An investigation into the sources of depreciations in Mongolian Tugrik exchange rate: A structural VAR approach

Ganbayar, Gunbileg Article An investigation into the sources of depreciations in Mongolian Tugrik exchange rate: A structural VAR approach Journal of Risk and Financial Management Provided in Cooperation with: MDPI – Multidisciplinary Digital Publishing Institute, Basel Suggested Citation: Ganbayar, Gunbileg (2021) : An investigation into the sources of depreciations in Mongolian Tugrik exchange rate: A structural VAR approach, Journal of Risk and Financial Management, ISSN 1911-8074, MDPI, Basel, Vol. 14, Iss. 11, pp. 1-16, https://doi.org/10.3390/jrfm14110529 This Version is available at: https://hdl.handle.net/10419/258632 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Journal of Risk and Financial Management Article An Investigation into the Sources of Depreciations in Mongolian Tugrik Exchange Rate: A Structural VAR Approach Gunbileg Ganbayar 1,2   Citation: Ganbayar, Gunbileg. 2021. An Investigation into the Sources of Depreciations in Mongolian Tugrik Exchange Rate: A Structural VAR Approach. Journal of Risk and Financial Management 14: 529. https:// doi.org/10.3390/jrfm14110529 Academic Editors: Aviral Kumar Tiwari and Shigeyuki Hamori Received: 8 September 2021 Accepted: 2 November 2021 Published: 6 November 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1Graduate School of Humanities and Social Sciences, Saitama University, 225 Shimo-Okubo, Sakura-ku, Saitama 338-8570, Japan; [email protected] 2Department of Finance, Business School, National University of Mongolia, Ulaanbaatar 214192, Mongolia Abstract: This paper empirically investigates the sources of fluctuations in real and nominal Mongolian Tugrik (MNT) exchange rates by estimating the structural vector autoregressive (SVAR) model over the period January 1994–May 2021 and decomposing the exchange rate series into stochastic components induced by real and nominal shocks under the assumption of the long-run neutrality of nominal shocks on the real exchange rate level. The empirical results show that the real MNT exchange rate movements are primarily due to the real shocks, while the nominal shocks have a major role in explaining nominal exchange rate movements in the short and long run. The nominal exchange rate shows a delayed over-shooting occurring between one and three years after a nominal shock hits the economy. The long-run effect of a monthly one standard deviation nominal shock on nominal MNT exchange rate is 2.5%, which results in a permanent divergence between real and nominal MNT exchange rate and causes non-cointegrated relation between real and nominal MNT exchange rates. The historical decomposition of forecast error indicates that the nominal shock plays a significant role in explaining the depreciation in nominal MNT exchange rate over the last three decades. Our recommendation is to stop “cash handling” policy, minimize monetary shock, and coordinate fiscal and monetary policies to avoid large nominal depreciation. Keywords: SVAR; long-run restriction; real shock; nominal shock; Mongolian Tugrik JEL Classification: C32; F31 1. Introduction Exchange rate may be the most discussed topic in international economics. Numerous papers haven been written in this field since the collapse of the Bretton Woods system, and the study dimensions have been expanding rapidly. One dimension we emphasize here is to disentangle nominal and real shocks in the exchange rate series. This paradigm emerged from the methodological (Blanchard and Quah 1989;King et al. 1991) and theoretical (Dornbusch 1976;Stockman 1987,1988) advancements in times series econometrics and exchange rate theory, respectively. Lastrapes (1992) first studied the exchange rate fluctuations of six advanced economies by decomposing exchange rate series into real and nominal components by identifying real and nominal shocks using a bivariate structural vector autoregressive (SVAR) model under the assumption of long-run neutrality of nominal shocks on the real exchange rate level. Further, this approach was applied for selected advanced economies (Enders and Lee 1997) , for selected transition economies (Dibooglu and Kutan 2001) , for emerging economies (Chowdhury 2004), for central and east European advanced and transition economies (Morales-Zumaquero 2006), for east Asian economies (Ok et al. 2010), for Saudi Arabia (Aleisa and Diboo ˆ glu 2002) and for India (Moore and Pentecost 2006) to explain the sources of exchange rate fluctuations in these countries. To the best of our knowledge, this type of study has not yet been conducted for the Mongolian Tugrik (MNT) exchange rate, which has depreciated substantially over the last three decades (Figure 1). J. Risk Financial Manag. 2021,14, 529. https://doi.org/10.3390/jrfm14110529 https://www.mdpi.com/journal/jrfm J. Risk Financial Manag. 