You are on page 1of 9

Available online at www.sciencedirect.

com

Borsa _Istanbul Review


_
Borsa Istanbul Review 19-1 (2019) 15e23
http://www.elsevier.com/journals/borsa-istanbul-review/2214-8450

Full Length Article

Dynamics of the impact of currency fluctuations on stock markets in India:


Assessing the pricing of exchange rate risks
Smita Mahapatra*, Saumitra N. Bhaduri
Madras School of Economics, Chennai, Tamil Nadu, India
Received 3 March 2017; revised 2 April 2018; accepted 28 April 2018
Available online 21 July 2018

Abstract

This paper studies the dynamics of the impact of currency fluctuation on Indian stock market by assessing the pricing of exchange rate risk
during the period 2005e2016, specifically before and after financial crises. Estimating a two-factor arbitrage pricing model, using a random
coefficient model, the paper presents evidence that stock returns react significantly to foreign exchange rate fluctuations in the post-crisis period.
Particularly, during the last four years of our sample, 2012e2016, the exchange rate risk factor is becoming a prominent determinant of stock
returns, indicating that Indian investors are increasingly expecting a risk premium on their investment for their added exposure to exchange rate
risk. This is also further corroborated by the study by highlighting the fact that higher the foreign exchange exposure of industry, measured by
trade balance (net inflows), higher is their sensitivity to exchange rate risk (bS). A plausible reason for such premium could be the inadequate
hedging by Indian firms to mitigate the exchange rate risk.
_
Copyright © 2018, Borsa Istanbul Anonim Şirketi. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-
ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

JEL classification: G01; G11; G12; G32


Keywords: Exchange rate risk; Arbitrage pricing theory; Stock returns; Risk premium; Hedging; Financial crisis

1. Introduction decades of which the most obvious example is the Asian


financial crisis in 1997e1998.
As the global economy is gradually regaining its balance The present study is focused on analyzing the impact of
after the 2007e08 recession and the European financial crisis, such exchange rate fluctuations in recent years on the stock
a gradual process of tightening monetary policy by the U.S. market of one of the important emerging markets, India.
and other large economies has been mounting pressure on the Specifically, the study examines the reaction of investors in
currencies of some major EMEs (Emerging Market Econo- terms of changes in premium demands due to the added
mies). This had led to a sell-off in several currencies in early exposure of risks associated with this kind of vulnerability.
2014 indicating a potential crisis in these countries that could For this analysis, the questions addressed are threefold:
destabilize the ongoing global recovery process. Many EMEs First, whether the exchange rate fluctuations impact stock
have been hit by currency crises over the past two to three markets such that it prices the exchange rate risk as a part of
the market premium. Secondly, how the shocks in foreign
exchange market affect this relationship. And finally, if the
currency crisis periods are more vulnerable to such changes in
* Corresponding author. Madras School of Economics, Gandhi Mandapam
Road, Behind Government Data Center, Kottur, Chennai, Tamil Nadu, 600025,
investor sentiments.
India. The financial theory argues that in a well-developed
E-mail addresses: mahapatra.smita93@gmail.com (S. Mahapatra), financial market, foreign exchange risk is a part of the un-
saumitra@mse.ac.in (S.N. Bhaduri). systematic risk that can be hedged away. According to the
_
Peer review under responsibility of Borsa Istanbul Anonim Şirketi.

https://doi.org/10.1016/j.bir.2018.04.004
_
2214-8450/Copyright © 2018, Borsa Istanbul Anonim Şirketi. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
16 _
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23

