Copyright © CC-BY-NC 2020, CRIBFB | AMFBR American Finance & Banking Review; Vol. 5, No. 1; 2020 ISSN 2576-1226 E-ISSN 2576-1234 Published by Centre for Research on Islamic Banking & Finance and Business, USA 27 The Dynamics of the Exchange Rate and Extension of Monetary Trilemma (Iran Economy Case Study)** Mahmoud Allahyarifard PhD Graduated from PhD in Economics Ferdowsi University of Mashhad, Iran E-mail: allahyarifard@gmail.com Mostafa Karimzadeh PhD Assistant Professor of Economics Ferdowsi University of Mashhad, Iran E-mail: m.karimzadeh@um.ac.ir Mohammad Ali Falahi PhD Professor of Economics Ferdowsi University of Mashhad, Iran E-mail: falahi@um.ac.ir Ali Akbar Naji Meidani PhD Associate Professor of Economics Ferdowsi University of Mashhad, Iran E-mail: naji@um.ac.ir Abstract Simultaneous making policy of interest rates, exchange rates and capital accounts can be extended to trilemma theory, contrary to its earlier theories, provided that the imbalances of the private sector, the government and the capital account adjusted through the policy variables such as the government expenditures, the interest rates on domestic deposits, the interest rates on domestic loans, effective exchange rates, foreign prices and foreign interest rates. On the other hand, the components of the extension of trilemma theory in the form of internal and external imbalances affect the exchange rate. In other words, if the real sector markets of the economy are not cleared through the aforementioned trilemma components, and policy variables, internal and external imbalances will be affected by opposite direction of net domestic assets (ΔNDA) and net foreign assets (ΔNFA) of the banking system. This is in accordance with the fundamental principles of the monetary approach balance of payments and exchange rate. Policy variables do not put pressure on the unofficial exchange rate as long as they have the same effect on the net changes in the domestic and foreign assets of the banking system. The purpose of this study is to consider the effect of internal and external imbalances on exchange rate through the simultaneous equations system, generating impulses in policy variables, and examining reactions in Iranian economy. In this paper, the monetary exchange rate determination model is analyzed and examined by using the extension of trilemma theory for macroeconomic data of Iran in the form of internal and external imbalances. The results of this study suggest that policy variables can stabilize the unofficial exchange rate (with other conditions being constant) through trading off internal and external imbalances. Thus, the economic policymaker can, while independently policing interest rates, capital accounts and government expenditures and other policy variables in this research, maintain exchange rate stability as a strategic variable and anchor the general level of prices. Keywords: Trilemma, Internal and External Equilibrium, Policy Variables Trade-Off, Iran. JEL Classification: F31, F32, F3. **This article is based on the Ph.D. dissertation of first author, entitled "Investigating the Relationship between Internal and External Imbalances through the Extension of the Monetary Trilemma (The Case Study of Iran Economy)" with the Supervisor of Dr. Mostafa Karimzadeh, Dr. Mohammad Ali Falahi and consulting of Dr. Ali Akbar Naji Meidani. mailto:Allahyarifard@Gmail.com mailto:m.karimzadeh@um.ac.ir mailto:falahi@um.ac.ir mailto:naji@um.ac.ir Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 28 1. Introduction The impossibility of simultaneously policing of the exchange rate, the capital account, and the interest rate in a small economy is called Trilemma theory. Assuming that policymaking for each of the aforementioned variables in the small assumed economy does not affect other countries (in case of being ceteris paribus), the key question is how the transitional mechanism of internal and external imbalances in the real and monetary sectors will effect on the exchange rate? What are the triggers on the exchange rate in the context of trilemma? Most studies in the field of trilemma emphasize on these three policy indicators, but policy indicators have been developed in some studies. The most important developmental theories of trilemma focused on foreign exchange reserves (Aizenman & Sengupta, 2013) so that internal and external imbalances are moderated through open market operations (OMOs) or foreign exchange operations (FXOs) by economic policymakers. The lack of sufficient foreign exchange reserves to balance the real and the monetary sectors on the one hand and the simultaneous policing of exchange rate, interest rate and capital account variables in a small country with an open economy on the other hand is a clear reason for creating balance of payments and foreign exchange crises (Krugman, 1979). The mechanism of the transition of internal and external imbalances to the foreign exchange sector through the system of simultaneous equations is examined in the context of the extension of trilemma theory. It should be noted that in trilemma theory, the emphasis is on exchange rate stabilization (exchange rate policy), which is one of the limitations of trilemma ignored when the exchange rate floats. Since exchange rate stabilization is important in some small economies as an anchor of price, any variable influencing exchange rate stabilization is important (Maurice, 2005). 2. Literature Review According to Bidabad (2005) the items affecting Balance of Payment (BOP), import and export prices, liquidity, interest rates, risk at both home and abroad, customs barriers and restrictions for export and import affect the exchange rate. In this study, Bidabad used the equilibrium condition of BOP and using the Fischer's quantity theory of money (QTM) (assuming the stability of velocity of money, prices and earnings) to determine exchange rate by the imbalances of money and goods markets at home and abroad under specific assumptions. According to Mundell (1963); Fleming (1962) monetary policy in terms of floating exchange rate and full capital mobility has the most effect on national income compared to fixed exchange rate (in general equilibrium). In sterilization conditions, the volume of the reserves is not effective on the monetary basis, in other words, if the monetary policy is expanded under the fixed exchange rate regime, reducing the reserves having a negative impact on the monetary base reduces the liquidity that should be offset by the injection of new money.The policy should be short-term because, by continuously injecting liquidity in order to counteract the decline in the reserves and its monetary effect, the reserves are continually diminished, whereas in non- sterilization and fixed exchange rate regimes, the reduction of foreign exchange reserves is effective on a monetary basis and the effect of monetary policy is neutralized. Fixed exchange rate and the effects of sterilization of the reserves to keep equilibrium is one of the reasons for limiting trilemma in concurrent interest rate and capital account policy making. Frankel (1983) view is another important monetary base theory in determination of exchange rate. In general, according to the theory, the factors influencing the exchange rate and BOP are the relative supply and demand for money. In this model, for the stability of money demand function, the QTM is used as the basic theory of money demand function (first assumption). According to this view, the economy is permanently in full employment as prices and wages adjusted rapidly (second hypothesis). Also the theory of purchasing power parity (PPP) (third assumption) is crucial to determine the exchange rate. In the monetary BOP model, the relative increase in the supply of money relative to the demand for money (the increase in the money supply compared to the trading country) is effective in raising the exchange rate. It is therefore inferred that any of the factors affecting money supply and demand (money supply sources including net domestic and foreign assets of the banking system) can lead to exchange rate volatility and force the policy maker to intervene, or without government intervention leading to turbulence, and the currency crisis in the economy as well. The current account deficit is one of the factors influencing the exchange rate in the theoretical literature of the monetary model1. 