Sage Journals: Discover world-class research

Abstract

Understanding which factors influence exchange rate movements is important for understanding economic development, trade patterns, investment decisions, and for designing economic policies. In this study, we allow for monetary, domestic and global financial variables to assess the relevant importance of each of the variables to exchange rate volatility in the case of Turkey. The paper investigates the dynamics of exchange rate volatility of the Turkish lira in a complementary perspective by employing both Generalized autoregressive conditional heteroskedasticity (GARCH) method and Lyapunov exponents over the period from March 1, 2019 to November 11, 2021. Firstly, decomposing the impact of domestic and global financial variables, the results of the GARCH model indicate that the exchange rate volatility is affected by Volatility index (VIX) and Credit default swaps (CDS). This result suggests that the exchange rate shocks experienced are mainly caused by global factors, therefore policymakers should focus on volatility spillovers caused by global financial markets. Secondly, detected positive maximal Lyapunov exponent shows that complexity in foreign exchange markets has been increased, market expectations lead to multiple-equilibria and diverging volatility eventually will generate recurrent spikes in currency value. These complementary findings have important implications for interventions of Central banks and preventing systemic risks, as well as portfolio and risk management.

JEL Classification: C32, G14, G15.

Plain Language Summary

The paper investigates the dynamics of exchange rate volatility of Turkish Lira in a complementary perspective by employing both Generalized autoregressive conditional heteroskedasticity (GARCH) method and Lyapunov exponents over the period from March 1, 2019 to November 11, 2021. Firstly, decomposing the impact of domestic and global financial variables, the results of the GARCH model indicate that the exchange rate volatility is affected by Volatility index (VIX) and Credit default swaps (CDS).

The importance of the subject: The exchange rate dynamics are known to influence financial stability, which has been targeted as a policy implication simultaneous to macroeconomic stability. Moreover, the exchange rate volatility stands out as the main source of financial vulnerability and the loss of macroprudential stability. Especially in emerging markets, the exchange rate movements affect sustainability of many other macroeconomic parameters, such as interest rates, trade balance, output and asset prices.

Contribution: Domestic and global financial causes for exchange rate volatility are identified. Volatility dynamics were decomposed by GARCH (1,1) method and global effects were identified. Chaotic dynamics are shown in Turkish foreign exchange rate markets, violating Efficient market hypothesis.

Methods: Stochastic time series analysis methods are applied. First, GARCH (1,1) is used. Second, we reconstruct phase space in order to observe multiple equilibria embedded in financial time series. We obtain multidimensional embedded space and then we compute the Lyapunov exponents (LE) of the associated dynamical system. The Tisean Package introduced by Hegger et al. and R programme is used for computations.

This result suggests that the exchange rate shocks experienced are mainly caused by global fact.

Keywords

exchange rate volatility systemic risk market efficiency GARCH Lyapunov exponent Chaos

Introduction

The financial systems, particularly foreign exchange markets are evolving complex systems. The complexity originates from multiple decision makers and counterparties. It has long been understood that the investigation of exchange rate behavior is particularly important for emerging and developing economies. This importance mostly arises in formulating conclusive monetary and fiscal policies to ascertain the less volatile exchange rate process. Even though many studies focused on the influences of exchange rate (Bilgili et al., 2021), what determines exchange rate alterations is an unsettled issue in the literature. Understanding which factors influence exchange rate movements is crucial for understanding economic development, trade patterns, investment decisions, and for recommending economic policies. Few studies focused on the process of exchange rate determination. This study aims to fill this gap. More recently, Raza and Afsan (2017) noted that the exchange rate movements stem from the changes in terms of trade, trade openness, economic growth, money supply, inflation rate. The authors found particularly trade openness to be connected with economic growth and exchange rates. In a related study the direct and indirect effects of the degree of openness in emerging and developing economies on exchange rate movements are shown by Keefe and Saha (2022). Another factor is revealed by the study Chen et al. (2020), where the authors determined the impact of economic policy uncertainty on exchange rates. However, these studies analyze the determinants of monthly or quarterly fluctuations in exchange rates. The present study investigates determinants of the exchange rate fluctuations in daily observed data. Since the governments are more likely to revise economic policies when the volatility is increasing (see Chen et al., 2020), the early determination and the length of time-span in which volatility occurs may affect the policy response’s effectiveness. Due to financial stability preference, the Central banks and governments especially in emerging markets would intervene to stabilize exchange rates during higher volatility.

The unforeseen fluctuations in exchange rates cause important changes on foreign trade dynamics, economic policies and variables such as the interest rate, inflation rate, domestic investment, and foreign direct investments (Bilgili et al., 2022; Grossmann & Orlov, 2022) and asset prices (Aghion et al., 2009; Cabral et al., 2016; Devereux & Lane, 2003; Frankel, 2010; Ganguly & Breuer, 2010; Hausmann et al., 2006). The process through which particularly investment decisions and asset prices influenced are examined recently by Van Greuning and Bratanovic (2020). The authors state that “Currency risk results from changes in exchange rates and originates in mismatches between the values of assets and liabilities denominated in different currencies. Currency risk comprises (1) transaction risk, or the price-based impact of exchange rate changes on foreign receivables and payables; (2) economic or business risk related to the impact of exchange rate changes; and (3) revaluation risk or translation risk, arising when a bank’s foreign currency positions are revalued in domestic currency. Other types of risk that often accompany currency risk include counterparty risk, settlement risk, liquidity risk, and currency-related interest rate risk.” As it is widely known, a major risk associated with foreign trade is the uncertainty of future exchange rates. The relative values of the two currencies could alter between the time the deal is made and the time payment is received. If there is not an appropriate protection, a depreciation or devaluation of the foreign currency could cause the investor to lose money. In this vein, for an investor, it is not the exchange rate level matters, but its variation over time. Higher exchange rate volatility augments the risk factor of domestic firms engaged in international trading, which may increase prices to hedge against the additional risk premium (Giannellis & Papadopoulos, 2011). This causes relatively higher welfare cost for emerging market economies (EMEs onwards). More generally, the investor behavior under risk and uncertainty may depend on the degree of risk-aversion (Asteriou et al., 2016; Bilgili et al., 2021). Therefore, the investments in international companies, businesses or foreign currencies are always exposed to currency risk. Specifically, currency risk in private equity is substantial, especially when the impact of currency movements on fund returns is analyzed (see Cornelius, 2011).

As in every developing country, economic and financial crises have determined the direction and dynamics of exchange rate volatility in Turkey. Turkey has recently experienced local crises such as 1994 and 2001 and global crises such as 1997 Asian crisis, 2008 Mortgage crisis, and Global 2019 COVID-19 Outbreak. It is hard to say that the Turkish development process has successfully managed capital flows. Most of the period, the economy has experienced high amplitude capital flows, but was not preferred by long-term investors in productive and cutting-edge technological sectors. In the last two decades, financial openness and trade openness have not sufficiently fed back each other. According to Bank for International Settlements (BIS) (2019) (https://www.bis.org/publ/arpdf/ar2019e2.pdf): While a sustained 1% depreciation pushed up inflation by 0.6 percentage points in the early 2000s, the long-run effect was just 0.3 percentage points through the decade. However, from 2014 to 2019, 1% depreciation caused 1.5% increase in price level. In Turkey, while exchange rate pass-through has declined, the inflationary outcome of exchange rate swings has not been eliminated. When the time evolution of Turkey’s economic structure is analyzed, it can be determined that exchange rate volatility is main source of financial vulnerability and the loss of macroprudential stability. Guimarães and Karacadag (2004) report that the exchange rate has remained an important determinant of inflationary outcomes and expectations even after the change in exchange rate regime from fix to float, and the policy makers have been vigilant against volatility and have resisted it, market conditions and international reserve levels permitting. As it has already been conjectured in previous studies (Calvo & Reinhart, 2002; Hausmann et al., 2001), we observed that over the period from 2019 to 2022 where policy credibility is lower, reserve level is low and pass-through from exchange rate movements to inflation is higher, policy makers’ increased pressure on Central Bank of the Republic of Türkiye (CBRT) to intervene in the foreign exchange markets. Following persistent exchange rate volatility, it is seen that macroeconomic stability has deteriorated with the amplified dollarization process, and inflation have augmented significantly in 2022Q2, reaching annually 85%. Therefore, in perspective of policy making, it is crucial to find the root factors that drive exchange rate volatility. Without it, policymakers face a strong incentive to build expansive regulatory regimes capable of influencing practices that may or may not truly reduce uncertainty in financial system (Hendricks et al., 2007).

