Capital account liberalization in morocco: is it compatible with fixed or flexible exchange rate regime? Elhadj Ezzahid and Brahim Maouhoub Chapitre du live Overture, productivité et croissance économique au Maroc , Édité par Chatri Abdellatif, Publié par Laboratoire d’Economie Appliquée (Mohammed V Univ.) & Policy Center for the New South, ISBN (WEB) : 978-9920-37-593-1 Citer ce document : Ezzahid, E. & Maouhoub, B. (2019). Capital account liberalization in morocco: is it compatible with fixed or flexible exchange rate regime?. In A. Chatri (éd). Ouverture, productivité et croissance économique au Maroc. Laboratoire d’Économie Appliquée & Policy Center for the New South. Rabat Lien vers l'article : http://www.labeamse.com/2019/04/OPCM12.html Copyright © 2019 Laboratoire d’Économie Appliquée, Policy Center for the New South & CNRST. Tous les droits sont réservés. CHAPITRE 12 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO : IS IT COMPATIBLE WITH FIXED OR FLEXIBLE EXCHANGE RATE REGIME ? Elhadj Ezzahid, Brahim Maouhoub Laboratoire d’Économie Appliquée, Université Mohammed V de Rabat E-mail de correspondance : ezzahidelhadj@yahoo.fr Abstract : This paper examines the opportunity of exchange rate regime flexibilization in Mo- rocco under the policy of capital account liberalization. Basing on our findings in E ZZAHID et M AOUHOUB (2014), we develop a new theoretical game model with four economic agents : monetary authorities, government, foreign firms and domestic firms. We explore the optimal ex- change rate regime for Morocco under new conditions such as the presence of a compensation fund’s effect, restrictions on capital outflows, etc. Starting with a first simulation based on ac- tual economic parameters, the results show that losses under a flexible exchange rate regime are lower than losses under a fixed exchange rate regime. Varying different parameters allows disco- vering the ‘appropriate level’ from which monetary authorities should move toward the flexible exchange rate. Key-words : Capital account liberalization, Exchange rate regime flexibilization, compensation fund, Real exchange rate, and game theory. 12.1 Introduction The financial liberalization theory predicts that capital account openness allows develo- ping countries to receive international capital flows and to boost investment, and therefore contribute to more efficient allocation of resources and to more financial development. Ouverture, productivité et croissance économique au Maroc, Éd. Chatri Abdellatif. 203 Copyright c 2019 Laboratoire d’Économie Appliquée & Policy Center for the New South. 204 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO The confrontation of theory predictions to developing countries’ experiences, renew the debate on capital account liberalization effects. Furthermore, to take advantages from ca- pital account liberalization, developing countries adopt various approaches and strategies appropriate to their specific circumstances. Capital account liberalization is not without risks. International capital flows can bring with them their own problems, such as real exchange rate appreciation and increase of in- flation under a fixed exchange rate (C ORBO & H ERNANDEZ, 1996) The experiences of Latin America and East Asia give us many lessons about the risk associated with capi- tal account liberalization under a fixed exchange rate. The important one is that preparing exchange rate flexibilization in parallel with capital account liberalization is required to ensure exchange rate adjustments. Furthermore, referring to the impossible trinity, it’s im- possible for an economy to attain simultaneously fixed exchange rate regime, free capital account and autonomy of monetary policy M UNDELL (1963). It is true that the fixed exchange rate regime ensures economic confidence and policy credibility, stabilizes nominal exchange rate and avoids volatility risks, but, it can result in a misallocation of resources and then in a low economic growth. The explanation is that in countries with fixed exchange rate regime and higher investment, the productivity grows more slowly than in countries with floating exchange rates, because nominal exchange rate is unable to be used as an adjustment mechanism (G HOSH, 1996). However, the empirical link between the choice of exchange rate regime and econo- mic growth is a controversial debate. Some studies show that less flexible exchange rate regimes are associated with slower growth in developing countries (see for example L EVY- Y EYATI et S TURZENEGGER (2003), and others studies indicate the absence of any robust relationship between the choice of exchange rate regime and economic growth in develo- ping countries (see for example D E V ITA et K YAW (2011)). On one hand, the exchange rate flexibilization found its argument on the hypothesis of F RIEDMAN (1953) and M UNDELL (1961), arguing that flexible exchange rates act as a ’shock absorber’ in a small open economy. Accordingly, flexible exchange rate regime is favourable for developing countries because it allows the adjustment process that stabi- lizes the macroeconomic variables when negative external shocks hit the economy (H OFF - MANN , 2007). Furthermore, the choice of exchange rate regime for developing countries depends on their policy orientations and their economic settings rather than