1 Based on the monetary model of exchange rate, one can see how changes in the money supply stimulate the exchange rate.The demand (Md) and supply (MS) of money equations in terms of relative equilibrium (based on QTM) are defined as follows for the domestic and foreign country: Md= kPy Md*= k*P*y k, k*>0 Where Md is money demand, k is income elasticity of money demand, P is local prices index, y is real national income, Ms is money supply and e is the exchange rate (The asterisk (*) represents foreign countries variables). In equilibrium case, equality of supply and demand of domestic and foreign currency, the following equations exist: Ms = Md Ms*=Md* Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 29 In the research of Mussa (1976) four main points are made in the development of the fundamental principles of the monetary base of BOP method:  The exchange rate is the relative value of national currencies rather than national product.  The exchange rate is strongly influenced by the future expectations of asset owners that are affected by the exchange rate, and these expectations are influenced by monetary policy.  Real factors, as same monetary factors as, influence the behavior of exchange rate.  The contradiction in the nature of fixed exchange rate policy (trilemma), though reduced by the advent of managed floating exchange rates, but it has not been entirely eliminated. This paper deals with some applications of rational expectations in exchange rates. According to Maurice (2005) studies, the choice of a fixed exchange rate regime in the periphery economies makes monetary policy serve to stabilize the exchange rate. In their work, the reasons for choosing a fixed exchange rate regime despite the monetary policy sacrifice are:  Unpredictable exchange rate volatility is detrimental, and most economists believe that exchange rate uncertainty will reduce international trade, reduce investment and harm human capital.  Fixed exchange rate avoids inflationary pressures on the economy, which may be due to government budget deficits, wage policy and pricing by private sector. Therefore, the fixed exchange rate prevents the incentives to follow the macroeconomic expansionary policies of the government.  The fixed exchange rate regime is seen as the anchor of prices after a period of price volatility in some economies to determine price expectations of the goods and services for the next economic activity period. Aizenman & Ito (2014) examined the potential effect of divergence on open door policies (capital mobility) in the context of the macroeconomic trilemma. They found that emerging countries in the past 15 years (since the date of the survey) have adopted a combination of policy indicators of trilemma with the least relative deviation of policies. They also found that countries most likely to deviate from policy indicators are likely to be in a foreign exchange or debt crisis. However, under developing and emerging countries with the most relative deviation from the policies are less likely to be disrupted in the event of a foreign exchange or banking crisis. Maurice (2015) examines resilience of emerging market economies to financial and monetary shocks of foreign origin through the monetary policies of these countries. In other words, how the emerging countries' interest rate policy is driven by foreign financial and monetary shocks is important in this study. The trilemma first brings to mind that the countries with a floating exchange rate have a much better position than those countries that sacrifice their monetary policy, although the floating exchange rate regime by itself does not protect countries' economies from foreign and monetary shocks. The study shows that, despite the potential economic benefits of global integration, globalization weakens economic management. In other words, the trade-off efficiency of monetary policy indicators for achieving multiple domestic goals is undermined. Ray's (2016) study also suggests that the exchange rate regime (whether fixed or floating) affects the correlation between the short-term interest rates of periphery and center countries, so that in the floating exchange rate regime, the interest rates of these countries are less correlated and in the fixed exchange rate regime, the interest rates are higher. Magas (2018) study shows that the resilience of the economy to external shocks depends not only on the structure of trilemma but also on other macroeconomic components. In this regard, the stability of the budget and the repayment of external debt is of particular importance. The problem of reaction to external impulses is not limited to trilemma. The debt crisis of Portugal and Greece is one of the recently confirmed examples of the use of the European currency (Euro Money) as domestic currency, whereas the Czech Republic, despite being a traditional conservative non-euro European member, has relative independence in interest rate policy, and it resists foreign shocks well through its floating exchange rate and interest rate margins. 3. Methodology In this paper, we try to analyze the effect of internal and external imbalances on exchange rate fluctuations in line with macroeconomic theories in real and monetary equilibrium conditions based on trilemma framework. Extending of the trilemma theory in determining the exchange rate, applying the economic relations governing the structure of general equilibrium in the k, k*>0 Given the relative supply and demand of money in terms of equilibrium (based on the above equations) and PPP, the following equation can be obtained: e = P/P* Ms/Ms*= k.P.y/ k*.P*.y* The above equation can be rewritten as the following equation by PPP equation and replacing it with the equilibrium equations of money supply and demand and solving its reduced form of exchange rate as follows: Ms/Ms*= k.y.e / k*.y* e = )Ms/ Ms*(/)k.y/ k*.y*( Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 30 markets for goods and services and money and how they relate to each other and having a purely theoretical framework based on the applicable principles of the National Accounts System (2008) and practical and applied monetary economic theories is achieved. Assuming that the real sector markets are not cleared under this model so that it affects the net changes in the assets (domestic and foreign assets) of the banking system multiplies velocity of money, and it effects on the exchange rate as well, the main question in this paper is, what is the role of policy variables in imbalances so as to stabilize the exchange rate? The exchange rate fluctuations resulting from the model presented in this paper separates the role of the trilemma factors into two parts of internal and external imbalances affecting the exchange rate. Internal imbalances, including the private sector and the public sector deficits increase unofficial exchange rate and external imbalances (balance of payments surplus) reduce it. In other words, this paper attempts to apply policy variables such as government expenditures, interest rates on domestic deposits, interest rates on domestic loans, export, import, capital inflows and outflows effective exchange rate, and analyzes opposite effects through trading off internal and external imbalances in the model. 3. Exchange Rate Determination 3.1 Real Economy Sector Equilibrium The equilibrium condition in the real sector of the economy is obtained by the following equation in the macro-economic: I – S + G – T + EX - IM – Rf = 0 (1) I: private sector investment costs, S: private savings, G: government expenditures (both investment and consumption), T: net indirect taxes, EX: export, IM: import and Rf: net foreign transfer payments. The equation shows that the balance of aggregate supply and aggregate demand, which means that the sum of the following items at macroeconomic level must be equal to zero:  Net (deficit or surplus) private sector savings or borrowings (I-S)  Government financial budget (deficit or surplus) (G - T)  Net (deficit or surplus BOP) external sector (EX - IM – Rf) 3.2 Monetary Sector Equilibrium The equilibrium condition in the monetary sector of the economy is obtained by the following equation: NFA + NCG + NCP + NK = M2 = TD + DD + CU (2) On the left side of the above equation, Liquidity Uses (M2) include Net Foreign Assets (NFA), Net Domestic Assets (NDA) (including Net Claims on Government (NCG), Net Claims on Private sector (NCP) and net capital account of the banking system (NK) (in the form of non-returnable assets). Liquidity items to the right of the equation include Time Deposits (TDs), Demand Deposits (DDs) and money (banknotes and coins) in circulation (CU). For convenience, the value of NK in the NCP can be considered and the following equation can be used to define liquidity: M2 = NCP + NCG + NFA (3) 3.3 The Relationship between the Real and