modern portfolio theory, only risks that cannot be diversified currency risk exposures in country equity index returns.
away (systematic risks) should be priced by the capital market Several other studies find significant foreign exchange expo-
i.e., shareholders must be paid a premium for bearing such sure using unconditional factor models on international data.
risks. The exchange rate fluctuations can influence firm per- For instance, Roll (1992) as one of the three explanatory in-
formance in two ways: The direct impact could be on net fluences as per his inferences, documents that exchange rate
foreign monetary and real domestic assets of the firm. Indi- fluctuations explain a major portion of the variation in equity
rectly, it may affect aggregate demand, the cost of traded in- index returns for developed countries. Ferson and Harvey
puts, competing imported goods and inflationary expectations. (1994) empirically examine multifactor asset pricing models
However, as these foreign exchange risks are perceived as for the returns on eighteen developed markets. The factors
insignificant or diversifiable through active hedging, these chosen, measure global economic risks and they conclude that
sources of risks should not be priced and hence should not foreign exchange risk is one of the most crucial factors in
impact firms’ cost of capital. However, the more recent explaining international equity index returns.
Arbitrage Pricing theory proposed by Ross (1976) suggests Since developing countries have a very different institu-
that if the economy can be represented by certain pervasive tional framework and economic environment, it is important to
factors, then these factors may be priced in the sense that in- study this aspect of these economies, distinctly. For instance,
vestors will be willing to pay a premium to avoid these sources within EMs there are commodity exporters (Brazil, South
of risk (Jorion, 1991). To date, a number of studies have Africa, Russia) and Commodity Importers (India, China,
investigated the possible sources of exchange rate exposure Turkey). Emerging markets like India, Brazil, Turkey, etc.
but not many have addressed the problem of empirically have been attracting enormous amounts of capital from in-
measuring whether this exposure commands a risk premium in ternational investors over the past decade to give a boost to
the stock markets. The ones that have, are mostly centered their rapidly growing economies. However, unlike large
around developed economies. economies, EMEs are faced with a threat of ‘capital flight’ if
In a pioneering study, Jorion (1991) examines such risk- there is any instability detected in the economy by investors.
rewards on the US stock markets from the year 1971e1987 Further, the impact of global changes is not of the same in-
using a multi-factor framework of Arbitrage Pricing models. tensity in all emerging markets. It depends on upon factors like
The relation between expected returns and the sensitivity to the liquidity in the market and the access to international
market and exchange rate movements is analyzed using a two- markets. Some investors and analysts are suspicious of a
factor model. The empirical results suggest that the uncondi- looming financial crisis in emerging markets while others
tional risk premium associated with foreign currency exposure argue that the root cause of vulnerability in exchange rates lie
appears to be insignificant and exchange rate risk is not priced in structural problems in the domestic economies of the EMEs
in the U.S. stock markets. On similar lines, the study by Chen, rather than the effect of external shocks.
Roll, and Ross (1986) finds that macroeconomic variables like At an aggregate level, there are various studies showing a
industrial production, expected and unexpected inflation, the relationship between exchange rate changes and stock returns.
spread between long and short interest rates, and the spread For instance, Chun Mun (2008) examines the effect of fluc-
between high and low-grade bonds are sources of risk that are tuations in foreign exchange rates on international stock
significantly priced. Whereas market portfolio, oil price risks market volatility and cross-market correlations between the
and aggregate consumption are not priced in the stock mar- U.S. stock markets and the appreciating/depreciating local
kets. Sweeney (1986) conducted a comparable study using currencies of the sample countries considered. These include
interest rate changes (yield on long-term U.S. government the Pacific-basin developing countries which depend, to a
bonds) as the second factor instead of exchange rates. It was great extent on international trade and equity flows through
empirically shown that most interest-sensitive stocks were in FDIs for their economic growth. The Asian financial crisis
the utility industries and there was reasonable evidence that indicates the period of financial turmoil and fluctuations in
the interest factor was priced in the sense of APT (two-factor exchange rates. The observations made by the empirical
model). Some studies find significant exchange rate risk using analysis suggested that fluctuating exchange rates significantly
conditional models. For example, De Santis and Gerard contribute to a higher local equity market volatility and the
(1998), using a parsimonious multivariate GARCH process, post-crisis contribution of exchange rate imbalance on the
have found that a significant fraction of the total risk premium U.S.-local market correlation was higher than what it was
is generally represented by the premium for bearing currency before the crisis. On similar lines, Aquino (2005) investigates
risk for four countries: Germany, Japan, the United Kingdom, the foreign exchange exposure faced by Philippine firms
and the United States. A similar study by Choi et al., 1998 around the Asian financial crisis period (1992e2001) and finds
reports that ‘conditional currency risk’ plays a crucial role in that while stock returns were not impacted by exchange rate
explaining the stock returns of firms in Japan. Patro, Wald, and fluctuations before the crisis, there was a significant impact of
Wu (2002) uses a GARCH framework with a panel approach the fluctuations in stock returns after the crisis i.e., after 1997.
and estimates a time-varying two-factor asset pricing model The analysis was based on the two-factor APT model which is
for weekly equity index returns of 16 OECD countries. Using the basis of the present study as well. The study also finds that
the trade-weighted basket of exchange rates and the MSCI during the post-crisis period, investors started expecting a
world market index as risk factors, they report significant premium on their investments for their added exposure to
_
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23 17

