stock prices and exchange rates a review article

Journal of Economic Studies
On the relation between stock prices and exchange rates: a review article
Mohsen Bahmani-Oskooee, Sujata Saha,

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Article information:
To cite this document:
Mohsen Bahmani-Oskooee, Sujata Saha, (2015) "On the relation between stock prices and
exchange rates: a review article", Journal of Economic Studies, Vol. 42 Issue: 4, pp.707-732, https://
doi.org/10.1108/JES-03-2015-0043
Permanent link to this document:
/>Downloaded on: 14 January 2018, At: 18:48 (PT)
References: this document contains references to 59 other documents.
To copy this document:
The fulltext of this document has been downloaded 3115 times since 2015*

Users who downloaded this article also downloaded:
(2013),"Returns and volatility spillover between stock prices and exchange rates: Empirical evidence
from IBSA countries", International Journal of Emerging Markets, Vol. 8 Iss 2 pp. 108-128 href=" />(2014),"The relationship between stock prices and exchange rates in Asian markets: A wavelet
based correlation and quantile regression approach", South Asian Journal of Global Business
Research, Vol. 3 Iss 2 pp. 209-224 <a href=" />doi.org/10.1108/SAJGBR-07-2013-0061</a>
Access to this document was granted through an Emerald subscription provided by emeraldsrm:561544 []

For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald
for Authors service information about how to choose which publication to write for and submission
guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.

About Emerald www.emeraldinsight.com

Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as
well as providing an extensive range of online products and additional customer resources and
services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the
Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for
digital archive preservation.
*Related content and download information correct at time of download.


The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0144-3585.htm

On the relation between stock
prices and exchange rates:
a review article
Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Mohsen Bahmani-Oskooee and Sujata Saha
The Center for Research on International Economics and
Department of Economics, The University of Wisconsin-Milwaukee,
Milwaukee, Wisconsin, USA

Stock prices
and exchange
rates
707
Received 10 March 2015
Revised 10 March 2015
Accepted 11 March 2015

Abstract
Purpose – While changes in stock prices are said to affect exchange rates, exchange rate changes are
also said to affect stock prices. The purpose of this paper is threefold. First, the authors review all
empirical literature by dividing them into two groups of univariate and multivariate studies. Second,
a table which summarizes the main features of each study is provided to help future researchers to
have easy access to summary of each study. Finally, a new direction for future research is proposed.
This new direction relies upon non-linear ARDL approach and shows how to investigate symmetric vs
asymmetric effects of exchange rate changes on stock prices.
Design/methodology/approach – The paper reviews existing published work and provides
suggestions for future research.
Findings – The paper reviews existing published work and provides suggestions for future research.
An application reveals that exchange rate changes have asymmetric effect on stock prices.
Originality/value – This is the first review paper on the relation between exchange rates and
stock prices.
Keywords Stock prices, Exchange rates, Review article, Asymmetry
Paper type Research paper

I. Introduction
One of the areas in financial economics that has received the largest attention is the link
between foreign exchange market and the stock market. The easiest way to infer the
link is by a reference to portfolio approach to exchange rate determination. Under this
approach wealth is one of the main determinants of the exchange rate. An increase in
stock prices usually increases public wealth. This in turn increases the demand for
money and therefore, interest rates. By attracting international investment, domestic
currency appreciates. On the other hand, depreciation of domestic currency can boost
exports and eventually profit of exporting firms. High profits once announced, can
cause share prices to rise. Furthermore, depreciation raises cost of imported inputs.
This can increase production cost even to firms that are not export oriented. If higher
costs result in lower profits or expectation of lower profits, share prices could be

affected. For this reason stock prices could move in either direction.
Clearly if one needs to test the link between stock prices and exchange rates, one has
to concentrate on the current floating exchange rate system that began in 1973. For this
reason, almost all studies have engaged in empirical analysis using data from post-1973
period. While some have concentrated on the link between stock prices and exchange
rates at bilateral level, some have included additional determinants of stock prices. For
this reason, we review the empirical literature between the two variables at bilateral
JEL Classification — F31, G15

Journal of Economic Studies
Vol. 42 No. 4, 2015
pp. 707-732
© Emerald Group Publishing Limited
0144-3585
DOI 10.1108/JES-03-2015-0043


JES
42,4

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

708

level in Section 2. In Section 3 we review studies that have included other variables
in their models. In Section 4 we propose a direction for future research and finally,
we concluded in Section 5. A table is provided in which all features of the reviewed
studies are summarized[1].
II. Review of bivariate studies
Perhaps the first study that concentrated on the relation between stock prices and

exchange rates is that of Aggarwal (1981) who used monthly data during the period
1974-1978 from the USA. By using an aggregate index of stock prices and effective
exchange rate of the dollar he showed that there is a positive relation between the two
variables, i.e., dollar depreciation or a decline in the effective exchange rate of the dollar
caused stock price to decline. The implication is that more firms were hurt by
depreciation than helped. Exactly opposite was concluded when Soenen and Hennigar
(1988) looked into the response of stock prices of seven industrial sectors in the USA to
changes in the value of the dollar. The seven sectors were selected on the belief that
they were affected heavily by international trade. The seven sectors were automobile,
computer, machinery, paper, textile, steel and chemical. The finding that the relation
between stock price of each sector and the value of the dollar was negative implied that
as dollar depreciates, every sector exports more and rips profit from trade.
None of the studies mentioned above accounted for integrating properties of the two
variables nor for cointegration between them. Thus, their findings could suffer from
spurious regression problem. To resolve the issue, Bahmani-Oskooee and Sohrabian
(1992) used monthly data from the period of 1973-1988 to show that index of S & P 500
and the effective exchange rate of the dollar are non-stationary variables. Application
of Engle and Granger (1987) cointegration analysis revealed that there is no long-run
relationship between the two variables. However, application of the Granger causality
test revealed that the two variables Granger cause each other in the short run.
The Asian financial crisis of 1997 triggered a renewed interest in studying the
interaction between exchange rates and stock prices, mostly in developing countries.
Granger et al. (2000) concentrated on East-Asian countries of Hong Kong, Indonesia,
Japan, South Korea, Malaysia, the Philippines, Singapore, Thailand and Taiwan and used
Granger causality test and Gregory-Hansen cointegration test to analyze the relationship
between stock prices and exchange rates. They used daily data for the period 1986-1997
and showed that exchange rate changes affect stock prices in Japan and Thailand. For
Taiwan, the relationship was reversed, that is, stock prices affected exchange rates. They
found bi-directional relationship between the two variables in Indonesia, South Korea,
Malaysia and the Philippines, a finding similar to that of Bahmani-Oskooee and

Sohrabian (1992) for the USA. However, Singapore failed to show any pattern of
relationship. Through Granger causality test it was inferred that exchange rates affected
stock prices in eight of the nine countries. Following the same path, Nieh and Lee (2001)
used daily data from the period 1993-1996 and explored the dynamic relationship
between stock prices and exchange rates in the G-7 countries (Canada, France,
Germany, Italy, Japan, UK and USA). Engle-Granger and Johansen maximum likelihood
methods of cointegration were applied. Their results, again, supported the findings of
Bahmani-Oskooee and Sohrabian (1992) and reported that there was no long-run
relationship between stock prices and exchange rates in all of the G-7 countries. The
results of VECM estimation suggests that the two financial variables do not have
predictive capabilities for more than two consecutive days and thus there is a short-run
significant relationship which lasts only for one day for certain G-7 countries.


Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Continuing the same line of research, Smyth and Nandha (2003) examined the
relationship between stock prices and exchange rates in four South-Asian countries of
Bangladesh, India, Pakistan and Sri Lanka using daily data from 1995 to 2001. Like
previous studies, using Engle-Granger as well as Johansen’s cointegration techniques,
they were unable to find any long-run equilibrium relationship between the two
variables in any of the four countries. Using Granger causality test they also concluded
that exchange rates Granger cause stock prices in India and Sri Lanka but for
Bangladesh and Pakistan they found no evidence of causality running in either
direction. During a crisis period the fluctuations in stock prices and exchange rates are
high. The Asian financial crisis of 1997 led to series of financial downfall. To consider
experiences of other countries in the crisis region, Lean et al. (2005) used weekly data
from 1991 to 2002 for Hong Kong, Indonesia, Singapore, Malaysia, South Korea, the
Philippines and Thailand to study the pre- and post-crisis scenario and the effect of
9-11 terrorist attack. Japan was used as a control. They applied both cointegration and

bivariate causality technique. For all of the countries except for the Philippines
and Malaysia, they found no evidence of Granger causality between stock prices and
exchange rates in the period before Asian financial crisis. During the crisis period, they
found evidence of causality between the two variables. Results show no cointegration
between the variables before or during the Asian crisis of 1997 but after the 9-11
terrorist attack, weaker cointegration relationships between the variables were found.
Phylaktis and Ravazzolo (2005) used monthly data from 1980 to 1998 for Hong Kong,
Indonesia, Malaysia, the Philippines, Singapore and Thailand. They analyzed the
short- and long-run relationships between exchange rates and stock prices and the
avenues through which exogenous shocks affect these two variables. They found that
exchange rates and stock prices are positively related using the method of cointegration
and Granger causality tests. US stock price is the causing variable which acts as a
channel that links the exchange rates of the five countries to their stock market indices.
Shifting to Europe, Obben and Shakur (2006) analyzed the relationship between
the performance of the stock market and exchange rates in New Zealand using a
cointegrating VAR approach using weekly data from 1999 to 2005. The five exchange
rates were those currencies that are used in the construction of New Zealand’s tradeweighted index series. They concluded that both in the short run and long run there is bidirectional causality between the five exchange rates and a couple of share price indices.
With regards to non-dollar exchange rates, Yau and Nieh (2006) used monthly data of
Japan and Taiwan from 1991 to 2005 to study the relationship among stock prices
of Taiwan and Japan and NTD/Yen exchange rate. They applied Granger causality test
and found that there is bi-directional causality between the stock prices of Taiwan and
Japan but there is no significant causal relationship between the NTD/Yen exchange rate
and the stock prices of Japan and Taiwan. From the Johansen method of cointegration it
was concluded that there was no long-run relationship among the three variables.
However, Yau and Nieh (2009) revisited the issue by testing for cointegration with
threshold effect between the stock prices and the exchange rates in Japan and Taiwan
and the effect of US exchange rate on the financial market of Taiwan. Using monthly
data from 1991 to 2008, they found evidence of a long-run equilibrium relationship
between NTD/JPY and the stock prices of Japan and Taiwan. There was no short-run
causal relationship between the two countries financial assets which means that

exchange rate and stock price movements do not affect each other significantly in the
short run. The results supported the traditional approach that a long-run positive
relationship runs from exchange rates of either Japan or USA to stock prices of Taiwan.

Stock prices
and exchange
rates
709


JES
42,4

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

710

For seven Asian countries (Hong Kong, Japan, Korea, Malaysia, Singapore, Taiwan
and Thailand), Pan et al. (2007) applied the methods of Granger causality and Johansen
cointegration test to examine the linkages between stock prices and exchange rates
using daily data from 1988 to 1998. They concluded that during the Asian financial
crisis period there is no long-run equilibrium relationship between exchange rates and
stock prices. For Hong Kong, Japan, Malaysia and Thailand there existed a significant
causal relationship from exchange rates to stock prices before the 1997 Asian financial
crisis and during the financial crisis period they found causal relationship from
exchange rates to stock prices for all countries except for Malaysia.
All studies reviewed above have assumed that the relation between a stock price
and an exchange rate is linear. Ismail and Isa (2009) uses Markov Switching VAR
model and assume the two variables to be regime dependent. They then studied the
non-linear relationship between exchange rates and stock prices in Malaysia using

monthly data from 1990 to 2005. The Johansen cointegration test suggested evidence of
no cointegration between the variables. Their analysis showed that a non-linear model
is more appropriate to model all the series than the linear model. They also found
evidence of common regime switching behavior between the variables. Whether linear
or non-linear relationship, it appears that no study finds evidence of long-run
relationship. This is also true of Rahman and Uddin (2009) who used monthly data
from 2003 to 2008 for Bangladesh, India and Pakistan and the method of Johansen
cointegration and Granger causality test. Not only they found evidence of no long-run
relationship between stock prices and exchange rates, they also found no causal
relationship in either direction between the variables. The implication is that market
participants cannot use information of one market to help to forecast the other market.
Considering the experience of Australia, using daily data from 2003 to 2006,
Richards et al. (2009) studied the relationship between the two variables. Using
Johansen cointegration test they showed that that both stock prices and exchange rates
are cointegrated in the long run. The method of Granger causality test supported the
portfolio balance model which says that changes in stock prices affect changes in
exchange rates. However, using weekly data from 1989 to 2006, Kutty (2010) was
unable to support cointegration in Mexico, though some evidence of short run. Granger
causality was reported. Considering the Chinese experience, the dynamic relationship
between exchange rates and stock prices was studied by Zhao (2010) using monthly
data from 1991 to 2009. Applying Johansen method of cointegration, the results showed
no stable long-run equilibrium relationship between the real effective exchange rate
and the stock price. The source and the magnitude of the spillovers were identified
through vector auto-regression and multivariate generalized autoregressive conditional
heteroskedasticity models. From the foreign exchange market to the stock market there
was no mean spillover effect but there was bi-directional volatility spillover effects.
Further attempt was made by some studies to consider the link between the two
variables by using updated data. To that end, Alagidebe et al. (2011) used monthly data
from 1992 to 2005 for Australia, Canada, Japan, Switzerland and UK. Again, they found no
long-run relationship between the variables. Through Granger causality test it was found

that in Canada, Switzerland and UK, there is causal linkage from exchange rates to stock
prices and in Japan there is causality running from stock prices to exchange rates. In the
same line, Harjito and McGowan (2011) used weekly data from 1993 to 2002 for Indonesia,
the Philippines, Singapore and Thailand and reported evidence of bi-directional causality
in Thailand and Singapore. They also found cointegration between exchange rates and
stock prices and cointegration among the stock markets of all four countries.


Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

By using weekly data from 1999 to 2010 for the countries of Australia, New Zealand,
Japan, Switzerland, USA, UK and Euro Zone, Katechos (2011) examined the
relationship between stock markets and exchange rates in the light of the global equity
market returns. The method of maximum likelihood regression with GARCH was
applied and results showed that there is a link between the exchange rates and the
global stock market returns but the characteristics of the currencies determine the sign
of the relationship. The value of currencies with higher rates of interest is positively
related to global equity returns and the value of currencies with lower rates of interest
is negatively related to global equity returns. Larger is the interest rate differential
more is the explanatory power of the model. Sticking to weekly data, Lean et al. (2011)
used weekly data from 1990 to 2005 for Hong Kong, Indonesia, Japan, Korea, Malaysia,
the Philippines, Singapore and Thailand and examined the interactions of exchange
rates and stock prices by allowing for structural breaks. They applied the methods of
panel Lagrange Multiplier (LM) cointegration, Gregory-Hansen test for cointegration
and Granger causality test to find little evidence of long-run equilibrium relationship
between exchange rates and stock prices. Only in Korea, exchange rates and stock
prices were cointegrated. The predictive power of the two variables is limited only to
short run, though not for all countries. Again, using weekly data during 2000-2008
period, Lee et al. (2011) considered the experience of Indonesia, Korea, Malaysia, the
Philippines, Taiwan and Thailand to examine the relationship between the two

variables and the effect on their correlation due to stock market volatility. They used
the method of Smooth Transition Conditional Correlation GARCH model to show
that in. Indonesia, Korea, Malaysia, Thailand and Taiwan there are significant price
spillovers from stock market to foreign exchange market. Stock market volatility does
affect the correlation between the stock and foreign exchange markets. For all the
countries except for the Philippines, the correlation becomes higher when the stock
market becomes more volatile.
Using rolling regression analysis, Kollias et al. (2012) studied the link between the
two variables. The advantage of using rolling regression is, with the sample size
remaining same, at a time, the sample period moves forward by one observation.
Hence it takes into account of the new information available. They used daily data
from 2002 to 2008 for European countries and showed that there is no long-run
relationship between the two variables. The direction of causality depends on the
condition of the market. There is causality running from exchange rates to stock
prices under normal situation whereas causality might run from stock prices to
exchange rates during crisis situations.
Deviating from standard approaches, Tsai (2012) employed quantile regression
approach on monthly data from 1992 to 2009 for Singapore, Thailand, Malaysia, the
Philippines, South Korea and Taiwan. The method of quantile regression helps to
study the relationship under different market conditions (“different quantiles of
exchange rates”). Exchange rates and stock prices are negatively related when the
exchange rates are extremely high or low. So depending on the conditions of the
market, the relationship can change. Considering exchange rates other than against
the US dollar, Wickremasinghe (2012) examined the relationship between stock
prices and the Sri Lankan exchange rates against the Indian rupee, the Japanese yen,
the British pound and the US dollar. The results produced no evidence of any
long-run relationship between any of the four exchange rates and stock prices in
Sri Lanka. There was only evidence of unidirectional causality running from stock
prices to Sri Lankan exchange rate against US dollar. Through variance

