Explaining the impact of exchange rate volatility on economic growth

(1)

Master’s Thesis

“EXPLAINING THE IMPACT OF EXCHANGE

RATE VOLATILITY ON ECONOMIC

GROWTH”

Author: Arnanda Hafhiza Firda Wibowo

Student Number: 10671463

Supervisor: Dr. Kostas Mavromatis

MSc Economics - Monetary Policy and Banking

15 August 2018

(2)

STATEMENT OF ORIGINALITY

This document is written by Arnanda H. F. Wibowo, who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the full contents.

(3)

ABSTRACT

This thesis studies the impact of exchange rate volatility on economic growth for eight East Asian1_{economies, using quarterly data for the period of 1999 – 2016. This paper builds on the} dynamic panel data model and the Generalized Method of Moments (GMM) is applied as the estimation method. Using a measure for exchange rate volatility, it was found that the volatility increases following a switch from a fixed to flexible exchange rate regime. Subsequently, we found a robust negative relationship between exchange rate volatility and long run economic growth. The impact of volatility on growth are investigated through three transmission channels: international trade, international capital market and macroeconomic stability. It was found that higher volatility slows down growth by lowering trade. The resulting coefficients on the international capital market and macroeconomic stability are however mixed and subject to further research. Our analysis remains robust under an alternative of volatility measure.

Key words: Exchange Rate Volatility, Economic Growth, East Asia.

1 Throughout this paper, “East Asia” refers to the eight countries for which we have complete data in the region (in alphabetical order): Hong Kong, Indonesia, Japan, South Korea, Malaysia, Philippines, Singapore and Thailand.

(4)

1. INTRODUCTION

Exchange rates are one of the crucial elements that bring about uncertainty in economic activities. The risk in exchange rates is therefore an important factor in international macroeconomic policy analysis as its potential linkages to economic variables are likely to affect long-run growth performance. This development has contributed to reshape the international monetary system as well as stimulated the policy debate over the appropriate exchange rate policy, making the understanding of the key drivers of exchange rate volatility and the channels of its manifestation both a theoretical and empirical matter at hand. The transmission of the volatility to domestic economy is of greater importance when the economies in question benefit substantially from cross border transaction activities (i.e. small open economies). In this thesis, we focus on the economies of eight East Asia countries following major currency devaluations in the region in the aftermath of the Asian Financial Crisis of 1998 (henceforth Asian crisis). The focus is on the notion of exchange rate volatility by placing East Asia’s growth experience in a broad and international perspective. This will allow us to better comprehend the key drivers behind the extremely rapid growth figures of these economies over the last decade. To conduct this study, the following research question is formulated: “To what extent have exchange rate volatility been responsible for the real

GDP growth in eight East Asia economies after the Asian Financial Crisis 1998?”.

Exchange rate volatility can be defined as persistent fluctuations in the value exchange rates, which can be measured by the standard deviation or variance of exchange rate movements (Alagidede & Ibrahim, 2017). These fluctuations are commonly caused by real and financial aggregate shocks and are widely used to measure risks, including risks in asset pricing, portfolio optimization, option pricing or risk management (Aghion et al., 2009). In other words, the volatility in the exchange rate presents an accurate example of risk measurement that is widely applicable in economic decision making. Having said that, in assessing the impact of exchange rate volatility on the real economy, growing literature has put forward the attention on its impact towards economic growth through different economic variables. Remarkably, exchange rate volatility has been found to affect international trade (Doyle, 2001; Clark et al., 2004; Tenreyro, 2007), inflation (Danjuma et al., 2013), investment (Serven, 2003), and all in all growth (Mundell, 1995; Holland et al., 2011). In addition, the volatility in the exchange rate can also affect the domestic value of debt payments (i.e. for debts denominated in foreign currency), and workers’ remittances, which will consequently affect domestic wages as well as employment, output and most importantly prices (Diebold &

(6)

Nerlove, 1989). More specifically, in the presence of exchange rate volatility, there is higher uncertainty exports and imports prices, which means that the value of international reserves and open position in the foreign exchange markets will also respond to these fluctuations accordingly.

A common supposition is that the main driving factor of exchange rate volatility is the choice of exchange rate regime: fixed or floating. Schnabl (2007) argues that a fixed exchange rate can contribute to macroeconomic stability and help to minimize, by eliminating exchange rate risks. On the contrary, flexible exchange rate allows for an easier adjustment when responding to asymmetric and country specific real shocks (Schnabl, 2007). It is trivial to think that the risks associated with exchange rate volatility would discourage economic agents from engaging in cross border transaction activities. However, growing evidence have also pointed out that there are good instruments to hedge against the volatility, meaning that the effect should in the end be immaterial (Tenreyro, 2007). In light of this view, the discussion remains on the growth effect of the volatility, that is the center of investigation in this study.

In this thesis, we consider the impact of exchange rate volatility on growth for eight East Asia economies: Hong Kong, Indonesia, Japan, South Korea, Malaysia, Philippines, Singapore and Thailand. The regional economies continue to fuel the debate among economists and policymakers from their remarkable growth figures that have outpaced the growth of other world regions in the last decades. Many factors2_{have been cited to be responsible for this} growth, among these factors are trade openness, high savings rate, human capital accumulation and macroeconomic policy (Bloom & Finlay, 2009). Nevertheless, one explanation that has been insufficiently examined is the role of exchange rate volatility. For the regional economies, the integration to the world economy is engaged from the promotion of international trade, where up to the Asian crisis3_{, the exceptional degree of stability in the exchange rate has been} regarded as an important constituent of the East Asian Miracle4_{(World Bank, 1993).}

As depicted in figure 1, prior the crisis, the best performers in the region grew of over 5.5% annually, on average higher than the figure for the rest of the world. Growth figures this high are proven to be unprecedented in history (Radelet, Sachs & Lee, 1997). Nonetheless, as the region moves towards a more flexible exchange rate regime following the crisis, one can

2_{These factors mostly accounted for the East Asian growth phenomenon during the 1960s-1990s}

3_{The Asian Financial crisis, also called the “Asian Contagion,” was a sequence of steep currency devaluations that began in the summer of 1997 and spread} through many Asian markets. The foreign exchange markets failed following the abandonment the peg of the local currency to the U.S. Dollar. The currency declines spread rapidly especially through Southeast Asia, which in turn caused major stock market declines, fall in import revenues and government upheaval.

4_{Refers to eight countries in East Asia – Japan, South Korea, Taiwan, Hong Kong, Singapore, Thailand, Malaysia, and Indonesia, because of their economies’} dramatic growth. The real per capita GDP rose twice as fast as in any other regional grouping between 1965 – 1990.

(7)

see that the average regional growth have declined, as can be seen in figure 2 (IMF World Economic Outlook, 2017). This can be seen as the starting argument of the interplay between exchange rate and economic growth.