on the theory predictions, the empirical studies and the economic experiences. The International Monetary Fund -as part of his annual consultations for developing countries, among them Morocco- recommends more flexibilization of their exchange rate regime. In this paper, we assess Moroccan experience and the opportunity of Morocco to move toward more flexible exchange rate regime. The point that we seek to determine is the ‘appropriate level’, based on the specificities of the Moroccan economy- from which monetary authorities should move to more flexible exchange rate regime. In other words, we try to determine the optimal exchange rate regime under actual conditions. The paper proposes a new theoretical model adapted to the Moroccan context and able to take into account the presence of real exchange rate misalignment, inflation targeting, gradual capital account liberalization and Compensation Fund 1 Assessments. To do this, we organize the paper as follow. Section 2 presents the theoretical framework. Section 3 1. A Fund created by Moroccan government to subside elementary goods such as liquid petroleum, Butane gas, sugar and national soft wheat flour, in order to stabilize prices  THEORETICAL AND CONCEPTUAL FRAMEWORK 205 is reserved for the model developments. The parameters are estimated in Section 4. Mo- del simulations and results are discussed in Section 5. The last Section summarizes the concluding remarks. 12.2 Theoretical and conceptual framework The issue of the optimality of the exchange rate regime is highly debated by A IZEN - MAN (1994), C HIN, M ILLER et al. (1995), D EVEREUX et E NGEL (2001), etc. An optimal exchange rate regime corresponds to a regime that minimizes risks (currency instability, capital volatility, etc.) and maximizes benefits (policy credibility, economic confidence, economic agent planning, etc.) associated with a given exchange rate regime. Thus, an op- timal exchange rate regime may be a fixed or a flexible exchange regime, all depends on economic conditions and policy objectives. To model the economic agents’ behaviour, we use the game theory. The game is re- presented as a situation where monetary authorities make decisions and other economic agents react to these decisions (Figure 1). All economic agents are conscious that their benefits depend not only on their decisions, but also on the reactions of others. To make an optimal decision, monetary authorities have to define their objective function and the objective-functions of other economic agents. Figure 1 : Economic agents and different choices in the base model The base model considered in this work (See the Base Model in Appendix) is foun- ded on the framework proposed by AGÉNOR (1991, 1994) for developing countries. We consider an open small economy producing tradable and non-tradable goods, where eco- nomic agents are monetary authorities and domestic firms interact through an optimization game. Each agent tries to minimize his loss-function using his own instruments : choice of an exchange rate regime for monetary authorities (fixed of flexible) and price fixation for domestic firms (increase or decrease). Many extended models are simulated for developing and emerging countries. For example Z HANG (2001) explored the Chinese case, S AMI (2006) explored the Tunisian case, and E ZZAHID et M AOUHOUB (2014) explored the Moroccan case. The aim of these simula- tions is to determine from which economic level the choice of flexible exchange rate regime becomes an optimal choice for monetary authorities. 206 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO 12.3 Model constructions The base model doesn’t take into consideration the specificities of the Moroccan eco- nomy such as : the existence of compensation effect that reduce inflation and the govern- ment strategy to supress progressively the compensation fund, foreign firms and the in- ternational capital returns, domestic firms and the presence of some restrictions on capital outflows, etc. introducing these facts in the model may affect the results significantly. For this reason, we construct a model adapted to Moroccan economic conditions and founded on the theoretical framework presented above. Consider a model of a small and open economy (Figure 2) where the monetary authori- ties (MA) have two options : stay under a fixed exchange rate regime or move to a flexible exchange rate regime. The choice of exchange rate regime is required in a context of capital account liberalization. Thus, to make an optimal choice ; the monetary authorities must de- fine the objective-functions of all economic agents and determine the variables that impact his decisions : monetary authorities’ objective-function, government objective-function, domestic firms’ objective-function and foreign firms’ objective-function : Figure 2 : Model developed for Morocco, authors’ design Monetary authorities Decisions/instruments Choice 1: Fixed regime Choice 2: Flexible regime ∆𝐸𝑛𝑡 =0 ∆𝑬𝒏 ≠𝟎 𝑾𝑴𝑨 Objective -functions 𝑾𝑫𝑭 𝑾𝑭𝑭 𝑾𝑮 𝑷𝑵𝒕 𝑷𝑻𝒕 Decisions/instruments and 𝑲𝑶𝒕 and 𝑲𝒊𝒕 𝑫𝒕 Domestic Foreign Economic agents Government firms firms The model developed is based on the following hypothesis : Hypothesis 1 : price general level determination (Πt ). The economy is composed by two-sectors ; tradables sector and non-tradable sector. The price general level is supposed to be determined as follow : 1−δ Πt = ΠδN t .