the Monetary Sector According to Fisher definition, the QTM is defined as the following equation (Mankio, 107: 2016): M.V=P.t = T (4) Here, M: supply (or demand for money), V: velocity of money, P: price level, t: quantity of goods or services traded, T: value of goods and services traded in the economy. In other words, Persons (2012) in his book, Purchasing Power of Money, describes the relationship between money and transactions in the economy, while other economists in QTM explain the relationship between income and money. In other words, the following can be said: M.V = P.t = T = P.as = P.ad = AS = AD (5) ad: aggregate demand in economy at constant price, as: aggregate supply in economy at constant price, AS: aggregate supply in nominal price, and AD: aggregate demand in economy in nominal price. According to the above equations, multiplying the amount of money by the velocity of money is equal to the value of transactions in the economy. Substituting Equation (3) into Equation (4) or (5) can be written as follows: Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 31 M.V = V[NCP + NCG +NFA] = P.t = P.as = AS (6) In order to connect a relationship between the real economy sector (goods and services) and the monetary sector in the economy, the above equation is differentiated. Assuming V remains unchanged, the following equation can be obtained: V.ΔM = V.ΔNCP + V. ΔNCG +V.ΔNFA (7) The right side of the Equation (7) can be divided into two groups of banking system assets: NDA and NFA. In other words, Equation (7) can be rewritten as Equation (8): V.ΔM = V.ΔNDA+V.ΔNFA (8) On the left side of the Equation (8), the changes in the supply or demand of money multiply by the velocity of money, can be divided into two parts the sum of the changes of the domestic (ΔMs) and foreign (ΔMs*) money supply (in the domestic money) multiply by the velocity of money (V). In other words, the left of Equation (8) can be written as: V .ΔM = V( ΔMs +ΔMs*) (9) Since the result of changes in domestic money supply multiply by the domestic velocity of money comes from the changes in net domestic assets of the banking system multiply by the velocity of money (V.ΔNDA), and the result of changes in the foreign money supply multiply by the foreign velocity of money caused by the changes in the net foreign assets of the banking system multiply by the velocity of money (in domestic money), so by using the Equation (8) and (9), we can write the following equations: V. ΔMs= V.ΔNDA (10) V.e.ΔMs*=V.ΔNFA (11) Assuming, the foreign prices stability, only one foreign currency in outside the world that can be converted into domestic currency by the exchange rate (e), given the stability of the domestic-to-foreign money demand ratio (k.y/ k*.y* = 1), purchasing power parity (PPP= P/P*) validity, so in accordance with the monetary model exchange rates, the following equations can be defined for the exchange rate. e=Ms/ Ms*/ k. y/ k*. y*= NDA/NFA (12) V.Δe= V.ΔNDA/ V.ΔNFA (13) By simplifying algebraic Equation (12) and taking the logarithm and the derivative from the both sides, the following equation is obtained: (14) The equation shows that the percentage of exchange rate changes, in case ceteris paribus mentioned in this study, is directly related to the percentage of NDA and inversely related to the percentage NFA of the banking system. It is now possible to examine each of the components of the liquidity utilization arises from what part of the real economy changes in equilibrium. By comparing the components of changes in the monetary sector to the imbalances in the real sector, one can infer the concept of the reasons for the imbalances in the money sector resulting from the imbalances in the real sector of the economy as follows: (I – S) ≡ V.ΔNCP (G – T) ≡ V.ΔNCG (15) (EX – IM – Rf) ≡ V.ΔNFA The above accounting equations mean that: e NDA NFA  Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 32  Private sector deficit (deficit or surplus), (I-S) by using the banking system resources, causes changes in NCP of banking system to the same extent.  Government budget surplus (or deficit), (G-T), causes changes in NCG of banking system to the same extent.  Surplus (or deficit) foreign sector, (EX-IM-Rf) in foreign currency, causes changes in NFA of the banking system to the same extent. Here, inserting the accounting Equation (15) into Equation (13) then we can arrive at Equation (16): V.Δe= (I-S) + (G-T) /)EX-IM-Rf) (16) The right-hand side of Equation (16) suggests the relationship between exchange rate, income, and BOP through absorption method (Pilbim, 1998). In other words, although internal and external imbalances in the case of non-trade-off the NDA and the NFA affect the liquidity, under these circumstances, it can have an opposite effect on the exchange rate. The Equation (16) highlights an important issue in determining of the exchange rate, so that the greater the foreign exchange reserves due to the BOP surplus (external imbalances) than the private and public sector imbalances (internal imbalances), the unofficial exchange rate expectations decrease, and conversely, the greater the internal imbalance than the external imbalance, the unofficial exchange rate market expectations increase as well. The above equation somehow illustrates the extension of the trilemma theory (Allahyarifard et al., 2019), the effect of internal (I-S) + (G-T) and external (EX-IM-Rf) imbalances on the effect of money changes multiplied by velocity of money2. Equation (17) shows the variables that influence the behavior of the principal components of Equation (16): V.Δe = (I (iL, IROLPV) – S (iD, Y)) + (G – T(Y))/($.)(eex.EX$((eex.P*/P), IRXOILD, y) – eem.IM$((eem.P*/P), (EX$ + TX$ -TM$)) + eexf.TX$((iD- eexf .iD*), P) – eemf.TM$(iD- eemf. iD*))) (17) Equation (17) shows the reduced form of the equilibrium exchange rate in terms of both the monetary and real sector economy equilibrium. In Equation (17) iL: interest rates on domestic loans, IROLPV: obligated loan, iD: interest rates on domestic deposits, Y national income or GDP at nominal prices, y: national income or GDP at constant prices (as a production capacity), eex: effective exchange rate of export, IRXOILD: oil export, eem: effective exchange rate of import, eexf: effective exchange rate of production factor export, eemf: effective exchange rate of production factor import, P: domestic prices index, P*: foreign prices index and iD*: interest rate on foreign deposits. The $ suffix is to express variables in foreign currency. 3.4 Exchange Rate Dynamics Based on Krugman's (1991) target zone regime, the exchange rate is determined by the following equation. (18) Where Ek is the (log of the) spot price of foreign exchange, m the domestic money supply, v a shift term representing velocity shocks, and the last term is the expected rate of depreciation. Since, Krugman (1991) considers the Equation (18) for the mathematical modeling of exchange rate under a target zone regime such that m changes caused by OMO and FXO are assigned to keep the minimum and maximum targeting intervals [ , ]e e . In the model, there are two fundamental factors that affect the exchange rate, money supply (m) to stabilize the exchange rate in the minimum and maximum intervals and velocity shift term (v). If the exchange rate is within the minimum and maximum range, the value of m is constant and exchange rate expectations are E [ds]/ dt =0. Under these conditions, the exchange rate fluctuations between the minimum and maximum depend on the transfer rate (v) and m, which is similar to (12) and (14). 4. Empirical Application To study the effect of internal and external imbalances on the exchange rate and policy influencing variables on the behavior of model endogenous variables in different scenarios and how they trade-off with one another in the Iran economy, it is necessary to use a system of simultaneous equations. The time series data of Iran's macroeconomic variables3 in the current study are from 1959 to 2016 AD (according to 1338-1396 Hijri Shamsi). The equations have been tested several times, and the best ones have 2 The extension of the trilemma (Allahyarifard et al., 2019) is deduced as follows: V .ΔM = )I – S) + )G – T) + )EX – IM – Rf( 3 The data in this study are derived from the Bank of Iran time series macroeconomic variables database.   /KE m v E ds dt   Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 33 been estimated. It should be borne in mind that endogenous and exogenous variables (policy variables) are divided into two parts: flow and inventory, so the balance between the left and right variables of the structural equations is considered in this study. 