Sosyoekonomi

ISSN: 1305-5577
DOI: 10.17233/sosyoekonomi.2024.01.09
RESEARCH
Date Submitted: 14.07.2023
ARTICLE
Date Revised: 10.01.2024
2024, Vol. 32(59), 207-222 Date Accepted: 14.01.2024

Effects of FX on ETF Prices: Evidence from BIST

Tayfur BAYAT (https://orcid.org/0000-0002-4427-0999), İnönü University, Türkiye; tayfur.bayat@inonu.edu.tr

Altuğ Murat KÖKTAŞ (https://orcid.org/0000-0002-0911-2143), Necmettin Erbakan University, Türkiye;
altugmkoktas@gmail.com
Selim KAYHAN (https://orcid.org/0000-0002-4838-6892), Necmettin Erbakan University, Türkiye;
skayhan@erbakan.edu.tr
Gökhan KONAT (https://orcid.org/0000-0002-0964-7893), Bolu Abant İzzet Baysal University, Türkiye;
gokhan.konat@inonu.edu.tr

Döviz Kurunun BYF’ler Üzerindeki Etkileri: BİST’ten Kanıtlar

Abstract
This study better investigates the possible relationship between exchange rates and ETF prices
in the BIST to understand ETF investors' behaviour in the Turkish economy. Conventional and
Fourier-based co-integration and causality analysis methods were employed to test models. According
to findings, although the exchange rate has no direct effect on ETF prices in Türkiye, it is effective on
ETF prices indirectly via the risk and share of foreign investors. The originality of the study lies in
models built with additional control variables. In doing so, we measure the direct and indirect effects
of the exchange rate on the Turkish economy.
Keywords : ETF, MSCI Türkiye, Fourier Toda-Yamamoto Causality Test.
JEL Classification Codes : F31, G12, G23.
Öz
Bu çalışma, Türkiye ekonomisinde BYF (Borsa Yatırım Fonu) yatırımcılarının davranışlarını
daha iyi anlamak için döviz kurları ile BİST'teki BYF fiyatları arasındaki olası ilişkiyi araştırmaktadır.
Modelleri test etmek için geleneksel ve Fourier tabanlı eş-bütünleşme ve nedensellik analizi yöntemleri
kullanılmıştır. Elde edilen bulgulara göre, döviz kurunun Türkiye'deki BYF fiyatları üzerinde
doğrudan bir döviz kuru bulunmamakla birlikte, yabancı yatırımcıların riski ve payı aracılığıyla BYF
fiyatı üzerinde dolaylı bir etkisi bulunmaktadır. Bu çalışmanın özgünlüğü ek kontrol değişkenleri ile
oluşturulan modellerde yatmaktadır. Kurulan model sayesinde döviz kurunun Türkiye ekonomisindeki
doğrudan ve dolaylı etkilerini ölçme şansı oluşmuştur.
Anahtar Sözcükler : BYF, MSCI Türkiye, Fourier Toda-Yamamoto Nedensellik Testi.
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

1. Introduction

ETF is a basket containing securities traded on an exchange like a single stock

(Geetha et al., 2020: 356). ETFs aim to replicate the performance of their underlying indices
(Williams, 2014: 392). Ramaswamy (2011) expresses the structure of ETFs in two steps.
First of all, the market maker buys shares of securities from the stock market. They build a
basket of securities through an ETF sponsor who creates shares for the ETF. In the following
step, ETFs are supplied to investors in the secondary market. From a different perspective,
ETFs are open-end funds comprising securities assembled according to an investment
objective and strategy.

When the ETFs are examined historically, the first examples can be seen in the early
1990s. Although Williams (2014) argues that the first launch was in 1996, according to
Geetha et al. (2020) and Shin and Soydemir (2010), the first ETF launched in 1993 was the
S&P Depository Receipts (SPDR). It tracks S&P500 stock indices. As of 2020, the global
volume of ETFs is around $7 trillion, and they are invested in equity, bonds, commodities,
and currencies (Todorov, 2021: 41). Increasing ETFs has made ETFs an important financial
tool. In this regard, it becomes important to investigate possible determinants of the price
and demand of ETFs in an economy.

The exchange rate is also an important indicator for countries with high current
account deficits where a fluctuating exchange rate regime is implemented. According to
Yıldız (2014), economic theory argues that changes in exchange rates affect the cash flows,
investments, and profitability of firms by making their inputs and outputs cheaper or more
expensive and have a significant effect on the stocks of the firms (Yıldız, 2014: 77). In this
regard, it is rational to think that exchange rate is effective on ETFs either.

Economists examine the relationship between stock prices and exchange rates to
understand better how they interact. Because the definition of relationship is crucial for
decision-makers responsible for monetary and fiscal policies, firms taking exchange rate
risks in their businesses, and investors in stock markets. However, the findings in the
literature need to be more conclusive. While results for stock prices are inconclusive, another
stock market tool, exchange-traded funds, needs to be investigated better in the context of
exchange rate relations.

Increasing volume and price efficiencies of ETFs bring a question into mind: "Is the
possible relationship between stock prices and exchange rates valid for ETFs, or does price
efficiency and increasing volume of ETFs due to the advantages of ETFs differentiate the
strength and direction of the relationship?”. These questions will also help to understand the
behaviour of ETF investors and whether they employ ETFs in the same way as single stock
shares. To answer these questions, we investigate the MSCI Türkiye ETF and Turkish lira
against the U.S. dollar between 2008 and 2022. In this study, we employ the fractional
frequency Engle-Granger co-integration test and fractional frequency Fourier Toda-

208
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

Yamamoto causality test in addition to conventional co-integration and causality tests to

better understand the relation in the case of Türkiye.

A theoretical framework is drawn in the next section, and a possible relation between
the exchange rate and stock and/or ETF price is summarised. In the third section, a literature
review is presented. In the fourth section, the model is built, and data belonging to the model
variables are presented. In the fifth section, empirical findings are summarised and
interpreted. At the end of the study, political and theoretical implications are made.