To examine the exchange rate volatility in developing countries, Aghion et al. (2009) suggest: “The volatility of real shocks relative to financial shocks—which features so prominently in the literature on developed country exchange rate regimes—also matters for developing countries. But because financial shocks tend to be greatly amplified in financially underdeveloped economies, one has to adjust calibrations accordingly.” In this study, we distinguish between the impact of global financial variables such as Credit default swap (CDS onwards), implied volatility index (VIX), dollar index, and the effect of domestic financial variables such as short-term interest rates, inflation expectations. Specifically, we apply GARCH(1,1) model to the data over the March 1, 2019 to November 11, 2021 period.

We make several important contributions to the existing literature. First, our paper is the first to study the effects of the key macroeconomic, global and domestic financial time series on the exchange rate volatility. Therefore, our approach enables to understand volatility spillovers from global markets to domestic markets (Badshah et al., 2013). There are similar studies focusing on exchange rate volatility in emerging market economies. However, some of these studies investigated the impact of domestic and external variables, which are low-frequency observations (Canales-Kriljenko & Habermeier, 2004; Devereux & Lane,2003; Ganguly & Breuer, 2010) and some others have mainly looked at the impact of foreign exchange rate intervention on exchange rate volatility (e.g., Baillie & Osterberg, 1997; Berganza & Broto, 2012; Dominguez, 1998; Guimarães & Karacadag, 2004). The global financial factors influencing financial markets of a country may also be considered as a mapping of external financial factors of that country. The literature on the effect of global financial factors is thin (Cairns et al., 2007; Smallwood, 2019). More recently Muhammed and Ramachandran (2018) examine 10 EMEs over the period from 1997Q1 to 2017Q1. The evidence reported by the authors indicates that exchange rate volatility significantly increases in response to the shocks in portfolio capital flows than to shocks in foreign direct inflows. In a similar study focusing on pandemic policies, Feng et al. (2021) conclude that the various policies adopted by governments in response to the pandemic, such as closing schools and public lockdown, also inhibit exchange rate volatility.

More generally, excessive volatility may be a symptom of disorderly markets, which involve a collapse of liquidity. However, it is often difficult to identify empirically dynamics of excessive exchange rate movements that are unwarranted (Cairns et al., 2007; Kriljenko, 2003). Our second contribution is to investigate the Lyapunov exponents in dynamics of foreign exchange markets. This enables to shed light on evolution of expectations in financial markets and to determine the orbit of equilibrium. We know that complex financial systems can exhibit the following dynamic features: non-linearity, path dependency, and sensitive dependence to initial conditions (see speech by Jean-Pierre Landau, Deputy Governor of the Bank of France, at the Conference on “The macroeconomy and financial systems in normal times and in times of stress,” jointly organized by the Bank of France and the Deutsche Bundesbank, Gouvieux-Chantilly, 8 June 2009). In the related literature, all these features are linked to chaotic dynamics of markets (Abhyankar et al., 1997; Das & Das, 2007; Hagtvedt, 2009; Orlando & Bufalo, 2022; Scheinkman & LeBaron, 1989; Vaidyanathan & Krehbiel, 1992; Vasilios et al., 2019). Moreover, detected Lyapunov exponents imply that the Efficient market hypothesis (EMH onwards) is violated, the market equilibrium diverges, predictability and arbitrage occurs (Brock et al., 1992; Gencay, 1998, Ozkaya, 2015). An increase in complexity decreases the available information and leads to augmented uncertainty in the foreign exchange market as demonstrated (Ehrmann et al., 2001). This causes volatility in exchange rates. In perspective of policy making, it important to prevent multiple-equilibria before volatility occurs (Cooper, 2005). Otherwise, long-term macroeconomic stability would be violated (Malkiel, 2003; Pesaran, 2010). More recently, Ahmad et al. (2019) investigate the dynamics of the real exchange rate for both developed and developing countries. In general, the results find a greater incidence of nonlinear dynamics for developing country exchange rates.

The paper is organized as follows. The next section provides a brief literature review; the third section explains the methodology utilized. The fourth section explains the data used in the study and presents the empirical results derived from GARCH and Chaos Theory methods. Fifth section concludes and finally we present policy recommendations.

Literature Review

We propose a review of related literature from the perspective of the variables used to explain exchange rate volatility. Although many different ways of modeling volatility have been mentioned in the literature, different opinions have been put forward as to which model is superior. Engle (1982) proposed the Autoregressive conditional variance (ARCH) model, which relates the variance of the error term to the squares of the previous term error terms. Bollerslev (1986) states the fundamental superiority of the GARCH model over the ARCH model, in which a more flexible lag structure is allowed. Following Bollerslev (1986), many authors as Figlewski (1997) and Engle and Patton (2007) mentioned the superiority of the GARCH model over other volatility measurement models. Among the GARCH models, the most widely used and accepted model was the GARCH (1,1) model (Alexander & Lazar, 2006; Doroodian & Caporale, 2001; Engle, 2001; Poon & Granger, 2005). For more recent applications see Paolella (2018) and Kim et al. (2021).

Beginning from 90s, many countries have implemented inflation targeting within their monetary policy framework and increasing number of emerging countries have adopted floating exchange rate regime, and changed their nominal anchor, from exchange rate to inflation (Grossmann et al., 2014). These countries have gained further credibility from communication (explicitly announcing the target), which helped to anchor and lower inflation expectations (Mishkin & Schmidt-Hebbel, 2007). For these reasons we include the difference between inflation expectation in Turkey and in United States in our GARCH model.

In this study, unlike other studies, the effects of both domestic and global financial time series on exchange rate volatility are investigated in order to find the causes of exchange rate volatility in Turkey amid pandemic crisis. One financial variable to measure global risk aversion, proxied by the implied volatility of the S&P index (VIX) as measure for global volatility spillover (Cairns et al., 2007) is included in our study. The emerging market currencies are evaluated as risky assets with respect to financial vulnerabilities. In addition, Turkey Credit default swap 5-year is included in this study as another global factor. CDS premiums, known as the global pricing of a country’s economic and political stance, national security and stability, are also included in the model.

Grossmann et al. (2014)“…the literature suggest that country specific variables, such as the chosen exchange rate regime, real variables, monetary variables as well as financial variables are linked to (overall) exchange rate volatility, while the effects of these variables seem to be more pronounced for developing countries.” Short-term nominal interest rates comprise information about the future economic conditions and the state of investment opportunities. Dornbusch (1976) has already explained the role of monetary variables on exchange rate volatility in the short run and long run. More recently, Ganguly and Breuer (2010) provide evidence that nominal interest rate changes, have a stabilizing effect on the residual volatilities of the real and nominal exchange rates in developing countries. Giannellis and Papadopoulos (2011) found that interest rates, which serve as a proxy for monetary shocks, impact the foreign exchange rate markets of some Eastern European and European Union economies. For these reasons, we included short-term interest rates in our GARCH model.

In the literature, generally Central bank foreign currency reserves have been introduced in the models. Investigating EMEs, Hviding et al. (2004) find that higher levels of reserves are associated with lower exchange rate volatility. On the other hand, Guimarães and Karacadag (2004) explain: “The single most important impediment to empirical work on intervention’s effectiveness is the lack of publicly available data on daily intervention. Attempts to use proxies of intervention—e.g., the change in the stock of central bank reserves—have not worked. Neely (2001) has shown that even for G-7 countries, changes in reserves are a poor proxy for intervention.” Over the period under examination, the foreign exchange reserve data is lagged and is not completely available because of low transparency. In this perspective, the GARCH model in this study includes Dollar index as an explanatory variable.