4.1 System of Simultaneous Equations Since the exchange rate equation (Equation 17) is a function of the components of internal and external imbalances derived from the extension of the trilemma theory. Thus, it can be shown in the context of a system of equations that the internal imbalances include the imbalance of the private sector investment-national savings, government deficit and the external sector imbalances, including BOP, affecting the unofficial exchange rate. 4.1.1 System of Equations The Equation (19) to (26) is the structural equation of the Equation (17). (19) The above equation confirms the effect of internal and external imbalances on the exchange rate. (20) Investment equation (Equation (20)) as a result of economics theory, the inverse function of the interest rate. (21) According to the Keynesian model, the savings equation is considered as a function of national income. (22) The equation of government tax revenue is considered as a function of national income according to economic theories. (23) In the above equation, the export at constant prices (in foreign currency) is shown as a function of the effective export exchange rate and the GDP at constant prices as a production capacity (24) (25)     IREM C(1) C(2)* IREM( 1) C(3)*D(IRIPV IRSV IRGV IRGRTV) C(4)*D(IRXV / IREEX IRMV / IREEM IRXFYV / IREEXFY IRMFYV / IREEMFY) IV : C, IREM 2 , IRIPV IRSV IRGV IRGRTV , IRXV / IREEX IRMV / IREEM IRXFYV / IREEXFY IRMFYV / IRE                   EMFY , IRIRD, IRIRL                       IRIPV  C 10 C 11 *IRIPV 1 C 12 *IROLPV C 13 * IRIRL IRIRL 1 C 14 *D92 IV : C, IRIPV 2 , IROLPV 1 , IRIRL IRIRL 1 , D92                  IRSV C(20) C(21)*IRIRD C(22)*IRGDPMV( 1) IV : C, IRIRD 1 , IRGDPMV                         IRGRTV C 30 C 31 *IRGRTV 1 C 32 *D IRGDPMV C 33 *D89 C 34 *D91 IV : C, IRGRTV 2 , IRGDPMV IRGDPMV 1 , D89, D91                  IRXDFCPI (C(40) C(41)*IRXDFCPI( 1) C(42)*(IRXOILD / FCPI IRXOILD( 1) / FCPI( 1)) C(43)*IREEX*FCPI / IRPGDPM C(44)*D(IRGDPM) IV : C, IRXDFCPI 2 , IRXOILD / FCPI, IREEX*FCPI 1 / IRPGDPM 1 , IRGDPM / IRGDPM 1 , IREEXFY, I              REEMFY         IRMDCIFP (C(50) C(51)*IRMDCIFP( 1) C(52)*(IREEM*FCPI / IRPGDPM) C(53)*D(IRXD IRXFYV / IREEXFY IRMFYV / IREEMFY) IV : C, IREEM 1 , FCPI 1 , IRPGDPM 1 , IRXD IRXFYV / IREEXFY IRMFYV / IREEMFV                                     IRXFYV C 60 C 61 *IRXFYV 1 C 62 * IRIRD FIR *IREEXFY C 63 *D91 C 64 *D88 C 65 *D93 C 66 *D92 IV : IRIRD 1 , FIR 1 , IREEXFY 1 , D91, D88, D93, D92             Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 34 Foreign currency imports at constant prices are considered as a function of the effective exchange rate of import and export earnings. Export and import of factor income from abroad (equations 25 and 26) according to Frankel (1983) are considered as a function of the difference between domestic interest rates. (26) Model accounting equations to tight the closure includes: IRXD ≡ IRXDFCPI*FCPI (27) The above accounting equation shows the relationship between nominal exports (million Dollars) and real exports (million Dollars). IRXV≡ (IREEX* IRXD)/1000 (28) The above equation indicates the relationship between the nominal exports (billion Rials) and the nominal exports (million Dollars). IRMD≡ IRMDCIFP*IRCIFP (29) The above equation indicates that the relationship between nominal imports (million Dollars) and real imports (million Dollars). IRMV≡ (IREEM* IRMD)/1000 (30) The above equation indicates the relationship between nominal imports (billion Rials) and nominal imports (million Dollars). IRGDPMV≡ IRCV+ IRIV+ IRGV+ IRXV- IRMV+ IRDISV+ IRNITV (31) The above equation illustrates the relationship between nominal GDP and its accounting components (billion Rials). IRPGDPM≡ IRGDPMV/IRGDPM (32) The above equation indicates the relationship between real GDP and nominal GDP, agregate price deflator. IRGDPM≡ (IRCV+ IRIV+IRGV+ IRXV- IRMV+ IRDISV+ IRNITV)/IRPGDPM (33) The above equation illustrates the relationship between real GDP and its accounting components (billion Rials). Endogenous variables 1 IREM Market exchange rate, Rials/Dollar 2 IRGRTV Government tax revenue, billion Rials 3 IRIPV Private investment at current prices, billion Rials 4 IRMD Import at current prices, million Dollar 5 IRMV Import at current prices, billion Rials 6 IRSV Private Saving at current prices, billion Rials 7 IRMDCIFP Real import, million Dollars 8 IRMFYV Nominal import of factor income from abroad, billion Rials          IRMFYV (C(70) C(71)*IRMFYV( 1) C(72)*(IRIRD FIR *IREEMFY) C(73)*D(IRGDPMV) IV : C, IRMFYV 2 , IRIRD 1 FIR 1 *IREEMFY 1 , IRGDPMV            Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 35 9 IRPGDPM Gross domestic product price deflator 10 IRXFYV Nominal export of factor income from abroad, billion Rials 11 IRXD Export, million Dollars 12 IRXDFCPI Real export, million Dollars 13 IRXV Nominal export, billion Rials 14 IRGDPM Real gross domestic product at market prices, billion Rials 15 IRGDPMV Nominal gross domestic products at market price, billion Rials Exogenous variables 1 FCPI Average OECD consumer price index 2 FIR Average OECD Interest rate (banking system) 3 IRCV Nominal private consumption, billion Rials 4 IRCPI Consumer price index 5 IRCIFP Import CIF price index 6 IRDISV Nominal discrepancies, billion Rials 7 IREEM Effective exchange rate for Import 8 IREEMFY Effective exchange rate for Import of factor income from abroad 9 IREEX Effective exchange rate for Export 10 IREEXFY Effective exchange rate for Export of factor income from abroad 11 IRGV Nominal government consumption, billion Rials 12 IRNITV Nominal net indirect taxes, billion Rials 13 IROLPV Government budget obligatory loans granted to private sector 14 IRIRD Saving deposits weighted average interest rate (banking system) 15 IRIRL Loans weighted average interest rate (banking system) 16 IRXOILD Export of oil, million Dollars 17 D88 Dummy variable for the 1388 18 D89 Dummy variable for the 1389 19 D91 Dummy variable for the 1391 20 D92 Dummy variable for the 1392 21 D93 Dummy variable for the 1393 4.1.2 Structural Equations Identification In general, before estimating the parameters of the model equations, it is necessary to identify the order (necessary condition) and the rank condition (necessary and sufficient condition) for consistent estimation4. As can be seen in Equations (19) to (33) there is a stochastic equation or an accounting equation for each endogenous variable in Equation (19). The absence of linear combination in the vectors of the instrumental variables and other independent variables determines the rank condition. In other 4 If the structural equation has an endogenous variable (y1) on the left, an endogenous variable (g1) and an independent variable (k1) on the right, so that the endogenous variable to the right of the equation is correlated with the error term, then the estimation of the model equation coefficients is inconsistent under the OLS method. If there is a structural equation for each right endogenous variable to describe its behavior then it is called a complete system of equations. Assuming that K independent variables exist in the system of equations, then the minimum number of independent variables necessary for consistent estimation in the structural equations is obtained through the order condition for identification. The order condition is obtained from 2 1k g provided k2 = K- k1; k2 is the number of the independent variables outside the equation examined for identification in the equation system that affect the behavior of the endogenous variables to the right of the equation. It should keep in mind that the order condition is necessary to identify and the rank condition is sufficient as well. Also, without using the matrices used to obtain the rank order, the two-stage least squares (2SLS) method can have a consistent estimate for the parameters, so that in the first step, the endogenous variables of the equation are obtained using the other independent variables affecting the behavior of the endogenous variable are estimated by the OLS method and secondly, by estimating the variables in the first step, the endogenous variable to the left of the equation is estimated. Therefore, the order condition identification requirement in the 2SLS method is sufficient for consistent estimation of model equations (Baltagi, 2008: pp.256: 277). In the model examined in this paper, Equation (19) has 7 endogenous variables on the right and 9 structural equations that each equation is estimated from at least one independent variable outside of Equation (19). Therefore, the order condition is over-identified which can be estimated by using the 2SLS method. The recent issue is also mentioned in Chapter15 of the book Econometrics Analyzes (Green, 2002: 393) that "It is unusual for a model to pass the order but not the rank condition". Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 36 words, if the model of the M equations and the M endogenous variables can only have at least one non-zero determinant of the matrix (M-1) * (M-1) related to the coefficients of the endogenous and predetermined variables outside the examined equation but included in other model equations is identified (Gujarati, 2004). The model has 8 random equations, 7 accounting equations, 15 endogenous variables (M = 15) and 28 pre-determined explanatory variables (K = 33)5. So the model equations (equation19 to 26) are over-identified. 4.1.3 Stationary of the Variables Consideration As can be seen in the table (1), some variables are at least first-order integrated I(1), and the first differences are used in the regression equations for being stationary. Therefore, in order to avoid spurious regressions in the system of equations, first-order difference is used. According to table (2), all the parameters of lagged variables of the structural equations are approximately equal to one, and mathematically if the lagged variable is moved to the left of the equation then the endogenous variable difference is obtained6. In other hand, after numerous and repeated reviews this solution can be solved first, I(1) and random walk by first order difference through insertion of first order lagged dependent variable to the right, then by using the residual inverse through Weighted Least Squares method (WLSs) eliminate variance heterogeneity of the residuals that have more than first-order integrated I (1)7. Table 1. Generalized Dickey-Fuller unit root test on the level of the model variables Variable Augmented Dickey-Fuller test statistic Prob.* t-Statistic IREM 1.579045 0.9707 -1.946764 IRIPV -4.655605 0.0000 -1.947975 IRSV 0.317437 0.7732 -1.947975 IRGV 6.668408 1.0000 -1.947520 IRXV 2.965434 0.9990 -1.947975 IRMV -.225549 0.5997 -1.947975 IRXFYV 3.373523 0.9997 -1.947975 IRMFYV -4.31828 0.0001 -1.947975 IROLPV -2.309784 0.0214 -1.946654 IRIRL 0.665009 0.8568 -1.946654 IRIRD 1.518375 0.9668 -1.946764 IRGDPMV 1.390017 0.9569 -1.947975 IRXDFCPI -0.930043 0.3095 -1.946654 IRXOILD -0.760627 0.3824 -1.946654 FCPI -1.157899 0.2222 -1.946996 IREEX 5.273733 1.0000 -1.946654 IRPGDPM 2.875961 0.9987 -1.947975 IRMDCIFP -1.575963 0.1074 -1.946764 IREEM 2.595083 0.9972 -1.947975 IRXD .177568 0.7341 -1.946654 IREEXFY 1.270808 0.9467 -1.946654 5 Number of independent variables in the model (26) + Number of lagged endogenous variables (7) = Number of explanatory and predetermined variables (K = 33). The 26 independent variables included: 5 dummy variables + 5 weighted variables (residuals of the model equations) to eliminate heteroskedasticity + 16 policy variables. 6 The stability of the endogenous variables of the structural equations of this model, without considering the lagged endogenous variable on the right of the equations, can be examined by Engle-Granger method. The stability results are discussed in table (3). 7 Economic theories refer to possible equilibrium relationships between variables, but do not explain the relevant adjustment processes at work. If there is an equilibrium relationship, then the variables specified in the relationship should be cointegrated. Testing for cointegration is, in fact, the test of equilibrium relation, and hence of whether the model is well defined. When the variables are cointegrated, the estimates of the long-run equilibrium parameters are consistent and highly efficient. Indeed these estimates are "super consistent", converging even more quickly in probability to the true parameter values than the least-squares estimator in the standard case. This consistency property does not require the absence of correlation between the right-hand-side variables and the error term, unlike consistency results in the usual classical regression-model context (Perman, 1991). Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 37 IREEMFY 7.721210 0.9999 -1.947248 FIR -0.574141 0.4643 -1.946654 Source: Obtained from results of the research Some economic events and structural failures in some variables were corrected through dummy variables. The statistics of F and its probability are highly acceptable such that the standard deviation error statistics, t, and significant probability close to zero confirm the parameters (see Table 2). 4.2 Estimation of the System of Equations by OLS, 2SLS and 3SLS Methods Since the explanatory variables in the simultaneous equations are correlated with the error term, so the OLS parameter estimation is not consistent with classical econometric issues8, so it is necessary to use the 2SLS or 3SLS (IV) variables to estimate the regression equations parameters. Three-Stage Least Squares method (3SLS) is a version of Two-Stage Least Squares (2SLS) in the Seemingly Unrelated Regression equations (SUR). On the other hand, if the structural equation error term has heteroskedasticity and the correlation between error terms with each other structural equation is recognized, then 3SLS is more efficient (less variance) than 2SLS9. If there is no correlation between the error terms of the simultaneous equations, then the results of the estimators in the 2SLS and 3SLS methods are the same. Another point is that since 2SLS is a single-equation estimation method it is not possible to examine the covariance between the residuals of the equations. Table 2. Results of Estimates and Statistics by OLS, 2SLS and 3SLS Methods Variable Type of estimates Coefficient Std. Error t-Statistic Prob. Intercept OLS 261.45 28.39 9.21 0.0000 2SLS 306.40 66.93 4.59 0.0000 3SLS 264.95 78.77 3.36 0.0008 8 The estimators are consistent when the assumption that Xt is non-random, or no correlation between the explanatory variables and the disturbance component is proved. That is, despite the autocorrelation with increasing sample size, the variance of the estimating parameters tends to zero, ie: 𝑝𝑙𝑖𝑚(�̂�) = 𝛽 9 2SLS is a single equation estimator that does not take into account the covariances between the residuals of simultaneous equations, so this method is generally not efficient. In contrast, 3SLS is a systematic method for simultaneous estimation of all coefficients of the model, after simultaneous estimation of all coefficients, the weights are formed, and the model is again estimated using the estimated weight matrix. System estimators such as 3SLS consider zero constraints of each structural equation as the variance-covariance matrix of the whole system of equations residuals. The first two stages of 3SLS are similar to 2SLS, and the third stage uses Feasible Generalized Least Squares (FGLS) in the same way as SUR. 3SLS uses the 2SLS residuals for consistent estimation of the structural equation (Σ) covariance matrices. In 3SLS, structural equations overlap like SUR, in mathematical terms. So consider: y = Z u  1 1 1 1 2 2 2 2 0 0 0 0 0 0 . ; ; ; . . . . . . . 0 0 0 G G G G y Z u y Z u y Z u Z u y                                                     where in: u with 0 mean and TI variance-covariance indicating the correlation between the residuals of the structural equations. ', ( )ij i j ij TE u u I     , and the marker represents the Kronecker variance - residual covariance. If there is no correlation between the residuals of the structural equations with each other then 0ij  , as a diagonal matrix, therefore the estimators of the least squares and ordinary least squares are similar (Baltagi, 2008: pp.256: 277). Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 38 IREM(-1) OLS 1.014 0.002 484.84 0.00000 2SLS 1.011 0.003 296.53 0.0000 3SLS 1.016 0.007 144.93 0.0000 D(IRIPV- IRSV+ IRGV- IRGRTV) OLS 0.001 0.00008 13.11 0.0000 2SLS 0.001 0.0001 9.42 0.0000 3SLS 0.0009 0.00002 4.01 0.0000 D(IRXV/IREEX) - (IRMV/IREEM) + (IRXFYV/IREEXFY)- (IRMFYV/IREEMFY) OLS -13.79 2.39 -5.77 0.0000 2SLS -17.55 5.64 -3.11 0.0030 3SLS -14.42 6.62 -2.18 0.0299 Intercept OLS 7067.74 7243.03 0.98 0.3337 2SLS 6309.48 7795.32 0.81 0.4220 3SLS 6284.52 7423.56 0.85 0.3984 IRIPV(-1) OLS 1.08 0.01 104.60 0.0000 2SLS 1.07 0.02 63.97 0.0000 3SLS 1.07 0.02 2567 0.0000 IROLPV OLS 0.56 0.06 9.12 0.0000 2SLS 0.69 0.14 5.13 0.0000 3SLS 0.69 0.13 5.38 0.0000 (IRIRL-IRIRL(-1)) OLS -12039.17 5279.18 -2.28 0.0267 2SLS -16725.79 6975.73 -2.40 0.0202 3SLS -16710.36 6655.21 -2.51 0.124 D92 OLS -467599.5 58073.07 -8.05 0.0000 2SLS -518314.8 77059.63 -6.73 0.0000 3SLS -518246.5 72536.69 -7.05 0.0000 Intercept OLS -250670.9 17986.51 -13.94 0.0000 2SLS -312893 33811.88 -9.25 0.0000 3SLS -244555.1 35909.34 -6.81 0.0000 IRIRD OLS 38276.33 2628.25 14.56 0.0000 2SLS 47386.45 4947.54 9.58 0.0000 3SLS 37959.78 5217.06 7.28 0.0000 IRGDPMV(-1) OLS 0.39 0.008 49.15 0.0000 2SLS 0.37 0.015 25.02 0.0000 3SLS 0.39 0.01 25.58 0.0000 Intercept OLS -892.62 2313.75 -0.39 0.7012 2SLS -962.83 2363.62 -0.41 0.6855 3SLS -893.82 2254.57 -0.40 0.6920 IRGRTV(-1) OLS 1.16 0.02 53.55 0.0000 2SLS 1.16 0.02 52.34 0.0000 3SLS 1.16 0.02 54.88 0.0000 D(IRGDPMV) OLS 0.04 0.007 5.27 0.0000 2SLS 0.04 0.007 4.99 0.0000 3SLS 0.04 0.007 5.24 0.0000 D89 OLS -94729.88 16077.64 -5.89 0.0000 2SLS -94958.33 16242.40 -5.85 0.0000 3SLS -95007.14 15496.55 -6.13 0.0000 D91 OLS -53079.03 16217.26 -3.27 0.0019 2SLS -53565.12 16387.80 -3.27 0.0019 3SLS -53941.74 15634.96 -3.45 0.0006 Intercept OLS -1520.59 160.35 -9.48 0.0000 2SLS -1395.74 154.00 -9.06 0.0000 3SLS -1763.34 452.68 -3.89 0.0001 Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 39 IRXDFCPI(-1) OLS 1/00 0.003 334.40 0.0000 2SLS 1/00 0.002 450.74 0.0000 3SLS 1/00 0.009 115.07 0.0000 IRXOILD/FCPI-IRXOILD(- 1)/FCPI(-1) OLS 0.83 0.1 57.13 0.0000 2SLS 0.83 0.01 71.37 0.0000 3SLS 0.84 0.03 25.25 0.0000 IREEX*FCPI/IRPGDPM OLS 0.17 0.03 5.69 0.0000 2SLS 0.14 0.03 4.52 0.0000 3SLS 0.22 0.09 2.36 0.0188 D(IRGDPM) OLS 0.03 0.003 13.46 0.0000 2SLS 0.003 0.001 23.32 0.0000 3SLS 0.03 0.007 4.46 0.0000 Intercept OLS 112846.9 5624.06 20.06 0.0000 2SLS 119149.1 6066.96 19.64 0.0000 3SLS 83608.23 20792.606 4.02 0.0001 IRMDCIFP(-1) OLS 0.91 0.02 47.64 0.0000 2SLS 0.90 0.02 43.47 0.0000 3SLS 0.97 0.06 15.83 0.0000 (IREEM*FCPI/IRPGDPM) OLS -14.16 0.65 -21.86 0.0000 2SLS -14.90 0.69 -21.39 0.0000 3SLS -10.62 2.44 -4.35 0.0000 D(IRXD+IRXFYV/IREEXFY- IRMFYV/IREEMFY) OLS 1.43 0.08 19.18 0.0000 2SLS 1.52 0.08 19.00 0.0000 3SLS 1.12 0.26 4.34 0.0000 Intercept OLS -369.85 212.73 -1.74 0.0883 2SLS -298.12 255.62 -1.17 0.2491 3SLS -311.13 239.03 -1.30 0.1937 IRXFYV (-1) OLS 0.96 0.02 50.32 0.0000 2SLS 0.95 0.03 31.04 0.0000 3SLS 0.95 0.03 33.17 0.0000 IRIRD-FIR*IREEXFY OLS -0.25 0.03 -7.85 0.0000 2SLS -0.24 0.06 -4.13 0.0001 3SLS -0.24 0.05 -4.50 0.0000 D91 OLS 28230.86 1302.04 21.68 0.0000 2SLS 28438.76 1337.61 21.26 0.0000 3SLS 28446.82 1251.62 22.73 0.0000 D88 OLS -15367.00 1305.66 -11.77 0.0000 2SLS -15224.69 1327.01 -11.47 0.0000 3SLS -15251.71 -1241.11 -12.29 0.0000 D93 OLS -22578.59 1645.75 -13.72 0.0000 2SLS -22255/00 1868.64 -11.91 0.0000 3SLS -22285.11 1748.68 -12.74 0.0000 D92 OLS 17001.52 1449.77 11.73 0.0000 2SLS 17405.01 1464.6 11.88 0.0000 3SLS 17316.56 1367.71 12.66 0.0000 Intercept OLS 5.96 0.15 38.51 0.0000 2SLS 5.62 0.18 30.84 0.0000 3SLS 4.93 1.32 3.73 0.0002 IRMFYV(-1) OLS 1.04 0.005 210.96 0.0000 2SLS 1.05 0.006 184.17 0.0000 3SLS 1.06 0.04 25.76 0.0000 (IRIRD - FIR*IREEMFY) OLS 0.008 0.001 40.61 0.0000 Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 40 2SLS 0.007 0.0002 31.35 0.0000 3SLS 0.006 0.001 5.56 0.0000 D(IRGDPM) OLS 0.001 0.00006 25.98 0.0000 2SLS 0.001 0.000006 202.62 0.0000 3SLS 0.001 0.00005 2.35 0.0191 Source: Obtained from results of the research According to the table (2), R-squared and Adjusted R-squared show very good explanatory power of the equations. The use of lagged first order endogenous variables on the right of the equations increases R-squared (R2). Dorbin-Watson's statistics show good quality model equations specifications and no serial correlation in the residuals. It should be noted that in cases of the high R-squared, it is not necessary to consider the Dorbin-Watson's statistic on the problem of the serial correlation of the residuals, this is not. However, acceptable either weak serial correlation or when the statistic falls within the uncertain limits of serial correlation in the Dorbin-Watson's table. Note that the Dorbin-Watson's statistic is not usable for the expression of the residuals due to the existence of an endogenous lagged variable on the right of the equation. It should keep in mind that if R2 is high, the estimates of the equations in the two methods OLS and 2SLS are very close to each other (Gujarati, 2004, p. 879). In order to examine the residual stability, prior to simulating the actual data, it is necessary to examine the Dickey-Fuller unit root test according to the table (3). Table 3. Group unit root test: Summary Series: RESID04, RESID06, RESID08, RESID09, RESID10, RESID11, RESID12, RESID13 Date: 05/27/19 Time: 02:37 Sample: 1338 1395 Exogenous variables: None Automatic selection of maximum lags Automatic lag length selection based on SIC: 0 to 8 Newey-West automatic bandwidth selection and Bartlett kernel Cross- Method Statistic Prob.** sections Obs Null: Unit root (assumes common unit root process) Levin, Lin & Chu t* -18.0355 0.0000 8 432 Null: Unit root (assumes individual unit root process) ADF - Fisher Chi-square 407.884 0.0000 8 432 PP - Fisher Chi-square 933.188 0.0000 8 443 ** Probabilities for Fisher tests are computed using an asymptotic Chi -square distribution. All other tests assume asymptotic normality. Source: Obtained from results of the research According to the statistics of the table (3) based on the Engle-Granger method, the residuals of the regression equations of the model were evaluated for stability, the results show both common or individual do not have any unit root. 4.3 Simulation After estimating the parameters, the model was solved statically for the Ex-post simulation by using the Broyden method. The time searies of the main equation of the proposed model for real (squared) and simulated (starred) information is shown in Figure (1). Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 41 Figure (1): An informal exchange rate trend chart based on real and simulated information Source: Obtained from results of the research 4.4 Model Evaluation To evaluate the model, the root mean square error (Rmsp (e)) method, which represents the percentage error of the simulated variable in the model, was used as follows: 2 1 ( ) ( ) 100. M t i t e Y Rmsp e M    (34) et = Yt – Ŷ: Simulated variable error Yt: The actual variable Ŷ: Simulated variable Table 4. Root Mean Square Error of Simulated Error (Rmsp (e)) 1998-2016 IRMFYV IRXFYV IRMV IRXV IRGRTV IRSV IRIPV IREM Type of estimation 34.09 41.25 48.64 15.64 17.88 40.28 10.79 13.58 OLS 34.27 41.25 50.61 15.73 17.87 46.86 10.64 14.02 2SLS 35.47 43.00 43.50 13.03 16.99 40.39 10.64 13.28 3SLS IRPGDPM IRMDCIFP IRXDFCPI IRMD IRXD IRGDPMV 0.000001 48.64 15.64 48.64 15.64 12.59 OLS 0.00001 50.61 15.73 50.61 158.73 12.97 2SLS 0.00001 43.50 13.02 43.50 13.02 10.98 3SLS Source: Obtained from results of the research Table (4) shows the root mean square error of the endogenous variables for the 18-year period ending in 2016, which illustrates the appropriate model explanation. 4.5 Impulse Analysis After simulation of endogenous variables through comparative static solution of the model and examining the validity of the model, by generating impulse in each of the exogenous variables, several scenarios can be defined with respect to impulse and response of the endogenous variables. The analysis shows that despite the imbalances in the real sectors of the economy, the 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 40 45 50 55 60 65 70 75 80 85 90 95 IREM IREM (Baseline) IREM Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 42 unofficial market exchange rate does not fluctuate if internal and external imbalances are cleared through the trade-off private, government and BOP imbalances. The response of the impulses considered varies according to the type of relationship of the exogenous variables. Various scenarios with the assumption that other conditions are stable (ceteris paribus) over the 18-year period ending in 2016 are shown in table (5). Table 5. Exchange of imbalances in real sector due to shocks (OLS / 2SLS / 3SLS 1% impulse foreign deposit interest rate 10% impulse foreign prices 10% Impulse effective exchange rare 1% Impulse domestic loan interest rate 1% impulse domestic deposit interest rate 10% impulse on government spending Type of estimation - - - -1.07 - - OLS IRIPV - - - -1.48 - - 2SLS - - - -1.48 - - 3SLS - - - - 2.70 - OLS IRSV - - - - 3.37 - 2SLS - - - - 2.78 - 3SLS - 0.62 2.11 - - 0.60 OLS IRGRTV - 0.54 2.11 - - 0.57 2SLS - 0.72 1.72 - - 0.63 3SLS - 11.11 13.53 - - 0.83 OLS IRXV - 10.97 13.53 - - 0.80 2SLS - 11.18 12.71 - - 0.80 3SLS - 3.25 -0.12 - - 1.74 OLS IRMV - 3.29 -0.47 - - 1.77 2SLS - 384 -5.63 - - 1.32 3SLS 5.36 - 1.87 - - - OLS IRXFYV 5.14 - 1.79 - - - 2SLS 5.24 - 1.83 - - - 3SLS -0.11 017 1.16 - - 0.36 OLS IRMFYV -0.10 0.11 1.10 - - 0.33 2SLS -0.09 0.21 0.73 - - 0.30 3SLS -0.03 -0.26 -1.30 - - 0.13 OLS IREM -0.03 -0.28 -1.68 - - 0.16 2SLS -0.03 -0.35 -1.10 - - 0.09 3SLS Source: Obtained from results of the research According to table (5), a 10% increase in government expenditures, after trading off internal and external imbalances, has an effect on the unofficial exchange rate about 0.09% to 0.16% (in case ceteris paribus). In other words, under general equilibrium as long as the government expenditures shocks are offset through external sector imbalances, it will not have an involuntary effect on the unofficial exchange rate. The impact of other policy variable shocks on unofficial exchange rates such as interest rates on domestic deposits, interest rates on domestic loans, effective exchange rates on exported-imported goods and services, expored- imported factor income from abroad, foreign prices, and foreign interest rates can be similarly analyzed. 