2. Theoretical Background

When a literature survey was made, it was seen that many researchers had
investigated the relationship between stock price and exchange rate volatility. Although
there has yet to be a consensus about the direction and strength of the relation, alternative
theoretical explanations are made in different ways. According to Zhao (2010), the linkage
between exchange rates and stock prices has taken two forms. One is the “flow-oriented”
model, which Dornbusch and Fischer (1980) suggested. According to them, changes in
exchange rates affect international competitiveness and trade balances. Another is the
“stock-oriented” model, which was indicated by Branson (1983) and Frankel (1983). This
model views the exchange rate as equating the supply and demand for assets such as stocks.
This approach determines exchange rate dynamics by giving the capital account a vital role
(Zhao, 2010: 104).

Yıldız (2014) classifies the possible relationship between exchange rates and stock
prices into four groups in the light of models suggested by Dornbusch and Fischer (1980),
Branson (1983) and Frankel (1983). These are conventional, portfolio balance, stock, and
asset market approaches.

Positive uni-directional causality running from exchange rate to stock prices can
occur, according to Naeem and Rashed (2004). A depreciation in local currency leads to an
increase in the competitiveness of local firms. Increasing competition would raise firms'
export volume. In the end, the value of firms will increase, and stock prices will rise. Of
course, this is valid only for firms with export capability (Naeem & Rasheed, 2004: 536). In
the case of importing firms which have been listed on the stock exchange, causation linkage
would be the opposite. The positive relation is called the “Conventional approach”.
According to this approach, volatility in the exchange rate would also affect firms’ future
payments in terms of foreign currency and business risk, even if it exports or imports (Yıldız,
2014: 79).

According to Naeem and Rashed (2004), the portfolio balance approach implies a
negative relationship between stock indices and exchange rates, and uni-directional causality
runs from stock price indices to exchange rates. Namely, individuals hold both domestic and
foreign assets in their portfolios. It includes stock indices, equities, bonds, and currencies.
With an increase in domestic stock prices, demand for it would rise. The investors would

209
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

exchange foreign and domestic assets to buy more stock indices. In the end, the local
currency would be appreciated. Moreover, increasing local asset prices induce an increase
in money demand. Increasing money demand would bring interest rate raises together.
Again, high interest rates would attract foreign capital, and capital inflows would appreciate
local currency due to the rising foreign currency supply. These two approaches are related
to “flow-oriented” models, as Zhao (2010) indicated.

The stock approach implies bi-directional causality between the exchange rate and
stock price. The exchange rate is a factor that balances the supply and demand of financial
assets. Since the values of capital assets are evaluated over the present value of future cash
flows, expectations regarding exchange rates play an important role in their price
fluctuations. Therefore, stock prices can affect or be affected by exchange rate dynamics
(Yıldız, 2014: 79). Moreover, any change in interest rate which affects binary
simultaneously, there might be a bi-directional causality.

The last approach, “Asset Market,” claims weak or no relationship between variables.
A depreciation in local currency might make exporting firms more competitive, but if they
import raw materials, that will increase costs and prices. That would induce a loss of
competitiveness advantage for firms, and there will be neutrality between variables.

Another explanation of neutrality was made by Naeem and Rasheed (2002).

According to them, the exchange rate is the price of an asset. Therefore, like prices of other
assets, the exchange rates are determined by expected future exchange rates. Any
news/factors that cause changes in the exchange rate will affect today’s exchange rate. These
factors may differ from those that cause stock price changes (Naeem & Rasheed, 2002: 536).
So, there would be no linkage between variables.

The theoretical explanations are valid for exchange rate - stock price relation, but it
needs to be clarified if one of them is valid for ETFs or not. Although ETFs are baskets of
stock indices, the behaviour of investors in ETFs can be different from that of stockholders.
That is why the situation for the ETFs needs to be investigated. In the conclusion section,
we compare results with the approaches summarised above.

3. Literature Review

The literature examining ETFs in the context of factors affecting ETF prices is
limited. Studies investigating ETFs are generally focused on the relationship between ETF
and its underlying assets, such as tracking errors, etc. (please see Shin & Soydemir, 2010;
Dedi & Yavas, 2016; Da & Shive, 2018; Tsalikis & Papadopoulos, 2019; Joshi et al., 2021).
When we summarise them, it is possible to conclude that they focus on the price efficiency
of ETFs in different financial systems.

A few researchers investigate the interaction between ETF and foreign exchange rate
volatility. One of them belongs to Shank and Vianna (2016). In their study, one of the ETF
types of behaviour of currency-hedged ETFs in U.S. financial markets is investigated. The

210
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

study's empirical results indicate an interaction between exchange rate shocks and investor
behaviours. Although this implication is an initial result for ETF - FX relation, due to the
type of ETF, results cannot be generated for all ETFs.

Geetha et al. (2020) investigate the sensitivity of ETFs to an exchange rate

fluctuation. They employ exchange rate and ETF in NASDAQ from 2013 to 2018, using
GARCH models, and find no direct relation between variables. The authors emphasise that
the relationship between ETF and exchange rate might change according to underlying
assets.

In another study, Sakarya and Ekinci (2020) analyse the relationship between ETF
and exchange rate in the context of uncertainty. They employ MSCI Türkiye ETF data and
the EGARCH method. Results suggest an asymmetric behaviour as outflows of ETFs are
followed by an exchange rate depreciation with less exchange rate uncertainty, while
significantly large inflows of ETFs lead to higher exchange rate uncertainty.

As can be seen, studies investigating ETFs focus on something other than ETF - FX
relation. To our knowledge, this is the first study including risk and share of foreign investors
in the stock market into the model to investigate the relation between ETF and FX. Also,
unlike existing studies, we employ Fourier-based co-integration and causality analysis
methods to get more robust results.

4. Data and Methodology

In this study, we investigate the possible interaction between stock price and
exchange rate differently. While we measure the direct effects of the exchange rate, we put
additional variables to capture indirect effects arising from risk. To do this, we employ credit
default swap premium (CDS, hereafter) as a measure of risk and share of foreign capital in
the stock market.