Methodology and Data

Chaos Theory has brought evidence of some determinism intermittency/regime switching between regular or laminar phases and chaotic phases (see Hołyst et al., 2001; Mastroeni et al., 2019; Orlando & Zimatore, 2018) which could explain the occurrence of spikes and extreme events. In similar studies, from the theoretical point of view deciding whether reality can be represented through a chaotic model, generally means that dynamics are deterministic and non-linearities are endogenous (Mastroeni et al., 2019; Orlando & Zimatore, 2018). A stochastic model, instead, implies exogenous randomness. Note that this does not mean that we cannot assume a chaotic model contaminated by noise (Kantz, 1994; Nychka et al., 1992). In Section “Lyapunov Exponents Results” it will be defined in detail. A second point is whether the observed time series is outcome of an unobserved dynamical system with higher dimensionality. This brings about more flexibility in modeling and imposes less assumptions on the unobserved dynamical system representing complex financial market. As it is already explained, the financial systems become more complex and have unobserved dynamics. Takens’ (1981) theorem enables to consider unobserved high-dimensional dynamical system and take observed data as its output in low-dimensional subspace. The key idea is to construct an appropriate phase-space and examine the embedded series on that space. This methodology has obvious advantages with respect to time-domain visualization and analysis. As an example, Marwan et al. (2007) explain that recurrence is a fundamental property of dynamical systems, which can be exploited to characterise the system’s behavior in phase-space. This includes the quantification of recurrence plots, which is highly effective to detect, for example, transitions in the dynamics of systems from time series and multiple-equilibria set where the system spends most of its time. The Lyapunov exponent enables to understand whether the system under examination diverges or converges. This is important also for predictability analysis, where if we know the rate of divergence/convergence then we compute distances in time-domain (Vanli et al., 2016; Zheng et al., 2018). On the other hand, it is common that the evolution of financial markets is inherently different from those of natural systems. The decision makers behavior and reaction to the upcoming events and related information influence the outcome. Therefore, we propose to consider and test the stochastic structure as well. In this vein, this study benefits both GARCH and Chaos Theory methods. For chaotic model of the study, see Section “Lyapunov Exponents Results.” With respect to linear stochastic time series models, where first and second moments are time invariant, GARCH method enables to model and estimate deviations from expected values. Despite the heteroskedasticity condition, the OLS coefficients are still unbiased. However, the standard errors and confidence intervals will probably be too narrow, and it causes forecasters to make some false assumptions about their estimations. Thanks to ARCH/GARCH models, heteroskedasticity does not cause a spurious regression, because these models can be set in the presence of changing variance condition.

In related exchange rate movement models, the percent return of the nominal exchange rate against the dollar is used as the dependent variable (Berganza & Broto, 2012). Table 1 provides the descriptive statistics for the variables. The dCds denotes the first difference of the Cds series, which will be introduced in GARCH(1,1) model. In Table 1 we present summary statistics for the variables. As can be seen, all series other than Central bank interest rate series are positively skewed. Moreover, all series show the presence of excess kurtosis. This implies that large changes occur more often than would be the case if these volatility changes series were normally distributed. Using Augmented Dickey Fuller (ADF) (Dickey & Fuller, 1979) and Phillips-Perron (PP) tests (Phillips & Perron, 1988) on the level data, we find that only in the case of the Cds there is evidence of stationarity, producing ADF and PP test statistics that is significant at the 1% level. For the other variables we cannot reject the null hypothesis of a unit root. The appropriate data obtained were inserted into the GARCH (1,1) model given in (1).

\begin{matrix} R_{t} = β_{0} + β_{1} RVI X_{t} + β_{2} (E [{INF}_{t}^{t + k}] - E [{INF}_{At}^{t + k}]) \\ + β_{3} Δ USD X_{t} + β_{4} Cd s_{t} + β_{5} Δ TLREF + ε_{t} \end{matrix}

(1)

ε_{t} N (0, h_{t})

where,

\begin{matrix} h_{t} = a_{0} + a_{1} ε_{t - 1}^{2} + a_{2} h_{t - 1} + a_{3} RVI X_{t} + a_{4} Δ USD X_{t} \\ + a_{5} Δ TLRE F_{t} + a_{6} Cd s_{t} + a_{7} (E [{INF}_{t}^{t + k}] - E [{INF}_{At}^{t + k}]) \end{matrix}

$P_{t}$ : USD/TRY currency value time-series, a unit of U.S. Dollar in terms of the Turkish lira

$R_{t}$ : the difference in logarithms of USD/TRY currency, $R_{t}$ = $\ln (\frac{P_{t}}{P_{t - 1}}) * 100$ ,

a positive $R_{t}$ implies a depreciation of the local currency against the dollar.

$RVI X_{t}$ : the difference in logarithms of Volatility Index, $RVI X_{t}$ = $\ln (\frac{\sqrt{VI X_{t}}}{\sqrt{VI X_{t - 1}}})$

$E [{INF}_{t}^{t + k}]$ : k-period ahead the daily inflation expectation in Turkey

$E [{INF}_{At}^{t + k}]$ : k-period ahead the daily inflation expectation in the United States

$Δ USD X_{t}$ : the difference in logarithms of U.S. Dollar index, $Δ USD X_{t}$ = $\ln (\frac{USD X_{t}}{USD X_{t - 1}})$

$Cd s_{t}$ : percentage of Credit default swap, $Cd s_{t} = \frac{CD S_{t}}{100}$

$Δ TLRE F_{t}$ : the difference in logarithms of reference interest rate of the Turkish lira = $\ln (\frac{TLRE F_{t}}{TLRE F_{t - 1}})$

In model (1), the main purpose is to reveal the causalities for the volatility shock in the USD/TRY exchange rate.

Table 1.

Descriptive Statistics.

Variables	M	Median	SD	Skewness	Kurtosis	Jarque-Bera	ADF	Zivot Andrews	PP
R	0.088	0.025	0.86	1.32	17.77	6,174.92	−24.643	−7.62	− 24.643
Cds	4.15	3.98	0.92	0.38	2.53	22.11	−2.275***	−4.52***	−2.454***
CBRTINT	0.0006	0	0.03	−0.07	12.21	2,331.04	−27.055	−17.72	−27.951
RVIX	0.00033	−0.0027	0.04	0.83	5.67	272.76	−28.691	−12.01	−28.701
USDX	−3.63	−0.0001	0.003	0.48	6.06	282.36	−22.832	−11.62	−22.908
INFDIF	0.0006	0.0005	0.0004	2.13	8.24	1,253.73	−4.186	−5.37	−4.242
dCds	0.002	−0.0033	0.1542	1.4079	14.904	4,103.47	−22.515	−10.851	−22.528

***

Denotes the existence of 1 unit-root within 99% significance level.

****

Denotes the existence of 1 unit-root in the presence of structural break within 99% significance level.

Empirical Results

GARCH Modeling Results

Before we set the GARCH (1,1) model, we have to test whether all the variables are stationary. As shared in detail in the literature review section, the GARCH model has certain parameter constraints. The explanatory variables imposed in the GARCH model should be stationary. If these variables are determined to exhibit non-stationary behavior, then by implementing suitable methods such series will be stationary as demonstrated by Said and Dickey (1984). The Cds series is shown to be non-stationary at 99% significance level. Structural break detected in Cds series may also because of drastic change in monetary policies and/or large Global economic shock amid COVID-19 outbreak. For this reason, the first difference in Cds data is taken to make the series stationary. In Table 1, test statistic implies that the dCds series is stationary. After that dCds can be included in the GARCH model. The installed GARCH (1,1) model and the parameters of this model are given in detail below.