exchange rate risk as perceived by them. Another interesting by Mishra, Malhotra, and Swain (2007). The results of the study
study by Chkili and Nguyen (2014) investigates the dynamic show that volatility in both markets is highly persistent and
relationship between stock market returns and exchange rates predictable based on past innovations.
for the BRICS nations between the years 1997 and 2013 using India had seen a near currency crisis situation in the year
a regime switching model. The results from the Markov 2013 which was mainly caused due to two reasons. Firstly,
Switching model applied on the data suggests that stock many EMEs including India were caught off guard with their
markets have more influence on exchange rates during both domestic currencies weakening against the US dollar when the
low and high volatility periods. Bailey and Chung (1995) US Federal Reserve floated the possibility of a taper in May
studies the impact of political and exchange rate risks on the 2013 and tightening of monetary policy. The second reason
risk premiums reflected in equity returns of individuals from being that on the domestic front, the Rupee depreciation sit-
Mexico and finds some evidence of equity market premiums uation as a result of stagnant reforms and declining foreign
for exposure to these risks. Another interesting study by Han, investments, was compounded by the government passing the
Xu, and Yin (2017) empirically investigates the impact of Food Security Bill in late 2013 as a populist pre-election
investor attention on the movement of exchange rates for 9 strategy. Thus, apart from a structural break due to the
countries. Google Search Volume of key terms such as “Dol- global financial crisis of 2008, it is expected that the period
lar”, “USD” etc. has been used as a proxy for ‘attention’. The around 2013 which seems to be a high volatility exchange rate
findings highlight that lagged investor attention significantly regime also exhibits a risk that could affect investor sentiments
influences currency returns although the effect is short term. in India to some extent. The current study empirically shows
The results also demonstrate the reverse analogy that changes that currency risk is increasingly being perceived as a risk that
in exchange returns have an impact on investor attention and is priced by investors in the Indian stock markets in the recent
that too a long-lasting one. years post the global financial crisis in 2008.
Overall, these investigations suggest that the risk factors The paper is arranged into four sections, the first of which
that are essentially ‘priced’ by investors may vary across space introduces the topic, purpose of the study and describes the
and time and hence the results and conclusions of these studies theoretical framework in which the analysis has been con-
cannot be generalized. The present study aims to fill the gap in ducted. Section 2 describes the model and hypothesis tested.
the literature by extending such analyses to the developing The third section on ‘Sample’ explains the construction of
world by focusing on India. variables along with their sources. This is followed by a sec-
Two major financial crises in the last decade e The Global tion on the ‘Methodology and Empirical Results’ of the
Financial crisis (2007e08) and the Eurozone Debt crisis various statistical tests performed, and the final chapter con-
(2010e11), that shook the world have brought about an cludes the study with some important inferences.
increasing interest among researchers to study about financial
markets with the backdrop of these crisis years and understand 2. Empirical model
the impact of such structural breaks. Several studies have sug-
gested that during crisis periods, there is an increased co- The hypothesis of the study and details of the two-factor
movement of markets and a rise in exchange rate volatility APT model employed have been explained in this section.
which in turn has significant implications for the risk manage- We attempt to test the hypothesis that the exchange rate risk is
ment strategies of international investors. A study along these not priced as a premium by investors during low exchange rate
lines by P. Dua and Tuteja (2016) examines the financial volatility regimes, however during (near) currency crisis pe-
contagion and cross-market correlation between the stock and riods foreign exchange rate risk essentially becomes a risk that
currency markets of China, Eurozone, India, Japan and the US is priced in the stock markets.
during the two global crisis periods mentioned above. The re- The model for the analysis has been adopted from a study
sults show significant contagion across the asset classes and by Jorion (1991) who has followed Ross's approach of Arbi-
demonstrate the non-existence of the benefits of portfolio trage Pricing Theory (APT).
diversification for stock markets as shocks after crises are The two-factor version of Ross (1976) APT implies a linear
transmitted internationally. These findings make it important for relationship between expected returns and the sensitivity to
one to study the impact of any such shock or instability on the market and exchange rate movements:
 
perception of investors about their expectation on risk premiums E R~i ¼ d0 þ d1 bmi þ ds bsi ð1Þ
and stock returns during such crisis periods. Our sample,
therefore, is justified as it covers these years in particular. where.
In the light of currency crises and other such events in the Returns can be defined as nominal returns in excess of the
EMEs in the recent past, a country like India which is one of the risk-free rate, Rit ¼ R*it  Rfi ,
fastest growing economies in the world makes for an appropriate
choice for our study. Given the increasing openness of the Indian bmi : ith stock's sensitivity to market movements and
economy, Indian companies can be substantially affected by bsi : ith stock's sensitivity to exchange rate movements.
movements in the value of the currency. A strong evidence of
this relationship and the volatility spillover between the Indian In this model, the market excess return could be interpreted
stock markets and foreign exchange markets has been examined as a transformation of the original factors which are not
18 _
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23

observable anyway. Since the market's sensitivity to exchange 9. Miscellaneous manufacturing (9)
s ¼ 0 by construction, this implies
rate movements bm 10. Diversified (11)
     
E Ri ¼ d0 þ E R~m  d0 bmi þ ds bsi
~ ð2Þ 11. Mining (6)
12. Electricity (15)
Assuming stationarity, the rate of return on an asset at time 13. Services - other than financial (102)
t can be statistically decomposed into an expected component 14. Construction and Real Estate (32)
EðR ~ it Þ and an innovation,
15. Financial Services (79)
    