2021,14, 529 2 of 16 The defining features of the current state of the Mongolian economy are a mining boom and external indebtedness (Ganbayar 2021). Following the recent huge mining developments, the Government of Mongolia (GoM) implemented several “cash handling” policies to distribute mining income to citizens (Dagys et al. 2020; Yeung and Howes 2015 ). As a result of these cash handling policies, the money supply in Mongolia increased substantially (Figure 2). On the other hand, the skyrocketed external debt ( Batsuuri 2015 ; Ganbayar 2021 ), which is primarily in US dollars, causes the Bank of Mongolia (BoM) to manage exchange rate depreciation to secure confidence in domestic currency despite its inflation targeting mandate because there might be an epidemic of the “fear of floating” (Calvo and Reinhart 2002;Taguchi and Gunbileg 2020). As a consequence, the de facto exchange rate arrangement of BoM was recently reclassified twice: (1) to crawl-like from floating, effective 18 September 2017, and (2) to other managed from crawl-like, effective 11 April 2018, according to the International Monetary Fund (IMF (2020)). Under this background, the identification of driving sources of MNT exchange rate depreciation is crucial for coordinating fiscal and monetary policy, participating effectively in exchange rate markets and understanding MNT exchange rate fluctuations. Hence, the main purpose of this study is to empirically identify the sources of substantial depreciation in MNT exchange rate over the last three decades by decomposing both real and nominal MNT exchange rate series into nominal and real components using a bivariate SVAR model, as applied in Lastrapes (1992) and Enders and Lee (1997). Our contribution is to enrich the international economic literature with the empirical evidence of the Mongolian economy, which has rarely been mentioned in international economic literature, and to make reasonable recommendations to the country’s exchange rate policy, especially for the coordination between fiscal and monetary policies. J. Risk Financial Manag. 2021, 14, x FOR PEER REVIEW 2 of 16 To the best of our knowledge, this type of study has not yet been conducted for the Mongolian Tugrik (MNT) exchange rate, which has depreciated substantially over the last three decades (Figure 1). The defining features of the current state of the Mongolian economy are a mining boom and external indebtedness (Ganbayar 2021). Following the recent huge mining developments, the Government of Mongolia (GoM) implemented several “cash handling” policies to distribute mining income to citizens (Dagys et al. 2020; Yeung and Howes 2015). As a result of these cash handling policies, the money supply in Mongolia increased substantially (Figure 2). On the other hand, the skyrocketed external debt (Batsuuri 2015; Ganbayar 2021), which is primarily in US dollars, causes the Bank of Mongolia (BoM) to manage exchange rate depreciation to secure confidence in domestic currency despite its inflation targeting mandate because there might be an epidemic of the “fear of floating” (Calvo and Reinhart 2002; Taguchi and Gunbileg 2020). As a consequence, the de facto exchange rate arrangement of BoM was recently reclassified twice: (1) to crawl-like from floating, effective 18 September 2017, and (2) to other managed from crawl-like, effective 11 April 2018, according to the International Monetary Fund (IMF (2020)). Under this background, the identification of driving sources of MNT exchange rate depreciation is crucial for coordinating fiscal and monetary policy, participating effectively in exchange rate markets and understanding MNT exchange rate fluctuations. Hence, the main purpose of this study is to empirically identify the sources of substantial depreciation in MNT exchange rate over the last three decades by decomposing both real and nominal MNT exchange rate series into nominal and real components using a bivariate SVAR model, as applied in Lastrapes (1992) and Enders and Lee (1997). Our contribution is to enrich the international economic literature with the empirical evidence of the Mongolian economy, which has rarely been mentioned in international economic literature, and to make reasonable recommendations to the country’s exchange rate policy, especially for the coordination between fiscal and monetary policies. Figure 1. Exchange rate evolution of country groups comparable to Mongolia and relative labor productivity. Source: author’s estimation based on publicly available data from IFS, IMF and APO productivity database 2020. Note: Relative labor productivity is the ratio of labor productivity of Mongolia to that of China. The country groups are based on United Nation’s country classification in 2020. We exclude Sudan, Suriname, Turkey and Zimbabwe due to the large depreciations in their currencies. Exchange rates are per Chinese Yuan in terms of specific country’s currency. DCs—Developing countries. TCs—Transition countries. ESEADCs—East and Southeast Asian developing countries. All series are normalized to 1 at 2000. - 0.50 1.00 1.50 2.00 2.50 3.00 3.50 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 MNT/CNY (2000=1) DCs' median exchange rate (2000=1) DCs' 75th percentile exchange rate (2000=1) TCs' median exchange rate (2000=1) ESEADCs' median exchange rate (2000=1) Relative labor productivity (MN/CN, 2000=1) Figure 1. Exchange rate evolution of country groups comparable to Mongolia and relative labor productivity. Source: author’s estimation based on publicly available data from IFS, IMF and APO productivity database 2020. Note: Relative labor productivity is the ratio of labor productivity of Mongolia to that of China. The country groups are based on United Nation’s country classification in 2020. We exclude Sudan, Suriname, Turkey and Zimbabwe due to the large depreciations in their currencies. Exchange rates are per Chinese Yuan in terms of specific country’s currency. DCs—Developing countries. TCs—Transition countries. ESEADCs—East and Southeast Asian developing countries. All series are normalized to 1 at 2000. J. Risk Financial Manag. 