This GARCH (1,1) model of the daily percentage change in the exchange rate, is estimated using data from 1 March, 2019, to 26 October, 2021. In Table 2 ARCH L1 and GARCH L1, the coefficients of both the lagged squared residual and lagged conditional variance ( $σ_{t - 1}^{2}$ ) terms in the conditional variance equation are individually statistically significant at 99% significant level. According to these results, the volatility that comes from the previous period has explanatory power on the volatility of the current period. According to Stock and Watson (2007), one way to test the persistence and long-term effects of variance $σ_{t - 1}^{2}$ and shocks $ε_{t - 1}^{2}$ in the GARCH model is to interpret the sum of these coefficients. As it can be seen in Table 2, this sum approximately (ARCH L1 and GARCH L1) amounts to 0.70, which is summing close to unity and indicates that changes in the conditional variance have a long-term effect. To put it more precisely, a large sum of these coefficients means that large positive and large negative shocks will have a large variance to be estimated in the future (Epaphra, 2017). This explanation also confirms the volatility clustering situation that should be in GARCH models and in chaotic dynamics (Barkoulas et al., 2012). Our findings show that volatility in the exchange rate is influenced by previous volatilities and shocks to the exchange rate. Table 2 depicts that only dCds and VIX coefficients are statistically significant at the 95% levels, meaning that change in Cds and VIX are influencing the volatility of the exchange rate. In a recent study Ozcelebi (2018) states for ten selected OECD countries that the factors causing a rise in short-term interest rates with respect to the U.S. bonds can lead to volatility in exchange rates. Our result refines the finding of this study. One another effect can be considered as follows. Grossmann and Orlov (2022), by using data from 1988 to 2014, find that relative PPP-based exchange rate misalignments (the dollar is highly undervalued) increase exchange rate volatility in developing countries.

Table 2.

GARCH(1,1) Model Estimation Results.

$R (t)$	Coefficient	SE	Z	p > \|Z\|
CBRTINT	0.0005	0.0055	0.1	.918
RVIX	0.0227	0.0049	4.60	.000
USDX	0.0081	0.059	1.38	.168
INFDIF	−0.3033	0.6411	−0.47	.636
dCds	0.0186	0.0012	14.69	.000
cons	0.0009	0.0005	1.96	.050
ARCH L1	0.3796	0.0528	7.18	.000
GARCH L1	0.3137	0.0807	3.89	.000
cons	0.000018	3.26e−06	5.54	.000

On the other hand, the effects of USDX, INFDIF, and TLREF series are not statistically significant. From the perspective of economic and financial decision-making, dCds implies the differential in the risk premium of Turkey. This shows that change in risk perception (higher pricing of the default risk on Turkish assets) coincides with higher volatility in the exchange rate. The presence of a significant constant term in the model indicates that volatility cannot be zero even if the variables determined in this model are assumed to be zero. This is particularly due to openness degree of the Turkish economy to global financial system. Studies find that more open economies experience lower exchange rate volatility (Hausmann et al., 2006; Kearney & Patton, 2000; Obstfeld & Rogoff, 2000). The result of these studies is supported by our results. More recently Calderón and Kubota (2018) investigate 82 countries from 1974 to 2013 and confirm empirically that financial openness matters for real exchange rate stabilization, as well. Furthermore, this evidence is consistent with the results in Eichler and Littke (2018). The authors show that an increase in the availability of information (increasing transparency) about monetary policy objectives decreases exchange rate volatility.

Lyapunov Exponents Results

The pioneering studies Takens (1981), Grassberger and Procaccia (1983), Wolf et al. (1985), Brock et al. (1987), Rosenstein et al. (1993), Kantz (1994) showed that nonlinear systems can follow complex and chaotic dynamics. In these studies various test tools have been proposed to distinguish between nonlinearity and linearity, and to identify chaotic dynamics in time series. The identification of these dynamical properties are crucial because, if the financial time series is chaotic, then it will violate efficient market hypothesis (Cochrane 2005; Das & Das, 2007, Pesaran, 2010). According to “common knowledge” argument (Aumann,1976), any deterministic structure in the series vanishes out once decision makers are informed about it. Therefore, departures from EMH can not be used to obtain excess return without taking proportional risk (Cochrane, 2005; Malkiel, 2003). However, one has to note that until last decade the empirical analysis for financial market dynamics has been limited with high-volume markets in developed countries. On the other hand the research for emerging markets was narrow and has been accelerated in recent years. During low volume transactions, the foreign currency markets in emerging markets can exhibit chaotic behavior and predictability may occur, which in turn yields trade-off possibilities. In this study, the method proposed in Kantz (1994) is used to detect maximal Lyapunov exponent. With respect to Wolf et al. (1985), this method requires low computational intensity and proposes higher test power per required parameters even with small data sets.

The dynamical system is, $f : R^{n} \to R^{n}$ , with the trajectory,

x_{t + 1} = f (x_{t}) + u_{t + 1}, t = 0, 1, 2, . . . .,

(2)

The Lyapunov exponents determine the average rate of divergence or convergence of an orbit. The $u_{t}$ denotes the error term (Nyhcka et al., 1992). One can rarely identify the state of the system at any period. It is assumed that there is an observation function $g : R^{n} \to R$ associated with the system in (2) and the series ${z_{t}}$ given in (3) can be observed.

z_{t} = g (x_{t})

(3)

Under the condition where embedding dimension $(m)$ is large enough, Takens’s theorem (Takens, 1981) implies that the reconstructed trajectory is an embedding of the original trajectory. To reconstruct embedding space, Gencay and Dechert (1992) proposed a sufficient condition and upper-worst case, $m \geq 2 n + 1$ . Accordingly, similar studies report that conditional on data length, embedding can be established even when m is lower than 2n + 1. In this vein, this analysis follows the approach suggested in Abarbanel (1995). Hegger and Kantz (1999) proposes following setup given in (4) and (5), which has widely used in the literature. Let ${z_{t}}$ be observed series with length N. Therefore, the orbit can be demonstrated as (4).

y_{i} = (z_{i}, z_{i + d}, . . ., z_{i + (m - 1) . d}) for all i \in (N - (m - 1) . d)

(4)

Based on observation interval on time-domain, time delay d is taken as 1. In (5) the distance between a reference trajectory and $ε -$ neighbor(s) following the iteration $τ$ , is shown.

D (.) : R^{m} \to R D (y_{t}, y_{t}^{(ε)}; τ) = | y_{t + τ} - y_{t + τ}^{(ε)} |

(5)

In order to measure the maximal Lyapunov exponent one should compute the average distances as a function of $τ$ . This is given in (6).

S (τ) = \frac{1}{T} \sum_{t = 1}^{T} \ln (\frac{1}{| U_{t} |} \sum_{y_{t}^{(ε)} ϵ U_{t}} D (y_{t}, y_{t}^{(ε)}; τ))

(6)

In (6) $| U_{t} |$ measures the cardinality of the set of $ε -$ neighbors of $y_{t}$ . The slope of $S (τ)$ yields the maximal Lyapunov exponent defined in Kantz (1994):

\frac{\partial S (τ)}{\partial τ} ≅ λ_{\max} (t) for any τ in the scaling region, τ \leq τ_{\max}

(7)

The results are presented in Figures 1 and 2. The Tisean Package which has been introduced in Hegger et al. (1999) is used to obtain the output of (7). All computations, Figures 1 and 2 are obtained by using R programme. Figure 1 depicts phase-space representation of volatility series, $R_{t}$ . This representation enables to visualize the evolution of market equilibria. As it can be seen in Figure 1, even though the volatility wastes most of the time around 0.01 to 0.03 interval, increasing number of spikes (extreme points) implies shift in expectations in the Turkish lira against U.S. Dollar. These shifts generate multiple-equilibria. If maximal Lyapunov exponent is positive, then the system eventually will visit these extreme points, shift in expectations will be agreed upon, realized and evaluated by the market.

Figure 1.

Phase-space representation of time evolution of volatility series, $R_{t}$ .

Figure 2.

The time evolution of initially nearby points—volatility series, $R_{t}$ .