Figure 1: Pre-crisis Real GDP Growth

Source: International Monetary Fund

Figure 2: Post-crisis Real GDP Growth

(8)

Subsequently, our main hypothesis is that higher exchange rate volatility is detrimental to long-run economic growth. We use a panel of eight countries in East Asia to analyze how economic growth is affected by the volatility in the exchange rate following a financial distress (i.e. Asian crisis). We aim to model this relationship using a de facto volatility measure, allowing us to spare the shortcomings and complications involved with classifying the different exchange rate regimes (both officially declared or as classified by International Monetary Fund). Against this backdrop, we refer to the exchange rate arrangements of the International Monetary Fund (IMF), which in this case differs from the official arrangements announced by the respective national central banks. This will allow us to focus our analysis on the volatility effects more comprehensively rather than on the exchange rate policies itself.

In this thesis, we apply a measure of exchange rate volatility as the standard deviation of the logarithm of the exchange rates. To address the issue of endogeneity across all variables in a growth analysis, we estimate the dynamic panel data model using the Generalized Method of Moments (GMM) methodology. A set of control variables as well as interaction terms are subsequently incorporated into the model to improve our analysis. Using the dataset over the period of 1999Q1 – 2016Q4, our estimation result suggests a significant negative relationship between exchange rate volatility and economic growth. The impact of the volatility is analyzed under three transmission channels through which volatility can affect growth: international trade, international capital market and macroeconomic stability as proxied by inflation. Our estimates remain robust under different model specifications and an alternative measure of volatility. This paper contributes by shedding light on a number of important issues, in particular regarding the implications of uncertainty in the exchange rate in East Asia following a number of structural changes and their increasing integration to the global market.

This thesis is organized as follows. Section 2 provides the literature review, which clarifies all the fundamental issues behind the relationship between exchange rate volatility and economic growth. Thereafter, section 3 presents the model and methodology, which contains the framework to measure volatility and the empirical specifications. Whereafter, the description of the data is presented in section 4 and the empirical results are shown in section 5. In addition, a robustness test is presented in section 6 and section 7 discusses the policy implications and recommendation. Finally, section 8 gives the conclusion.

(9)

2. LITERATURE REVIEW

The purpose of this section is to provide an overview of the literature regarding the impact of exchange rate volatility on economic growth. We begin with the theoretical framework behind the choice of exchange rate regime and the resulting volatility depending on the regime choice. In addition, we identify a number of drivers of the volatility. We extend our analysis to focus on the flexibility of the exchange rate in eight East Asian countries and the management of the exchange rate by the national monetary authority. The next subsection will provide a review on the current literature on the subject, followed by an overview of the transmission channels through which volatility affects growth.

2.1. THEORY: EXCHANGE RATE REGIME AND VOLATILITY

One of the important decisions that the monetary authority have to make is the decision with respect to choosing the optimal exchange rate regime. Their main consideration must be based on what they perceive to be in the best interest of the country both economically and politically. This decision will in turn determine the path of their intervention and the degree of monetary (central bank) independence. For simplicity, the IMF, in official terms, ranks the exchange rate arrangements on the basis of their degree of flexibility and the formal or informal commitments to the exchange rate paths (IMF, 2016). These arrangements assess the implications of the regime choice as well as the degree of monetary policy independence (IMF, 2016). Subsequently, exchange rate arrangements can be classified into four broad categories: 1. Fixed regime, also known as hard peg (i.e. currency board arrangements); 2. Soft pegs (i.e. crawling pegs, stabilized arrangements and craw-like arrangements); 3. Floating regime (i.e. managed floating and free floating); 4. Residuals (IMF, 2016).

Fixed Exchange Rate

On the one hand, a nation may adapt a ‘fixed’ regime, also widely known as a hard peg. Martin (2017) defines a hard peg as meaning that the value of the currency is pegged to another currency or a basket of currencies, and this can take into several forms. One is that the rate is set by law and without special provisions to defend the value of the currency (Martin, 2017). Alternatively, they may as well create a “currency board”, who is responsible to hold sufficient reserves to convert the domestic currency into the designated reserve currency and at a predetermined exchange rate (Martin, 2017). These reserves are used to intervene in the FOREX market when necessary in order to maintain the fixed rates. The main advantage of

(10)

keeping the value of the exchange rate fixed is that exchange rate risks are eliminated accordingly (Martin, 2017). This will provide for more stability for business and investment activities between the two currencies. However, this also means that the pegged exchange rate is subject to higher exposure of speculative attack. Not only that, the monetary authority also loses the autonomy of domestic monetary policy (i.e. they could either target inflation or liberalize international capital movements). Contemporary economic theory asserts that a nation cannot simultaneously maintain a fixed exchange rate, liberalize capital movements and to have an independent monetary policy (Thirwall, 2013). This means that the intervention of the authority is limited when they have to respond to adverse (an unanticipated) shocks.

Floating Exchange Rate

On the other hand, when the value of a currency is allowed to fluctuate following other major currencies in the foreign exchange (henceforth FOREX) market, the nation can be referred to have adopted a “free-floating” regime. As a result, the value of the currency is fundamentally decided by the supply and demand of the currency in the FOREX market (IMF, 2016). One main advantage of this arrangement is that the monetary authority is allowed to have more autonomy over domestic monetary policy (Martin, 2017). Its main drawback, however, is related with stability issue. Allowing the value of the currency to fluctuate ‘freely’ in the market will expose the country to a greater exchange rate risks when it comes to international transactions, which could potentially destabilize both the balance sheets and international capital flows (Martin, 2017). On a lesser extreme, monetary authority is allowed to intervene in the event of economic disturbances. They do so by designating a band around a determined exchange rate and intervene in the FOREX market when the currency hits the limits of its upper or lower value (Martin, 2017). In such instance, the country is said to have adopted a “managed floating” regime. The monetary authority in this regard will mainly intervene at some level in order to maintain their macroeconomic stability as well as minimizing the volatility impact (Martin, 2017).

Recent literature on this issue remains divided with respect to the optimal exchange rate regime for an economy. The issue becomes more prominent the fact that stability remains in the center of macroeconomic policymaking. On the one hand, evidence have pointed out that hard pegging the exchange rate contributes to sound fiscal position and low inflation, which promote investment and subsequently boost long – term growth. On the other hand, it should still be noted the risk of capital flight that could result in a currency collapse that underlines the vulnerability of a fixed exchange rate (Ghosh and Ostry, 2009). In general, exchange rate

(11)

can be said to exhibit higher volatility under a more flexible exchange rate regime. Mussa (1986) was among the first to show that in the short-term, there is higher variability in the exchange rate under a floating regime. This is because the change in the real exchange rate would then be more persistent. Furthermore, the increase in volatility is also observed following a regime switch from a fixed into a floating one (Rose, 1996). A common explanation is that the volatility is ‘bottled up’ in a fixed regime as the value of the exchange rate is pegged to a certain level, implying that there is virtually no exchange rate risk (Rose, 1996). As a result, when the system collapses into a floating one, the aforementioned volatility is released, giving more fluctuating movements in the fundamental value of the exchange rate in the FOREX market.