2021,14, 529 3 of 16 J. Risk Financial Manag. 2021, 14, x FOR PEER REVIEW 3 of 16 Figure 2. Exchange rate and relative money supply. Source: Author’s estimation based on publicly available data from National Statistical Office, the BoM, and IFS, IMF. Note: MNT/CNY-Nominal exchange rate per Chinese Yuan in terms of Mongolian Tugrik. The BoM’s NEER and REER series are inverted to be consistent with our other measures. Relative money supply is the ratio of monetary aggregates of Mongolia to those of China. The remainder of the paper is organized as follows. In the next section, we provide background facts related to our context. Section 3 reviews past literature. In Section 4, we present our empirical model and shock identification strategy. Section 5 provides our data and preliminary analysis. In Section 6, we present our estimation results. Section 7 provides robustness analysis for the empirical results in Section 6. In Section 8, we make conclusions and discuss policy implications of empirical findings. 2. Country Background Since Mongolia transformed the economic system from a centrally planned economy to a market-based economy in the early 1990s, the Bank of Mongolia (BoM) started implementing a flexible exchange rate regime. The BoM had been adopting monetary aggregate targeting with its reserve money being an operational target until 2006. As the linkage between reserve money and inflation became unstable, the BoM introduced an inflation targeting framework in 2007. In this framework, the BoM, like other inflation targeting central banks, equipped the policy mandates of announcing a mid-term targeted inflation rate to the public and of taking every possible measure to maintain the inflation rate within its targeted range (Taguchi and Gunbileg 2020). According to the principle of the “policy trilemma”, an economy has to give up one of three goals: fixed exchange rate, independent monetary policy, and free capital flows (Fleming 1962; Mundell 1963). Thus, given the free capital mobility in Mongolia, the BoM faces a trade-off in their policy targets between exchange-rate stability and price stability. While the de jure exchange rate arrangement of Mongolia is floating, the de facto exchange rate arrangement was recently reclassified twice: (1) to crawl-like from floating, effective 18 September 2017, and (2) to other managed from crawl-like, effective 11 April 0 1 2 3 4 5 6 7 8 1991M10 1992M07 1993M04 1994M01 1994M10 1995M07 1996M04 1997M01 1997M10 1998M07 1999M04 2000M01 2000M10 2001M07 2002M04 2003M01 2003M10 2004M07 2005M04 2006M01 2006M10 2007M07 2008M04 2009M01 2009M10 2010M07 2011M04 2012M01 2012M10 2013M07 2014M04 2015M01 2015M10 2016M07 2017M04 2018M01 2018M10 2019M07 2020M04 2021M01 MNT/CNY(2000=1) Real MNT/CNY(2000=1) NEER (2000=1) REER (2000=1) Relative M2 (MN/CN, 2000=1) Relative M1 (MN/CN, 2000=1) Figure 2. Exchange rate and relative money supply. Source: Author’s estimation based on publicly available data from National Statistical Office, the BoM, and IFS, IMF. Note: MNT/CNY-Nominal exchange rate per Chinese Yuan in terms of Mongolian Tugrik. The BoM’s NEER and REER series are inverted to be consistent with our other measures. Relative money supply is the ratio of monetary aggregates of Mongolia to those of China. The remainder of the paper is organized as follows. In the next section, we provide background facts related to our context. Section 3reviews past literature. In Section 4, we present our empirical model and shock identification strategy. Section 5provides our data and preliminary analysis. In Section 6, we present our estimation results. Section 7 provides robustness analysis for the empirical results in Section 6. In Section 8, we make conclusions and discuss policy implications of empirical findings. 