Theoretically, risk is an important factor affecting investment decisions. In this

regard, risk will reduce capital investments even if they belong to domestic or foreign
investors. In light of empirical evidence in the literature, it is possible to empirically mention
the bi-directional causation linkage between exchange rate and CDS premium (Lu et al.,
2009; Yang et al., 2010). So, an increase in a country's CDS premium would accelerate
exchange rate volatility, and increasing volatility in the exchange rate will affect stock
market investments.

CDS premiums would also directly affect stock prices. An increase in CDS premium
would reduce stock prices due to increasing risks in the related economy, and investors
prefer to stay more liquid. So, the relation between variables runs from CDS premium to
stock price. An increase in CDS premium will reduce ETF demand and price (Please see
Ngene & Hassan, 2012; Mateev, 2019).

211
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

Another variable is the share of foreign capital in the stock market. An increase in
the share of portfolio investments of foreign investors in a stock market would increase the
performance of the stock market (Singh & Weisse, 1998; Gümüş & Güngör, 2013; Ali &
Hussain, 2013). Also, stock market performance might affect foreign capital (Haider et al.,
2017). In this regard, a positive relation exists between share of foreign capital and stock
price. On the other hand, the exchange rate might affect the share of foreign capital in the
stock market. According to Ogundipe et al. (2019), Fidora et al. (2007), and Aydoğan and
Vardar (2020), exchange rate volatility is one of the driving factors of foreign portfolio
investments. So, an increase in exchange rate volatility would lower foreign portfolio
investments in the stock market and reducing investments will reduce stock prices.

In the light of all theoretical explanations, direct and indirect interaction between
exchange rate and ETF price can be predicted as follows.

Figure: 1
Relation Between Variables

In the Turkish economy, the behaviour of stock market investors is affected by

exchange rate volatility and behaviour of them affects exchange rate volatility (Çiçek &
Öztürk, 2007: 98-102). The primary source of this relation is the net reserve position of the
economy. In this regard, a change in the investment amount of foreign investors to ETF may
also change the exchange rate.

After the 1980s, financial liberalisation started, financial markets improved, and new
financial instruments were used. In 2008, Morgan Stanley established an ETF including
stock shares from the Borsa Istanbul. We use data belonging to MSCI Türkiye ETF from
March 2008 to June 2022. The sources of data are listed in the following Table.

Table: 1
The Source of Data
Variable Acronym Source
Exchange Traded Fund price ETF MSCI Türkiye
Exchange Rate (U.S. dollar) EXC International Monetary Fund, International Financial Statistics
Credit Default Swap Premium (5-year) CDSTR Bloomberg
Share of Foreign Investors in BIST SHARE Banking Regulation and Supervision Agency

To investigate the relation, two separate models were built. These are:

212
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

𝑀𝑜𝑑𝑒𝑙 1: 𝐸𝑋𝐶𝑡 = 𝛽0 + 𝛽1 𝐸𝑇𝐹𝑡 + 𝛽2 𝐶𝐷𝑆𝑇𝑅𝑡 + 𝛽3 𝑆𝐻𝐴𝑅𝐸𝑡 + 𝑒𝑡 (1)

𝑀𝑜𝑑𝑒𝑙 2: 𝐸𝑇𝐹𝑡 = 𝛼0 + 𝛼1 𝐸𝑋𝐶𝑡 + 𝛼2 𝐶𝐷𝑆𝑇𝑅𝑡 + 𝛼3 𝑆𝐻𝐴𝑅𝐸𝑡 + 𝑢𝑡 (2)

In models, 𝛽0 and 𝛼0 are constant terms, 𝛽𝑖 and 𝛼𝑖 are slope parameters for 𝑖 =
0,1,2,3. t is defined as follows: 𝑡 = 2008: 03,2008: 05,2008: 06, … ,2022: 06 and 𝑒𝑡 and 𝑢𝑡
are error terms representing variables which were not included in the model. The sources of
data belonging to variables are presented in Table 1. Sources of data are listed also in the
same table.

In the first model, the dependent variable is the exchange rate (EXC, hereafter), and
the independent variables are ETF, credit default swap premium of the Turkish economy
(CDSTR, hereafter), and foreign investment share in the BIST (SHARE, hereafter). In the
second model, the dependent variable is ETF, and the independent variables are EXC,
CDSTR and SHARE.

In the first step, unit root tests are applied. To test structural breaks in series, we
employ both conventional unit root tests, unit root tests taking structural breaks into account
and Fourier-based unit root tests taking smooth transition breaks into account. According to
unit root test results, all variables are stationary in the first differences, and we accept
variable I (1).

In the second step, we test the co-integration relation between variables in the long
run by using the conventional Engle-Granger (1987) (EG, hereafter) co-integration test and
residual-based co-integration test with a Fourier approximation. The residual-based co-
integration test with a Fourier approximation (FEG, hereafter) is developed by Yılancı
(2019). It aims to adapt to an unknown number and shape of structural breaks in the presence
of soft structural changes with the help of Fourier functions. Yılancı (2019) has shown in his
study that it is a test with good size and strength properties in the presence of fractures. In
addition, the causal relationships between the variables were also investigated. Traditional
Toda-Yamamoto (1995) and Fourier Toda-Yamamoto causality Nazlioglu et al. (2016) tests
were used.

Descriptive statistics of the variables considered for both models in the study are
presented in Table 2, and the time path graph of the variables is shown in Figure 2.