In view of this, it can be argued that different behaviors of exchange rate volatility are observed under different regime. To support this, a number of studies have yielded similar conclusion. Applying a Threshold GARCH (TGARCH) model on the sample of 5 East European countries, Kocenda and Valachy (2006) found that volatility is in fact greater under a floating regime. As noted previously, Rose (1996) supports this notion as he showed empirically that there is a positive and significant relationship between exchange rate band and the volatility. He interprets this finding as meaning that the regime choice does matter in explaining for the volatility. However, it is worth to note that the volatility patterns are different across countries and that the impact of interest rate differential on volatility are likely to be higher under a more flexible regime, albeit not significant (Giannellis & Papadopoulos, 2011). This is because under a fixed regime, the foreign “anchor” country is the one that now sets the domestic interest rates as the monetary policy is no longer independent (Giannellis & Papadopoulos, 2011).

2.2. THEORY: DRIVERS OF EXCHANGE RATE VOLATILTY

It is worth to note that there is no consensus in the literature on the factors affecting exchange rates and their volatility. One of the main reasons is the difficulty in predicting the movements of exchange rates in the short-run (Meese & Rogoff, 1983). The theory points out that since exchange rate is an endogeneous variable, its volatility is likely to depend on the volatility of economic fundamentals, meaning that whenever the actual rates deviate from the long-run (sustainable) values, the volatility is expected to be higher (Giannellis & Papadopoulos, 2011). That being said, exchange rate volatility corresponds to the short-run fluctuations in its value around its long-run trends. Given the structure of the economy, a number of variables are

(12)

considered as the contributing factors, including monetary, real and financial variables. Financial variables for example, are found to be more important when it comes to developing countries (Giannellis & Papadopoulos, 2011). It is argued that when there are higher financial linkages, the volatility is lowered, in which this is the case in developing countries (Giannellis & Papadopoulos, 2011). In addition, Barkoulas et al. (2002) employed a model for exchange rate volatility as a result of three components: 1. Changes in economic fundamental factors that could drive exchange rate process (i.e. consumer purchasing power); 2. Microstructure changes in foreign exchange market; 3. Signal of future policy changes (w.r.t. inflation, interest rate, and/or money supply). In recent times, increase in cross-border flows together with the upward trend towards capital account liberalization, currency speculations and rapid technological advancement are also named to cause higher fluctuations in the exchange rate.

Broadly speaking, variations in exchange rates can also arise from domestic real shocks that affect the supply and demand. Or, by external real and nominal shocks, which affects money supply in the economy. Referring to the standard Dornbusch (1976) model, the unanticipated monetary policy shocks can generate large variations in the exchange rate. This means that the nominal shocks can indeed affect the real exchange rate especially in the short-run, and that the real rate consequently deviates from its long-run equilibrium value. The popular explanation relates to the fundamentals of supply and demand, where the former relies on the level of output capacity that is consistent with the Balassa-Samuelson (1964) hypothesis. The Balassa-Samuelson hypothesis infers that a productivity increase in the tradable sectors will essentially push the wages in all sectors (including the non-tradable), and therefore in the economy as a whole (Balassa, 1964). Since the productivity in the non-tradable sectors remains unchanged in response to higher wage push, the prices of non-tradable goods will increase, and likewise its relative price to the tradable goods (Balassa, 1964). This will put an appreciating pressure on the domestic exchange rate. The demand factors, on another note, can be associated with the role of government spending, while the external shocks will reflect to changes in the terms of trade, trade openness and capital flows (Alagiedede & Ibrahim, 2016).

To analyze the effects of real shocks on volatility, Clarida and Gali (1994) refer to fluctuations in business cycle. This finding is supported by Gauthier and Tessier (2002) who found that movements in exchange rates are highly triggered by supply shocks. In addition, Hausmann et al. (2006) argue that the real exchange rate in the developing countries tend to be more volatile than that of their developed counterparts. The reasoning for this is that developing countries tend to have higher exposure to real and nominal shocks, making the real rates to be more sensitive to greater variations in shocks. In spite of this fact, it was found that the real

(13)

exchange rate volatility in developing countries to be reduced by external financial liabilities as external debt contracts financial constraints, which in the end lowers the implied volatility because they are less sensitive to shocks now(Devereux and Lane, 2003). Amid these factors, it is worth to note the driver of exchange rate volatility beyond the traditional economic theory. One of which is in a finding by Hausmann et al. (2006) that there is a strong correlation between exchange rate volatility and the quality of institutions, as proxied by the rule of law.

2.3. EXCHANGE RATE FLEXIBILITIES IN EAST ASIA

According to the International Monetary Fund (IMF), monetary authorities in East Asia can be said to have adopted a variety of exchange rate policies, although it is observed that there is gradual switch towards somewhat greater flexibility in the regional economies. The regime varies from Hong Kong’s currency board system, linking the exchange rate to the U.S. Dollar, to the free-floating regime of the Japanese exchange rate. Categorizing a government’s exchange rate policy can be complicated, especially when the economy is under pressure of going through a period of financial distress. Subsequently, the stability in East Asian currencies was also subject to complication prior to the Asian crisis as there was a wide variety of officially declared exchange rate regime and lack of transparency with regard to how these regimes were actually operated.

Before the Asian crisis, Hong Kong, Indonesia, South Korea, Malaysia, Philippines, Singapore and Thailand had pegged their exchange rate to the U.S. Dollar. Although these economies used a variety of exchange rate systems, their common peg to the Dollar provided an informal common monetary standard that has increasingly enhanced macroeconomic stability in the region (McKinnon & Schnabl, 2004). The outbreak of the crisis has shown however, that de facto hard pegging to the U.S. Dollar had not coped very well especially when faced with massive capital inflows (Kwan et al., 1998). With the exception of Hong Kong, debtor nations such as Indonesia, South Korea, Malaysia, Philippines and Thailand were forced to abandon their peg (McKinnon & Schnabl, 2004). Not only that, creditor country such as Singapore, also experienced moderate depreciations even though the Singaporean Dollar was not attacked. Japan, on the other hand, let the yen to float downward somewhere between 1997 to 1998, which further aggravated the crisis for the rest of the countries in the region (McKinnon & Schnabl, 2003). The subsequent floats of the currency have brought about the very sharp fluctuations and greater uncertainty. The post crisis behavior of foreign exchange rate in East Asia has in general changed following the crisis-hit countries shifting to more open

(14)

economic policies by implementing capital and foreign exchange market liberalization (Ogawa & Yang, 2004).