Stock prices
and exchange
rates
711


JES
42,4

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

712

decomposition analysis it was inferred that most of the variance of the stock price is
explained by Indian rupee[2].
Most of the papers reviewed so far concentrated either on developed or on
developing countries. However, Buberkoku (2013) considered both developed and
developing countries and used monthly data from 1998 to 2008 to study the
relationship between stock prices and exchange rates for countries like Australia,
Canada, England, Germany, Japan, Singapore, South Korea, Switzerland and Turkey.
The methods used were Engle-Granger and Johansen cointegration test and Granger
causality test. The results showed that in the long run there is no relationship between
the variables in the considered countries, except for Singapore. In the short run, stock
prices affect exchange rates in Canada, Switzerland and Turkey. Causality runs from
exchange rates to stock prices for Singapore and South Korea. But for Australia,
England, Germany and Japan there was no causal relationship in either directions.
To test for sensitivity of the results to data frequency, Tsagkanos and Siriopoulos
(2013) used both daily and monthly data from 2008 to 2012 for European Union and
USA to study the relationship between the two variables during the financial crisis of
2008 to 2012. They applied methods of structural non-parametric cointegrating

regression, Johansen cointegration test and Granger causality test. They found that
movements in stock prices affect movements in exchange rates in EU in the long run
and in USA in the short run. Paying special attention to the crisis period, Caporale et al.
(2014) focussed on the banking crisis period of 2007-2010 to analyze the connections
between stock prices and exchange rates. For Canada, Euro area, Japan, Switzerland,
UK and USA, they used weekly data which were sub-divided into time periods: the precrisis period (2003-2007) and the crisis period (2007-2011). Using Bivariate UEDCCGARCH models they found that in the short run there is unidirectional Granger
causality from stock returns to exchange rate changes in the USA and the UK; in the
opposite direction in Canada, and for the Euro area and Switzerland there is
bi-directional causality. Causality-in-variance from stock returns to exchange rate
changes is found in the USA and for the Euro area and Japan it is in opposite direction,
while there is evidence of bi-directional feedback in Switzerland and Canada. During
the recent financial crisis, dependence between the two variables has increased.
Finally, in this bivariate models Yang et al. (2014) used daily data from 1997 to
2010 for India, Indonesia, Japan, Korea, Malaysia, the Philippines, Singapore, Taiwan
and Thailand to study the relationship between stock returns and exchange rates.
They applied Granger causality test in quantiles and they found that during the
Asian financial crisis, all the countries except for Thailand there are feedback
relations between exchange rates and stock prices and in Thailand, stock returns lead
exchange rates. The causal effects are heterogeneous across different quantiles and
different periods and most of the stock and foreign exchange markets are negatively
correlated.
The findings by bivariate studies reviewed above could be biased due to other
omitted variables from the models. The next group of studies try to address this issue
by including other macro variables in their model.
III. Multivariate models
Over the past recent years due to the trend of the globalization, capital flows among
different international markets has increased which has also led to the increase in close
relationship between the stock markets and the foreign exchange markets. Empirical
studies have also focussed on examining the effect of different macroeconomic

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

variables on stock market. The analysis of the effects of different macroeconomic
variables like inflation, GDP, industrial production index, money supply, oil prices,
interest rates, foreign capital, exchange rates, etc. are important since they can help
policy makers of an economy to better formulate policies. Investors find important and
interesting to see how and which variables cause the stock prices to fluctuate. Using an
APM model, Chen et al. (1986) examined the effect of different macroeconomic variables
(industrial production, inflation, risk premia, etc.) on the stock returns of the USA and
found that macroeconomic variables have significant effect on expected stock returns.
Fama and French (1993) examined four to five factors that affect stock returns and
stock prices.
Tian and Ma (2010) studied the relationships among stock prices and macroeconomic
variables like exchange rates, money supply, industrial production and consumer price
index using monthly data from 1995 to 2009 for China. They employed the ARDL
method of cointegration. Their results show that prior to financial liberalization of 2005,
no cointegration exists between the major foreign exchange rates and the Shanghai
stock price index. After the liberalization, cointegration exists. Money supply and
exchange rates affect stock prices with positive correlation in China and also previous
month CPI Granger causes stock prices. Using Johansen method of cointegration,
Chortareas et al. (2011), for countries like Egypt, Kuwait, Oman and Saudi Arabia
examined the role of oil prices as a link between the stock markets and exchange rates.
They used monthly data from 1994 to 2006 and their results show that when oil price is
not considered, there is no long run cointegration between exchange rates and stock
prices. Inclusion of oil prices show no cointegration between exchange rates and stock
prices when full sample period is considered. Before the oil price shock of 1999, no
cointegration among the variables was found. After the shock, exchange rates, stock
prices and oil prices are cointegrated in Egypt, Oman and Saudi Arabia. But for
Kuwait, there is long-run relationship only between stock prices and oil prices. Real

exchange rates are positively related to stock prices in Egypt and Oman and in
Saudi Arabia they are negatively related. Oil prices have long-run positive effect on
stock prices. The model employed by Liu and Tu (2011) included exchange rate and
foreign capital as determinants of stock prices. They used daily data from 2001 to 2007
from Taiwan to study the relationships among the variables and to analyze whether in
these markets the properties of asymmetric volatility switching and mean-reverting
exists or not. They found that the movements of the exchange rate and the stock price
index are affected by overbuying and overselling rates of foreign capital. All of the
three conditional means exhibit asymmetric mean-reverting behavior (negative returns
reverting quicker than positive returns). The volatility of the three markets exhibited
GARCH effects.
The model employed by Parsva and Lean (2011) included like interest rates, inflation
rates and oil prices as main determinants of stock prices in Egypt, Iran, Jordan, Kuwait,
Oman and Saudi Arabia. Using monthly data from 2004 to 2010 they estimated their
model using Johansen method of cointegration and Granger causality test. They found
that in the long run, all variables are cointegrated. Both in short run and long run there is
bi-directional causality between stock prices and exchange rates for Egypt, Iran and
Oman before the crisis. In Kuwait causality runs from exchange rates to stock prices in
the short run. Comparing the pre- and post-crisis periods, there was not much distinction
in the behavior of exchange rates and stock returns. Oil price was also included in a
model by Basher et al. (2012) who used monthly global data from 1988 to 2008 to examine
the relationship among stock prices in emerging markets. Additionally, they included

Stock prices
and exchange
rates
713


JES

42,4

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

714

global real economic activity as one of the variable which affects oil prices. Using a
structural VAR model and through the analysis of impulse response function they found
that positive shock to oil prices decreases emerging markets stock prices and US dollar
exchange rates in the short run. Exchange rates respond to changes in oil prices in the
short run, a positive shock to oil prices leads to decrease in trade-weighted exchange
rates. Oil price decreases with increase in oil production but a positive shock to real
economic activity increases the price of oil. Along similar lines, Eita (2012) employed
Johansen’s method and quarterly data from 1998 to 2009 for Namibia to examine the
determinants of stock prices. The results showed that stock prices are affected by
economic activity, exchange rates, inflation, interest rates and money supply. Stock
prices increase with increase in economic activity and money supply and stock prices
decrease with increase in inflation and interest rates. Exchange rates, GDP, money
supply and inflation move stock market away from equilibrium. Similarly, Inegbedion
(2012) considers the experience of Nigeria by using data from 2001 to 2009. By applying
Cochran-Orcutt Autoregressive Model, the results show that exchange rates and stock
prices are negatively related. The relationship of stock prices with interest rates and
inflation, respectively are not significant. But the joint effect of all the variables on stock
prices is significant[3].
Foreign reserves and interest rates were added as additional variables into the
relation between stock price and exchange rate to explore the effects of portfolio
adjustment by Lin (2012). Using monthly data during 1986-2010 and the ARDL
approach, the model was estimated for Asian emerging countries of India, Indonesia,
Korea, Philippines, Taiwan and Thailand. During crises periods, in terms of long run
cointegration and short-run causality, the co-movement between exchange rates and

stock prices became stronger. Spillover effect was mostly from stock price shocks to
exchange rates. Further analysis showed that the co-movement is generally driven by
capital account balance than the trade balance. Separately, Pakistan was the country of
concern by Aslam and Ramzan (2013) who studies the effects of the real effective
exchange rate index, CPI, per capita income and discount rate on the stock prices.
Applying NLS and ARMA techniques revealed that while discount rates and inflation
negatively affected Karachi stock price index, per capita income and real effective
exchange rate index affected positively. Discount rate impacted stock index the most.
This study helps to understand how effectively a country can control its
macroeconomic variables for better performance of the stock market. In the same
vein, commodity prices are introduced into the relation between the exchange rate and
stock market by Groenewold and Paterson (2013) who employed monthly data during
1979-2010 from Australia. Their results showed that when commodity prices are not
considered, there is no cointegration between exchange rates and stock prices. With the
inclusion of commodity prices, all the three variables are cointegrated in the long run.
When only exchange rates and stock prices are considered, there is no causality
between them in either direction. In the short run, exchange rates affect commodity
prices and commodity prices in turn affect stock prices.
Different macro variables in Pakistan were also considered by Khan et al. (2013) who
used monthly data from 1998 to 2008. The macroeconomic variables considered were
market returns, CPI, risk-free rate of return, industrial production and M2. The results
showed that both stock prices and exchange rates affect each other in the short run but
there is no long run association between the variables. In the long run, market return
and risk-free return are not related to stock prices but there is some association of
industrial production and stock prices. There exists both short run and long-run


Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

relationship between stock prices and inflation and money supply. Boonyanam (2014)

explored the relationship between different monetary variables with stock prices.
The monetary variables included were nominal bilateral exchange rate in terms of Baht
per US dollar, CPI, narrow money and 14 days repurchase rate and the methods used
were multivariate cointegration, VECM and variance decomposition analysis. Monthly
data from 1999 to 2012 was used for Thailand and the results show evidence of
long-run relationship between stock prices and monetary variables. In the short run,
narrow money and interest rate affect stock prices. There is one way causality from
exchange rates to stock prices and from interest rates to stock prices. There was also
positive relationship between CPI and stock price.
Rather than using stock prices, Moore and Wang (2014) examines the source of the
relationship between stock return differentials and real exchange rates using monthly
data for Australia, Canada, Indonesia, Japan, the Philippines, Malaysia, Singapore,
South Korea, Thailand and the UK. At the first stage, the dynamic conditional
correlation (DCC) is derived between the two variables and then the derived DCC is
used to regress on the interest rate differentials and the trade balance. With the help of
bivariate GARCH model with DCC they found that there is a negative relationship
between the relative stock prices and real exchange rates. There exists time-varying
correlation between stock return differentials and real exchange rate changes. The US
stock market influences the foreign exchange market and local stock market. Trade
balance is the major determinant of the dynamic correlation for Asian market and
interest rate differential is the key factor for developed countries. For the countries
where capital mobility is low, economic integration acts as the cause of the linkage and
thus it supports the flow-orientated model. But where capital mobility is more, financial
integration acts as the cause of the linkage which in turn favors the stock-oriented
model. Finally, the case of Turkey is considered by Tuncer and Turaboglu (2014) who
used quarterly data from 1990 to 2008 to examine the short run and long-run
relationships between stock prices and GDP, treasury bills rates and exchange rates.
They employed the method of Johansen test for cointegration to study the long-run
relationship and found evidence of long-run relationship between stock prices and the
other variables. In the short run, stock prices and real effective exchange rate affect

GDP but there is no causality relationship from treasury bills to GDP. There is
causality from real effective exchange rates to stock prices. All the variables do not
affect exchange rates in the short run hence exchange rate is comparatively an
exogenous variable.
In sum, the literature on the relation between stock prices and exchange rate is vast.
From the review of more recent studies it is clear that the link between the two
variables is dependent on the data frequency and period chosen, the countries studied,
and other macro variables, etc. But in general most of the papers concluded that in the
short run, stock prices and exchange rates are related but there is no relationship
between them in the long run. Other macroeconomic variables like, CPI (inflation rate),
interest rates, discount rates, oil prices, money supply, industrial production, GDP and
foreign capital also are found to affect stock prices. Table I provides the main features
of each study reviewed.
IV. A new direction for future research
The models reviewed in the previous section and all studies listed in Table I have one
common feature. They have all assumed that the effects of exchange rate changes on
stock prices are symmetric, i.e., if depreciation has positive effects on stock prices,

Stock prices
and exchange
rates
715


Gregory Hansen
Exchange rates
cointegration test,
Granger causality test

Exchange rates

Engle-Granger two
steps and Johansen
maximum likelihood
cointegration test,
Vector Error
Correction Model
(VECM)
Exchange rates
Engle-Granger two
steps, Johansen
cointegration method,
Granger causality test
Cointegration based on Exchange rates
OLS and Granger
causality test

Johansen cointegration Exchange rates
test and Granger
causality test

Granger et al. (2000)

Nieh and Lee (2001)

Phylaktis and
Ravazzolo (2005)

Lean et al. (2005)

Findings

Daily; January 3,
For Japan and Thailand, exchange rates affect stock
1986-November 14, prices (a positive relation). For Taiwan, stock prices affect
1997
the exchange rates (a negative correlation). Indonesia,
Korea, Malaysia and Philippines show feedback effect.
Singapore fails to show any pattern. Based on the
Granger causality, exchange rate affects stock prices in
eight of the nine countries
No long-run relationship between the variables. There is
Daily; October1,
1993-February 15, short-run relationship that lasts for only one day for
1996
certain G7 countries

Data period

(continued )

Daily; January 2,
No long-run equilibrium relationship between the two
1995-November 23, variables. Exchange rates affect stock prices in India and
2001
Sri Lanka, but no evidence of causality in Bangladesh and
Pakistan
Weekly; January 1, Before the Asian financial crisis, all of the countries
Hong Kong,
1991-December 31, except Philippines and Malaysia have no evidence of
Indonesia, Japan,

2002
Granger causality between stock prices and exchange
Singapore,
rates. No cointegration relationship between the two
Malaysia, Korea,
variables before or during the 1997 Asian crisis but after
Philippines and
the 9-11 terrorist attack some weaker cointegration
Thailand
relationship exists
Monthly; 1980-1998 US stock market is a causing variable which acts as a
Hong Long,
(depends on data
channel through which the foreign exchange market and
Malaysia,
stock market is linked. Positive association between the
Singapore, Thailand availability)
stock market and the real exchange rate
and Philippines

Bangladesh, India,
Pakistan and Sri
Lanka

Hong Kong,
Indonesia, Japan,
South Korea,
Malaysia,
Philippines,
Singapore,

Thailand, Taiwan
Canada, France,
Germany, Italy,
Japan, UK and the
USA

Countries

716

Smyth and Nandha
(2003)

Methodology

Table I.
Main features of
studies reviewed

References

Independent variables

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

JES
42,4


Yau and Nieh (2009)

Richards et al. (2009)

Rahman and Uddin
(2009)

Ismail and Isa (2009)

Pan et al. (2007)

Yau and Nieh (2006)

non-linear MS-VAR
model, Johansen
cointegration test
Johansen cointegration
test, Granger causality
test
VAR model, Johansen
cointegration test,
Granger causality test
TECM Granger
causality test,
EG-ES threshold
cointegration test
Australia

Japan and Taiwan

Exchange rates

Bangladesh, India
and Pakistan

Malaysia

Hong Kong, Japan,
Korea, Malaysia,
Singapore, Taiwan
and Thailand

Taiwan and Japan

New Zealand

Countries

Exchange rates

Exchange rates

Exchange rates

Granger causality test, Exchange rates
vector autoregressive
analysis, Johansen’s
maximum likelihood
method

Johansen cointegration Exchange rates

test, Granger causality
test
Johansen cointegration Exchange rates
test, non-linear test to
incorporate structural
break, Granger
causality test

Obben et al. (2006)

Independent variables

Methodology

References

Findings

Monthly; January
1991-March 2008

(continued )

Exchange rate and stock prices are cointegrated in the
long run. No short-run causal relationship between the
two countries financial assets. There is long run positive
causal relationship either from Japan or US exchange rate
to stock prices of Taiwan

Weekly; January

For short run and long run, found bi-directional causality
1st, 1999-2005/2006 between the five exchange rates and a couple of share
market indices
Monthly; January
No long-run relationship among the three variables (stock
1991-July 2005
prices of Japan and Taiwan and exchange rate of NTD/
Yen). Bi-directional causality between stock prices of
Taiwan and Japan. No significant short-run relationship
between the NTD/Yen exchange rate and the stock prices
of Taiwan or Japan
Daily; January
No long-run equilibrium relationship between exchange
1988-October 1998 rates and stock prices in few countries. Before the Asian
crisis, there is significant causal relationship from
exchange rates to stock prices for Hong Kong, Japan,
Malaysia and Thailand. During the crisis, all countries
except Malaysia show causal relationship from exchange
rates to stock prices
No long-run relationship between change in stock indices
Monthly; January
and exchange rate changes. Found evidence of common
1990-December
regime switching behavior between the variables
2005
Monthly; January
No long-run relationship between stock prices and
2003-June 2008
exchange rates. No way causal relationship between the
variables

Daily; January 2,
The variables are cointegrated in the long run. Causality
2003-June 30, 2006 exists from stock prices to exchange rates

Data period

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Stock prices
and exchange
rates
717

Table I.