R~it ¼ E R~it þ bmi R~mt  E R~mt þ bsi F~st þ ~εit ð3Þ
For the variable, R ~ mt which is the market excess returns
where, F ~st : residual of the regression of the exchange rate series, data on the monthly closing prices were collected from
movement against the rate of return on the stock market: the official Bombay Stock Exchange (BSE) website. S&P BSE
  500 index represents almost 93% of the total market capitali-
F~st ¼ R~st  g b 1 R~mt
b0 þ g ð4Þ zation on BSE. It can be considered as a good proxy for the
Given that there will be some correlation between the ex- market index since the underlying 500 companies cover all the
change rate and the market, one cannot simply use the ex- major industries of the economy and BSE is the oldest stock
change rate as the second factor. Without orthogonalization, exchange in India. The risk-free rate was calculated as the log
the second factor could appear to be priced because of a of ten-year bond yield monthly rate collected from RBI. The
nonzero correlation with a priced market. exchange rate return series (R~ st Þ was calculated from the USD-
1
Under rational expectations, substituting (2) in (3) yields: INR* monthly average currency rates collated from the RBI
    website. The exchange rate can be defined as the domestic
R~it ¼ d0 1  bmi þ ds bsi þ bmi R~mt þ bsi F~st þ ~εit ð5Þ
currency price of a unit foreign currency. The second inde-
This is the restricted model to be tested. pendent variable used in the test equation is F ~ st , which is the
Exchange rate exposure orthogonal (statistically indepen- residual of the regression of the exchange rate movement
dent) to the market, is priced if the coefficient ds is non-zero, against the rate of return on the stock market.
which would imply a rejection of the mean-variance efficiency ‘Returns’, as the rate of change of each of the above vari-
of the market. ables were calculated as the natural log of the ratio of the
The unrestricted equation is given by prices of one period to its previous period. The benefit of using
R~it ¼ ai þ bmi R~mt þ bsi F~st þ ~εit ð6Þ returns over prices is in ‘normalization’ (measuring all vari-
ables in a comparable metric) that it provides. If prices are
Given the restricted and unrestricted models, the cross- assumed to be distributed lognormally, then log returns will
section restriction is tested by a likelihood ratio test. follow a normal distribution and that is a major advantage
3. Data since most of the classical statistics presume normality.
The timeline for this study is between January 2005 and
The study uses secondary data for analysis. The dependent January 2016 (Returns data begin from February 2005 since it is
variable in the test equation (6), ‘R~ it ’ is a series of returns calculated as a ratio of current to the previous time period). The
(excess of risk-free rate) on companies' common stocks. This entire period of study has been divided into three sub-periods
data was collected using Prowess which is a database of the with the purpose of accommodating the possible structural
financial performance of Indian companies maintained by breaks caused due to the financial crisis or near-crisis situations
Centre for Monitoring Indian Economy (CMIE). All the that would help us comprehend whether and if so, then in what
companies listed in the S&P BSE 500 index have been way investors reacted to such relative instability (Table 1).
included in the analysis. The data on the monthly adjusted GDP growth rates depicted in Fig. 1, highlights the possible
closing prices for the period January 2005eJanuary 2016 were structural breaks considered.
collected for each of these 500 companies and fifteen portfo-
lios (number of companies in each portfolio given in paren- 4. Methodology and Empirical Results
thesis) were created based on the pre-defined sets given in
Prowess. The average of the company stock prices for each The market excess return series and exchange rate series have
month was taken as the portfolio averages on which monthly been tested for stationarity using the ADF test. The correlation
returns were calculated. The portfolios are: between the excess market return series and the exchange rate
series is calculated to be 0.125 which is a reasonable explanation
1. Food & agro-based products (26)
to justify the use of the residuals of the regression between these
2. Textiles (16)
two variables in order to ensure orthogonality. Fig. 2 shows a
3. Chemicals and chemical products (78)
basic time series plot of the changes in exchange rate series to
4. Consumer goods (21)
5. Construction material (24)
1
6. Metals and metal products (24) It must be noted that using USD-INR as the exchange rate measure reflects
7. Machinery (31) 80% of the foreign exchange market in Inda. An alternative measure, INR
NEER which takes 36 currencies into account based on their weights, might
8. Transport equipment (26)
provide a broder base proxy and can be considered for future research.
_
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23 19

Table 1
Explanation of subsample periods.
Sub-period Feb 2005eDec 2007 Jan 2008eDec 2011 Jan 2012eJan 2016
Rationale Pre-crisis period of presumed The global financial crisis A near-crisis situation in FY 2012e13 with
overall economic stability and period with the initial plunge another plunge in GDP growth rate and a volatile
high GDP growth rates. in growth rate followed by gradual recovery. rupee value, followed by gradual recovery.

highlight the fluctuations in this variable in the period consid- Squares Seemingly Unrelated Regression estimation technique
ered. The substantial volatility of the exchange rates in the to simultaneously estimate the betas and pricing parameters.
speculated currency crisis years of 2008 and 2013 can be Due to the advantages that Random coefficient models
inferred on observing the line graph. provide, especially in terms of degrees of freedom and
The exchange rate exposure represented by the variable, F ~st simplified computation, the present study chooses to use this
is calculated first. The parameters, g ^
b 0 and g1 in equation (4) estimation technique. These models are called as random co-
are estimated using Ordinary Least Squares and the F ~st se- efficient models because they assume that the individual spe-
ries is constructed which is used as the second independent cific regression parameters are random, i.e. each represents a
variable in our final test equation (5). draw from a population. All the variants of these type of models
The Hausman test between random and fixed effects sug- are a particular case of the more general class of linear mixed
gested that a random effects model would fit the data better. effects models. Linear mixed effects models can be thought of
However, according to Hsiao and Pesaran (2004), “Conven- as extensions of linear regression models for the cases where
tional models do not allow the interaction of the individual data needs to be summarized in groups. In Swamy's Random
specific and/or time-varying differences in the included Coefficient linear regression model, rather than only the inter-
explanatory variables.” This was the rationale behind using cept varying across groups, all the coefficients are allowed to
Random Coefficient models for estimation with the kind of vary. The ‘xtrc’ function in the software Stata, that uses the
panel data we have. Aquino (2005) uses the Generalized Least Swamy Random Coefficient model (Swamy, 1970) has been
used to estimate the beta coefficients (bm and bs) in the unre-
stricted equation (6) which gives us the market and exchange
GDP Growth (Annual %) rate exposure coefficients for the fifteen industry portfolios. The
12
possibility of a non-linear relationship between stock returns
10 and foreign exchange exposure has been explored by including
% GDP growth rate

8 the quadratic term of exchange rate variable in the regression.