5. Concluding Remarks The main purpose of this paper is to provide a theoretical framework for examination of the relationship between internal and external imbalances and the unofficial exchange rate. Based on the extension of the trilemma theory in general equilibrium framework, the trade-off between the real sector imbalances of the economy reduces the impact of internal and external imbalances on the unofficial exchange rate. In other words, this paper shows that the policy variables affecting the real economy components so that internal and external imbalances have moderated each other, it is then possible to avoid high volatili ty of the Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 43 exchange rate as one of the key elements of the trilemma theory. Trading off financial imbalances in the private, public and foreign sectors of the economy enables the equilibrium of the real sector to be maintained without involuntary shock (with other factors remaining constant), and the unofficial exchange rate is less volatile. Failure to clear internal and external imbalances in the real sector of the economy, the imbalances in the form of changes in the net domestic and foreign assets of the banking system is transferred to the unofficial exchange rate. The use of simultaneous equations for time series data of macroeconomic variables of Iran proves the above theory so that any impulse in the policy variables confirms the effect on the exchange rate. In other words, a 10 percent increase in government expenditures, will increase the unofficial exchange rate fluctuation by 0.9-0.16 percent. Any impulse is driven by other policy variables, such as rising domestic interest rates (which lead to increased national savings or reduced demand for real balances) or increasing tax revenues or affecting the external sector, then the imbalances in the real sector of the economy can be moderated by the increase in government expenditures. The issue of trade-off between the imbalances can be analyzed in the form of different impulses in policy variables and the response of real-sector endogenous variables in different scenarios. Simply put, the purpose of this article is to illustrate the important point that, as long as macroeconomic policy-making has a neutral effect on the financial markets of the private sector, the government sector and the external sector, it does not cause real and currency fluctuations. From a theoretical point of view, this can be seen as a new approach to the development of the trilemma theory that has a realistic and practical view of equilibrium in the real, monetary and foreign exchange rate sectors. The econometric study of the empirical application in this paper suggests a strong confirmation of the recent expression. References Aizenman, J., & Sengupta, R. (2013). Financial Trilemma in C hina and a Comparative Analysis with I ndia. Pacific Economic Review, 18(2), 123-146. Aizenman, J., & Ito, H. (2014). Living with the trilemma constraint: Relative trilemma policy divergence, crises, and output losses for developing countries. Journal of International Money and Finance, 49, 28-51. Allahyarifard, M., Falahi, M. A., Karimzadeh, M., & Najimeydani, A. A. (2019). The Internal and External Disequilibrium Relationship in Economy and Extending of Monetary Trilemma (The Case Study of Iran Economy). Econometric Modelling, 4(2), 77-108. Baltagi, B. H. (2008). Econometrics. 4th ed., Berlin, Springer. Bidabad, B. (2005). Macroeconometric Model of Iran. 5th ed., Tehran: Monetary & Banking Research Institute of Iran. Fleming, J.M. (1962). Domestic Financial Policies Under Fixed and Under Floating Exchange Rates. IMF Econ Rev, 9(3), 369–380. Frankel, J. A. (1983). Monetary and portfolio-balance models of exchange rate determination, in economic interdependence and flexible exchange rates. JS Bhandari and BH Putnam (eds.), MIT Press, Cambridge MA. Greene, W. H. (2002). Econometric Analysis, 5th Edition. Prentice Hall: Upper Saddle River, NJ. Gujarati, D. (2004). Basic Econometrics. Translated by Abrishami, Hamid, Tehran University Publications. Krugman, P. (1979). A Model of Balance-of-Payments Crises, Journal of Money, Credit and Banking, 11(3), 311-325. Krugman, P. R. (1991). Target Zones and Exchange Rate Dynamics, The Quarterly Journal of Economics, 106(3), 669-682. Magas, I. (2018). Financial Adjustment in Small, Open Economies in Light of the “Impossible Trinity” Trilemma. Financial and Economic Review, 17(1), 5–33. Mankiew, N. G. (2016). Macroeconomics, 9th Edition, NY, Worth Publishers. Mundell, R. 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(4), 475-485. Maurice, O. (2015). Trilemmas and Tradeoffs: Livinga with Financial Globalization, in Global Liquidity, Spillovers to Emerging Markets and Policy Responses, Vol. 20, eds. by C. Raddatz, D. Saravia, and J. Ventura of Central Banking, Analysis, and Economic Policies Book Series(Central Bank of Chile), Chapter 2, pp. 13–78. Mussa, M. (1976). The Exchange Rate, the Balance of Payments and Monetary and Fiscal Policy under a Regime of Controlled Floating, The Scandinavian Journal of Economics, 78(2), 229-248, Proceedings of a Conference on Flexible Exchange Rates and Stabilization Policy. Perman, R. (1991). Cointegration: An Introduction to the Literature, Journal of Economic Studies, 18(3), 3-30. Persons, W. M. (2012). Fisher's the Purchasing Power of Money, Publications of the American Statistical Association, 12(96), 818-829. Pilbeam, Keith. (1998). International Finance. City University, London, Macmillan Business. Pilbeam, K. (1998). International Finance. City University, London, Macmillan Business. Rey, H. (2016). International Channels of Transmission of Monetary Policy and the Mundellian Trilemma. IMF Economic Review, 64(1). Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 44 System of National Accounts. (2008). Prepared under the auspices of the Inter-Secretariat Working Group on National Accounts, 1, 456-458. Appendices 0 10,000 20,000 30,000 40,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IREM (Baseline) IREM 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRGDPM (Baseline) IRGDP M 0 4,000,000 8,000,000 12,000,000 16,000,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRGDPMV (Baseline) IRGDP MV 0 200,000 400,000 600,000 800,000 1,000,000 1,200,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRGRTV (Baseline) IRGRTV -1,000,000 0 1,000,000 2,000,000 3,000,000 4,000,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRIPV (Baseline) IRIP V 0 50,000 100,000 150,000 200,000 250,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRMD (Baseline) IRMD 0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRMDCIFP (Baseline) IRMDCIFP 0 10,000 20,000 30,000 40,000 50,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRMFYV (Baseline) IRMFYV 0 1,000,000 2,000,000 3,000,000 4,000,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRMV (Baseline) IRMV 0 1 2 3 4 5 6 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRPGDPM (Baseline) IRP GDP M -2,000,000 0 2,000,000 4,000,000 6,000,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRSV (Baseline) IRSV 0 40,000 80,000 120,000 160,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRXD (Baseline) IRXD 0 50,000 100,000 150,000 200,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRXDFCPI (Baseline) IRXDFCP I -20,000 0 20,000 40,000 60,000 80,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRXFYV (Baseline) IRXFYV 0 1,000,000 2,000,000 3,000,000 4,000,000 40 45 50 55 60 65 70 75 80 85 90 95 Actuals IRXV (Baseline) IRXV OLS-Simulation System: SYS3SLSH Estimation Method: Three-Stage Least Squares Date: 05/26/19 Time: 14:35 Sample: 1339 