FIIN
12-Jul-00
10-Aug-01 4-Aug-00
7-Sep-01 30-Aug-00
4-Oct-01
CHART A2.3
CHART A2.2
25-Sep-00
31-Oct-01
19-Oct-00
27-Nov-01
14-Nov-00
26-Dec-01
Sensex (Close)
Sensex (Close)
7-Dec-00
21-Jan-02
2-Jan-01
14-Feb-02
25-Jan-01
12-Mar-02
20-Feb-01
10-Apr-02
19-Mar-01
7-May-02
12-Apr-01
31-May-02
2400
2600
2800
3000
3200
3400
3600
3800
4000
3000
3200
3400
3600
3800
4000
4200
4400
4600
4800
5000
Sensex Sensex
APPENDIX 2: Approximation of Daily Market Capitalisation ICRA BULLETIN
Notation:
qijt : No. of outstanding shares of the i-th company on the j-th Money
day of month t &
pijt : Price of a share of the i-th company on the j-th day of Finance
month t
APRIL–SEPT. 2002
mcijt = pijt q ijt : market capitalisation of the i-th company’s
shares on the j-th day of month t
MC jt = ∑ mcijt : aggregate market capitalisation on the j-th
i
day of month t
We want to find approximate value of MC jt .
We have
∑p ijt qijt ∑p i0 q ijt ∑p i0 q ijt
MC jt = ∑ p ijt qijt = i i
∑p q i 0 = Π Pjt i
∑p qi 0 (1)
∑p ∑p ∑p
i0 i0
i i0 qijt i0 qi 0 i i0 qi 0 i
i i i
where Π Pjt is the Paasche share price index for the j-th day of
month t with respect to some base date 0,
∑p i0 qijt
i
= Q jt is an index of volume of outstanding shares
∑pi
i0 qi 0
for the j-th day of month t with respect to the base date,
51