In Figure 2 the output of (7) is presented. Figure 2 demonstrates the time evolution of initially nearby points of $R_{t}$ . In Figure 2, y-axis denotes $S (τ)$ , and x-axis denotes $τ$ values. The embedding dimension value is $m = 5$ . The negative values arise from logarithm of the distances lying the interval (0,1]. The $τ$ value where the slope of the curves approximates zero is denoted by $τ_{\max}$ and signifies the last step of scaling range, implying that the dynamic system is still predictable. For the series, $τ_{\max}$ = 11. For the iterations exceeding this value $τ^{*} > τ_{\max}$ , the points are no more affected by the chaotic attractor, thus the system investigated is said to jump through an unpredictable state. One can find a scaling range to compute the maximal Lyapunov exponent. From Figure 2 and equation (7), it is straightforward that the series show chaotic behavior associated with positive maximal Lyapunov exponent. This result is in contrast with the EMH and shows that both currency value and return can be predictable within a short-term interval, and thus trade-off can occur. The predictability of exchange rates, stock markets in emerging economies has already been reported in Vasilios et al. (2019). The foreign exchange market in Turkey is not efficient, and financial agents can earn excessive returns with respect to risk taken. In this sense the result of this analysis gives support to above-mentioned studies.

We can summarize the main reasons for the observed chaotic behavior of foreign exchange market in Turkey. From the perspective of macroeconomic developments, policy making and financial decision are dominated by government objectives, which is considered as unorthodox policy set. Therefore, frequency of implicit Central bank interventions have been augmented. This has been evaluated by financial agents as a problem of confidence in Turkish assets in global financial markets. This leads to a gradual increase in CDS and large fluctuations occur in portfolio flow. Similarly, domestic financial agents engaged in dollarization and foreign currency savings excessively increased. Within the period under examination, the total amount of foreign currency deposits exceeds the Turkish lira deposits in banks and reached 60% to 40%, respectively (https://evds2.tcmb.gov.tr/index.php?/evds/serieMarket). In EMEs, in stress times financial agents hold foreign reserve currency to protect their savings from financial loss through future price fluctuations and expected high inflation. Amid the COVID-19 pandemics crisis, expansionary monetary and fiscal policies have been implemented across the world, which lead to sudden shifts in inflation expectations and fluctuations in pricing behaviors in global financial markets. This can be observed from VIX series, and decreased portfolio inflows to emerging markets, and hence this is the case for Turkey.

Conclusion and Discussion

It is observed that recent studies in the related literature mainly focus on comparison between (seemingly random) deterministic and stochastic models in perspective of their explanatory powers (Orlando & Bufalo, 2022; Orlando et al., 2022). However, the methodology and results of the present study point to the importance of evaluating the outcome of both approaches in a complementary manner. More specifically, this presents itself as follows: in contrast to the chaotic model proposed in Orlando et al. (2022), the setup proposed in this paper (see equation (2)) enables to assume noise in original unobserved dynamical system. This examination of chaotic dynamics contributes to understanding the influence of economy policies and unpredictable actions of policy makers, which drastically diverge from beliefs and expectations of financial agents.

This study also is in line with the studies using GARCH (1,1) modeling of exchange rates. For example, for the case of the Turkish lira, Bilgili et al. (2021) introduced exchange rate volatility as an explanatory variable, whereas the present study aims to find its determinants. The present study supports the importance of financial factors in understanding exchange rates in emerging market economies as demonstrated in Aghion et al. (2001) and Eichengreen (2002). To contribute macroeconomic perspective of the governments, Central banks across the world aim to maintain price stability along with sustaining financial stability. However, the exchange rate volatility is the main source violating financial stability which has been targeted as a policy implication simultaneous to macroeconomic stability. In this study, CBRT overnight reference rate, USD index, Turkey’s CDS premiums, VIX rate, and difference in inflation expectations between Turkey and United States are the explanatory financial time series in the model. In the model, VIX and CDS series represent global dynamics. Note that CDS series reflect risk perception as a function of country’s external variables, such as external debt. In this study, it is found out that among others, the impact of both VIX and CDS on volatility is significant. The analysis showed that exchange rate volatility was influenced by global dynamics rather than domestic dynamics and it was observed that volatility was affected by previous shocks and volatilities, and that these effects were permanent in the long term. It is important to note that in a wavelet coherence analysis of monthly data from 2011 to 2018, Kuskaya et al. (2022) report strong co-movements between the levels of USDX and USD/TRY. On the other hand, in the present study, the log difference of USDX is considered as an explanatory variable. Different from the period under examination of the present study, the results of Kuskaya et al. (2022) may also be due to capital inflows during FED’s Quantitative Easing 2 period. Cairns et al. (2007) report that, emerging market currencies generally depreciate in an environment of elevated global volatility and that the Turkish lira stood out for its high sensitivity to global volatility over the period 2000 to 2006. The results of the present study coincide with the findings of the authors. During the period of pandemics crisis 2020 to 2021, a deep depreciation in the Turkish lira against U.S. Dollar was observed. This finding is in accord with the findings in the literature pointing out that economies offering higher short-term interest rates tend to see their currencies depreciate against lower-yielding currencies in periods of rising capital market volatility (Hofman et al., 2021). This is also consistent with Irving Fisher’s view that the higher-yielding currency would tend to depreciate, over time, against the lower-yielding currency, offsetting the yield advantage. The accord of our results with the previous findings in the literature leads us back to “this time is different syndrome” (Reinhart & Rogoff, 2011). This implies that while countries do weather their financial storms, Reinhart and Rogoff prove that short memories make it all too easy for crises to recur. Similar to the dynamics observed in episodes of higher global volatility over 2000 to 2006 period, it is seen that the uncertainty occurred amid 2019 pandemics outbreak caused global capital market volatility which in turn has increased volatility in the Turkish lira. The linkage between two similar periods may be found in fear of floating (Calvo & Reinhart, 2002), which gives rise to policy maker’s incentives to intervene.

As it is known, currency risks are relevant for international project financing. Currency risks may arise to the extent revenues, operating expenses, and fuel costs are denominated in a currency different from the one for financing. Typically, debt financings for large international projects are denominated in major currencies such as euros, U.S. dollars, and Japanese yen. In the near-future, this type of currency risk will be more material. Former U.S. Treasury Secretary Summers predicts that substantial increases will occur in investment demand caused by energy transition investments in the private sector and that sustainability projects for green transformation may require firms across the world to increase debt issuance (Larry Summers, 2023). On the other hand, Raikar and Adamson (2020) state “certain currencies, especially in emerging markets, do not have the depth in the commercial swaps market to absorb the size of the deals for the tenor-matching financing. Therefore, sponsors may need to access cross-currency swap facilities available from multilateral institutions.” This bottleneck in financing will persist for emerging markets as well as the Turkish economy. The Lyapunov exponents can be used in debt stock sustainability analysis.

The Bank of England Foreign Exchange Joint Standing Committee Turnover Survey data reported in its semi-annual survey of the London foreign exchange market, the world’s biggest, that average daily trading volumes of the Turkish lira against the U.S. dollar fell to $33.5 billion in April 2019 from $35.2 billion in October 2018 and $56.2 billion from April 2018. That translated to a fall in market share of overall forex trading volumes to 1.2% from 2.1%. More recent data point out percentage shares of average daily turnover from 0.7% in October 2021 to 0.3% in April 2022. The low volume of trade is in accordance with violation of EMH and with existence of positive maximal Lyapunov exponent (Ozkaya, 2015). Accordingly, exchange rate volatility is greatly influenced by capital flows in an open economy (Gelman et al., 2015). The low-volume trading also associates with durable capital flows. It is well-known fact that financial markets with daily low volumes can be difficult to sell because there may be little buying interest or a dominant trader capable of trade-off. Additionally, low-volume markets can be quite volatile because the spread between the ask and bid price tends to be wider. Even in more high-depth markets, transactions can be stopped because of intraday low-trading and fear of trade-off possibility (https://www.ft.com/content/6c8385bf-b477-41f6-bfdc-10044f5ec0d7). When it comes to globally low-traded currencies, the foreign currency-forward operations (FX-forward) may have greater influence on daily price fluctuations. The reason is that the existence of the FX-forward may give rise to the synthetic loans. Banking operations in Euromarkets are considered offshore operations, taking place essentially outside the jurisdiction of national banking authorities (outside EU). In case of the lack of appropriate credit lines, and in case of low-trading volume, a desired amount of loan (i.e., USD funds) is not available for a national bank. A major reason for this is due to higher CDS primes and/or rise in global volatility. In this case, national banks can use foreign currency markets judiciously to receive identical cash flow and thus create a synthetic money market loan (Neftci, 2008). This synthetic loan creation system may lead to volatile short-term capital flow amid the shortage of loanable funds. Specifically, a sharp decrease in available loanable funds cause economic uncertainty which yields further spikes in exchange rates. This is in line with the observation noted in Bilgili et al. (2021). In future agenda, the effects of such alternative indebting process on short-term FX debt accumulation should be carefully examined.