Given the rapid and highly volatile flows of capital in the years before the crisis, foreign exchange developments among regional central banks remain high in many instances. According to IMF (2016), adopting fixed exchange rate policy means that the monetary authority chooses to use the exchange rate as the unique monetary anchor. On the contrary, countries with de facto floating exchange rate arrangements, they could use a variety of monetary anchor, one of which is the inflation targeting framework (IMF, 2016). This framework involves a public announcement of numerical targets for inflation, where the central banks set through a strong commitment in order to achieve these targets, typically over a medium-term horizon (IMF, 2016). In that sense, all inflation targeting central banks use interest rates as the operating monetary policy target, where greater importance is placed with regard to the inflation effect on wage and price stabilization as the main indicator of macroeconomic stability. Accordingly, the main objective of inflation targeting is to induce public’s expectations of inflation to converge with the central bank’s inflation target level as indicated prior the public announcement and successful attainment of that target level (IMF, 2016).

The exchange rate policy in Indonesia, South Korea, Japan, Philippines and Thailand can be categorized as the floating regime. The national central banks implemented the inflation targeting framework of monetary policy in the year of 2005, 2001, 2013, 2002 and 2000 respectively. (IMF, 2017). Adopting this policy framework means that the national central banks aim at the stability of the exchange rate around its fundamental from stable figures of inflation. To implement this, the monetary authority holds power to conduct monetary policy through the establishment of monetary targets (i.e. money supply or interest rates), in order to achieve their primary goal of keeping inflation close to their target. Nevertheless, monetary authority in this regard, can still intervene in the FOREX market at some level in order to keep the rate from undergoing excessive fluctuations in its value (Rajan, 2010). On the operational level, the aforementioned monetary objectives rely on the use of various instruments. As an example, in the case of the Indonesian exchange rate management, the central bank uses the instruments that include open market operations on FOREX markets, discount rate decision, credit regulation and the level of reserve requirements (Bank Indonesia, 2013).

On another note, Singapore, Malaysia and Hong Kong do not implement inflation targeting framework as their main monetary policy framework. In Singapore, the Monetary Authority of Singapore (MAS) adapted a managed floating regime with monetary policy

(15)

framework of exchange rate anchor to composite. This means that unlike its neighboring countries, MAS gives up control over domestic interest rates and money supply. The Singaporean Dollar (SGD) is managed against a basket of currencies of its major trading partners (Rajan, 2010). Furthermore, rather than kept to a fixed value, the trade weighted exchange rate is allowed to fluctuate within an undisclosed policy band. The band provides the flexibility to accommodate the short-term fluctuations in the foreign exchange markets as well as some buffer in the estimation of the country’s equilibrium exchange rate, which tends to be unknown (Monetary Authority of Singapore, 2001).A distinguished case is the regime choice in Hong Kong. Its exchange rate is one where it pegs the exchange rate to another currency. The Hong Kong Dollar (HKD), in this instance, links its value with respect to the U.S Dollar, such that the monetary policy framework is to adopt exchange rate anchor to USD (Hong Kong Monetary Authority, 2018). It is worth to note that unlike a fixed exchange rate system, the Hong Kong Monetary Authority (HKMA) does not actively interfere in the FOREX market by controlling the supply and demand of HKD (Hong Kong Monetary Authority, 2018). Instead, the rate is maintained through an exchange mechanism, whereby the HKMA authorizes note-issuing banks to issue new banknotes provided that they deposit an equivalent value of USD with the HKMA (Hong Kong Monetary Authority, 2018). Such operation of a strict and robust Currency Board system requires the stock and flow of the monetary base to be fully backed by foreign reserves (Rajan, 2010).

2.4. EXCHANGE RATE VOLATILITY AND ECONOMIC GROWTH: A REVIEW

It has been a long-standing view that large swings in the value of exchange rates can be costly to the domestic economy. Obstfeld and Rogoff (1998) were among the first to point out that the direct effect of fluctuations in the exchange rate movements make it difficult for economic agents to smooth consumption, therefore changing the trend for consumption and leisure to fall. This view however, is contended by Devereux and Engel (2003), who argue that the welfare effects of the volatility in the end are conditional upon the manner of price determination mechanism. If prices are fixed to the foreign currency, domestic consumption is more likely to be unchanged.

Empirical Evidence

The empirical literature on the subject is yet to find the consensus on the impact of exchange rate volatility on economic growth. Dollar (1992) was among the first to empirically investigate

(16)

the growth effect of volatility in developing economies. Using a sample of 95 developing countries with dataset overseeing the period of 1976-1985, he finds that distortions in the real exchange rate lowers the growth of per capita GDP in these nations (negative relationship). This was observed after controlling for the effects of exchange rate variability and investment level. In light of this view, Bosworth et al. (1996) show that countries with a more stable exchange rends tend to grow faster, implying a negative relationship between exchange rate volatility and long-term growth. The reasoning behind this view is that higher exchange rate volatility reduces total productivity factor and therefore economic growth in the long-term. Furthermore, in a paper by Kandil (2004), it was discussed that depending on the degree of openness, higher exchange rate volatility is detrimental to economic growth through its impact on output growth and inflation. He found that in the long run, anticipated exchange rate movements significantly increase inflation and lower output growth.

More recently, Holland et al. (2011) study the impact of real exchange rate volatility on long run economic growth by using a sample of both emerging and advanced economies over the period of 1970 – 2009. The study concludes that after controlling for a model containing both the level of exchange rate and exchange rate misalignments, the stability of exchange rate is more crucial in propelling long-run growth than the impact of its misalignments. Moreover, Aghion et al. (2009) support the notion on the negative relationship between exchange rate volatility and economic growth. Exploiting a panel of 83 countries, spanning the years of 1960 – 2000, Aghion et al. (2009) find that real exchange rate volatility can have a significant impact on long-term productivity growth, but this is conditional on the level of financial development of each country. Such that for countries with relatively low levels of financial development, exchange rate volatility lowers growth, while no significant effect was observed with advanced financial developments. The reasoning behind this view is that in advanced economies, firms are more likely to internalize exchange rate fluctuations as they have the tools to effectively hedge against the risks of these fluctuations. Continuing on this consensus, Benhima (2011) develops a model with a credit-constrained firms facing liquidity shocks to study the impact of liability dollarization on the relationship between exchange rate volatility and growth. Using a panel of 76 countries over the period of 1995 - 2004, he empirically shows that a more negative impact of volatility is observed when the degree of dollarization is higher.