1395 Included observations: 57 Total system (unbalanced) observations 451 Linear estimation after one-step weighting matrix Coefficient Std. Error t-Statistic Prob. C(1) 257.4782 78.53644 3.278456 0.0011 C(2) 1.016612 0.006980 145.6551 0.0000 C(3) 0.000946 0.000237 4.000208 0.0001 C(4) -13.79725 6.597782 -2.091194 0.0371 C(10) 6502.746 7431.449 0.875031 0.3821 C(11) 1.073701 0.015968 67.24165 0.0000 Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 45 C(12) 0.693183 0.129148 5.367339 0.0000 C(13) -16667.99 6654.533 -2.504757 0.0126 C(14) -518010.6 73536.49 -7.044265 0.0000 C(20) -247194.3 35944.07 -6.877193 0.0000 C(21) 38446.01 5208.951 7.380759 0.0000 C(22) 0.382802 0.014729 25.98960 0.0000 C(30) -893.5623 2254.571 -0.396334 0.6921 C(31) 1.159795 0.021133 54.88123 0.0000 C(32) 0.036973 0.007056 5.239906 0.0000 C(33) -95011.44 15496.55 -6.131134 0.0000 C(34) -53945.09 15634.96 -3.450286 0.0006 C(40) -1752.126 454.1764 -3.857810 0.0001 C(41) 1.008546 0.008766 115.0503 0.0000 C(42) 0.840617 0.033307 25.23851 0.0000 C(43) 0.218115 0.093692 2.327999 0.0204 C(44) 0.030519 0.006828 4.469704 0.0000 C(50) 84002.14 20788.60 4.040778 0.0001 C(51) 0.967116 0.061199 15.80282 0.0000 C(52) -10.65886 2.439674 -4.368968 0.0000 C(53) 1.121065 0.257734 4.349702 0.0000 C(60) -287.3049 238.2357 -1.205969 0.2285 C(61) 0.949651 0.028643 33.15526 0.0000 C(62) -0.243415 0.053905 -4.515581 0.0000 C(63) 28374.06 1249.203 22.71372 0.0000 C(64) -15191.66 1239.338 -12.25789 0.0000 C(65) -22132.75 1743.549 -12.69408 0.0000 C(66) 17320.84 1366.833 12.67224 0.0000 C(70) 4.929394 1.323247 3.725226 0.0002 C(71) 1.067234 0.041424 25.76360 0.0000 C(72) 0.006378 0.001278 4.990199 0.0000 C(73) 0.001141 0.000485 2.352898 0.0191 Determinant residual covariance 1.45E-06 Equation: (IREM)/RESIDIREM^2= (C(1)+C(2)*IREM(-1)+ C(3)*D(IRIPV- IRSV +IRGV-IRGRTV)+ C(4)*D(IRXV/IREEX-IRMV/IREEM+ IRXFYV/IREEXFY -IRMFYV/IREEMFY))/RESIDIREM^2 Instruments: C IREM(-2) (IRIPV-IRSV+IRGV-IRGRTV) (IRXV/IREEX- IRMV /IREEM+ IRXFYV/IREEXFY-IRMFYV/IREEMFY) IRIRD IRIRL Observations: 56 R-squared 0.999889 Mean dependent var 0.060141 Adjusted R-squared 0.999882 S.D. dependent var 0.183909 S.E. of regression 0.001996 Sum squared resid 0.000207 Durbin-Watson stat 1.950318 Equation: IRIPV = C(10) + C(11)*IRIPV(-1) + C(12)*IROLPV+ C(13)* (IRIRL -IRIRL(-1)) + C(14)*D92 Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 46 Instruments: C IRIPV(-2) IROLPV(-1) (IRIRL-IRIRL(-1)) D92 Observations: 56 R-squared 0.997424 Mean dependent var 481023.0 Adjusted R-squared 0.997222 S.D. dependent var 977225.2 S.E. of regression 51503.05 Sum squared resid 1.35E+1 1 Durbin-Watson stat 2.124041 Equation: IRSV/RESIDIRSV^2 = (C(20) + C(21)*IRIRD + C(22)*IRGDPMV( -1))/RESIDIRSV^2 Instruments: C IRIRD(-1) (IRGDPMV) Observations: 57 R-squared 0.998510 Mean dependent var 0.000158 Adjusted R-squared 0.998455 S.D. dependent var 0.000581 S.E. of regression 2.28E-05 Sum squared resid 2.82E-08 Durbin-Watson stat 1.560718 Equation: (IRGRTV) = (C(30) + C(31) *IRGRTV(-1)+ C(32)*D(IRGDPMV))+ C(33)*D89+ C(34)*D91 Instruments: C IRGRTV(-2) (IRGDPMV-IRGDPMV(-1)) D89 D91 Observations: 56 R-squared 0.994865 Mean dependent var 97808.20 Adjusted R-squared 0.994462 S.D. dependent var 212364.7 S.E. of regression 15803.64 Sum squared resid 1.27E+1 0 Durbin-Watson stat 3.253614 Equation: (IRXDFCPI)/(RESIDIRXDFCPIF^2)= (C(40)+ C(41)*IRXDFCPI(-1) +C(42)*(IRXOILD/FCPI-IRXOILD(-1)/FCPI(-1))+ C(43)*IREEX*FCPI /IRPGDPM+ C(44)* D(IRGDPM))/(RESIDIRXDFCPIF^2) Instruments: C IRXDFCPI(-2) IRXOILD/FCPI IREEX(-1)*FCPI(- 1)/IRPGDPM( -1) (IRGDPM)/IRGDPM(-1) IREEXFY IREEMFY Observations: 56 R-squared 1.000000 Mean dependent var 1.273640 Adjusted R-squared 1.000000 S.D. dependent var 8.750115 S.E. of regression 0.001857 Sum squared resid 0.000176 Durbin-Watson stat 1.871593 Equation: (IRMDCIFP)/(RESIDIRMDCIFPF^2)=(C(50)+C(51)*IRMDCIFP(-1) +C(52)*(IREEM*FCPI/IRPGDPM)+C(53)*D(IRXD+IRXFYV/IREEXFY -IRMFYV/IREEMFY))/(RESIDIRMDCIFPF^2) Instruments: C IREEM(-1) FCPI(-1) IRPGDPM(-1) (IRXD+IRXFYV/IREEXFY -IRMFYV/IREEMFY) Observations: 57 R-squared 0.998256 Mean dependent var 0.000613 Adjusted R-squared 0.998158 S.D. dependent var 0.001492 S.E. of regression 6.41E-05 Sum squared resid 2.17E-07 Durbin-Watson stat 2.095896 Equation: IRXFYV =C(60)+ C(61)*IRXFYV (-1)+C(62)* (IRIRD- FIR*IREEXFY) Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 47 + C(63)*D91+C(64)*D88 + C(65)*D93 + C(66)*D92 Instruments: IRIRD(-1) FIR(-1) IREEXFY(-1) D91 D88 D93 D92 C Observations: 57 R-squared 0.995999 Mean dependent var 9075.868 Adjusted R-squared 0.995519 S.D. dependent var 18647.67 S.E. of regression 1248.256 Sum squared resid 7790720 0 Durbin-Watson stat 1.553226 Equation: (IRMFYV)/RESID07^2= (C(70) + C(71)*IRMFYV(-1) + C(72) *(IRIRD - FIR*IREEMFY)+ C(73)*D(IRGDPMV))/RESID07^2 Instruments: C IRMFYV(-2)/RESID07^2 (IRIRD(-1) - FIR(-1)*IREEMFY(-1)) /RESID07^2 (IRGDPMV)/RESID07^2 Observations: 56 R-squared 0.999999 Mean dependent var 72.69083 Adjusted R-squared 0.999999 S.D. dependent var 389.7440 S.E. of regression 0.431112 Sum squared resid 9.664583 Durbin-Watson stat 1.749032 Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 48 -2 -1 0 1 2 40 45 50 55 60 65 70 75 80 85 90 95 D(IREM)/RESIDIREM Residuals -300,000 -200,000 -100,000 0 100,000 200,000 300,000 40 45 50 55 60 65 70 75 80 85 90 95 IRIPV Residuals -8 -6 -4 -2 0 2 4 40 45 50 55 60 65 70 75 80 85 90 95 IRSV/RESID03 Residuals -120,000 -80,000 -40,000 0 40,000 80,000 40 45 50 55 60 65 70 75 80 85 90 95 IRGRT V Residuals -.015 -.010 -.005 .000 .005 .010 40 45 50 55 60 65 70 75 80 85 90 95 IRXDFCPI/RESID05^2 Residuals -400,000 -200,000 0 200,000 400,000 600,000 40 45 50 55 60 65 70 75 80 85 90 95 IRMDCIFP Residuals -4,000 -2,000 0 2,000 4,000 6,000 8,000 40 45 50 55 60 65 70 75 80 85 90 95 IRXFYV Residuals -3 -2 -1 0 1 2 3 4 40 45 50 55 60 65 70 75 80 85 90 95 IRMFYV/RESID07^2 Residuals 3SLS Structural Equation residual correlation table RESID04 RESID06 RESID08 RESID09 RESID10 RESID11 RESID12 RESID13 RESID04 1.00 0.00 -0.17 0.05 -0.18 -0.03 -0.16 -0.01 RESID06 0.00 1.00 -0.28 -0.37 0.05 -0.18 0.07 0.04 RESID08 -0.17 -0.28 1.00 0.07 0.25 -0.03 0.09 -0.11 RESID09 0.05 -0.37 0.07 1.00 0.08 -0.06 -0.24 0.00 RESID10 -0.18 0.05 0.25 0.08 1.00 -0.05 0.29 0.17 RESID11 -0.03 -0.18 -0.03 -0.06 -0.05 1.00 -0.01 0.02 RESID12 -0.16 0.07 0.09 -0.24 0.29 -0.01 1.00 -0.01 RESID13 -0.01 0.04 -0.11 0.00 0.17 0.02 -0.01 1.00 Copyright © CC-BY-NC 2020, CRIBFB | AMFBR www.cribfb.com/journal/index.php/amfbr American Finance & Banking Review Vol. 5, No. 1; 2020 49 0 5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 40 45 50 55 60 65 70 75 80 85 90 95 IREM IREM (Baseline) IREM 3SLS Copyrights Copyright for this article is retained by the author(s), with first publication rights granted to the journal. This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/). 0 10,000 20,000 30,000 40,000 1340 1350 1360 1370 1380 1390 Actuals IREM (Baseline) IREM 0 500 ,000 1 ,000 ,00 0 1 ,500 ,00 0 2 ,000 ,00 0 2 ,500 ,00 0 3 ,000 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRGDPM (Baseline) IRGDPM 0 2 ,500 ,00 0 5 ,000 ,00 0 7 ,500 ,00 0 10,000,0 00 12,500,0 00 15,000,0 00 1340 1350 1360 1370 1380 1390 Actuals IRGDPMV (Baseline) IRGDPMV 0 200 ,000 400 ,000 600 ,000 800 ,000 1 ,000 ,00 0 1 ,200 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRGRTV (Baseline) IRGRTV -1 ,000 ,00 0 0 1 ,000 ,00 0 2 ,000 ,00 0 3 ,000 ,00 0 4 ,000 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRIPV (Baseline) IRIPV -200 ,000 -100 ,000 0 100 ,000 200 ,000 300 ,000 400 ,000 1340 1350 1360 1370 1380 1390 Actuals IRMD (Baseline) IRMD -500 ,000 0 500 ,000 1 ,000 ,00 0 1 ,500 ,00 0 2 ,000 ,00 0 2 ,500 ,00 0 3 ,000 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRMDCIFP (Baseline) IRMDCIFP 0 10,000 20,000 30,000 40,000 1340 1350 1360 1370 1380 1390 Actuals IRMFYV (Baseline) IRMFYV -4 ,000 ,00 0 -2 ,000 ,00 0 0 2 ,000 ,00 0 4 ,000 ,00 0 6 ,000 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRMV (Baseline) IRMV 0 2 4 6 8 1340 1350 1360 1370 1380 1390 Actuals IRPGDPM (Baseline) IRPGDPM -1 ,000 ,00 0 0 1 ,000 ,00 0 2 ,000 ,00 0 3 ,000 ,00 0 4 ,000 ,00 0 5 ,000 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRSV (Baseline) IRSV 0 40,000 80,000 120 ,000 160 ,000 1340 1350 1360 1370 1380 1390 Actuals IRXD (Baseline) IRXD 0 50,000 100 ,000 150 ,000 200 ,000 1340 1350 1360 1370 1380 1390 Actuals IRXDFCPI (Baseline) IRXDFCPI -20 ,000 0 20,000 40,000 60,000 80,000 1340 1350 1360 1370 1380 1390 Actuals IRXFYV (Baseline) IRXFYV 0 1 ,000 ,00 0 2 ,000 ,00 0 3 ,000 ,00 0 4 ,000 ,00 0 1340 1350 1360 1370 1380 1390 Actuals IRXV (Baseline) IRXV 3SLS Simulation