Table: 2
Descriptive Statistics of Variables
𝑬𝑿𝑪 𝑬𝑻𝑭 𝑪𝑫𝑺𝑻𝑹 𝑺𝑯𝑨𝑹𝑬
𝐴𝑣𝑒𝑟𝑎𝑔𝑒 3.866 41.365 284.926 24.433
𝑀𝑒𝑑𝑖𝑎𝑛 2.664 41.832 243.796 24.508
𝑀𝑎𝑥𝑖𝑚𝑢𝑚 16.684 74.043 803.610 25.082
𝑀𝑖𝑛𝑖𝑚𝑢𝑚 1.159 17.663 117.809 23.532
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣. 3.172 14.968 128.495 0.367
𝑆𝑘𝑒𝑤𝑛𝑒𝑠𝑠 1.966 0.111 1.293 -0.520
𝐾𝑢𝑟𝑡𝑜𝑠𝑖𝑠 6.997 1.922 4.429 2.270
𝐽𝑎𝑟𝑞𝑢𝑒 − 𝐵𝑒𝑟𝑎 225.342 (0.000) 8.678 (0.013) 62.563 (0.000) 11.582 (0.003)
Note: Values in parentheses show probability values.

213
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

In Table 2, results imply that series belonging to variables do not show normal
distribution.

Figure: 2
Time Path Graph of The Variables
20 80

70
16

60
12

50
8
40

4
30

0
20

-4 10
08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22

EXC FOURIER_EXC ETF FOURIER_ETF

900 25.2

800 25.0

24.8
700
24.6
600
24.4
500
24.2
400
24.0
300
23.8

200 23.6

100 23.4
08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22

CDSTR FOURIER_CDSTR SHARE FOURIER_SHARE

3.5 15

3.0
10
2.5
5
2.0

1.5 0

1.0 -5

0.5
-10
0.0
-15
-0.5

-1.0 -20
08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22

DIFF_EXC DIFF_FOURIER_EXC DIFF_ETF DIFF_FOURIER_ETF

214
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

300 .4

.3
200
.2

.1
100
.0

-.1
0
-.2

-100 -.3

-.4

-200 -.5
08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22

DIFF_CDSTR DIFF_FOURIER_CDSTR DIFF_SHARE DIFF_FOURIER_SHARE

As shown in Figure 2, after taking the difference of the variables, they return to the
mean and their appropriate frequencies are seen.

5. Empirical Findings

The first step to investigate the co-integration relation between variables for both
models is unit root analysis. The unit root test results are presented in the following tables.

Table: 3
ADF and PP Unit Root Test Results
Variables ADF Test Stat. P-P Test Stat.
𝐸𝑇𝐹 -0.934 (0.775) [10] -1.631 (0.465)
∆𝐸𝑇𝐹 -5.745 (0.000) [9]*** -11.775 (0.000)***
𝐸𝑋𝐶 3.692 (0.998) [12] 6.455 (0.999)
∆𝐸𝑋𝐶 0.576 (0.998) [7] -10.564 (0.000)***
Constant
𝐶𝐷𝑆𝑇𝑅 -1.033 (0.741) [0] -0.903 (0.786)
∆𝐶𝐷𝑆𝑇𝑅 -13.167 (0.000) [0]*** -13.220 (0.000)***
𝑆𝐻𝐴𝑅𝐸 -2.031 (0.273) [1] -1.739 (0.409)
∆𝑆𝐻𝐴𝑅𝐸 -10.703 (0.000) [0]*** -10.499 (0.000)***
𝐸𝑇𝐹 -3.077 (0.115) [6] -2.883 (0.171)
∆𝐸𝑇𝐹 -6.007 (0.000) [9]*** -11.741 (0.000)***
𝐸𝑋𝐶 3.089 (0.997) [12] 5.240 (0.998)
∆𝐸𝑋𝐶 -11.264 (0.000) [0]*** -11.264 (0.000)***
Constant and Trend
𝐶𝐷𝑆𝑇𝑅 -2.251 (0.458) [0] -2.294 (0.434)
∆𝐶𝐷𝑆𝑇𝑅 -13.254 (0.000) [0]*** -13.371 (0.000)***
𝑆𝐻𝐴𝑅𝐸 -2.730 (0.226) [1] -2.484 (0.336)
∆𝑆𝐻𝐴𝑅𝐸 -10.706 (0.000) [0]*** -10.509 (0.000)***
Note: *** indicates significance at the 1% level. Values in parentheses present probability, and values in brackets show suitable lag length.

In Table 3, the unit root test results of Dickey-Fuller (1981) (ADF, hereafter) and
Phillips-Perron (1988) (PP, henceforth) are presented. According to results obtained from
ADF and PP, the series are stationary in both models with constant and constant trends when
we consider the first difference of the series.

Table 4 presents the unit root test results with structural breaks developed by Lee-
Strazicich (2003). Results obtained from conventional unit root tests and unit root tests with
structural breaks are similar. Model A allows two structural breaks in the model with a

215
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

constant, and Model B allows two structural breaks in the model with a constant and
constant, and the trend is stationary in the series of the first difference.

Table: 4
Lee-Strazicich (2003) Unit Root Test Results
Critical Value
Test Variable Test Stat. Break Date 1% 5% 10%
𝐸𝑇𝐹 -3.238 2009:11,2018:07 -4.087 -3.579 -3.321
∆𝐸𝑇𝐹 -9.805*** 2014:02,2017:03 -4.087 -3.579 -3.322
𝐸𝑋𝐶 -0.925 2011:12,2020:08 -4.087 -3.579 -3.321
∆𝐸𝑋𝐶 -12.247*** 2016:08,2020:12 -4.087 -3.579 -3.321
LS (Model A)
𝐶𝐷𝑆𝑇𝑅 -3.195 2012:06,2020:02 -4.087 -3.579 -3.321
∆𝐶𝐷𝑆𝑇𝑅 -10.076*** 2009:12,2020:12 -4.087 -3.579 -3.321
𝑆𝐻𝐴𝑅𝐸 -2.640 2018:07,2020:02 -4.087 -3.579 -3.321
∆𝑆𝐻𝐴𝑅𝐸 -8.900*** 2014:02,2016:01 -4.087 -3.579 -3.321
𝐸𝑇𝐹 -4.082 2009:08,2013:06 -6.552 -5.801 -5.424
∆𝐸𝑇𝐹 -11.923*** 2010:09,2018:09 -6.299 -5.737 -5.427
𝐸𝑋𝐶 -5.063 2016:08,2021:01 -6.552 -5.801 -5.424
∆𝐸𝑋𝐶 -13.233*** 2017:05,2018:08 -6.172 -5.611 -5.321
LS (Model C)
𝐶𝐷𝑆𝑇𝑅 -4.816 2018:05,2020:11 -6.179 -5.617 -5.327
∆𝐶𝐷𝑆𝑇𝑅 -13.864*** 2012:06,2020:08 -6.299 -5.737 -5.427
𝑆𝐻𝐴𝑅𝐸 -3.906 2009:08,2012:08 -6.179 -5.617 -5.327
∆𝑆𝐻𝐴𝑅𝐸 -10.749*** 2012:06.2013:12 -6.414 -5.720 -5.408
Note: *** presents a 1% significance level.