(17)

2.5. EXCHANGE RATE VOLATILITY AND ECONOMIC GROWTH: TRANSMISSION CHANNELS

2.5.1. INTERNATIONAL TRADE

One of our hypotheses is built upon the economic policy recommendation to remove exchange rate volatility in order to foster trade, which in turn will increase growth. In light of this view, the impact of exchange rate volatility on trade echoes more general arguments of both microeconomic and macroeconomic dimension. From a microeconomic perspective, volatility in the exchange rate is highly associated with slower GDP growth in due fact to the increase in transaction costs (Schnabl, 2007). These transaction costs arise from risks of higher uncertainty and that to hedge these risks is costly. One can also interpret the increase in the volatility as meaning to lower transparency in international prices. Higher uncertainty will make prices less transparent, meaning that consumers will find it hard to compare prices between different currencies (Schnabl, 2007). All that said, efficiency, productivity and welfare in trade activities are more likely to be enhanced when the volatility in the exchange rate is lower. On the other hand, the macroeconomic perspective holds the view that long-term exchange rate fluctuations affect the country’s trade competitiveness with respect to export and import in competing industries (Schnabl, 2007). Especially in developing countries, the outward-oriented policies, reflected in the real exchange rate level is said to foster the growth of exports by pushing for the development in the tradable sectors (Dollar, 1992).

On balance, it remains unclear the extent to which international trade is adversely affected by the volatility in the exchange rate. This is based on the change in the volatility itself and on the degree of firms’ sensitivity toward exchange rate risks and their capability to hedge these risks as efficient as possible in terms of cost saving (Clark, 2004). From the view of firms’ profitability, the theoretical evidence has suggested that with the increasing fraction of international transactions undertaken by firms, the volatility in the exchange rate is likely to adversely affect world trade (Clark, 2004). Associating higher trade volume with positive economic growth, higher volatility in the exchange rate would imply a negative growth rate over the long-term horizon.

2.5.2. INTERNATIONAL CAPITAL MARKET

International capital markets have played an increasingly important role in the discussion on the topic of exchange rate volatility and growth (McKinnon and Schnabl, 2004a, DeGrauwe

(18)

and Schnabl, 2005a). In light of this view, the focus has been on the center of physical capital flows, namely Foreign Direct Investment (FDI), where growing evidence have suggested that the volatility impact depends largely on the time horizon of the investment (i.e. in the long or short-term horizon). In addition, Froot and Stein (1991) provide an example where under the presence of asymmetric information, the impact of exchange rate volatility on capital flows depends on the information regarding the intensity of the flow itself. This means that if FDI flows are more information-sensitive than portfolio investment, then they are more sensitive to exchange rate fluctuations. More recently, Dixit and Pindyck (1994) outline the relationship between the two variables using the “options to wait” approach. This approach entails that under uncertainty condition, firms will invest only if the expected value of their investment exceeds a certain threshold level. Furthermore, since exchange rate risks makes the expected returns indeterminate, one should look carefully at the value of an investment strategy based on the certain and uncertain elements involved in it. Dixit et al. (1994) interpret this as meaning that exchange rate uncertainty restrains capital inflows even if the investors are not risk-averse. On the one hand, high exchange rate volatility can be interpreted as higher risk for growth since they affect the balance sheet of banks and enterprises since foreign debt tend to be denominated in foreign currency (Eichengreen and Hausmann, 1999). This increases the likelihood of sharp depreciations that would inflate the liabilities in terms of domestic currency, creating the risk for default or even a crisis. For this reasoning, in debtor countries with highly dollarized financial sectors, the incentive to avoid such fluctuations is even stronger (Chmelarova and Schnabl, 2006). Growing evidence further point out to the disincentive that investors face in times of high exchange rate volatility (see amongst others Dixit and Pindyck, 1994). This would imply that investors prefer to wait as the option value of waiting increases with the increasing uncertainty in order to avoid losses from volatile exchange rate. In addition, there is also the chance to engage in the investment activities later in the future in the absence of such volatility, hence FDI inflows are lowered now (Furceri and Borelli, 2008).

On the contrary, a number of literature attempt to present a distinctive and favourable impact of exchange rate volatility on FDI inflows. Goldberg and Kolstad (1995) were among the first to argue that firms will in fact buy an option to shift the investment when exchange rates are uncertain. This builds on the premise that the option value is shown to be positively correlated with the risk attached to the exchange rate, meaning that uncertainty will in the end lead to higher FDI inflows. Against this backdrop, it is worth to note that volatility in the exchange rate is not the sole factor determining the level of FDI flows. Del Bo (2009), amongst others, suggests other determining factors affecting the level of FDI flows, which include the

(19)

risk of insecurity of venture, economic environment of destination country and macroeconomic (in)stability.

2.5.3. MACROECONOMIC STABILITY

With regard to macroeconomic stability, exchange rate is one of the macroeconomic fundamentals that significantly affect consumer price index, and thereby the inflation rate. Gopinath (2015) argues that the relationship between fluctuations in the exchange rate and inflation varies considerably between countries. Using the sample of both developed and developing economies, she found that the currency used to set international prices has large and asymmetric effects on whether exchange rate fluctuations affect domestic prices. The main finding from her study is that when a large fraction of a country’s trade volume is denominated in foreign currencies, the likelihood that exchange rate fluctuations to affect inflation rate is higher. Moreover, it has been a long-standing view that this strong and negative impact is caused since it is costly for firms to adjust prices. If price adjustments were nonetheless costless, such denomination would rather be irrelevant (Gopinath, 2015). Consequently, the detrimental impact of exchange rates on domestic prices will eventually affect the economy, mainly from aggregate demand and output gap.

Against this backdrop, the discussion returns to the choice of exchange rate regime. Given one of the advantages of adapting a more flexible exchange rate regime is its ability to be able to better respond to adverse macroeconomic shocks. This view is however, contended by McKinnon (1963), who argue on the benefits of fixed exchange rate regimes for small open economies when challenged with nominal shocks. It is assumed that for small open economies, the international price level is given and traded goods make up a high share of domestically consumed goods. This means that lower variability in the exchange rate ensures domestic price stability, and therefore emerges in macroeconomic stability. The welfare effect is that it provides a favourable environment for investment and consumption (McKinnon, 1963). Correspondingly, when exchange rate fluctuations are smoothed, growth is enhanced from a more stable macroeconomic environment.