2. Country Background Since Mongolia transformed the economic system from a centrally planned economy to a market-based economy in the early 1990s, the Bank of Mongolia (BoM) started implementing a flexible exchange rate regime. The BoM had been adopting monetary aggregate targeting with its reserve money being an operational target until 2006. As the linkage between reserve money and inflation became unstable, the BoM introduced an inflation targeting framework in 2007. In this framework, the BoM, like other inflation targeting central banks, equipped the policy mandates of announcing a mid-term targeted inflation rate to the public and of taking every possible measure to maintain the inflation rate within its targeted range (Taguchi and Gunbileg 2020). According to the principle of the “policy trilemma”, an economy has to give up one of three goals: fixed exchange rate, independent monetary policy, and free capital flows (Fleming 1962;Mundell 1963). Thus, given the free capital mobility in Mongolia, the BoM faces a trade-off in their policy targets between exchange-rate stability and price stability. While the de jure exchange rate arrangement of Mongolia is floating, the de facto exchange rate arrangement was recently reclassified twice: (1) to crawl-like from floating, effective 18 September 2017, and (2) to other managed from crawl-like, effective 11 April 2018, as reported in IMF (2020). Even though the BoM has been conducting inflationtargeting monetary policies since 2007, the significant depreciation in MNT exchange J. Risk Financial Manag. 2021,14, 529 4 of 16 rate (see Figure 1) receives greater attention from Mongolian policy makers because there might be an epidemic case of the “fear of floating” (Calvo and Reinhart 2002) that comes from a lack of confidence in currency value, especially given that their high external debt is primarily denominated in US dollars because of the “original sin” hypothesis (Eichengreen and Hausmann 1999) . In fact, the BoM intervenes in foreign exchange markets by organizing auctions between commercial banks every Tuesday and Thursday. However, despite the substantially increased external debt in Mongolia (Batsuuri 2015; Ganbayar 2021), the data show that there is insignificant fear of floating in the monetary policy rule of the BoM (Taguchi and Gunbileg 2020). Since the BoM has started managing MNT rate recently according to the IMF (2020), determining the sources of exchange rate fluctuation is vital to participate in foreign exchange rate market effectively. Another point we mention here is the recent mining boom in the Mongolian economy. Following the giant mining developments, the Government of Mongolia (GoM) implemented cash handling policies to distribute mining income (tax, royalty, and dividend) across Mongolian citizens (Dagys et al. 2020;Yeung and Howes 2015). The GoM conducted cash handling policies by creating new vehicles. The largest is the “Human Development Fund”, through which every Mongolian citizen received the benefits of 1 million MNT between 2010 and 2012. The most recent indirect cash handling action backed by mining income was the one-time forgiveness of pension-backed debts in January 2020. At the time of loan cancelation, there were 229.4 thousand pensioners who had taken loans worth 763.3 billion MNT (Ganbayar 2021). These direct and indirect cash handling policies implemented by the GoM increase money supply (nominal shock) substantially and may cause large depreciations in nominal MNT exchange rate (see Figure 2). Figure 2shows the relative money supply of Mongolia compared to that of China, which is the most integrated country with Mongolia in terms of international trade measures. In fact, 90% of Mongolia’s total export and 30% of its total import (based on average share over the period 2005–2020, see Supplementary File) belong to China. Both M2 and M1 aggregates indicate that the money supply increased significantly in Mongolia over the last two decades. More specifically, relative M2 and M1 aggregates have increased by 6.8 and 4.2 times since 2000, respectively. These stylized facts imply that the nominal shocks may have a significant effect on the depreciation in nominal MNT exchange rate over the last three decades. The nominal MNT/CNY (per Chinese Yuan in terms of Mongolian Tugrik) exchange rate depreciated by a factor of 3.4 times over the same period (Figure 2). This pattern also works in nominal effective exchange rate (NEER); the NEER is depreciated by 2.45 times. Therefore, it is crucial to determine nominal shocks’ contribution to the evolution of MNT exchange rates. In addition to nominal factors mentioned above, real factors such as productivity have some role in exchange rate movements (Balassa 1964;Samuelson 1964). Figure 1 shows the relative labor productivity of Mongolia to China. Relative labor productivity of Mongolia decreased by four times over the last three decades. This productivity decline implies that the traded sectors’ competitiveness is deteriorated significantly relative to China, causing real depreciation in MNT exchange rate due to the decrease in non-traded sector’s prices. In contrast, the mining sector’s boom in Mongolia tends to result in real appreciation in MNT exchange rate due to the Dutch disease effect (Corden 1984;Ganbayar 2021). If the productivity effect dominates the Dutch disease effect, Mongolia experiences real depreciation and vice versa, ceteris paribus. Therefore, it is inevitable to determine the net effects of real shocks, including productivity and mining shocks, on the movements of MNT exchange rate. 3. Literature Review Huizinga (1987) first empirically decomposed the real US dollar (USD) exchange rate series into transitory and permanent components using Beveridge–Nelson decomposition (Beveridge and Nelson 1981). The results showed that the real exchange rate is a meanreverting unit root process and that most of the variance of changes in real USD exchange J. Risk Financial Manag. 