Table 5 presents conventional KPSS (Kwiatkowski, 1992) and Fourier KPSS (Becker
et al., 2006) unit root test results. Fourier KPSS unit root test results show that the series
presents a stationary level for some models. However, it is possible to conclude that all series
are stationary in the first difference when considering the financial series' long-run memory
features. In the light of all results, we include series belonging to variables into analyses with
their first difference.

Table: 5
KPSS and Fourier KPSS Unit Root Test Results
Critical Values
KPSS Test Stat. Freq Min. SSR Fourier KPSS Test Stat. F Test Stat.
1% 5% 10%
𝐸𝑇𝐹 1.106 1 11645.310 0.186 193.505*** 0.269 0.172 0.131
∆𝐸𝑇𝐹 0.069*** 4 2924.735 0.698 1.508 0.722 0.459 0.347
𝐸𝑋𝐶 1.340 1 969.015 0.623 65.492*** 0.269 0.172 0.131
∆𝐸𝑋𝐶 0.883 1 19.102 0.468 5.004** 0.269 0.172 0.131
Constant
𝐶𝐷𝑆𝑇𝑅 0.920 1 1478988.000 0.326 76.810*** 0.269 0.172 0.131
∆𝐶𝐷𝑆𝑇𝑅 0.263*** 4 428214.6 0.318 1.803 0.722 0.459 0.347
𝑆𝐻𝐴𝑅𝐸 0.774 1 8.315 0.202 150.192*** 0.269 0.172 0.131
∆𝑆𝐻𝐴𝑅𝐸 0.146*** 4 1.911 0.166 1.979 0.722 0.459 0.347
𝐸𝑇𝐹 0.228 1 10080.67 0.037 78.133*** 0.071 0.054 0.047
∆𝐸𝑇𝐹 0.031*** 4 2920.924 0.034 1.449 0.217 0.147 0.118
𝐸𝑋𝐶 0.337 1 186.631 0.081 154.464*** 0.071 0.054 0.047
∆𝐸𝑋𝐶 0.227 1 17.028 0.062 6.585** 0.071 0.054 0.047
Constant and Trend
𝐶𝐷𝑆𝑇𝑅 0.307 1 1134401.000 0.049 57.045*** 0.071 0.054 0.047
∆𝐶𝐷𝑆𝑇𝑅 0.044*** 4 423979.9 0.060 1.612314 0.217 0.147 0.118
𝑆𝐻𝐴𝑅𝐸 0.316 1 7.149 0.044 115.212*** 0.071 0.054 0.047
∆𝑆𝐻𝐴𝑅𝐸 0.023*** 4 1.899 0.028 1.858 0.217 0.147 0.118
Note: *, ** and *** show significance levels of 10%, 5% and 1%, respectively. In a model with constant, F test statistics is used to test KPSS test
statistics and significance of trigonometric terms and F statistics critical values for 1%, 5%, and 10% are 0.739, 0.463, 0.347 and 6.730, 4.929, 4.133,
respectively. In a model with constant and trend, the critical values of the F test used to test the significance of KPSS test statistics and trigonometric
terms are 0.216, 0.146, 0.119 and 6.873, 4.972, 4.162, respectively.

216
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

The second step is co-integration analysis. Table 6 presents conventional Engle-

Granger and Fourier Engle-Granger co-integration test results. According to conventional
co-integration test results, neither model is statistically significant. On the other hand, the
Fourier Engle-Granger co-integration test results imply that in Model 2, variables are co-
integrated when we consider smooth structural breaks. In light of this result, it is possible to
conclude that there is a long-run relation between variables. To determine the size of the
long-run relation, fully modified ordinary least squares (FMOLS, hereafter), dynamic
ordinary least squares (DOLS, hereafter) and canonical co-integration regression (CCR)
methods are employed in the third step.

Table: 6
Engle-Granger and Fourier Engle-Granger Co-integration Analysis Results
Test Stat. Min. KKT 𝒌 Critical Values 1% 5% 10%
𝐸𝐺 -2.762 (0.545) - -
Model 1
𝐹𝑜𝑢𝑟𝑖𝑒𝑟 − 𝐸𝐺 -4.166 389.038 4 -5.271 -4.605 -4.252
𝐸𝐺 -0.817 (0.992) - -
Model 2
𝐹𝑜𝑢𝑟𝑖𝑒𝑟 − 𝐸𝐺 -4.977** 849.237 1 -5.596 -4.957 -4.640
Note: ***, ** and * denote significance levels 1%, 5% and 10%, respectively; values in parentheses present probability values.

In Table 7, results belonging to FMOLS, DOLS and CCR analyses are presented.
Long-term estimates of results obtained from FMOLS, DOLS and CCR methods are similar.
According to these results, all variables, except the exchange rate variable and the
trigonometric terms included in the model, are statistically significant. In addition, it is found
that the 𝑅2 value, which represents the model's explanatory power, is relatively high in all
three methods.