3. MODEL AND METHODOLOGY

3.1. MEASURING EXCHANGE RATE VOLATILITY

In recent literature on exchange rate volatility, there is no clear consensus on the appropriate method to measure the volatility. One of the most widely used measure of exchange rate

(20)

volatility is the standard deviation of the first difference of logarithms of the exchange rate. In this thesis, we follow this methodology as outlined in a paper by Clark et al. (2004)5_{. The main} feature of this method is that the volatility is equal to zero if it follows a constant trend, which implies that there is no strong evidence for the source of uncertainty (Clark et al., 2004). Referring to the practice in most literature, the change in the exchange rate is computed over a one-month period from applying the end-of-month data. Subsequently, the standard deviation is computed over a quarterly window to indicate for our measure of exchange rate volatility:

EXCHANGERATEVOLATILITY_i,t = Std. Dev [ln(e_i,m) − ln(e_i,m−1)] , m = 1, … , 12

EXCHANGERATEVOLATILITYi,t : Volatility of exchange rate ei,m: Monthly exchange rate

e_i,m−1: Previous month exchange rate

Furthermore, it should be noted the distinction between using the nominal and real exchange rate. The choice depends in part on the time dimension of the economic decision being taken. For example, if the economic decisions are made on a short-run basis, the costs of production and prices for export and import are more likely to be known. This means that the exposure of the exchange rate movements to firms’ economic decision is a function of a nominal exchange rate (Clark et al., 2004). In the longer run, the decisions to engage in international transactions will be of longer horizon with more uncertainty, meaning that production costs, export and import prices denominated in foreign currencies will vary (Clark et al., 2004). As a result, real exchange rates are more appropriate to use. Having said that, since assumed prices to be sticky, the two exchange rates tend to move in a similar path (Clark et al., 2004). This can be interpret as meaning that the choice between the two is less likely to significantly affect the magnitude of the volatility in this analysis. Nonetheless, this thesis investigates the long run dynamics of economic growth in response to fluctuations in exchange rates, making the real exchange rates to be preferable on theoretical grounds.

Using the measure of exchange rate volatility, figure 3 and 4 present the comparison between the volatility of exchange rate in East Asian currencies before and after the crisis. As expected, the exchange rate fluctuations are more pronounced in the after-crisis period.

5_{See also Brodsky (1984), Kenen and Rodrick (1986), Frankel and Wei (1993), Dell’Ariccia (1999), Rose (2000), and}

(21)

Furthermore, more significant movements are observed in the crisis-hit currencies: Indonesia, South Korea, and Thailand. These nations experienced heavy depreciations in their currency after abandoning the peg and switching to floating regime6_{. Continuing on the discussion on} the sources of volatility, this confirms that the increase in volatility are often observed following a switch towards a more flexible exchange rate regime.

Figure 3: Pre-crisis Exchange Rate Volatility

Source: Bank for International Settlements & Author’s calculations

6_{These nations experienced nominal currency depreciation of more than 50% from July 1997 to early 1998. The offshore} nominal interest rates (primarily determined by forward exchange rates) or onshore rates reached at least 25% between June 1997 and January 1998.

(22)

Figure 4: Post-crisis Exchange Rate Volatility

Source: Bank for International Settlements & Author’s calculations

3.2. EMPIRICAL MODEL

Our main hypothesis is that higher exchange rate volatility is detrimental to long-run growth. We use a panel of eight East Asia countries and quarterly observations over the period of 1999Q1 – 2016Q4. It is imperative to note that endogeneity remains a growing concern in growth analysis as economic variables are to some extent affected by growth performance. Considering this, we use a dynamic Generalized Methods of Moments (GMM) specifications developed in Arellano and Bond (1991), Arellano and Bover (1995) and Blundell and Bond (1997). Our benchmark specification follows that of Aghion et al. (2009)7_{. In their} methodology, exchange rate volatility was specified as the main explanatory variable and a set of control variables as well as the interaction terms were added into the model, while the dependent variable corresponds to the productivity growth instead of economic growth as we measure from growth rate of real GDP in this thesis.

Given these specifications, the dynamic panel data estimator was considered to estimate the impact the impact of exchange rate volatility on long term growth under a GMM context. This method outlines the time-series dimension of the panel data while addresses the issue of endogeneity of all the explanatory variables in a dynamic formulation of the data (Aghion et

7_{The same approach was also applied, amongst others, by Benhima (2011) to find the relationship between volatility and} productivity growth through the role of financial dollarization. Schnabl (2007) also applied the same estimation to find for the growth impact of volatility both in EU periphery and East Asia.

(23)

al., 2009). In addition, this approach also deals with potential biases from the non-observable country specific effects (Aghion et al., 2009). What is more, our panel of the country and time-period observations are strongly balanced.

We begin by constructing a dynamic panel data set with real GDP growth as the dependent variable. To account for the implied volatility of country "𝑖" at time "𝑡", the indicator for the measure of volatility is used. In addition, a different set of control variables are incorporated, where we focus on the macroeconomic side of growth enhancing factors and a set of interaction terms. Using this benchmark, we first start by examining the direct effect of exchange rate volatility on long-run growth, augmented by initial growth condition. The model also controls for time-invariant effects that differ across countries by including country fixed effects, denoted by 𝑢𝑖. In addition, this model also controls for country-invariant effects, which changes over time, represented by 𝜂_𝑡, and lastly, 𝛿_𝑖,𝑡 and 𝜀_𝑖,𝑡 correspond to dummy variable and error term respectively. The following linear regression equation models the relationship between exchange rate volatility and economic growth:

𝐺𝐷𝑃𝐺𝑅𝑂𝑊𝑇𝐻𝑖,𝑡= 𝛼0 + 𝛼1𝐺𝐷𝑃𝑖,𝑡−1 + 𝛼2𝐸𝑋𝐶𝐻𝐴𝑁𝐺𝐸𝑅𝐴𝑇𝐸𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝑖,𝑡−1

+ 𝛽1𝐼𝑁𝑇𝐸𝑅𝑁𝐴𝑇𝐼𝑂𝑁𝐴𝐿𝑇𝑅𝐴𝐷𝐸𝑖,𝑡−1+ 𝛽2𝐹𝐷𝐼𝑖,𝑡−1+ 𝛽3𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑖,𝑡−1 + 𝛽4𝐼𝑁𝑇𝐸𝑅𝐸𝑆𝑇𝑅𝐴𝑇𝐸𝑆𝑖,𝑡−1+ 𝛾1𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑖,𝑡−1

+ 𝛾2𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝑇𝑅𝐴𝐷𝐸𝑖,𝑡−1+ 𝛾3𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝐹𝐷𝐼𝑖,𝑡−1+ 𝑢𝑖+ 𝜂𝑡+ 𝛿𝑖,𝑡+ 𝜀𝑖,𝑡

Our dependent variable 𝐺𝐷𝑃𝐺𝑅𝑂𝑊𝑇𝐻_𝑖,𝑡 corresponds to the quarterly real GDP growth rates. The variable 𝐸𝑋𝐶𝐻𝐴𝑁𝐺𝐸𝑅𝐴𝑇𝐸𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌_{𝑖,𝑡−1} corresponds to the measure of volatility. The variable 𝐺𝐷𝑃_{𝑖,𝑡−1}represents the initial level of GDP, that is standard to growth analysis. Our control variables include 𝐼𝑁𝑇𝐸𝑅𝑁𝐴𝑇𝐼𝑂𝑁𝐴𝐿𝑇𝑅𝐴𝐷𝐸_{𝑖,𝑡−1}, 𝐹𝐷𝐼_{𝑖,𝑡−1}, 𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁𝑖,𝑡−1, 𝐼𝑁𝑇𝐸𝑅𝐸𝑆𝑇𝑅𝐴𝑇𝐸𝑆𝑖,𝑡−1, which correspond to international trade as a percentage of GDP, foreign direct investment as a percentage of GDP, inflation rate, and money market rates respectively. All of the variables are represented in their lagged period. Lastly, we also model the interaction terms between volatility and the transmission channels as represented by the variables 𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝐼𝑁𝐹𝐿𝐴𝑇𝐼𝑂𝑁_{𝑖,𝑡−1} , 𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝑇𝑅𝐴𝐷𝐸_{𝑖,𝑡−1}, 𝑉𝑂𝐿𝐴𝑇𝐼𝐿𝐼𝑇𝑌𝐹𝐷𝐼_{𝑖,𝑡−1}. The inclusion of the dummy variable corresponds to the inflation targeting framework. Adopting this framework means the national monetary authority does not target the value of exchange rate as their main policy objection.