Chortareas et al. (2011)

Exchange rates
Australia, Canada,
Johansen and
Japan, Switzerland
Saikkonen-Lutkepohl
and UK
cointegration test,
Granger causality test,
Hiemstra-Jones nonparametric causality
test
Johansen cointegration Exchange rates and oil Egypt, Oman, Saudi
test

prices
Arabia and Kuwait

Alagidede et al. (2011)

China

Exchange rates

VAR, multivariate
GARCH, Johansen
method of
cointegration

Zhao (2010)

China

Mexico

Countries

Exchange rates, M1,
CPI, industrial
production

Exchange rates

Independent variables

Monthly; 1994-2006
(varies with
countries
depending on data
availability)

Monthly; January
1992-December
2005

Monthly; January
1991-June 2009

Weekly; January
1989-December
2006
Monthly; December
1995-December
2009

Data period

(continued )

When oil price is not considered, no long-run
cointegration between exchange rates and stock prices.
Inclusion of oil prices show no cointegration between
exchange rates and stock prices when full sample period
is considered. Before the oil price shock of 1999, no
cointegration among the variables. After the shock,

exchange rates, stock prices and oil prices are
cointegrated in Egypt, Oman and Saudi Arabia. But for
Kuwait, there is long-run relationship only between stock

In the short-run stock prices Granger causes exchange
rates. No long-run relationship between these two
variables
Prior to financial liberalization of 2005, no cointegration
exists between the major foreign exchange rates and the
Shanghai stock price index. After the liberalization,
cointegration exists. Money supply and exchange rates
affect stock prices with positive correlation in China and
also previous month’s CPI Granger causes stock prices
No stable long-run equilibrium relationship between
Renminbi real effective exchange rate and stock price. No
mean spillover effect from foreign exchange market to
stock market. There exists bi-directional volatility
spillover effects between foreign exchange and stock
markets
No long-run relationship between exchange rates and
stock prices. In Canada, Switzerland and UK, there is
causal linkage from exchange rate to stock price. Weak
causality from stock price to exchange rate is found only
for Switzerland

Findings

718

Tian and Ma (2010)

Methodology

Granger causality test,
Engle-Granger method
of cointegration
ARDL method to
cointegration, ECM
model

Kutty (2010)

Table I.

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

JES
42,4


VAR ANST GARCHM Model

Liu and Tu (2011)

Exchange rates and
foreign capital

Taiwan

Findings

Daily; January 3,
2001-December 31,
2007

(continued )

There are significant price spillovers from stock market to
foreign exchange market in Indonesia, Korea, Malaysia,
Thailand and Taiwan. Stock market volatility affects the
correlation between stock market and foreign exchange
market
The movements of the exchange rate and the stock price
index are affected by overbuy and oversell rates of
foreign capital. All of the three conditional means exhibit
asymmetric mean-reverting behavior (negative returns
reverting quicker than positive returns). The volatility of
the three markets exhibits GARCH effects

prices and oil prices. Real exchange rates are positively
related to stock prices in Egypt and Oman and in Saudi
Arabia they are negatively related. Oil prices have longrun positive effect on stock prices
Weekly; January
Bi-directional causality is present in Thailand and
Singapore. Cointegration exists between stock prices and
1993-December
exchange rates. Cointegration exists among the stock
2002

markets in all the four countries
Weekly; January
There is a link between the exchange rates and the global
1999-August 2010 stock market returns, the sign depends on the nature of
the currencies. Value of higher yielding currencies is
positively related to global stock market returns. Value of
lower yielding currencies is negatively related to global
stock market returns
Weekly; January 1, Little evidence of a long-run equilibrium relationship
1990-June 30, 2005 between exchange rates and stock prices. The predictive
power of the two variables is restricted to the short run,
even then it does not hold for all countries

Data period

Weekly; January
Indonesia, Korea,
2000-April 2008
Malaysia,
Philippines, Taiwan,
Thailand

Exchange rates
Panel Lagrange
Multiplier (LM)
cointegration test,
Gregory-Hansen test
for cointegration,
Granger causality test
Bivariate STCCExchange rates

EGARCH model

Lean et al. (2011)

Lee et al. (2011)

Hong Kong,
Indonesia, Japan,
Korea, Malaysia,
Philippines,
Singapore, Thailand

Maximum likelihood
regression with
GARCH

Katechos (2011)

Exchange rates

Indonesia,
Philippines,
Singapore and
Thailand
Australia, New
Zealand, Japan,
Switzerland, USA,
UK, Euro Zone

Granger causality test, Exchange rates

Johansen cointegration
test

Countries

Harjito and McGowan
(2011)

Independent variables

Methodology

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Stock prices
and exchange
rates
719

Table I.


Table I.

Methodology

Independent variables

Structural VAR

Johansen test for
cointegration

Cochran-Orcutt
Autoregressive Model

Rolling Granger
causality test, rolling
cointegration test

Basher et al. (2012)

Eita (2012)

Inegbedion (2012)

Kollias et al. (2012)

Exchange rates

Exchange rates,
inflation rate, interest
rate

Exchange rates,
economic activity
(income-GDP), interest
rates, inflation, money

supply

Europe

Nigeria

Namibia

Exchange rates and oil Global analysis
prices. (and global real
economic activity)

Egypt, Iran, Jordan,
Kuwait, Oman,
Saudi Arabia

Countries

Findings

(continued )

In the long run, all the variables are cointegrated. Both in
short run and long run there is bi-directional causality
between stock prices and exchange rates in Egypt, Iran
and Oman before the crisis. In Kuwait causality runs from
exchange rates to stock prices in the short run. Between
the pre- and post-crisis periods, there was not much
distinction in the behavior of exchange rates and stock
returns

Through the analysis of impulse response function, it is
Monthly; January
1988-December
found that positive shock to oil prices decreases emerging
2008
markets stock prices and US dollar exchange rates in the
short run. Exchange rates respond to changes in oil prices
in the short run, a positive shock to oil prices leads to
decrease in trade-weighted exchange rates. Oil price
decreases with increase in oil production but a positive
shock to real economic activity increases the price of oil
Quarterly; 1998:Q1- Stock prices are affected by economic activity, exchange
2009:Q4
rates, inflation, interest rates and money supply. Stock
prices increase with increase in economic activity and
money supply and stock prices decrease with increase in
inflation and interest rates. Exchange rates, GDP, money
supply and inflation move stock market away from
equilibrium
2001-2009
Exchange rates and stock prices are negatively related.
The relationship of stock prices with interest rates and
inflation are not significant. But the joint effect of all the
variables on stock prices is significant
Daily; January
No long-run relationship between the variables. Exchange
2002-December
rate affects stock returns during normal times. During
2008
crisis, stock returns might affect exchange rates

Monthly; January
2004-September
2010

Data period

720

Parsva and Lean (2011) Johansen cointegration Exchange rates,
test, Granger causality interest rate, oil price
test
and inflation rate

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

JES
42,4


Quantile regression
approach

Tsai (2012)

Engle-Granger
cointegration test

NLS and ARMA
techniques

Abidin (2013)

Aslam and Ramzan
(2013)

(continued )

Discount rates and inflation negatively affect Karachi
stock price index. Per capita income and real effective
exchange rate index positively affect Karachi stock price
index. Discount rate impacts stock index the most

Annual; 1991-2012

Daily; January
2006-December
2008

No long-run relationship between exchange rates and
stock prices. Uni-directional causality runs from stock
prices to US dollar exchange rates only. No causality is
found when exchange rates for Indian rupee, Japanese
yen and UK pound are considered. From variance
decomposition analysis (to examine out-of-sample causal
relation), it is inferred that most of the variance of ASPI
(All Share Price Index) is explained by Indian rupee
No significant long-run relationship between stock

markets and exchange rates

Monthly; January
1986-December
2004

Monthly; January
1992-December
2009

During crises periods, in terms of long-run cointegration
and short-run causality, the co-movement between
exchange rates and stock prices became stronger.
Spillover effect is mostly from stock price shocks to
exchange rates. Analysis of industry causality showed
that the co-movement is generally driven by capital
account balance than that of trade. Volatilities of changes
in foreign reserves and interest rates are more during the
crisis and market liberalization period
Exchange rates and stock prices are negatively related
when the exchange rates are extremely high or low. The
relationship changes depending on the market conditions

Monthly; January
1986-December
2010

India, Indonesia,
Korea, Philippines,
Taiwan, Thailand

Singapore,
Thailand, Malaysia,
Philippines, South
Korea, Taiwan
Sri Lanka

Findings

Data period

Countries

Australia,
Hong Kong,
Indonesia, Japan,
New Zealand, South
Korea, Thailand
Real effective exchange Pakistan
rate index, CPI, per
capita income and
discount rate

Exchange rates

Exchange rates
Wickremasinghe (2012) Johansen’s
cointegration test,
Granger causality test,
variance

decomposition
analysis

Exchange rates

Granger causality test, Exchange rates,
interest rates and
ARDL method of
foreign reserves
cointegration

Lin (2012)

Independent variables

Methodology

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Stock prices
and exchange
rates
721

Table I.