Table: 7
Long Run Analysis Results
Method Coefficient Standard Error Statistical value
𝐸𝑋𝐶 -0.008 0.161 -0.047 (0.962)
𝐶𝐷𝑆𝑇𝑅 0.017 0.007 2.573 (0.011) **
𝑆𝐻𝐴𝑅𝐸 42.416 2.442 17.370 (0.000) ***
FMOLS 𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡 -999.791 61.074 -16.370 (0.000)***
𝑆𝐼𝑁 6.304 0.556 11.329 (0.000)***
𝐶𝑂𝑆 4.434 0.610 7.263 (0.000)***
𝑅2 0.98
𝐸𝑋𝐶 0.061 0.208 0.293 (0.770)
𝐶𝐷𝑆𝑇𝑅 0.023 0.009 2.481 (0.014)**
𝑆𝐻𝐴𝑅𝐸 44.299 3.263 13.577 (0.000)***
DOLS 𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡 -1047.595 81.824 -12.803 (0.000)***
𝑆𝐼𝑁 6.337 0.597 10.610 (0.000)***
𝐶𝑂𝑆 4.471 0.668 6.698 (0.000)***
𝑅2 0.98
𝐸𝑋𝐶 -0.002 0.164 -0.014 (0.989)
𝐶𝐷𝑆𝑇𝑅 0.020 0.008 2.587 (0.011)**
𝑆𝐻𝐴𝑅𝐸 43.012 2.608 16.495 (0.000)***
CCR 𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡 -1015.008 65.388 -15.523 (0.000)***
𝑆𝐼𝑁 6.361 0.561 11.345 (0.000)***
𝐶𝑂𝑆 4.456 0.599 7.436 (0.000)***
2
𝑅 0.97
Note: *** and ** indicate significance at the 1% and 5% levels and represent the probability values in parentheses, respectively.

Long run analysis results show that 𝐶𝐷𝑆𝑇𝑅 and 𝑆𝐻𝐴𝑅𝐸 variables are positively
effective on ETF price. So, 𝐶𝐷𝑆𝑇𝑅 and 𝑆𝐻𝐴𝑅𝐸 variables create a positive effect on ETF

217
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

price. When we interpret results, theoretically, findings belonging to the SHARE variable
are significant. Because an increase in portfolio investments from abroad will increase ETF
prices; on the other hand, results show that an increase in credit default swap premiums will
increase ETF prices. This is opposite to theoretical expectations. In this regard, the share of
foreign investments in BIST is the dominant variable affecting ETF price in the long run.

In the fourth step, we made an error correction analysis. Results are also listed in
Table 8. ECT represents the error correction mechanism, and according to the results
presented in Table 8, it is seen that the error correction mechanism works because it is
statistically significant and is between -1 and 0. In addition, it is found that the short-term
imbalances in the system are resolved approximately in 18 (1/0.055) months. The causality
tests of the variables were carried out, and the traditional Toda-Yamamoto (1995) analysis
results are presented in Table 9 below.

Table: 8
Short Run Analysis Results
Coefficient Standard Error t-Statistic Prob.
𝐸𝑋𝐶 -0.533 0.562 -0.948 0.344
𝐶𝐷𝑆𝑇𝑅 -0.006 0.005 -1.338 0.183
𝑆𝐻𝐴𝑅𝐸 31.251 1.974 15.833 0.000***
𝐶𝑜𝑛𝑠𝑡𝑎𝑛𝑡 0.067 0.170 0.393 0.695
𝐸𝐶𝑇 -0.055 0.029 -1.873 0.063*
Note: *** and * indicate significance at the 1% and 10% levels.

Table: 9
Toda-Yamamoto Causality Analysis Test Results
𝐻0 Lag Length Wald Test Stats. Asymp. Prob Value Bootstrap Prob. Value
𝐸𝑇𝐹 ↛ 𝐸𝑋𝐶 8 4.877 0.771 0.733
𝐸𝑋𝐶 ↛ 𝐸𝑇𝐹 8 5.321 0.723 0.720
𝐸𝑇𝐹 ↛ 𝐶𝐷𝑆𝑇𝑅 1 2.883 0.089* 0.094*
𝐶𝐷𝑆𝑇𝑅 ↛ 𝐸𝑇𝐹 1 0.024 0.877 0.822
𝐸𝑇𝐹 ↛ 𝑆𝐻𝐴𝑅𝐸 8 7.794 0.454 0.457
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐸𝑇𝐹 8 6.151 0.630 0.625
𝐸𝑋𝐶 ↛ 𝐶𝐷𝑆𝑇𝑅 1 0.380 0.537 0.484
𝐶𝐷𝑆𝑇𝑅 ↛ 𝐸𝑋𝐶 1 0.293 0.588 0.559
𝐸𝑋𝐶 ↛ 𝑆𝐻𝐴𝑅𝐸 8 18.366 0.019** 0.024**
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐸𝑋𝐶 8 5.337 0.721 0.674
𝐶𝐷𝑆𝑇𝑅 ↛ 𝑆𝐻𝐴𝑅𝐸 2 8.618 0.013** 0.015**
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐶𝐷𝑆𝑇𝑅 2 1.629 0.443 0.445
Note: * and ** present 10% and 5% significance levels, respectively. Number of bootstrap simulations is 10.000.

Figure: 3
Conventional Toda-Yamamoto Causality Analysis Results

218
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

According to conventional Toda-Yamamoto causality test results, there is a uni-

directional causality running from 𝐸𝑇𝐹 to 𝐶𝐷𝑆𝑇𝑅, from 𝐸𝑋𝐶 to 𝑆𝐻𝐴𝑅𝐸 and 𝐶𝐷𝑆𝑇𝑅 to
𝑆𝐻𝐴𝑅𝐸. Results show that exchange rate and CDS premium are effective on share of foreign
investors in BIST. Statistical significance is 5% for both variables. On the other hand,
statistical significance of effects of ETF on CDSTR is low and theoretically it is almost
insignificant. Results are also seen in Figure 3.