2021,14, 529 5 of 16 rate is attributed to permanent components. Since Huizinga (1987) employs the univariate decomposition method, this approach omits some important information contained in other macroeconomic variables, such as nominal exchange rate. Blanchard and Quah (1989) developed a new approach to decompose output series into transitory (demand) and permanent (supply) components, assuming long-run neutrality of demand shocks on the output level. Using the Blanchard and Quah (1989) method, Lastrapes (1992) first disentangled real and nominal shocks in the exchange rate series using the bivariate SVAR model under the assumption of the long-run neutrality of nominal shocks on the real exchange rate level. This decomposition approach was developed based upon both the equilibrium (Stockman 1987,1988) and disequilibrium (Dornbusch 1976) theory of exchange rate. According to the equilibrium theory, the real shocks have a power to explain changes in real and nominal rates in both the short and long run. On the other hand, the disequilibrium theory emphasizes the role of nominal shocks in the evolution of real and nominal rates as the exchange rate market responds to the monetary disturbances promptly due to the price rigidity in goods and service markets. Because of this difference in the adjustment speed of these markets to the monetary disturbances, the nominal exchange rate shows an over-shooting dynamic according to the disequilibrium theory developed by Dornbusch (1976). As reported in Lastrapes (1992), the fluctuations of nominal and real exchange rates in advanced economies between March 1973 and December 1989 are due primarily to real shocks. Clarida and Gali (1994) modified Lastrapes (1992)’s approach by setting triangular long-run restriction on a trivariate SVAR model and decomposed the real exchange rate series into three sources induced by supply, demand, and monetary shocks. The literature related to Clarida and Gali (1994) is primarily focused on the drivers of real exchange rate fluctuations. Therefore, these studies exist in a different vein of literature from the objective of our study because we aim to decompose both real and nominal exchange rates into real and nominal components. Therefore, we concentrate here on the literature applying the same approach as Lastrapes (1992), which are discussed in the following paragraphs. Enders and Lee (1997) decomposed US dollar real and nominal exchange rate movements into the components induced by real and nominal shocks. They found that the real shocks explain a majority of real and nominal exchange rate fluctuations and that there is little evidence of exchange rate over-shooting in bilateral exchange rates between the US and Canada, Japan, and Germany over the period January 1973–April 1992. Dibooglu and Kutan (2001) conducted a similar study for two transition economies, Poland and Hungary. Using monthly real exchange rate and inflation series between January 1990 and March 1999, they found that nominal shocks had a larger influence in explaining real exchange rate movements in Poland, while real shocks had more influence on real exchange rate movements in Hungary. The empirical results imply that sticky-price disequilibrium models explain the behavior of the real exchange rate of Poland, while equilibrium exchange rate models are more suitable for Hungary. Aleisa and Diboo ˆ glu (2002) estimated a bivariate SVAR model using monthly data from January 1980 to February 2000 to investigate the sources of real exchange rate movements in Saudi Arabia. Their results suggest that real shocks play a significant role in explaining real exchange rate movements, while nominal shocks play a significant role in explaining price level movements in Saudi Arabia. Moore and Pentecost (2006) conducted the same study on the Indian Rupee against the US dollar over the period from March 1993 through January 2004. The authors found that real and nominal Rupee exchange rates are driven mostly by real shocks and suggested that Indian policy makers need to focus on the real side of the economy in order to stabilize the foreign exchange market. Chowdhury (2004) studied the sources of real and nominal exchange rate fluctuations in six emerging market countries: Chile, Columbia, Malaysia, Singapore, South Korea, and Uruguay. The author employed exactly the same SVAR model applied in Lastrapes (1992) using monthly observations from January 1980 to December 1996. The empirical results showed that real shocks result in long-run real and nominal appreciation, while J. Risk Financial Manag. 