Johansen test for

Exchange rates, market Pakistan
cointegration, Englereturn, risk-free rate of
Granger causality test return, CPI, industrial
production, M2

Exchange rates
Structural nonparametric
cointegrating
regression (SNCR),
Johansen test for
cointegration, Granger
causality test

Khan et al. (2013)

Tsagkanos and
Siriopoulos (2013)

(continued )

When commodity prices were not considered, no
cointegration between exchange rates and stock prices.
With the inclusion of commodity prices, all the three
variables are cointegrated in the long run. When only
exchange rates and stock prices are considered, there is
no causality between them in either direction. In the short
run, exchange rates affect commodity prices and
commodity prices in turn affect stock prices
Monthly; July 1998- Both stock prices and exchange rates affect each other in
June 2008

the short run but there is no long run association between
the variables. In the long run, market return and risk-free
return are not related to stock prices but there is some
association of industrial production and stock prices.
There exists both short- and long-run relationship
between stock prices and inflation and money supply
Daily, Monthly;
Movements in stock prices affect movements in exchange
January 2, 2008rates in EU in the long run and in USA in the short run
April 30, 2012

No long-run relationship between the variables in the
considered countries except for Singapore. In the short run,
stock prices affect exchange rates in Canada, Switzerland
and Turkey. In the short run, exchange rates affect stock
prices in Singapore, South Korea. No causal relationship
exists in Japan, Germany, England and Australia

Findings

722

European Union
(EU), USA

Exchange rates and
Johansen test for
cointegration, Granger commodity prices
causality test

Groenewold and
Paterson (2013)

Monthly; April
Japan, Canada,
1998-April 2008
England,
Switzerland,
Germany, Australia,
Singapore,
South Korea,
Turkey
Australia
Monthly; December
1979-December
2010

Exchange rates

Johansen and EngleGranger cointegration
tests, Granger
causality test

Buberkoku (2013)

Data period

Countries

Independent variables

Methodology

Table I.

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

JES
42,4


Multivariate
cointegration, VECM,
Granger causality
(Wald test/block
exogeneity test),
variance
decomposition
Bivariate VARGARCH model, EngleGranger and Johansen
trace test for
cointegration, Granger
causality test

Moore and Wang (2014) Bivariate GARCH
model with dynamic
conditional correlation
(DCC), linear
regression

Caporale et al. (2014)

Boonyanam (2014)

Exchange rates, trade
balance, real interest
rate differential,
measures of financial
development

There is long-run relationship among stock prices,
exchange rates and oil prices. In the long run, exchange
rates and oil prices Granger cause stock prices but oil
prices and stock prices do not affect exchange rates. In the
short run, there is bi-directional causality between oil
prices and stock prices
Long-run relationship exists between stock prices and
monetary variables. In the short run, narrow money and
interest rate affect stock prices. There is one way
causality from exchange rates to stock prices and from
interest rates to stock prices. Positive relation between
CPI and stock price

Findings

(continued )

Using Bivariate UEDCC-GARCH models they found that
in the short run there is uni-directional Granger causality

from stock returns to exchange rate changes in USA and
UK; in the opposite direction in Canada, and for the Euro
area and Switzerland there is bi-directional causality.
Causality-in-variance from stock returns to exchange rate
changes is found in the USA and in the Euro area and in
Japan it is in opposite direction, while there is evidence of
bi-directional feedback in Switzerland and Canada.
During the recent financial crisis, dependence between the
two variables has increased
There is a negative dynamic relationship between the
Monthly; for
Australia, Canada,
developed country- relative stock prices and real exchange rates. There exists
Japan, UK,
Indonesia, Malaysia, 1973-2006 and for time-varying correlation between stock return
emerging markets differentials and real exchange rate changes. US stock
South Korea,
market influences the economies foreign exchange and
depends on data
Philippines,
local stock markets. Trade balance is the major
Singapore, Thailand availability.
determinant of the dynamic correlation for Asian market
(focussing on the

Weekly; August 6,
2003-December 28,
2011

Exchange rates

USA, UK, Canada,
Japan, the Euro
Area, Switzerland

Monthly; January
1999-December
2012

Thailand
Exchange rates, CPI,
narrow money, 14 days
re-purchase rates

Data period

Panel cointegration
Exchange rates and oil Indonesia, Malaysia, Monthly; January
(Engel-Granger),
prices
the Philippines,
2006-December
Granger causality test
Singapore, Thailand 2012

Countries

Unlu (2013)

Independent variables

Methodology

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Stock prices
and exchange
rates
723

Table I.


Table I.

Methodology

Yang et al. (2014)

India, Indonesia,
Japan, Korea,
Malaysia,
Philippines,
Singapore, Taiwan,
Thailand

Turkey

Exchange rates, GDP
and treasury bills rate

Granger causality test Exchange rates
in quantiles

Countries

Independent variables

Daily; January 1,
1997-August 16,
2010

and interest rate differential is the key factor for
developed countries

floating/managed
floating regime
period)
Quarterly; 1990:Q12008:Q2

There is long-run relationship between stock prices and
the other variables. In the short run, stock prices and real
effective exchange rates affect GDP but there is no
causality relationship from treasury bills to GDP. There is
causality from real effective exchange rates to stock
prices. All the variables do not affect exchange rates in
the short run
During the Asian financial crisis, all the countries except

for Thailand there are feedback relations between
exchange rates and stock prices and in Thailand, stock
returns lead exchange rates. The causal effects are
heterogeneous across different quantiles and different
periods. Most stock and foreign exchange markets are
negatively correlated

Findings

Data period

724

Tuncer and Turaboglu Johansen test for
(2014)
cointegration,
multivariate Vector
Error Correction
models (VEC)

References

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

JES
42,4


Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

appreciation has the opposite effect. This need not be the case. When domestic
currency appreciates (or foreign currency depreciates), cost of imported inputs decline
resulting in more profit and hence, a positive effect on stock prices. However, when
domestic currency depreciates which results in an increased cost of imported input,
in order to maintain their market share, some domestic producers may absorb the
increased cost by giving up their profit margin. In this case, stock prices may not react
to depreciation, implying that the effects of exchange rate changes on stock prices
could be asymmetric. In what follows, we pick a standard model from the literature and
demonstrate how asymmetry could be introduced and tested within existing
framework and with the help of recent advances in time-series modeling.
Let us consider the following long-run specification used by previous research
(e.g. Boonyanam 2014; Moore and Wang, 2014):
LnSP t ¼ a þ bLnEX t þ cLnI PI t þ dLnCPI t þ eLnM 2t þ et

(1)

In Equation (1) in addition to nominal effective exchange rate (EX ), a measure of output
proxied by Index of Industrial Production (IPI ), the price level measured by the
Consumer Price Index (CPI ), and the money supply measured by nominal M2 are
identified to be the determinants of stock prices, SP. Estimate of Equation (1) by any
method will result in the long-run coefficient estimates. In order to infer the short-run
effects, the common practice is to specify Equation (1) in an error-correction format as
in Equation (2):
DLnSP t ¼ a þ
þ

n1
X
k¼1
n4

X
k¼0

bk DLnSP tÀk þ

n2
X

yk DLnCPI tÀk þ

dk DLnEX tÀk þ

k¼0
n5
X

n3
X

Fk DLnI PI tÀk

k¼0

pk DLnM 2tÀk þ l1 LnSP tÀ1

k¼0

þ l2 LnEX tÀ1 þ l3 LnI PI tÀ1 þ l4 LnCPI tÀ1 þ l5 LnM 2tÀ1 þ mt

(2)