(24)

4. DATA DESCRIPTION

We use quarterly observations over the period of 1999Q1 – 2016Q4. We deliberately started in the first quarter of 1999 to mark the aftermath of the Asian Crisis. We collected the data from the International Monetary Fund data-stream, specifically from the International Financial Statistics (IFS) and the Balance of Payments (BoP) account. The data set taken from the IFS include nominal GDP, GDP deflator, money market rates, investment and government expenditure. While the data set taken from the BoP account include the figures for export, import and direct investment as a measure of foreign direct investment (net inflow). Subject to data availability, the short-term money market rates have been preferred as an indicator for interest rates, which capture the movements in the money market. The figures for the initial level of GDP were calculated by taking the nominal GDP divided by the GDP deflator.

The figures for the quarterly real GDP growth were obtained from the Asian Development Bank database. These figures are the year-on-year percentage change, which correspond to the change in real GDP figures from the same quarter of the previous year. We also extracted the data from the Bank of International Settlements database. The data set include the monthly nominal exchange rate and consumer price index (CPI), subsequently used to compute the real exchange rate. These figures are presented in U.S. Dollar to make the value comparable across sample countries. We use inflation as a proxy for macroeconomic (in)stability, where the data on the monthly CPI were again obtained from the Bank of International Settlements, and taken as quarterly average. The inflation rates were calculated by taking the percentage change from the current period CPI to the previous period. Table 1 in Appendix B presents the summary of descriptive statistics for all of our observations. We consider a sample of eight countries in East Asia. Most of the data of the variables are available for each country, with the exception of the data on the Foreign Direct Investment of Indonesia for the year 1999 – 2003. Although Taiwan can be well categorized in the sample countries, due to data availability, is excluded from the analysis.

5. EMPIRICAL RESULTS

Table 2 in Appendix C presents the GMM estimation results for our sample countries. The table displays a result for four regressions, which we started parsimoniously and then added sets of variables to see the extent to which each addition affected the explanatory power of the regression. The first regression estimates the direct effect of exchange rate volatility on economic growth, augmented by initial growth condition. In the second regression estimation,

(25)

we added a set of control variables, such as international trade and foreign direct investment. Whereafter, we added more control variables such as inflation and interest rate in the third regression. The fourth regression estimates the growth effect of volatility with the inclusion of the interaction terms to model the transmission channels.

The estimation results provide a robust evidence of a negative impact of exchange rate volatility on economic growth. These results are confirmed with the coefficients on our measure of volatility that are statistically significant at 5% significance level. This confirmed our hypothesis that higher volatility is detrimental to growth, which remains robust following the inclusion of additional control variables. In addition, our estimation result is in line with the findings of, amongst others, Schnabl (2007), Aghion et al. (2009) and more recently Benhima (2011), who conclude a negative impact of volatility on long-run economic growth. The coefficient on the initial growth condition suggests that higher real GDP contributes to higher growth, in which this result is confirmed at a 1% significance level under all specifications. The resulting coefficients on the control variables are mixed and often significant at the common level. The coefficient on international trade enters positively and significantly under all specifications. This is as expected, for which these nations put high emphasize on the cross-border transaction activities so that international trade contributes to long run growth positively. The effect of FDI on growth is positive, as expected, although the results are not robust. The negative coefficient on inflation is as expected, meaning that higher inflation is detrimental to growth. This result is however, not significant as the implied z value is low. Interestingly, the addition of control variable such as interest rates, led us to infer that higher interest rates boost growth, albeit not statistically significant.

Accounting for the interaction terms, our coefficient of exchange rate volatility turns positive, albeit not significant. One important point to note is that the interaction between volatility and international trade is strongly negative. One can interpret this such that higher volatility affects growth negatively from lower trade. The estimation result of the interaction between of exchange rate volatility and FDI is however, not as expected. The result pointed out that volatility enhances growth through FDI inflows. Lastly, the interaction terms between volatility and inflation suggests that volatility affects growth by lowering macroeconomic instability as proxied by inflation, although the coefficient shows no significant result. Last but not least, the coefficients on the dummy variable, represented by the inflation targeting regime suggest a positive relationship between inflation targeting framework of monetary policy and growth, although we fail to confirm these results.

(26)

5.1. ENDOGENEITY ISSUE

As pointed out before, endogeneity remains a growing concern and hence is of capital important to be addressed accordingly in a growth literature. To do so, we look at two different approaches: one is to test the exogeneity of our instruments within the GMM methodology applied in the empirical analysis and the second approach is to find further evidence on the determinants of exchange rate volatility in current literature. The latter will provide the theoretical ground for the use of an external instrument for the exchange rate volatility variable. The application of the GMM methodology in particular, deals with the issue of endogeneity by incorporating the lagged value of dependent variable as the explanatory and instrumental variables (Aghion et al., 2009). It should be noted, however, that this can only be the case under the assumption that remaining regressors are predetermined, and that they are uncorrelated with the current and future realizations of the error term (Aghion et al., 2009). In other words, the rest of the independent variables should be weakly exogenous. To prove this, we refer to the Sargan test of over-identifying restrictions that tests for the quality of the instrumental variables. Our estimation results show that our equations are indeed well identified as the coefficients found suggest that we fail to reject the null hypothesis that our instruments are uncorrelated with the error term. This confirms for the validity of our instruments. Arellano and Bond (1991), Arellano and Bover (1995) and Blundell and Bond (1997) interpret this result as meaning that there is no endogeneity problem, which confirms the validity and robustness of our internal instruments, that is the case under all specifications.