Table: 10
Fourier Toda-Yamamoto Causality Test Results
𝐻0 Lag length Frequency Wald Test Stat. Asymp. Prob. Value Bootstrap Prob. Value
𝐸𝑇𝐹 ↛ 𝐸𝑋𝐶 8 1 3.402 0.907 0.878
𝐸𝑋𝐶 ↛ 𝐸𝑇𝐹 8 1 2.559 0.959 0.954
𝐸𝑇𝐹 ↛ 𝐶𝐷𝑆𝑇𝑅 2 1 2.475 0.290 0.298
𝐶𝐷𝑆𝑇𝑅 ↛ 𝐸𝑇𝐹 2 1 1.442 0.486 0.483
𝐸𝑇𝐹 ↛ 𝑆𝐻𝐴𝑅𝐸 8 1 9.169 0.328 0.342
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐸𝑇𝐹 8 1 6.073 0.639 0.630
𝐸𝑋𝐶 ↛ 𝐶𝐷𝑆𝑇𝑅 8 1 17.546 0.025** 0.072*
𝐶𝐷𝑆𝑇𝑅 ↛ 𝐸𝑋𝐶 8 1 7.493 0.484 0.455
𝐸𝑋𝐶 ↛ 𝑆𝐻𝐴𝑅𝐸 8 1 18.742 0.016** 0.018**
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐸𝑋𝐶 8 1 5.150 0.741 0.711
𝐶𝐷𝑆𝑇𝑅 ↛ 𝑆𝐻𝐴𝑅𝐸 2 1 10.426 0.005*** 0.007***
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐶𝐷𝑆𝑇𝑅 2 1 0.805 0.669 0.668
𝑆𝐻𝐴𝑅𝐸 ↛ 𝐶𝐷𝑆𝑇𝑅 2 1.629 0.443 0.445
Note: * and ** present 10% and 5% significance levels, respectively. Number of bootstrap simulations is 10.000.

Fourier Toda-Yamamoto causality analysis results powered by Fourier terms are

similar to conventional Toda-Yamamoto causality test results. Uni-directional causation
linkages running from 𝐸𝑋𝐶 to 𝑆𝐻𝐴𝑅𝐸 and 𝐶𝐷𝑆𝑇𝑅 to 𝑆𝐻𝐴𝑅𝐸 are found. Statistical
significance level is 1% for both results. When we compare conventional analysis,
significance level is higher than conventional one. On the other hand, there is a uni-
directional causality running from 𝐸𝑋𝐶 to 𝐶𝐷𝑆𝑇𝑅. This is different from conventional
analysis. Also, significance level is 10%. Although statistical significance is low,
theoretically increasing exchange rate is risky for both financial and real sectors. Increasing
risk will accelerate CDS premium. In the light of results obtained from causality analyses,
exchange rate does not have direct effect on ETF prices. But it is effective via share of
foreign investors and CDS premium. Results are also presented in Figure 4.

Figure: 4
Fourier Toda-Yamamoto Causality Test Results

219
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

6. Conclusion

In recent years, the ETF has been one of the most popular passive investment
instruments in the financial markets. As mentioned in the introduction section, the share of
ETFs in the stock markets increases yearly. Increasing interest in ETFs made it an important
instrument to investigate. There are few studies analysing different aspects of ETFs. The
tracking error of ETF with the underlying assets, return and volatility correlation and effect
of uncertainty on ETF performance are all tested in different financial markets. A few studies
analyse the possible interaction between ETF prices and exchange rates.

This study investigates the relationship between exchange rate and ETF prices in the
Turkish economy between 2008 and 2022 via advanced and conventional econometric
methods. According to the results, risk premiums and shares of foreign investors effectively
affect ETF prices in the long run. On the other hand, the coefficient belonging to the
exchange rate is not significant. That means the exchange rate is ineffective on ETF prices
in the Turkish economy. Moreover, only the share of foreign investors is effective on ETF
prices in the short run. The short and long-run analyses imply that the exchange rate is
effective on ETF prices neither in the short nor long run. However, risk premiums and shares
of foreign investors are effective on ETF prices.

When we apply causality analysis, it is seen that risk premium and exchange rate
affect the share of foreign investors in the BIST. Fourier's causality analysis also indicates a
unidirectional relation between the exchange rate and the share of foreign investors in BIST.
Combining all these results makes it possible to conclude that there is no direct effect of the
exchange rate on ETF price, but it is effective on the share of foreign investors. Also, foreign
investors' shares affect ETF prices in the short- or long-term. The transmission mechanism
works from the exchange rate to the share of foreign investors in BIST and from the share
of foreign investors in BIST to ETF prices. Another mechanism works from exchange rate
to risk premium, risk premium to share of foreign investors in BIST and from share of
foreign investors in BIST to ETF prices.

We can compare results with the theoretical explanations summarised before. As

mentioned earlier, four approaches are trying to explain the relationship between stock prices
and exchange rates. Results are similar to the “Asset market” approach, claiming no or weak
relation between ETF and exchange rate. But, different from this approach, we have findings
indicating an indirect relation. So, it is possible to imply that the relationship between ETF
prices and the exchange rate is somewhat different from the stock price exchange rate nexus.
This result may come from the structure of the Turkish capital market. As the Turkish
economy is an emerging market, the financial system is still on the way to developing. For
this reason, a relatively fordable market does not reflect exchange shocks efficiently and
induce possible indirect relations, as found in the study.