The error-correction model outlined by Equation (2) follows Pesaran et al. (2001) ARDL
approach to cointegration. Variables are said to be cointegrated if linear combination of
lagged-level variables as a proxy for lagged error term from Equation (1) are jointly
significant. Pesaran et al. (2001) propose applying the F-test with new critical values
that they tabulate. If lagged-level variables are jointly significant, estimates of λ2-λ5
normalized on λ1 will yield the long-run effects of exogenous variables on stock prices.
The short-run effects are then inferred by the coefficient estimates attached to firstdifferenced variables[4].
Specification Equation (1) or (2) that are now widely used in this area or other
areas in economics assume effects of any of the exogenous variables are symmetric.
Concentrating on the exchange rate, we try to test this by separating depreciations
from appreciations to determine whether these new variables have the same effects.
To this end, following recent studies in other areas, e.g. Apergis and Miller (2006),
Delatte and Lopez-Villavicencio (2012), Verheyen (2013), Bahmani-Oskooee and
Fariditavana (2014) and Bahmani-Oskooee and Bahmani (2015) we decompose the
movement of the LnEX variable into its negative (depreciation) and positive
þ
(appreciation) partial sums as LnEX ¼ LnEX 0 þ LnEX tþ þ LnEX À
t where LnEX t
À
and LnEX t are the partial sum process of positive and negative changes in LnEX.

Stock prices
and exchange
rates
725


JES
42,4

More precisely:
POS ¼ LnEX tþ ¼

t
t
X
X
À
Á
DLnEX jþ ¼
max DLnEX j ; 0 ;
j¼1

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

726

j¼1

t
t
X
X
À
Á
DLnEX À
min DLnEX j ; 0
N EG ¼ LnEX À
t ¼

j ¼
j¼1

(3)

j¼1

The next step is to go back to error-correction Model (2) and replace LnEX by POS and
NEG variables as in Equation (4):
DLnSP t ¼ a þ

n1
X

bk DLnSP tÀk þ

k¼1

þ

n4
X
k¼0

Fk DLnI PI tÀk þ

n2
X

d1;k DPOS tÀk þ

k¼0
n5
X

n3
X

d2;k DN EGtÀk

k¼0

yk DLnCPI tÀk þ

k¼0

n6
X

pk DLnM 2tÀk

k¼0

þ l1 LnSP tÀ1 þ l2 POS tÀ1 þ l3 N EGtÀ1 þ l4 LnI PI tÀ1
þ l5 LnCPI tÀ1 þ l6 LnM 2tÀ1 þ mt

(4)

Specification Equation (4) is usually called non-linear ARDL model proposed by Shin
et al. (2014) who have demonstrated that Pesaran et al.’s (2001) bounds testing approach

is equally applicable to Equation (4). Non-linearity is introduced through partial sum
or cumulative sum concept included in generating the new variables POS and NEG.
Once Equation (4) is estimated, estimates of δ1,k and δ2,k will be used to judge short run
symmetry or asymmetry effects of exchange rate changes and estimates of normalized
λ2 and λ3 will be used for judging the long-run symmetry or asymmetry.
For demonstrative purpose, we estimate both the linear and non-linear ARDL models
(i.e. Equations (2) and (4)) using monthly US data over the period 1973M1-2014M3[5].
A maximum of 6 lags is imposed on each first-differenced variable and Akainke’s
Information Criterion is used to select the optimum model. The results are reported
in Table II.
For both models, there are three panels. While Panel A reports the short-run
estimates, Panel B reports the long-run estimates. Finally diagnostic statistics are
reported in Panel C. From the estimates of the linear model and panel A we gather that
for each first-differenced variable there is at least one coefficient that is significant at
the 10 percent level, except the money supply. Concentrating on our variable of
concern, dollar depreciation seems to result in an increase in stock prices in the short
run. However, this short-run effect is not translated into the long run since normalized
long-run coefficient from Panel B is not significant. Indeed, none of the variables are
significant in the long run. This is also reflected by the low and insignificant value of
the F-test[6]. A few additional diagnostics are also reported in Panel C. Using
normalized long-run estimates from Panel B and Equation (1) we calculate the error
term and call it error correction term, ECM. After replacing the linear combination of
lagged-level variables by ECMt-1, we re-estimate the model after imposing the same
optimum lags. A significantly negative coefficient will support convergence and speed
with which variables adjust. Clearly, 2 percent of the adjustment takes place within


Stock prices
and exchange
rates

Section I: estimates of linear Model (2)

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

Panel A: short run
Variables
0
ΔlnSPt
ΔlnEXt
−0.50
ΔlnIPIt
−0.26
ΔlnCPIt
−1.47
ΔlnM2t
−0.02

Lags
1
(3.80)
(0.84)
(2.29)
(1.15)

Panel B: long run
lnEX
lnIPI
−1.05 (0.99)
1.10 (0.71)

2

4

5

727

0.65 (2.21)

lnCPI
2.71 (1.67)

Panel C: diagnostics
F
ECMt-1
LM
2.24
−0.02 (3.36)
12.13
Section II: estimates of non-linear Model (4)
Panel A: short
Variables
ΔlnSPt
ΔPOSt
ΔNEGt
ΔlnIPIt
ΔlnCPIt
ΔlnM2t

3

lnM2
−1.05 (1.08)

Constant
25.95 (0.98)

RESET
17.64

Adjusted R2
0.06

CUSUM (CUSUM2)
S (S)

lnCPI
2.78 (2.77)

lnM2
−2.39 (2.25)

Constant
58.42 (2.03)

RESET
10.58

Adjusted R2
0.08

CUSUM (CUSUM2)
S (S)

run
−1.07
−0.02
−0.21
−1.43
−0.09

(4.71)
(1.15)
(0.69)
(2.23)
(2.17)

0.62 (2.12)

Panel B: long run
POS
NEG
0.64 (0.77)
−0.63 (1.08)

lnIPI
0.84 (0.82)

Panel C: diagnostics
F
ECMt-1
2.69
−0.04 (4.03)

LM
11.49

one month. The LM statistic is also reported to test for autocorrelation. It has a χ2
distribution with 12 degrees of freedom since data are monthly. Given its critical value
of 21.03 at the 5 percent level, the LM statistic is insignificant, implying absence of
serial correlation in the optimum model. We have also reported Ramsey’s RESET
statistic to judge misspecification. This statistic is also distributed as χ2 but with only
one degree of freedom. Clearly, it is significant indicating a misspecified model[7].
Finally, to establish stability of short- and long-run coefficient estimates we apply the
well-known CUSUM and CUSUMSQ tests to the residuals of the optimum model. All
coefficients are stable and this is indicated by “S”[8].
Next we move to estimate of non-linear ARDL model outlined by Equation (4). These
results are reported in Section 2 of Table II. From the short-run results in Panel A, we
gather that while ΔPOS variable carries a negative and significant coefficient, ΔNEG
variable does not. This supports the fact that exchange rate changes have short-run
asymmetric effects on stock prices in the USA. More precisely, when dollar appreciates

Table II.
Full-information
estimates of
Models (2) and (4)
for the USA

JES
42,4

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

728

(or foreign currencies depreciate in terms of dollar), the balance sheet of US
multinational firms deteriorate in terms of the dollar and this exerts adverse negative
effect on stock prices of multinationals and eventually on overall stock price index in
the USA. All other variables also seem to have short-run effects on the stock prices
in the USA. Thus, the non-linear model seems to yield more significant short-run
results than the linear model. This is also the case in the long run. From the long-run
results in Panel B, the CPI and M2 carry significant coefficient. The variables are not
cointegrated by the F-test but they are by the ECMt-1.
V. Summary and conclusion
It is not too difficult to link macro variables to each other and try to understand the
feedback effects that exist among them and the link between stock market and foreign
exchange market in every country is no exception. An increase in stock prices over time
is said to increase the wealth leading to an increase in the demand for money, hence
interest rates. High interest rates in turn can attract foreign capital, causing currency
appreciation. When domestic currency appreciates or foreign currency depreciates, the
balance sheet of domestic multinational firms deteriorates in terms of domestic
currency which could be bad news for their shareholders and their share prices. On the
other hand appreciation of domestic currency or depreciation of foreign currency could
be good news for domestic produces due to cheap imported inputs. This will have a
favorable impact on these firms share prices. Overall, exchange rate changes can move
stock prices in either direction.
The main purpose of this paper is to review the literature pertaining to the relation

between stock prices and exchange rates. The literature was divided into two groups.
In the first group we reviewed studies that investigate the link between the two
variables without including other variables in their model. The second group includes
studies that they consider multivariate models by including other variables. No matter
which group we consider, the overall conclusion is that the findings are sensitive to
frequency of data used, study period chosen, the country considered, and other macro
variables included such as inflation, money supply, domestic production, capital flows,
etc. However, a general conclusion is that any relation that exists between the two
markets is short run. In most studies, no long-run relationship was found between
stock prices and exchange rates. A table which summarizes the main features of each
study is also provided.
In addition to reviewing the literature, we also identified one shortcoming of all studies
and proposed a direction for future research. All studies in the literature have assumed
that the effects of exchange rate changes on stock prices are symmetric. Using the US
data and non-linear ARDL approach we demonstrated that this may not be the case.