There are some limitations of the methodology that is used in this study. Firstly, according to Orlando and Bufalo (2022) (a) GARCH imposes over identifying restrictions on the dynamics of the volatility; (b) Makes it difficult to determine whether volatility shocks are persistent. Secondly, instead of computing the maximal Lyapunov exponent, computing all of the Lyapunov exponents gives a better understanding of the evolution of underlying nonlinear system (Kantz & Schreiber 2003). For future studies, the effect of an adverse shock on the value of maximal Lyapunov exponent can be examined in detail. This will strengthen our knowledge on the relationship between adverse stochastic shocks and deterministic divergence of the systems from equilibrium sets.

Policy Recommendations

It is a clear fact that political stability is unsustainable in an environment where government borrowing is reacting, growth is achieved through current account deficit or government spending, and taxes are constantly rising. From this perspective, investigating chaotic dynamics contributed to study for revealing systemic risks generated by policy makers’ objectives and policy choices. In the conducted studies, it has been observed that the policies expected to reduce volatility achieve their purpose under certain conditions. In December 2021, the Turkish Treasury authorities and CBRT imposed a financial instrument “FX-protected Turkish lira” in order to stabilize exchange rate movements. Thanks to investor demand and associated macroprudential measures which control the excess of FX in national banks, daily USD/TRY volatility has been under 1% threshold. This measure was indeed an exogenous shock having adverse effect (eliminating) on divergence of the FX process. However, inflation targeting fell out of the scope of monetary policy and the eco-political preferences singled out controlled-FX regime. The reason is, probably the sensitivity of voting behavior of the Turkish citizens is higher, as the public confidence surveys show.

For future studies, the framework of chaos theory and nonlinearity analysis give some leeway to identify these endogenous factors (Orlando et al., 2022). The examination of chaotic dynamics contributes to understanding the influence of unpredictable interventions of policy makers, which drastically diverge from common knowledge and expectations of financial agents. In this vein, Brito and Bystedt (2010) report evidence against the positive role of inflation targeting in emerging countries. There may be a trade-off between duration of inflation targeting and policy credibility. Even though the effectiveness of inflation targeting to bring down the volatility still remains obscure, this policy framework has been more durable than other monetary policy strategies (Mihov & Rose, 2008). Finally, the results of present study give support to the findings in Eichler and Littke (2018). The authors propose strong evidence showing that transparent monetary policy reduces exchange rate volatility. They report a robust policy implication: “By reducing the public’s uncertainty about the central bank’s policy objectives (i.e., by increasing central bank transparency), the central bank can reduce the volatility of inflation expectations.”

Carvajal and Bebczuk (2019) report that the development of local sources of long-term finance through capital markets requires several key preconditions to be in force: a certain level of development of the financial sector, including a strengthened and efficient banking sector, institutional investors, and financial openness. Accordingly, a robust institutional environment, predictability based on the rule of law, including a system to warrant the protection of investors. In developing countries, the degree of completeness of these features vary. Financial sector and capital market preconditions limit Turkey’s options to promote long-term finance. The bank-centric financial sector should be reorganized and credit experts should endow with knowledge of future cutting-edge technologies to guide credit channels. Otherwise they are exposed to substantial refinancing, liquidity, and credit risks. Secondly, domestic investor base should be enlarged by ensuring both length of maturity of investments and return expectations from productivity sectors. This will deepen capital market in various dimensions of financial system. Otherwise, in terms of financial development Turkey will stay behind its peers in EME group, despite demonstrating basic preconditions in financial development and openness. Third, curbing the domestic credit growth may avoid the pressure on the Turkish lira and the divergence between foreign and domestic demand. For another, an important factor that has a negative impact on exchange rate volatility is foreign exchange reserves (Eichler & Littke, 2018). As it is previously noted, currency swap with major Central banks should be a policy tool. For further policy recommendations in the long-run, manufacturing industries should be reorganized such that higher technological goods would be produced, branding practices be broadened and export of these products be overwhelmed import goods both in value, differentiation, and variation. In addition, monetary policy and forward guidance have also been put forward as factors affecting exchange rates of developing countries (Mueller et al., 2017). Whether they are communicated or not, the Central bank interventions are deterministic shocks to foreign exchange markets. Moreover, these interventions are coordinated depending upon swap-accords between the Central banks. Therefore, we conjecture that the detected maximal Lyapunov exponent can be used to measure the effect of the Central bank interventions on exchange rate volatility in episodes of excessive movements in exchange rates. We have to note that one implication of the Lyapunov exponents may be determining policy makers’ actions as an endogenous source for systemic risks (Hendricks et al., 2007).

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ata Ozkaya

Data Availability Statement

Research data is available upon to request

References

Abarbanel

H. D. I.

(1995). Analysis of observed chaotic data. Springer.

Abhyankar

Copeland

L. S.

Wong

(1997). Uncovering nonlinear structure in real time stock market indexes: The S&P 500, the DAX, the Nikkei 225, and the FTSE-100. Journal of Business & Economic Statistics, 15(1), 1–14.

Aghion

Bacchetta

Banerjee

(2001). Currency crises and monetary policy in an economy with credit constraints. European Economic Review, 47, 1121–1150.

Aghion

Bacchetta

Ranciere

Rogoff

(2009), Exchange rate volatility and productivity growth: The role of financial development, Journal of Monetary Economics, 56, (4), 494–513

Ahmad

Chien

L. M.

Staveley-O’Carroll

(2019). Nonlinearities in the real exchange rates: new evidence from developed and developing countries. Applied Economics, 51(25), 2731–2743. https://doi.org/10.1080/00036846.2018.1558354

Alexander

Lazar

(2006). Normal mixture GARCH (1,1): applications to exchange rate modelling. Journal of Applied Econometrics, 21(3), 307–336.

Asteriou

Masatci

Pilbeam

(2016). Exchange rate volatility and international trade: International evidence from the MINT countries. Economic Modelling, 58, 133–40. https://doi.org/10.1016/j.econmod.2016.05.006

Aumann

R. J.

(1976). Agreeing to disagree. The Annals of Statistics, 4(6), 1236–1239.

Badshah

I. U.

Frijns

Tourani-Rad

(2013). Contemporaneous spill-over among equity, gold, and exchange rate implied volatility indices. Journal of Future Markets, 33, 555–572. https://doi.org/10.1002/fut.21600

10.

Baillie

R. T.

Osterberg

W. P.

(1997). Why do central banks intervene? Journal of International Money and Finance, 16(6), 909–919.

11.

Barkoulas

J. T.

Chakraborty

Ouandlous

(2012). A metric and topological analysis of determinism in the crude oil spot market. Energy Economics, 34(2), 584–591. https://doi.org/10.1016/j.eneco.2011.10.004.

12.

Berganza

J.C.

Broto

. (2012). Flexible inflation targets, forex interventions and exchange rate volatility in emerging countries. Journal of International Money and Finance, 31(2), 428–444

13.

Bilgili

Ulucak

Soykan

M. E.

Erdogan

(2021). Can exchange rate volatility influence the export positively? Evidence from Turkey under the regime shifts. Global Business Review, 22(3), 588–611.

14.

Bilgili

Ünlü

Gençoğlu

Kuşkaya

(2022).Modeling the exchange rate pass-through in Turkey with uncertainty and geopolitical risk: A Markov regime-switching approach. Applied Economic Analysis, 30(88), 52–70.

15.

Bollerslev

(1986). Generalized autoregressive conditional heteroscedasticity. Journal of Econometrics, 31, 307–327. https://doi.org/10.1016/0304-4076(86)90063-1

16.

Brito

Bystedt

(2010). Inflation targeting in emerging economies: Panel evidence. Journal of Development Economics, 91, 198–210.