6
Our main findings remain invariant post such robustness test.
Most of the industry portfolios except for few industries such as
4
‘Electricity’, ‘Construction and Real Estate’ and ‘Financial
2 Services’, we do not observe non-linearity.
0 The coefficients across industries seem to be very different
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 in all the time periods. Most of the industries have recorded a
Year
significant negative exposure which might imply that these
industries tend to import a considerable portion of their factor
Fig. 1. Annual Growth Rate of GDP at Market Prices based on Constant
Rupee. Note: Based on official government statistics, the new base year is
inputs and are vulnerable to currency depreciation. On the
2011/12. The report was made using the data from the official website of the other hand, an industry like “Financial Services” shows a
Reserve Bank of India. significant positive exposure for the entire time period of
analysis and the second sub-period chosen. This indicates
export-oriented foreign operations and that the industry tends
Exchange rate fluctuaons to benefit from currency depreciation. A theoretical intuition
0.12 along these lines could be that higher sectoral import intensity
0.1 can influence the impact of exchange rate changes on national
Return on exchange rate

0.08
stock returns. To validate this hypothesis, the exchange rate
0.06
0.04
exposure (bs) for each industry portfolio was plotted against its
0.02 trade balance (Refer Fig. 3 in Annexure). Except for two
0 sectors (Financial services and Misc. manufacturing), it could
Aug-05

Aug-06

Aug-07

Aug-08

Aug-09

Aug-10

Aug-11

Aug-12

Aug-13

Aug-14
Feb-05

Feb-06

Feb-07

Feb-08

Feb-09

Feb-10

Feb-11

Feb-12

Feb-13

Feb-14

Feb-15

-0.02
-0.04
be observed that industries with higher imports had a ‘sig-
-0.06 nificant’ negative exposure. Future research is needed to
Month
elaborate on this by using alternate measures of international
exposure.
Fig. 2. Exchange Rate Fluctuations (February 2005eJanuary 2016). Note:
Time series line plot of the exchange rate returns series to highlight the Tables 2e5 report the systematic risk and the exchange rate
fluctuations in the period considered. Monthly data on the USD-INR exchange exposure of the industry portfolios. The exposure coefficients
rates were obtained from the official website of Reserve Bank of India. reported, after projection on the market risk factor and the
20 _
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23

Fig. 3. Exchange rate exposure of Industry portfolios with respect to their Trade Balance. Note: X axis e Trade Balance (Net Inflows) of each Industry portfolio (in
US billion $) for FY 2015e16. Y axis e Estimated Exchange rate exposure coefficient (bsi ) from the equation: R~it ¼ ai þ bm ~ s~
i Rmt þ bi F st þ ~
εit . *The highlighted
industries (shaded red) represent the portfolios that are significant at 5% los. Data Source: Trade data retrieved from ‘India's Foreign Trade 2015e16: An Appraisal’
report prepared by the Economic Division of Credit Analysis and Research (CARE).

exchange rate factor are denoted as bm and bs respectively. A second sub-period (Jan 2008 to Dec 2011) and the third sub-
test of the joint significance was conducted using the Wald test period (Jan 2012 to Dec 2016) show that the null hypothesis
with the null hypothesis that coefficients are jointly equal to is rejected for the market return but not rejected for exchange
zero, results of which have been reported below each table. rate exposure (at 5% level of significance).
The null is rejected for the market return but not rejected for To statistically test the difference in exposure coefficients
exchange rate exposure when tested for the entire study period across the industry portfolios, the t-statistics for equality of the
(Feb 2005 to Jan 2016). For the first sub-period (Feb 2005 to coefficients had been performed. The null hypothesis of equal
Dec 2007), the null is rejected for both, the market return exchange rate betas is rejected for all the periods except the
series and for exchange rate exposure. The results for the first sub-period (Feb 2005 to Dec 2007). This implies that