Secondly, literature has looked for the sources of exchange rate volatility, notably macroeconomic performance and the choice of exchange rate regime. With regard to the first factor, inflation and interest rates are arguably the important drivers of volatility. These two variables, are included in the regression specifications as control variables. In addition, growing literature has also found that international trade can also affect volatility. Similarly, this variable is also included in our regression as control variable, hence the issue of endogeneity in this manner is dealt with. What is more, it was found by Hausman et al. (2006) that growth has a significantly positive relationship with exchange rate volatility, meaning that the issue of reverse causality arises and thus biases the effect of exchange rate volatility upwards. Benhima (2011) holds the view that as long as the coefficient for exchange rate volatility remains significantly negative after accounting for this positive bias, the estimation result will remain robust. Furthermore, as discussed in section 2 of theoretical framework, the choice of the exchange rate regime has an important implication on the implied volatility of

(27)

the exchange rate. This factor in particular is taken into greater consideration using the inclusion of our dummy variables as the floating adapter implement the inflation targeting framework of monetary policy.

6. ROBUSTNESS TEST

The results of our analysis discussed above, while mixed, were broadly consistent with the existing literature and empirical evidence. Hence, this section will provide a robustness test to examine our results and their sensitivity to alternative specifications. We do so by first including additional traditional control variables to our growth regression. We propose to incorporate the proportion of government expenditure to GDP as well as investment-to-GDP ratio to our analysis. In addition, an alternative measure for exchange rate volatility are calculated and added into our regression. Under these specifications, we apply the same estimation method of GMM for dynamic panel data.

6.1. ALTERNATIVE MEASURE OF EXCHANGE RATE VOLATILITY

For robustness test, the alternative to measure exchange rate volatility is applied. Thorbecke (2008) point out that exchange rate volatility can be defined as the coefficient of variation (CV) of the monthly exchange rate during the year. This paper adapts this benchmark to use the coefficient of variation as a measure of exchange rate volatility by taking the standard deviation of Real Effective Exchange Rate (REER) divided by the mean of the REER. Similarly, the coefficient of variation is computed over a quarter rolling window. More specifically:

𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡 𝑜𝑓 𝑉𝑎𝑟𝑖𝑎𝑡𝑖𝑜𝑛 =𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑟𝑒𝑎𝑙 𝑒𝑥𝑐ℎ𝑎𝑛𝑔𝑒 𝑟𝑎𝑡𝑒 𝑚𝑒𝑎𝑛 𝑟𝑒𝑎𝑙 𝑒𝑥𝑐ℎ𝑎𝑛𝑔𝑒 𝑟𝑎𝑡𝑒

The REER is used in this regard as it takes into account the country’s competitiveness in the international market. According to IMF definition, the REER can be used as an indicator of the value of a currency against a weighted average of several foreign currency. In that sense, an increase in REER (appreciation) can be translated into a loss in trade competitiveness, as exports will be more expensive for foreign nations demanding domestic currency.

(28)

6.2. SENSITIVITY ANALYSIS

Table 3 in Appendix D presents the estimation results for our sensitivity analysis. In the first regression, we apply our measure for exchange rate volatility with the addition of control variables, such as government expenditure and investment ratio, in which the resulting coefficient on our measure of volatility remains negative and robust at the 5% significance level. The resulting coefficients on government expenditure is negative, albeit not significant. This suggests that an increase in government expenditure slows down growth. This can be the case when the expenditure is financed through increase in taxes, in which in this case lowers household’s consumption and subsequently growth. The coefficient on investment is not as expected, as higher investment is predicted to enter positively. The resulting coefficient, however, is not statistically significant. The table also presents the estimation result from applying the coefficient of variation as the measure of volatility. Similar to our previous analysis, in the second regression analysis, we started by estimating the direct effect of volatility on growth, augmented by initial growth condition. The resulting coefficient is negative and is confirmed at the 5% significance level. The analysis remains robust after adding the control variables as shown in the third, fourth and fifth regression. This suggests that from applying an alternative of exchange rate volatility, the impact of volatility on growth remains negative and robust under all specifications. However, the addition of the interaction terms in the fifth regression, did not suggest any significant and meaningful implication to add to our analysis.

7. DISCUSSION AND POLICY IMPLICATIONS

It is worth to note that there are important caveats to apply to our estimation results. Empirical analysis presented in this thesis addressed the impact of volatility on growth through three transmission channels, one of which it was found that volatility negatively affects long-run growth through international trade. Against this backdrop, it should be noted that the stability of the exchange rate over the medium to long-term is preferred in order to support for cross-border trade activities and subsequently growth. As observations across countries have also demonstrated that countries with favourable export growth tend to experience higher growth rates, in which in this case is experienced by the East Asian economies. This implies that the negative impact of volatility can be stronger as the share of export to GDP in these nations is relatively significant. Nonetheless, it should also be noted of the possibility that this interplay could might as well be a manifestation of other factors that could simultaneously raise volatility

(29)

and reduce trade. The unfolding of the Asian crisis can be seen to have led to major deterioration of trade volume (both export and import) and at the same time significant movements (steep depreciations) in the exchange rates following a regime switch. In this case, the increasing volatility at the time was not the only factor responsible for the trade decrease, but instead was also due to the significant fall in consumer confidence and domestic demand. Based on these considerations, our results do not provide clear policy guidance. Although it is the case that higher volatility is associated with lower trade flows, this does not directly imply that trade position would significantly improve if authorities stabilized the exchange rates to a significant extent. In light of this view, Clark et al. (2004) offer an alternative for which to focus on the variables with larger variations in the data on bilateral trade and exchange rates, which will allow monetary authority to base their policy judgements more comprehensively.

In relation with capital flows in the international capital market, the coefficient of the interaction between the two does not successfully provide significant estimation and analysis. This could be due to the fact that we exclusively focus on the inflows of FDI to represent the movement in the international capital market. Such that, many other factors are also of significant importance in the discussion as how exchange rate volatility affects capital flows. Regardless of this limitation, it has been argued that the instability in the exchange rates would have discouraged investors from borrowing in foreign currency, as it would expose them to higher exchange rate risks. Although growing evidence have suggested that such risk can be shifted or that there are sufficient instruments to hedge against these risks, the practice is much more difficult especially when a financial crisis is judged to be imminent. Nonetheless, it is worth pointing out that despite the evidence in favour of the negative growth effect of volatility, the relationship between the two is not linear. It worth remembering that lower variability, for instance, could potentially encourage for international investment by attracting capital into the domestic economy. But this in turn may encourage speculative capital inflows and possibly investment overheating, as was experienced prior the Asian crisis. In this context, given the increasing participation to the international capital market, the monetary authority should, if necessary issue a policy to address the risk of excessive capital flight when the exchange rate movements are perceived to be favourable.

With respect to macroeconomic stability, the role of national monetary authority is more important. In East Asia, the issue of stability remains an important ground for policy analysis. However, for nations adapting inflation targeting monetary policy framework, the role of monetary authority in affecting the value of the currency is limited. But since the fluctuations in the exchange rate can have impact on inflation, the central banks should still

Explaining the impact of exchange rate volatility on economic growth

Master’s Thesis