220
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

References
Ali, S. & J. Hussain (2013), “Impact of Portfolio Investments on Stock Market Performance”, SSRN
Discussion Paper, <https://ssrn.com/abstract=2334119>, 28.02.2013.
Aydoğan, B. & G. Vardar (2021), “Portfolio Flows - Exchange Rate Volatility: Is There a Puzzling
Relationship?”, Journal of Economic and Administrative Sciences, 37(4), 611-642.
Becker, R. et al. (2006), “A Stationarity Test in the Presence of an Unknown Number of Smooth
Breaks”, Journal of Time Series Analysis, 27(3), 381-409.
Branson, W.H. (1983), “Macroeconomic Determinants of Real Exchange Risk”, in: R.J. Herring
(ed.), Managing Foreign Exchange Risk, Cambridge University, Cambridge.
Çiçek, M. & F. Öztürk (2007), “Yabancı Hisse Senedi Yatırımcıları Türkiye’de Döviz Kuru
Volatilitesini Şiddetlendiriyor mu?”, Ankara Üniversitesi SBF Dergisi, 62(4), 83-107.
Da, Z. & S. Shive (2018), “Exchange Traded Funds and Asset Return Correlations”, European
Financial Management, 24(1), 136-168.
Dedi, L. & B.F. Yavas (2016), “Return and Volatility Spillovers in Equity Markets: An Investigation
Using Various GARCH Methodologies”, Cogent Economics & Finance, 4(1), 1266788.
Dickey, D.A. & W.A. Fuller (1981), “Likelihood Ratio Statistics for Autoregressive Time Series
with a Unit Root”, Econometrica, 49(4), 1057-1072.
Dornbusch, R. & S. Fischer (1980), “Exchange Rates and the Current Account”, American Economic
Review, 70(5), 960-971.
Fidora, M. et al. (2007), “Home Bias in Global Bond and Equity Markets: The Role of Real
Exchange Rate Volatility”, Journal of International Money and Finance, 26(4), 631-655.
Frankel, J.A. (1983), “Monetary and Portfolio-Balance Models of Exchange Rate Determination”, in:
J.S. Bhandari & B.H. Putnam (eds.), Economic Interdependence and Flexible Exchange
Rate, MIT, Cambridge.
Geetha, E. et al. (2020), “Are Global Exchange Traded Fund Pretentious on Exchange Rate
Fluctuation? A Study Using GARCH Model”, Investment Management and Financial
Innovations, 17(4), 356-366.
Gümüş, G.K. & B. Güngör (2013), “The Relationship Between Foreign Portfolio Investment and
Macroeconomic Variables”, European Scientific Journal, 9(34), 209-226.
Haider, M.A. et al. (2017), “The Impact of Stock Market Performance on Foreign Portfolio
Investment in China”, International Journal of Economics and Financial Issues, 7(2),
460-468.
Joshi, A. et al. (2021), “Dynamic Analysis Between ETF and Underlying Asset Using Tracking
Error”, in: S. Chakrabarti et al. (eds.), In Interdisciplinary Research in Technology and
Management (82-85), CRC Press, London.
Kwiatkowski, D. et al. (1992), “Testing the Null Hypothesis of Stationarity Against the Alternative
of a Unit Root”, Journal of Econometrics, 54(1), 159-178.
Lee, J. & M. Strazicich (2003), “Minimum Lagrange Multiplier Unit Root Test with Two Structural
Breaks”, The Review of Economics and Statistics, 85(4), 1082-1089.
Lu, L. et al. (2009), “Long Term Performance of Leveraged ETFs”, SSRN Discussion Paper,
<https://ssrn.com/abstract=1344133>, 28.02.2013.

221
 Bayat, T. & A.M. Köktaş & S. Kayhan & G. Konat (2024), “Effects of FX
on ETF Prices: Evidence from BIST”, Sosyoekonomi, 32(59), 207-222.

Mateev, M. & E. Marinova (2019), “Relation Between Credit Default Swap Spreads and Stock
Prices: A Non-Linear Perspective”, Journal of Econ Finance, 43(1), 1-26.
Muhammad, N. & A. Rasheed (2002), “Stock Prices and Exchange Rates: Are They Related?
Evidence from South Asian Countries”, The Pakistan Development Review, 41(4), 535-
550.
Nazlioglu, S. et al. (2016), “Oil Prices and Real Estate Investment Trusts (REITs): Gradual-Shift
Causality and Volatility Transmission Analysis”, Energy Economics, 60, 168-175.
Ngene, G. & K. Hassan (2012), Momentum, Nonlinearity in Cointegration and Price Discovery:
Evidence from Sovereign CDS and Equity Markets of Emerging Countries,
<http://cbt2.nsuok. edu/kwok/conference/submissions/swfa2013_submission_200.pdf>,
28.04.2023.
Ogundipe, A.A. et al. (2019), “Exchange Rate Volatility and Foreign Portfolio Investment in
Nigeria”, Investment Management and Financial Innovations, 16(3), 241-250.
Phillips, P. & P. Perron (1988), “Testing for a Unit Root in Time Series Regression”, Biometrika,
75(2), 335-346.
Ramaswamy, S. (2011), “Market Structures and Systemic Risks of Exchange-traded Funds”, BIS
Working Papers, (343), 01.04.2011.
Sakarya, B. & A. Ekinci (2020), “Exchange-traded Funds and FX Volatility: Evidence from
Turkey”, Central Bank Review, 20(4), 205-211.
Shin, S. & G. Soydemir (2010), “Exchange-traded Funds, Persistence in Tracking Errors and
Information Dissemination”, Journal of Multinational Financial Management, 20(4-5),
214-234.
Singh, A. & B.A. Weisse (1998), “Emerging Stock Markets, Portfolio Capital Flows and Long-term
Economic Growth: Micro and Macroeconomic Perspectives”, World Development,
26(4), 607-622.
Toda, H.Y. & T. Yamamoto (1995), “Statistical Inference in Vector Autoregressions with Possibly
Integrated Processes”, Journal of Econometrics, 66(1-2), 225-250.
Todorov, K. (2021), “The Anatomy of Bond ETF Arbitrage”, BIS Quarterly Review, March, 41-53.
Tsalikis, G. & S. Papadopoulos (2019), “ETFS-Performance, Tracking Errors and Their
Determinants in Europe and the USA”, Risk Governance & Control: Financial Markets
& Institutions, 9(4), 67-76.
Williams, S.O. (2014), “Country ETFs, Currencies and International Diversification”, Journal of
Asset Management, 25(6), 392-414.
Yang, J. et al. (2010), “Nonlinearity, Data-Snooping, and Stock Index ETF Return Predictability”,
European Journal of Operational Research, 200(2), 498-507.
Yıldız, A. (2014), “Döviz Kuru ile Hisse Senedi Endeksleri Arasındaki İlişki”, Finans Politik ve
Ekonomik Yorumlar Dergisi, 51(593), 77-91.
Zhao, H. (2010), “Dynamic relationship between exchange rate and stock price: Evidence from
China”, Research in International Business and Finance, 24(2), 103-112.

222