17.

Brock

W. A.

Dechert

W. D.

Scheinkman

(1987). A test for independence based on the correlation dimension (SSRI workshop paper 8702). University of Wisconsin–Madison.

18.

Brock

W. A.

Lakonishok

LeBaron

(1992). Simple technical trading rules and the stochastic properties of stock returns. Journal of Finance, 47, 1731–1764.

19.

Cabral

Carneiro

F. G.

Mollick

A. V.

(2016). Inflation targeting and exchange rate volatility in emerging markets (Policy Research Working Paper No. 7712). World Bank. https://openknowledge.worldbank.org/handle/10986/24629

20.

Cairns

Corrinne

McCauley

(2007). Exchange rates and global volatility: Implications for Asia-Pasific currencies. BIS Quarterly Review.

21.

Calderón

Kubota

(2018). Does higher openness cause more real exchange rate volatility? Journal of International Economics, 110, 176–204. https://doi.org/10.1016/j.jinteco.2017.08.002

22.

Calvo

Reinhart

C. M.

(2000). When capital inflows come to a sudden stop: Consequences and policy options. In Kenen

Swoboda

(Eds.), Key issues in reform of the international monetary and financial system (pp. 134–159). International Monetary Fund.

23.

Calvo

Reinhart

(2002). Fear of floating. Quarterly Journal of Economics, 117, 379–408.

24.

Canales

J.K.

Habermeier

. (2004). Structural Factors Affecting Exchange Rate Volatility: A Cross-Section Study, No 2004/147, IMF Working Papers, International Monetary Fund

25.

Carvajal

A. F.

Bebczuk

(2019). Capital markets development causes, effects, and sequencing. World Bank.

26.

Chen

(2020). Impact of economic policy uncertainty on exchange rate volatility of China. Finance Research Letters, 32, 101266,

27.

Cochrane

J. H.

(2005). Asset pricing (rev. ed.). Princeton University Press.

28.

Cooper

R. W.

(2005). Estimation and identification of structural parameters in the presence of multiple equilibria. Eastern Economic Journal, 31(1), 107–130.

29.

Cornelius

(2011). International investments in private equity. Academic Press.

30.

Das

(2007). Chaotic analysis of the foreign exchange rates. Applied Mathematics and Computation, 185, 388–396.

31.

Devereux

M. B.

Lane

P. R.

(2003). Understanding bilateral exchange rate volatility. Journal of International Economics, 60, 109–132.

32.

Dominguez

K. M.

(1998). Central bank intervention and exchange rate volatility. Journal of International Money and Finance, 17(1), 161–190.

33.

Dornbusch

(1976). Expectations and exchange rate dynamics. Journal of Political Economy, 84, 1161–1176.

34.

Dickey

D. A.

Fuller

W. A.

(1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366), 427–31. https://doi.org/10.2307/2286348.

35.

Doroodian

Caporale

(2001). Central bank intervention and foreign exchange volatility. International Advances in Economic Research, 7, 385–392. https://doi.org/10.1007/BF02295768

36.

Ehrmann

Fratscher

Rigobon

(2011). Stocks, bonds, money markets and exchange rates: Measuring international financial transmission. Journal of Applied Econometrics, 26, 948–974.

37.

Eichengreen

(2002). When to dollarize. Journal of Money, Credit and Banking, 34, 1–24.

38.

Eichler

Littke

H. C. N.

(2018). Central bank transparency and the volatility of exchange rates. Journal of International Money and Finance, 89, 23–49, https://doi.org/10.1016/j.jimonfin.2018.07.006

39.

Engle

R. F.

(1982). Autoregressive conditional heteroskedasticity with estimates of the variance of United Kingdom Inflation. Econometrica, 50, 987–1000.

40.

Engle

R. F.

(2001). GARCH 101: The use of ARCH/GARCH models in applied econometrics. Journal of Economic Perspectives, 15(4), 157–168.

41.

Engle

R. F.

Patton

A. J.

(2007). What good is a volatility model? In John

Stephen

(Eds.), Forecasting volatility in the financial markets (pp. 47–63). Elsevier.

42.

Epaphra

(2017). Modeling exchange rate volatility: Application of the GARCH and EGARCH models. Journal of Mathematical Finance, 7, 121–143. https://doi.org/10.4236/jmf.2017.71007

43.

Feng

G. F.

Yang

H.-C.

Gong

Chang

C.-P.

(2021). What is the exchange rate volatility response to COVID-19 and government interventions? Economic Analysis and Policy, 69, 705–719. https://doi.org/10.1016/j.eap.2021.01.018

44.

Figlewski

(1997). Forecasting volatility. Financial Markets, Institutions & Instruments, 6, 1–88. https://doi.org/10.1111/1468-0416.00009

45.

Frankel

J.A.

(2010) Monetary Policy in Emerging Markets: A Survey. National Bureau of Economic Research, Working Paper Series, 16125.

46.

Ganguly

Breuer

J. B.

(2010). Nominal exchange rate volatility, relative price volatility, and the real exchange rate. Journal of International Money and Finance, 29, 840–856.

47.

Gelman

M. A.

Jochem

Reitz

M. P.

(2015). Taylor real financial market exchange rates and capital flows. Journal of International Money Finance, 54, 50–69. 10.1016/j.jimonfin.2015.02.004

48.

Gencay

(1998). The predictability of security returns with simple technical trading rules. Journal of Empirical Finance, 5, 347–359.

49.

Gencay

Dechert

W. D.

(1992). Algorithm for the n Lyapunov exponents of an n-dimensional unknown dynamical system. Physica D, 59, 142–157.

50.

Giannellis

Papadopoulos

A. P.

(2011). What causes exchange rate volatility? Evidence from selected EMU members and candidates for EMU membership countries. Journal of International Money and Finance, 30, 39–61.

51.

Grassberger

Procaccia

(1983). Dimension and entropies of strange attractors from a fluctuating dynamics approach. Physica D, 13, 34–54.

52.

Grossmann

Love

Orlov

A. G.

(2014). The dynamics of exchange rate volatility: A panel VAR approach. Journal of International Financial Markets, Institutions and Money, 33, 1–27.

53.

Grossmann

Orlov

A. G.

(2022). Exchange rate misalignments, capital flows and volatility. The North American Journal of Economics and Finance, 60, 101640.

54.

Guimarães

Karacadag

(2004). The empirics of foreign exchange intervention in emerging market countries: The cases of Mexico and Turkey (IMF Working Paper No. 04/123). IMF.

55.

Hagtvedt

(2009). Stock return dynamics and the CAPM anomalies. Applied Economics Letters, 16(16), 1593–1596.

56.

Hausmann

Panizza

Rigobon

(2006). The long-run volatility puzzle of the real exchange rate. Journal of International Money and Finance, 23, 93–124.

57.

Hausmann

Panizza

Stein

(2001). Why do countries float the way they float? Journal of Development Economics, 66, 387–414.

58.

Hegger

Kantz

Schreiber

(1999). Practical implementation of nonlinear time series methods: The TISEAN package. Chaos, 9, 413.

59.

Hendricks

Kambhu

Mosser

(2007). New directions for understanding systemic risk. Federal Reserve Economic Policy Review, 13, 2.

60.

Hofmann

Shim

Shin

H.S.

(2021). Emerging Market Economy Exchange Rates and Local Currency Bond Markets Amid the COVID-19 Pandemic. Available at SSRN: https://ssrn.com/abstract=3761875 or http://dx.doi.org/10.2139/ssrn.3761875

61.

Hołyst

Żebrowska

Urbanowicz

(2001). Observations of deterministic chaos in financial time series by recurrence plots, can one control chaotic economy? European Physical Journal B, 20(4), 531–535.

62.

Hviding

Nowak

Ricci

L. A.

(2004). Can higher reserves help reduce exchange rate volatility? [IMF Working Paper No: WP/04/189].

63.

Kantz

(1994). A robust method to estimate the maximal Lyapunov exponent of a time series. Physics Letters A, 185, 77–87.

64.

Kantz

Schreiber

(2003). Nonlinear time series analysis (2nd ed.). Cambridge University Press. https://doi.org/10.1017/CBO9780511755798

65.