2021,14, 529 6 of 16 nominal shocks generally cause a nominal depreciation. Latin American exchange rates appear to be more sensitive to real and nominal shocks than East Asian exchange rates, according to Chowdhury (2004), because real and nominal shocks have stronger influences on the exchange rates of Latin American countries. Morales-Zumaquero (2006) investigated sources of real exchange rate fluctuations for selected advanced economies—Canada, Japan, US, UK, France, Italy, and Germany—and Central and Eastern European transition economies—Czech Republic, Hungary, Poland, Romania, and Slovenia. Based on the findings of this study, for advanced economies, real shocks account for the majority of exchange rate fluctuations over the period from January 1973 to December 1990, while nominal shocks play a central role in explaining exchange rate movements over the period from January 1991 to January 2000. The author further stated that nominal shocks have a major role in explaining real exchange rate fluctuations for Euro Zone countries over the subperiod January 1991—January 2000 due to the “single monetary policy” of these countries. As the transition economies have different initial conditions and economic policies, the real exchange rate movements in some transition economies (Czech Republic, Hungary, and Slovenia) are explained mostly by real shocks, while in others (Poland and Romania), nominal shocks play a larger role in explaining real exchange rate movements. Ok et al. (2010) investigated the sources of fluctuations in real and nominal US dollar exchange rates in Cambodia and Laos. Their findings were that real shocks have a significant impact on real and the nominal exchange rates, while nominal shocks induce long-run nominal changes and short-run real changes. Laos experiences relatively larger responses to the nominal shock than Cambodia because the exchange rates in an economy with higher inflation respond to the nominal shock more extremely, according to Ok et al. (2010). 4. The Empirical Model The empirical model identifying real and nominal shocks from the observed real and nominal exchange rate data is given as a bivariate SVAR model in Equation (1) if log of real (y1 t) and log of nominal (y2 t) exchange rates are not cointegrated. ∆yt=B0∆yt+B1∆yt−1+· · · +Bkyt−k+εt(1) where ∆yt=∆y1 t∆y2 t0 is 2 × 1 column vector, ( ∆y1 t ) and ( ∆y2 t ) are the first differences of log of real and nominal exchange rates, respectively, B0=0b12 b21 0 is 2 × 2 square matrix modeling the contemporaneous relation between ∆y1 t and ∆y2 t , B1 , B2 , . . . , Bk are unrestricted parameter matrix, and k is the sufficient lag length ensuring that structural shocks εt=ε1 tε2 t0 are serially uncorrelated processes with diagonal variance–covariance matrix: Eε1 tε2 t 0=H=h11 0 0h22  . If we find all coefficient matrixes, B0 , B1 , . . . Bk , the real ( ε1 t ) and nominal ( ε2 t ) shocks will be identified. However, we are not able to recover parameters from observed data using (1) due to the simultaneous equation bias. Hence, we transform (1) to the reduced form (2) by multiplying both sides of (1) with (I−B0)−1 , where Iis an identity matrix. ∆yt=A1∆yt−1+A2∆yt−2+· · · +Akyt−k+ut(2) where Ai=(I−B0)−1Bi and ut=(I−B0)−1εt . We can apply ordinary least square (OLS) estimator on Equation (2) and estimate reduced form parameters, A1 , A2 , . . . , Ak , and variance–covariance matrix Σ=Eutu0 t=σ11 σ12 σ21 σ22  . Without one more restriction on the structural model (1), we are not able to recover four structural parameters ( b12 , b12 , h11 and h22 ) from the estimated three reduced form parameters ( σ11 , σ12 =σ21 , σ22) . A simple choice for that restriction is to set b12 or b21 equal to zero. This is called short-run restriction in time series literature in Sims (1980). As Sims (1980) identification is not theoretically J. Risk Financial Manag. 2021,14, 529 7 of 16 rational for our context, we apply long-run restriction established in Blanchard and Quah (1989). To set long-run restriction on SVAR model (1), the reduced form model (2) needs to be written in moving average (MA) form by inverting the reduced form model (1) as (3). ∆yt=I−A1L−A2L2. . . −AkLk−1ut(3) where I is a 2 × 2 identity matrix and L is lag operator. It is worth noting that the reduced form model (2) could be written in MA form (3) given only that ∆y1 t and ∆y2 t are stationary. Further, we can define I−A1L−A2L2. . . −AkLk−1=C(L) , which is a matrix of infinite order lag polynomials, as (4). ∆yt=C(L)ut=C11(L)C12(L) C21(L)C22(L) u1 t u2 t(4) where u1 t and u2 t are reduced form residuals. Equation (4) is the MA representation of reduced form (2). We can transform first difference form (4) to level form (5) by multiplying both sides of (4) with (1−L)−1as follows: yt=(1−L)−1C(L)ut=(1−L)−1C(L)(I−B0)−1εt. (5) From the level form (5), the long-run effect of structural shocks on ytis lim s→∞ ∂yt ∂εt−s =C(1)(I−B0)−1εt. (6) The neutrality assumption of nominal shocks on the long-run value (level) of real exchange rates requires that the matrix C(1)(I−B0)−1 be lower triangular. As a result of this triangular restriction, we have b12 =−[C12(1)/C11(1)]. (7) Since we identify b12 from long-run restriction, the remaining three structural parameters are just identified from the following system of three equations by substituting the value of b12 into the system (8a)–(8c). σ11 −2b12σ12 +b2 12σ22 −h11 =0 (8a) −b21σ11 +(1+b12b21)σ12 −b12σ22 =0 (8b) b2 21σ11 −2b21σ12 +σ22 −h22 =0 (8c) 5. The Data During the early years of transition from centrally planned economy to market-based economy, MNT nominal exchange rate depreciated substantially (Figure 2) and consumer