Table 2
~ mt þ bs F
~ it ¼ ai þ bm R ~
Market and exchange rate exposure of industry portfolios. February 2005eJanuary 2016 (monthly data). Model: R i i st þ ~
εit .
S.No. Industry bm
i bsi
Coeff z-stat p-value Coeff z-stat p-value
1 Mining 1.0934 17.66 0.000 0.2889a 2.41 0.016
2 Food and Processing 0.8000 20.89 0.000 0.0709 0.67 0.506
3 Textiles 0.9743 18.12 0.000 0.1217 1.05 0.292
4 Chemical 0.8178 25.73 0.000 0.1503 1.50 0.135
5 Consumer 0.7433 16.01 0.000 0.1838 1.65 0.100
6 Construction material 0.9856 19.21 0.000 0.1765 1.54 0.122
7 Metal and metal products 1.0599 27.04 0.000 0.3879a 3.62 0.000
8 Machinery 1.0554 24.62 0.000 0.3633a 3.31 0.001
9 Transport Equipment 0.9676 21.12 0.000 0.1070 0.96 0.336
10 Misc. manufacturing 0.9097 18.04 0.000 0.1384 1.22 0.224
11 Diversified 1.0677 24.04 0.000 0.3469a 3.14 0.002
12 Electricity 1.3849 16.73 0.000 0.6579a 4.89 0.000
13 Construction and real estate 1.3747 24.95 0.000 0.6739a 5.80 0.000
14 Financial Services 0.6571 24.57 0.000 0.2588a 2.77 0.006
15 Services (other than financial) 0.9594 26.34 0.000 0.0348 0.33 0.741
Test for joint significance: chi2(1) ¼ 292.56 chi2(1) ¼ 3.18.
Prob > chi2 ¼ 0.0000 Prob > chi2 ¼ 0.0747.
a
Significance at 5% level. All coefficients are significant for the market coefficient, bim and are hence not marked for significance specifically.
_
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23 21

Table 3
~ mt þ bs F
~ it ¼ ai þ bm R ~
Market and exchange rate exposure of industry portfolios. February 2005eDecember 2007 (monthly data). Model: R i i st þ ~
εit .
S.No. Industry bm
i bsi
Coefficient z-stat p-value Coefficient z-stat p-value
1 Mining 1.1758 8.16 0.000 0.4366a 2.34 0.020
2 Food and Processing 0.8832 11.64 0.000 0.2902 1.42 0.155
3 Textiles 1.0924 11.01 0.000 0.3262 1.64 0.101
4 Chemical 0.8810 12.06 0.000 0.1248 0.61 0.544
5 Consumer 0.6929 6.36 0.000 0.1314 0.67 0.503
6 Construction material 1.0631 9.08 0.000 0.3536 1.83 0.068
7 Metal and metal products 1.0786 14.32 0.000 0.6704a 3.29 0.001
8 Machinery 0.9204 9.04 0.000 0.3192 1.61 0.107
9 Transport Equipment 1.0684 11.12 0.000 0.4122a 2.06 0.039
10 Misc. manufacturing 0.8371 8.91 0.000 0.2648 1.32 0.187
11 Diversified 1.1567 13.71 0.000 0.4386a 2.16 0.031
12 Electricity 1.3447 8.54 0.000 0.5096a 2.79 0.005
13 Construction and real estate 1.2532 10.38 0.000 0.5035a 2.62 0.009
14 Financial Services 0.5672 11.61 0.000 0.1184 0.63 0.531
15 Services (other than financial) 0.9005 9.58 0.000 0.3264 1.63 0.104
Test for joint significance: chi2(1) ¼ 153.27 chi2(1) ¼ 4.44.
Prob > chi2 ¼ 0.0000 Prob > chi2 ¼ 0.0352.
a
Significance at 5% level. All coefficients are significant for the market coefficient, bim and are hence not marked for significance specifically.

Table 4
~ mt þ bs F
~ it ¼ ai þ bm R ~
Market and exchange rate exposure of industry portfolios. January 2008eJanuary 2012 (monthly data). Model: R i i st þ ~
εit .
S.No. Industry bm
i bsi
Coefficient z-stat p-value Coefficient z-stat p-value
1 Mining 1.0811 16.87 0.000 0.3519 1.81 0.071
2 Food and Processing 0.7964 17.08 0.000 0.1081 0.62 0.533
3 Textiles 0.9914 0.69 0.000 0.2591 1.3 0.194
4 Chemical 0.8494 23.68 0.000 0.0906 0.60 0.549
5 Consumer 0.8009 15.04 0.000 0.2729 1.49 0.135
6 Construction material 0.9024 15.59 0.000 0.0239 0.13 0.899
7 Metal and metal products 1.1139 23.61 0.000 0.5337a 3.07 0.002
8 Machinery 1.0673 22.12 0.000 0.4232a 2.41 0.016
9 Transport Equipment 0.9266 17.80 0.000 0.0215 0.12 0.906
10 Misc. manufacturing 0.9938 14.07 0.000 0.3640 1.81 0.070
11 Diversified 1.0776 20.67 0.000 0.6169a 3.40 0.001
12 Electricity 1.3022 15.70 0.000 0.7721a 3.62 0.000
13 Construction and real estate 1.3676 20.04 0.000 0.8220a 4.13 0.000
14 Financial Services 0.6937 16.60 0.000 0.4586a 2.79 0.005
15 Services (other than financial) 1.0255 24.22 0.000 0.0926 0.56 0.576
Test for joint significance: chi2(1) ¼ 336.17 chi2(1) ¼ 2.32.
Prob > chi2 ¼ 0.0000 Prob > chi2 ¼ 0.1274.
a
Significance at 5% level. All coefficients are significant for the market coefficient, bim and are hence not marked for significance specifically.