Kearney

Patton

A. J.

(2000). Multivariate GARCH modeling of exchange rate volatility transmission in the European monetary system. The Financial Review, 35(1), 29–48.

66.

Keefe

H. G.

Saha

(2022). Threshold effects of openness on real and nominal effective exchange rates in emerging and developing economies. The World Economy, 45, 1386–1408. https://doi.org/10.1111/twec.13212

67.

Kim

J.-M.

Kim

D. H.

Jung

(2021). Estimating yield spreads volatility using GARCH-type models. The North American Journal of Economics and Finance, 57, 101396.

68.

Kuskaya

Toguc

Bilgili

(2022). Wavelet coherence analysis and exchange rate movements. Quality & Quantity, 56, 4675–4692.

69.

Kriljenko

MJIC

. (2003). Foreign exchange intervention in developing and transition economies: Results of a survey (IMF Working Paper 03/95). International Monetary Fund.

70.

Larry Summers (2023, January 20). Davos. https://www.weforum.org/agenda/2023/01/larry-summers-central-banks-soften-monetary-policy-inflation-davos-2023/

71.

Malkiel

(2003). The efficient market hypothesis and its critics. Journal of Economic Perspectives, 17(1), 59–82.

72.

Marwan

Romano

M. C.

Thiel

Kurths

(2007). Recurrence plots for the analysis of complex systems. Physics Reports, 438(5–6), 237–329.

73.

Mastroeni

Vellucci

Naldi

(2019). A reappraisal of the chaotic paradigm for energy commodity prices. Energy Economics, 82, 167–178.

74.

Mihov

Rose

(2008). Is old money better than new? Duration and monetary regimes. Economics: The Open-Access, Open-Assessment E-Journal, 2, 1–24.

75.

Mishkin

Schmidt-Hebbel

(2007). Does inflation targeting make a difference? In Mishkin

Schmidt-Hebbel

(Eds.), Monetary policy under inflation targeting (pp. 291–372). Banco Central de Chile.

76.

Muhammed Rafi

O. P. C.

Ramachandran

. (2018). Capital flows and exchange rate volatility: experience of emerging economies. Indian Economic Review, 53(1/2), 183–205.

77.

Mueller

Tahbaz-Salehi

Al.

Vedolin

(2017). Exchange rates and monetary policy uncertainty. The Journal of Finance, 72, 1213–1252.

78.

Neely

(2001). The practice of central bank intervention: Looking under the hood. Federal Reserve Bank of St. Louis Review, 83, 1–10.

79.

Neftci

N. S.

(2008). Principles of financial engineering (2nd ed.). Academic Press Advanced Finance.

80.

Nychka

Ellner

Gallant

A. R.

McCaffrey

(1992). Finding chaos in noisy system. Journal of the Royal Statistical Society Series B, 54, 399–426.

81.

Obstfeld

Rogoff

(2000). New directions in open macroeconomics. Journal of International Economics, 50, 117–153.

82.

Orlando

Bufalo

(2022). Modelling bursts and chaos regularization in credit risk with a deterministic nonlinear model. Finance Research Letters, 47, 102599.

83.

Orlando

Bufalo

Stoop

(2022). Financial markets’ deterministic aspects modeled by a low-dimensional equation. Scientific Reports, 12, 1693. https://doi.org/10.1038

84.

Orlando

Zimatore

(2018). Recurrence quantification analysis of business cycles. Chaos Solitons Fractals, 110, 82–94.

85.

Ozcelebi

(2018). Analysis of the role of exchange rate volatility in monetary policy conduction in OECD countries: Empirical evidence from panel-VAR Models. Chapter 6, 94–97, In Bobek

, (Ed.). Trade and global market (pp. 84–97). InTech. https://doi.org/10.5772/intechopen.70966

86.

Ozkaya

(2015). A model of active trading by using the properties of chaos. Digital Signal Processing, 39, 15–24.

87.

Paolella

M. S.

(2018). Linear models and time-series analysis: regression, ANOVA, ARMA and GARCH. John Wiley & Sons.

88.

Pesaran

M. H.

(2010). Predictability of asset returns and the efficient market hypothesis (Cesifo Working Papers, No. 3116).

89.

Phillips

P. C. B.

Perron

(1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346. https://doi.org/10.2307/2336182

90.

Poon

S. H.

Granger

(2005). Practical issues in forecasting volatility. Financial Analysts Journal, 61(1), 45–56. https://doi.org/10.2469/faj.v61.n1.2683

91.

Raza

S. A.

Afshan

(2017). Determinants of Exchange Rate in Pakistan: Revisited with Structural Break Testing. Global Business Review, 18(4), 825–848. https://doi.org/10.1177/0972150917692210

92.

Raikar

Adamson

(2020). Renewable energy finance: Theory and practice. Academic Press. https://doi.org/10.1016/C2017-0-04830-9

93.

Reinhart

C. M.

Rogoff

K. S.

(2011). This time is different: Eight centuries of financial folly. Princeton University Press.

94.

Rosenstein

M. T.

Collins

J. J.

De Luca

C. J.

(1993). A practical method for calculating largest Lyapunov exponents from small data sets. Physica D, 65, 117.

95.

Said

S. E.

Dickey

D. A.

(1984). Testing for unit roots in autoregressive-moving average models of unknown order. Biometrika, 71(3), 599–607. https://doi.org/10.2307/2336570

96.

Scheinkman

J. A.

LeBaron

(1989). Nonlinear dynamics and stock returns. Journal of Business, 62, 311–337.

97.

Stock

J. H.

Watson

M. W.

(2007). Why has U.S. Inflation become harder to forecast? Journal of Money, Credit and Banking, 39, 3–33. https://doi.org/10.1111/j.1538-4616.2007.00014.x

98.

Smallwood

A. D.

(2019). Analyzing exchange rate uncertainty and bilateral export growth in China: A multivariate GARCH-based approach. Economic Modelling, 82, 332–344.

99.

Takens

(1981). Detecting strange attractors in turbulence. In Rand

Young

LS.

(Eds.), Dynamical Systems and Turbulence, Warwick 1980. Lecture Notes in Mathematics, vol 898, pp. 366–381. Springer. https://doi.org/10.1007/BFb0091924

100.

Vaidyanathan

Krehbiel

(1992), Does the S&P 500 futures mispricing series exhibit nonlinear dynamics dependence across time? Journal of the Future Market, 12, 659–677.

101.

Van Greuning

Bratanovic

S. B

. (2020). Analyzing banking risk (fourth edition): A framework for assZivot-Andreessing corporate governance and risk management (pp. 261-279). World Bank Publications.

102.

Vanli

N. D.

Tunc

Donmez

M. A.

Kozat

S. S.

(2016). Growth optimal investment in discrete-time markets with proportional transaction costs. Digital Signal Processing, 48, 226–238.

103.

Vasilios

Rangan

Gil-Alana

L. A.

Wohar

M. E.

(2019). Are BRICS exchange rates chaotic? Applied Economics Letters, 26(13), 1104–1110. https://doi.org/10.1080/13504851.2018.1537473

104.

Wolf

Swift

J. B.

Swinney

H. L.

Vastano

J. A.

(1985). Determining Lyapunov Exponents from a time series. Physica D, 16(3), 285–317.

105.

Zheng

Liu

Shen

(2018).Very short-term maximum Lyapunov exponent forecasting tool for distributed photovoltaic output. Applied Energy, 229, 1128–1139.

106.

Zivot

Andrews

D. W. K.

(1992). Further evidence on the great crash, the oil-price shock, and the unit-root hypothesis. Journal of Business & Economic Statistics, 10(3), 251–270.

Domestic and Global Causes for Exchange Rate Volatility: Evidence From Turkey

Abstract

Plain Language Summary

Keywords

Introduction

Literature Review

Methodology and Data

Empirical Results

GARCH Modeling Results

Lyapunov Exponents Results

Conclusion and Discussion

Policy Recommendations

Footnotes

Declaration of Conflicting Interests

Funding

ORCID iD

Data Availability Statement

References