price index (CPI) skyrocketed due to the government abandoning control of price and exchange rate. Therefore, we exclude data before January 1994, because the exchange rate movements during that period were mostly driven by transition effects. The data used in our study are from the International Financial Statistics (IFS), IMF. We collected seasonally unadjusted monthly nominal exchange rates and seasonally unadjusted monthly consumer price indexes of Mongolia (MN) and China (CN) from January 1994 to May 2021. The nominal exchange rate is the monthly average of MNT per unit of CNY. The real exchange rate is the nominal exchange rate times the ratio of CPI of CN to CPI of MN. In other words, the real exchange rate is the per consumption basket of CN in terms of the consumption basket of MN. Our measure of real exchange rate closely mimics the REER estimated by the BoM between January 2000 and May 2021 (Figure 2), because J. Risk Financial Manag. 2021,14, 529 8 of 16 China weighs heavily in the REER calculation. We take the logarithm of these two variables before making preliminary analysis and estimating SVAR model. First of all, we perform unit root and cointegration tests in order to ensure that the SVAR is the proper model in our context. Table 1presents the results of unit root tests. Lag lengths of unit root tests are based on Akaike information criteria (AIC). The test statistics of the augmented Dickey–Fuller (ADF) and the Philips–Perron (PP) tests indicate that all of the variables have a unit root in level, as the null hypotheses of unit root are not rejected at the conventional significance level. However, both ADF and PP test statistics reject the presence of unit roots in all variables in the first differences at the 1% significance level. Since the real and nominal exchange rates are I(1) processes, we further need to test cointegrated or long-run relations between these two variables. The Engle–Granger (EG) and Johansen cointegration (JC) test statistics are shown in Table 1. Both EG and JC test statistics indicate that the null hypotheses of no cointegration are not rejected at a conventional significance level. We will discuss why these two series are not cointegrated in the following section. Therefore, we can apply the SVAR model, which is preferred to the structural vector error correction (SVEC) model, in our case. Table 1. Unit root and cointegration test. Unit Root Test Test Type Specification Level First Difference L. of R. Rate L. of N. Rate L. of R. Rate L. of N. Rate ADF None 0.775 4.109 −9.383 *** −4.254 *** Constant −2.371 −1.881 −9.381 *** −5.422 *** C. and T. −2.409 −2.582 −9.441 *** −5.529 *** PP None 0.850 4.338 −15.954 *** −15.228 *** Constant −2.008 −1.765 −123.990 *** −23.148 *** C. and T. −2.002 −2.559 −181.241 *** −18.223 *** Cointegration test EG Dependent var. tau-statistic Probability z-statistic Probability L. of N. rate −0.745 0.939 −1.696 0.944 L. of R. rate −1.095 0.883 −4.113 0.800 JC Unrestricted cointegration rank test based on trace statistic # of CE(s) Eigenvalue Trace statistic 0.05 C.V. Probability None 0.029 10.156 15.495 0.269 At most 1 0.004 1.204 3.841 0.273 Unrestricted cointegration rank test based on maximum eigenvalue # of CE(s) Eigenvalue M.E. Statistic 0.05 C.V. Probability None 0.029 8.952 14.265 0.290 At most 1 0.004 1.204 3.841 0.273 Source: Authors’ estimation. Note: L.—logarithm, R.—real, N.—nominal, ADF—augmented Dickey–Fuller test (Dickey and Fuller 1979), PP—Philips–Perron test (Phillips and Perron 1988), EG—Engle–Granger single equation test (Engle and Granger 1987), JC—Johansen cointegration test (Johansen 1991), var.—variable, #—number, CE(s)—cointegration equation(s), C.V.—critical value and M.E.—maximum eigenvalue. *** indicates 1% significance level. 6. Estimation Result The reduced form bivariate VAR model (2) is estimated for log difference of real and nominal exchange rate series over the period January 1994–May 2021. Starting with a maximum lag of 24, the optimal lag length (k) is 24 according to three criteria: likelihood ratio test statistic at the 5% significance level, final prediction error, and AIC (see Supplementary File). We test serial correlations in the reduced form residuals using Ljung–Box Q statistics, which affirms that there are no significant serial correlations in reduced form residuals (see Supplementary File). However, the residuals of the reduced form VAR model are correlated with each other such that the contemporaneous pairwise correlation between u1 t and u2 t is 0.81 and the corresponding t-statistic is 23.86 (see Supplementary File). This J. Risk Financial Manag. 2021,14, 529 15 of 16 References Aleisa, Eisa, and Sel Diboo ˆ glu. 2002. Sources of real exchange rate movements in Saudi Arabia. Journal of Economics and Finance 26: 101–10. [CrossRef] Balassa, Bela. 1964. The purchasing-power parity doctrine: A reappraisal. Journal of Political Economy 72: 584–96. [CrossRef] Batsuuri, Haltar. 2015. Original Sin: Is Mongolia Facing an External Debt Crisis? The Northeast Asian Economic Review 3: 3–15. Beveridge, Stephen, and Charles R. Nelson. 