relative stability of exchange rates is a feature of the first sub- from the previous estimation results of equation (6). The
period while the others show variability in sensitivity to ex- restricted equation to be tested is a linear form equation with
change rate fluctuations. The instability reported in the post- four parameters d0, ds, bm s
i and bi . The beta coefficients
crisis periods and the entire period indicates the importance within the restriction, represented in square brackets in
of the present study. It is evident that market exposure is equation (5) are treated as dependent variables. The values
significant as the null of equal market betas is strongly rejected for d0 and ds along with their statistics and p-values are re-
in all the time periods considered. We thus focus on the pricing ported in Table 6.
of the exchange rate factor since a difference has been The null hypothesis that the constant risk premium coeffi-
observed in the time periods under consideration. Overall, it cient, representing the pricing of exchange rate exposure,
can be said that sorting by industries generates a substantial d0 ¼ 0 is rejected for all the periods considered. This implies
dispersion in exposure coefficients, which is an important pre- that the model in use accounts for the total differential in
requisite for pricing tests. excess returns across the different time periods.
The final test equation (5) has been estimated by regres- The null hypothesis that ds ¼ 0 is rejected (at 5% signifi-
sion after plugging in the beta coefficient values obtained cance level) for the whole period (2005e2016) and for the
22 _
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23

Table 5
~ mt þ bs F
~ it ¼ ai þ bm R ~
Market and exchange rate exposure of industry portfolios. January 2012eJanuary 2016 (monthly data). Model: R i i st þ ~
εit .
S.No. Industry bm
i bsi
Coefficient z-stat p-value Coefficient z-stat p-value
1 Mining 0.9119 10.77 0.000 0.1476 1.17 0.241
2 Food and Processing 0.8193 9.22 0.000 0.0244 0.19 0.847
3 Textiles 0.8117 7.88 0.000 0.0168 0.13 0.896
4 Chemical 0.7086 9.99 0.000 0.0948 0.76 0.445
5 Consumer 0.7244 8.19 0.000 0.0292 0.23 0.817
6 Construction material 1.1812 12.34 0.000 0.2922a ¡2.29 0.022
7 Metal and metal products 0.8285 9.58 0.000 0.1451 1.15 0.250
8 Machinery 1.1805 13.75 0.000 0.2891a ¡2.29 0.022
9 Transport Equipment 0.9979 10.09 0.000 0.1337 1.05 0.296
10 Misc. manufacturing 0.7987 9.66 0.000 0.0634 0.50 0.614
11 Diversified 0.9366 9.25 0.000 0.0702 0.55 0.584
12 Electricity 1.3149 12.21 0.000 0.3651a ¡2.82 0.005
13 Construction and real estate 1.3437 17.40 0.000 0.4177a ¡3.34 0.001
14 Financial Services 0.7337 16.94 0.000 0.0529 0.49 0.627
15 Services (other than financial) 0.8193 13.06 0.000 0.0779 0.64 0.523
Test for joint significance: chi2(1) ¼ 185.96 chi2(1) ¼ 1.17.
Prob > chi2 ¼ 0.0000 Prob > chi2 ¼ 0.2784.
a
Significance at 5% level. All coefficients are significant for the market coefficient, bim and are hence not marked for significance specifically.

Table 6
~ mt þ bs F
~ it ¼ ½d0 ð1  bm Þ þ ds bs  þ bm R ~
Tests of Pricing of Exchange Rate Exposure. (Monthly data). Two Factor Model: R i i i i st þ ~
εit .
Factor Prices (t-statistics) [P>jtj] Test of Fit [p-value]
Period d0 ds c2
Feb 2005eJan 2016 0.0258696 a
0.0144305 a
0.95
(2.49) [0.013] (2.29) [0.022] [0.3302]
Feb 2005eDec 2007 0.0568258a 0.0025197 14.12
(-3.54) [0.000] (0.27) [0.787] [0.0002]
Jan 2008eDec 2012 0.0860055a 0.0143851 24.38
(4.12) [0.000] (1.67) [0.096]b [0.0000]
Jan 2012eJan 2016 0.0239179a 0.0213261a 7.23
(2.58) [0.010] (-2.05) [0.040] [1.0000]
Note: t-statistics between parentheses, marginal significance levels between brackets. d0 and ds are the pricing coefficients in the test equation. The chi-square
statistic tests the cross-sectional restrictions in the two-factor model.
a
Significant at the 5% level.
b
Significant at the 10% level.