Notes
1. Franck and Young (1972) and Ang and Ghallab (1976) are two studies that have employed
information from pre-1973 period and investigated the effects of devaluation on stock prices.
No uniform results were found by these studies.
2. Again using updated daily data from 2006 to 2008 for Australia, Hong Kong, Indonesia,
Japan, New Zealand, South Korea and Thailand Abidin (2013) employed Engle-Granger
cointegration test and showed no evidence of long run cointegration relationship between
stock markets and exchange rates.


Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

3. Unlu (2013) is another study that considers oil prices in studying the link between exchange
rates and stock prices using monthly data from 2006 to 2012 for countries like Indonesia,

Malaysia, the Philippines, Singapore and Thailand. Using panel cointegration and Granger
causality tests, Unlu finds evidence of long-run relationship among stock prices, exchange
rates and oil prices. In the long run, exchange rates and oil prices Granger cause stock
prices but oil prices and stock prices do not affect exchange rates. In the short run, there is
bi-directional causality between oil prices and stock prices.
4. Note that the advantage of this method is that the critical values that are tabulated by
Pesaran et al. (2001) account for integrating properties of all variables. Indeed, they
demonstrate that under this method variables could be I(0), I(1) or combination of the two.
For more on this and normalization procedure see Bahmani-Oskooee and Tanku (2008).
5. All data come from the International Financial Statistics of the IMF except the S & P 500
index and the nominal effective exchange rate. The former comes from Yahoo Finance and
the latter from BIS.
6. The upper bound critical value of the F-statistic when there are four exogenous variables is
4.02 at the 5 percent significance level and 3.52 at the 10 percent significance level. These
figures come from Pesaran et al. (2001, Table CI (iii) Case III on page 300).
7. Critical value is 3.84 at the usual 5 percent significance level.
8. For a graphical presentation of the CUSUM and CUSUMSQ tests see Bahmani-Oskooee et al.
(2005).

References
Abidin, S. (2013), “Cointegration between stock prices and exchange rates in Asia-Pacific
countries”, Investment Management & Financial Innovations, Vol. 10 No. 2, pp. 142-146.
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. 2 No. 3, pp. 7-12.
Alagidede, P., Panagiotidis, T. and Zhang, X. (2011), “Causal relationship between stock prices
and exchange rates”, The Journal of International Trade & Economic Development: An
International and Comparative Review, Vol. 20 No. 1, pp. 67-86.
Ang, J.S. and Ghallab, A. (1976), “The impact of U.S. devaluations on the stock prices of
multinational corporations”, Journal of Business Research, Vol. 4 No. 1, pp. 25-34.
Apergis, N. and Miller, S. (2006), “Consumption asymmetry and the stock market: empirical

evidence”, Economics Letters, Vol. 93 No. 3, pp. 337-342.
Aslam, M.T. and Ramzan, M. (2013), “Impact of consumer price index, real effective exchange rate
index, per capita income and discount rate on Pakistan’s stock market index”, International
Journal of Research in Commerce, Economics and Management, Vol. 3 No. 5, pp. 10-14.
Bahmani-Oskooee, M. and Bahmani, S. (2015), “Nonlinear ARDL approach and the demand for
money in Iran”, Economics Bulletin, Vol. 35 No. 1, pp. 381-391.
Bahmani-Oskooee, M. and Fariditavana, H. (2014), “Do exchange rate changes have symmetric
effect on the S-curve?”, Economics Bulletin, Vol. 34 No. 1, pp. 164-173.
Bahmani-Oskooee, M. and Sohrabian, A. (1992), “Stock prices and the effective exchange rate of
the dollar”, Applied Economics, Vol. 24 No. 4, pp. 459-464.
Bahmani-Oskooee, M. and Tanku, A. (2008), “The black market exchange rate vs the official rate
in testing ppp: which rate fosters the adjustment process”, Economics Letters, Vol. 99 No. 1,
pp. 40-43.

Stock prices
and exchange
rates
729


JES
42,4

Bahmani-Oskooee, M., Economidou, C. and Goswami, G.G. (2005), “How sensitive are Britain’s
inpayments and outpayments to the value of the British pound”, Journal of Economic
Studies, Vol. 32 No. 6, pp. 455-467.
Basher, S.A., Haug, A.A. and Sadorsky, P. (2012), “Oil prices, exchange rates and emerging stock
markets”, Energy Economics, Vol. 34 No. 1, pp. 227-240.

Downloaded by UNIVERSITY OF ECONOMICS HO CHI MINH At 18:48 14 January 2018 (PT)

730

Boonyanam, N. (2014), “Relationship of stock price and monetary variables of Asian small open
emerging economy: evidence from Thailand”, International Journal of Financial Research,
Vol. 5 No. 1, pp. 52-63.
Buberkoku, O. (2013), “The relationship between stock prices and exchange rates: evidence from
developed and developing countries”, ISE Review, Vol. 13 No. 52, pp. 1-16.
Caporale, G.M., Hunter, J. and Ali, F.M. (2014), “On the linkages between stock prices and
exchange rates: evidence from the banking crisis of 2007-2010”, International Review of
Financial Analysis, Vol. 33, pp. 87-103.
Chen, N.F., Roll, R. and Ross, S.A. (1986), “Economic forces and the stock market”, The Journal of
Business, Vol. 59 No. 3, pp. 383-403.
Chortareas, G., Cipollini, A. and Eissa, M.A. (2011), “Exchange rates and stock prices in the
MENA countries: what role for oil?”, Review of Development Economics, Vol. 15 No. 4,
pp. 758-774.
Delatte, A.-L. and Lopez-Villavicencio, A. (2012), “Asymmetry exchange rate pass-through:
evidence from major countries”, Journal of Macroeconomics, Vol. 34 No. 3, pp. 833-844.
Eita, J.H. (2012), “Modelling macroeconomic determinants of stock market prices: evidence from
Namibia”, Journal of Applied Business Research, Vol. 28 No. 5, pp. 871-884.
Engle, R.F. and Granger, C.W.J. (1987), “Cointegration and error correction: representation,
estimation, and testing”, Econometrica, Vol. 55 No. 2, pp. 251-276.
Fama, E.F. and French, K.R. (1993), “Common risk factors in the returns on stocks and bonds”,
Journal of Financial Economics, Vol. 33 No. 1, pp. 3-56.
Franck, P. and Young, A. (1972), “Stock price reaction of multinational firms to exchange
realignments”, Financial Management, Vol. 1 No. 3, pp. 66-73.
Granger, C.W.J., Huang, B.N. and Yang, C.W. (2000), “A bivariate causality between stock prices
and exchange rates: evidence from recent Asian flu”, The Quarterly Review of Economics
and Finance, Vol. 40 No. 3, pp. 337-354.
Groenewold, N. and Paterson, J.E.H. (2013), “Stock prices and exchange rates in Australia:

are commodity prices the missing link?”, Australian Economic Papers, Vol. 52 Nos 3/4,
pp. 150-170.
Harjito, D.A. and McGowan, C.B. Jr (2011), “Stock price and exchange rate causality: the case of
four ASEAN countries”, Southwestern Economic Review, Vol. 34 No. 1, pp. 103-114.
Inegbedion, H.E. (2012), “Macroeconomic determinants of stock price changes: empirical evidence
from Nigeria”, Indian Journal of Finance, Vol. 6 No. 2, pp. 19-23.
Ismail, M.T. and Isa, Z.B. (2009), “Modeling the interactions of stock price and exchange rate in
Malaysia”, The Singapore Economic Review, Vol. 54 No. 4, pp. 605-619.
Katechos, G. (2011), “On the relationship between exchange rates and equity returns:
a new approach”, Journal of International Financial Markets, Institutions and Money,
Vol. 21 No. 4, pp. 550-559.
Khan, F., Muneer, S. and Anuar, M.A. (2013), “Relationship between stock prices and
economic variables: sectoral analysis”, Actual Problems of Economics, Vol. 143 No. 5,
pp. 544-553.