1981. A new approach to decomposition of economic time series into permanent and transitory components with particular attention to measurement of the ‘business cycle’. Journal of Monetary Economics 7: 151–74. [CrossRef] Blanchard, Olivier Jean, and Danny Quah. 1989. The Dynamic Effects of Aggregate Demand and Supply Disturbances. The American Economic Review 79: 655–73. Calvo, Guillermo A., and Carmen M. Reinhart. 2002. Fear of Floating. The Quarterly Journal of Economics 117: 379–408. [CrossRef] Chowdhury, Ibrahim S. 2004. Sources of exchange rate fluctuations: Empirical evidence from six emerging market countries. Applied Financial Economics 14: 697–705. [CrossRef] Clarida, Richard, and Jordi Gali. 1994. Sources of real exchange-rate fluctuations: How important are nominal shocks? CarnegieRochester Conference Series on Public Policy 41: 1–56. [CrossRef] Corden, Warner Max. 1984. Booming sector and Dutch disease economics: Survey and consolidation. Oxford Economic Papers 36: 359–80. [CrossRef] Dagys, Kadirbyek, W. J. M. Heijman, Liesbeth Dries, and Bakyei Agipar. 2020. The mining sector boom in Mongolia: Did it cause the Dutch disease? Post-Communist Economies 32: 607–42. [CrossRef] Dibooglu, Selahattin, and Ali M. Kutan. 2001. Sources of real exchange rate fluctuations in transition economies: The case of Poland and Hungary. Journal of Comparative Economics 29: 257–75. [CrossRef] Dickey, David A., and Wayne A. Fuller. 1979. Distribution of the Estimators for Autoregressive Time Series with a Unit Root. Journal of the American Statistical Association 74: 427–31. [CrossRef] Dornbusch, Rudiger. 1976. Expectations and Exchange Rate Dynamics. Journal of Political Economy 84: 1161–76. [CrossRef] Eichengreen, Barry, and Ricardo Hausmann. 1999. Exchange Rates and Financial Fragility. Working Paper Series, Issue 7418. Cambridge: National Bureau of Economic Research, Inc. [CrossRef] Enders, Walter, and Bong-Soo Lee. 1997. Accounting for real and nominal exchange rate movements in the post-Bretton Woods period. Journal of International Money and Finance 16: 233–54. [CrossRef] Engle, Robert F., and C. W. J. Granger. 1987. Co-Integration and Error Correction: Representation, Estimation, and Testing. Econometrica 55: 251–76. [CrossRef] Fleming, J. Marcus. 1962. Domestic financial policies under fixed and under floating exchange rates. Staff Papers 9: 369–80. [CrossRef] Ganbayar, Gunbileg. 2021. Analysis of External Debt Sustainability in Mongolia: An Estimated DSGE Approach. Sustainability 13: 8545. [CrossRef] Huizinga, John. 1987. An empirical investigation of the long-run behavior of real exchange rates. Carnegie-Rochester Conference Series on Public Policy 27: 149–214. [CrossRef] International Monetary Fund (IMF). 2020. Annual Report on Exchange Arrangements and Exchange Restrictions 2019. Washington, DC: IMF. Johansen, Soren. 1991. Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive Models. Econometrica 59: 1551–80. [CrossRef] King, Robert G., Charles I. Plosser, James H. Stock, and Mark W. Watson. 1991. Stochastic Trends and Economic Fluctuations. The American Economic Review 81: 819–40. Lastrapes, William D. 1992. Sources of fluctuations in real and nominal exchange-rates. Review of Economics and Statistics 74: 530–39. [CrossRef] Lütkepohl, Helmut, and Markus Krätzig. 2004. Applied Time Series Econometrics. Cambridge: Cambridge University Press. Moore, Tomoe, and Eric J. Pentecost. 2006. The Sources of Real Exchange Rate Fluctuations in India. Indian Economic Review 41: 9–23. Morales-Zumaquero, Amalia. 2006. Explaining Real Exchange Rate Fluctuations. Journal of Applied Economics 9: 345–60. [CrossRef] Mundell, Robert. A. 1963. Capital mobility and stabilization policy under fixed and flexible exchange rates. Canadian Journal of Economics and Political Science/Revue Canadienne de Economiques et Science Politique 29: 475–85. [CrossRef] Ok, Seiha, Makoto Kakinaka, and Hiroaki Miyamoto. 2010. Real shock or nominal shock? Exchange rate movements in cambodia and lao pdr. The Singapore Economic Review 55: 685–703. [CrossRef] Phillips, Peter C. B., and Pierre Perron. 1988. Testing for a Unit Root in Time Series Regression. Biometrika 75: 335–46. [CrossRef] Samuelson, Paul A. 1964. Theoretical Notes on Trade Problems. The Review of Economics and Statistics 46: 145–54. [CrossRef] Sims, Christopher A. 1980. Macroeconomics and Reality. Econometrica 48: 1–48. [CrossRef] Stock, James H., and Mark W. Watson. 2001. Vector Autoregressions. Journal of Economic Perspectives 15: 101–15. [CrossRef] Stockman, Alan C. 1987. The equilibrium approach to exchange rates. FRB Richmond Economic Review 73: 12–30. Stockman, Alan C. 1988. Real exchange-rate variability under pegged and floating nominal exchange-rate systems: An equilibrium theory. Carnegie-Rochester Conference Series on Public Policy 29: 259–94. [CrossRef] J. Risk Financial Manag. 2021,14, 529 16 of 16 Taguchi, Hiroyuki, and Ganbayar Gunbileg. 2020. Monetary Policy Rule and Taylor Principle in Mongolia: GMM and DSGE Approaches. International Journal of Financial Studies 8: 71. [CrossRef] Yeung, Ying, and Stephen Howes. 2015. Resources-to-cash: A cautionary tale from Mongolia. Development Policy Centre Discussion Paper, 42. [CrossRef]