latest sub-period (2012e2016). It is rejected (at 10% loss) for investments for their perceived added exposure to exchange
the second sub-period i.e. between 2008 and 2012. The null is rate risk and it has become more prominent in the recent years.
not rejected for the first sub-period before any crisis which
shows that exchange rate risk was not priced in this period of 5. Conclusion
stability.
The chi-square goodness of fit statistic is computed using The impact of foreign exchange fluctuations on stock
the restricted and unrestricted models and reported in the last returns is increasingly becoming a prominent issue to in-
column of Table 6. The likelihood ratio statistics for the entire vestors, financiers and policymakers. The paper examines the
period and the most recent subperiod (2012e2016) suggest exposure of fifteen Indian industry groups (comprising the 500
that the model is reasonably well specified and seems to fit the firms listed in the S&P BSE 500 index) to the monthly USD-
data in the periods in which the pricing coefficients are INR exchange rate fluctuations for different time periods with
significant. the overall period being eleven years from February 2005 to
The observations indicate that the results for the pricing January 2016. The analysis is done in the two-factor APT
coefficients are significant for the periods considered post- framework.
2008 which suggest that after the onset of the global finan- Our choice of sample period also allows examining the
cial crisis, there was a slight change in the perception of in- effect of exchange rates on stock markets and the overall
vestors. In the last four years (2012e2016) however, the investor sentiment in India in the last decade which has not
exchange rate risk factor is becoming prominent as being been explored so far, in best of our knowledge. This is an
‘priced’ in the stock market. This essentially means that in- important aspect of the currency crisis that is seen imminent in
vestors are progressively expecting a risk premium on the forthcoming years and is affecting the investor
_
S. Mahapatra, S.N. Bhaduri / Borsa Istanbul Review 19-1 (2019) 15e23 23

communities in EMEs gradually. Thus, it is expected that the Chen, N. F., Roll, R., & Ross, S. A. (1986). Economic forces and the stock
more recent the period under study, the greater the chances are market. Journal of Business, 59(3), 383e403.
Chkili, W., & Nguyen, D. K. (2014). Exchange rate movements and stock
of capturing the change in investor sentiments. For a country market returns in a regime-switching environment: Evidence for BRICS
like India, that is going through a transitional phase in its countries. Research in International Business and Finance, 31, 46e56.
economy with its present government pushing for greater Choi, J. J., Hiraki, T., & Takezawa, N. (1998). Is foreign exchange risk priced
foreign investments for boosting the country's industrial and in the Japanese stock market? Journal of Financial and Quantitative
infrastructural growth, it is necessary to pay attention to how Analysis, 33, 361e382.
Chun Mun, K. (2008). Effects of exchange rate fluctuations on equity market
the investors and the stock markets will react to these changes. volatility and correlations: Evidence from the Asian financial crisis.
Empirical evidence suggests that investors are expectant of a Quarterly Journal of Finance and Accounting, 47(3), 77e102.
risk premium on their investment owing to the increased risk De Santis, G., & Gerard, B. (1998). How big is the premium for currency risk?
exposure caused by exchange rate fluctuations, especially since Journal of Financial Economics, 49, 375e412.
the last few years beginning from the crisis period. This implies Dua, P., & Tuteja, D. (2016). Financial crises and dynamic linkages across
international stock and currency markets. Economic Modelling, 59,
inadequate hedging by firms for the exchange rate risk and in the 249e261.
larger macroeconomic sense, it displays market inefficiencies in Ferson, W. E., & Harvey, C. R. (1994). Sources of risk and expected returns in
the stock market and/or foreign exchange market. global equity markets. Journal of Banking & Finance, 18, 775e803.
The study has qualitatively inferred from the results ob- Han, L., Xu, Y., & Yin, L. (2017). Does investor attention matter? The
tained, that Indian investors now require compensation for attention-return relationships in FX markets. Economic Modelling. https://
doi.org/10.1016/j.econmod.2017.06.015.
bearing exchange rate risk. However, the quantitative aspects Hsiao, C., & Pesaran, M. (2004). Random coefficient panel data models.
such as the difference between the risk premiums demanded CESifo Working Paper No., 1233.
by investors in the various time periods are beyond the scope Jorion, P. (1991). The pricing of exchange rate risk in the stock market.
of the present paper and are left for future research. Journal of Financial and Quantitative Analysis, 26(3), 363e376.
Mishra, A. K., Malhotra, D. K., & Swain, N. (2007). Volatility spillover be-
tween stock and foreign exchange markets: Indian evidence. International
Conflict of interest Journal of Business, 12(3).
Patro, D. K., Wald, J. K., & Wu, Y. (2002). Explaining exchange rate risk in
None. world stock markets: A panel approach,. Journal of Banking & Finance,
26, 1951e1972.
References Roll, R. (1992). Industrial structure and the comparative behaviour of inter-
national stock Market indices. The Journal of Finance, 47, 3e41.
Ross, S. A. (1976). The arbitrage theory of capital asset pricing,. Journal of
Aquino, R. Q. (2005). Exchange rate risk and Philippine stock returns: Before Economic Theory, 13, 341e360.
and after the Asian financial crisis. Applied Financial Economics, 15, Swamy, P. A. V. B. (1970). Efficient inference in a random coefficient
765e771. regression model. Econometrica, 38(2), 311e323.
Bailey, W., & Chung, Y. P. (1995). Exchange rate fluctuations political risk and Sweeney, R. (1986). The pricing of interest-rate risk: Evidence from the stock
stock returns: Some evidence from an emerging market. Journal of market. The Journal of Finance, 41(2), 393e410.
Financial and Quantitative Analysis, 30(4), 541e561.

You might also like