(ΠT t .EN t ) (Eq. 1.1) With ΠN t is price level of non-tradable goods, ΠT t is price level of tradable goods, EN t is nominal effective exchange rate and (1 − δ) is economic openness rate. In growth rate terms, the equation (eq.1.1) takes the following expression : Πt = δΠN t + (1 − δ) (ΠT t + ent ) (Eq. 1.2) We suppose that πt is the inflation rate hors compensation fund effect.  MODEL CONSTRUCTIONS 207 Hypothesis 2 : price level determination of tradables ΠT t and non-tradables ΠN t . The price level of tradables ΠT t is determined in international markets. We suppose that the price level of tradables ΠT t equalizes 2 the international price general level Π∗t : ΠT t = Π∗t (Eq. 1.2.1) In growth rate terms, the equation (eq. 1.2.1) gives the world inflation rate : Π∗t = ΠT t (Eq. 1.2.2) The price level of non-tradables ΠN t is domestically determined by the following me- chanism :  ρ Ert ΠN t = ◦ .Mtν (Eq. 2.1) Ert ◦ With Ert is equilibrium real effective exchange rate, Mt is money supply, ρ is elasticity of non-tradables price level to the real exchange rate misalignment and ν is elasticity of non-tradables price level to the money supply. The prices of non- tradable goods ΠN t are determined by two factors. The first is the deviation of the real exchange rate from its equilibrium level. Thus, a real depreciation increases the external competitiveness of the economy. As a consequence, we observe an increase of the tradable goods’ demand (i.e. increase in exports). Therefore, domestic firms shift resources from non-tradable goods sector to the tradable goods sector. This resources transfer lowers the production of non- tradable goods (decrease of supply), which leads to an increase of their prices. A real exchange rate appreciation produces exactly the opposite effects. In growth rate terms, the equation (eq. 2.1) takes the following expression : ΠN t = ρ (ert − e◦rt ) + νmt (Eq. 2.2) Hypothesis 3 : The compensation fund effect on prices (Dt ). The government reduces prices by the Compensation Fund Effect Dt . Thus, we suppose that the price level of tra- dables and non-tradables are reduced by the following mechanism : ΠN t ΠT t .EN t pN t = and pT t = (Eq. 3.1 and eq. 3.2) Dt a Dt 1−a With pN t is the subsided price level of non-tradables, pT t is the subsided price level of tradables, Dt is Compensation Fund effect, a is the proportion of compensation fund effect that reduces price level of non-tradables and 1 − a is the proportion of compensation fund effect that reduces price level of tradables. From (eq. 1.1), (eq.3.1) and (eq.3.2), we express the subsided general price level pt in function of subsided price level of non-tradables pN t and subsided Price level of tradables pT t as follow : 2. We suppose that the international price level may be expressed as follow : Π∗t = Π∗i N t .ΠT t 1−i with Π∗N t is the international price level of non-tradable goods, ΠT t is the international price level of tradable goods i.e. Price level of tradable goods and 1 − i is the international economic integration. Thus, when international economy is fully interacted : (1 − i) = 1, so ΠT t = Π∗t 208 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO  δ  1−δ Πt ΠN t ΠT t .EN t p= = . (Eq. 4.1) Dt aδ+(1−a)(1−δ) Dt a Dt 1−a δ 1−δ Or pt = [pN t ] .[pT t .EN t ] (Eq. 4.2) In growth rate terms, the equations (eq.4.1) and (eq.4.2) give the compensated inflation rate : Or pt = δ[ΠN t − a.dt ]+ (1 − δ) [ΠT t − (1 − a)dt + ent ] (Eq. 5.1) Or pt = δpN t + (1 − δ) (pT t + ent ) (Eq.5.2) = Πt − (aδ + (1 − a) (1 − δ)) dt We can write dt as follow : Πt − p t dt = aδ + (1 − a)(1 − δ) [δΠN t + (1 − δ) (ΠT t + ent )] − [δpN t + (1 − δ) (pT t + ent )] = (Eq. 6.1) aδ + (1 − a)(1 − δ) δ (1 − δ) = (πN t − pN t ) + [(πT t + ent ) − (pT t + ent )] aδ + (1 − a)(1 − δ) aδ + (1 − a)(1 − δ) The parameters b and c are introduced to measure the extent of compensation fund’s effect via the gap between non-subsided prices and subsided prices. As result, (eq. 6) be- comes : δb (1 − δ) c dt = pN t + (pT t + ent ) (Eq. 6.2) aδ + (1 − a)(1 − δ) aδ + (1 − a)(1 − δ) We consider the equation (eq. 6.2) as the government function reaction to non-tradable prices and tradable prices. Hypothesis 4 : Real effective exchange rate definition. In this model, we define real effective exchange rate as Gobb-Douglas function : Ent .PT t Ert = (Eq. 7.1) PN t The Ent is nominal effective exchange rate. In growth rate terms, the equation (eq. 7.1) takes the following expression : ert = ent + pT t − pN t (Eq. 7.2) Hypothesis 5 : capital account restrictions. We denote capital inflows by Kit and capital outflows by Kot . Moroccan monetary authorities impose restrictions on capital outflows  MODEL CONSTRUCTIONS 209 Kot to avoid any situation of capital flight. Nevertheless, monetary authorities don’t impose any restrictions on capital inflows Kit and they look for higher capital inflows. However, we suppose that capital inflows must not exceed domestic financial requirement in order to avoid any capital influx. We express the capital account 3 equation as follow : KAt = Kit ∗ Ent − Kot (Eq. 8.1) We suppose that restrictions on capital outflows Kot are expressed as follow : n Kot = Kit ∗ Ent (Eq. 8.2) With nR∗+ is a restriction on capital outflows. In other words, the amount of capital authorized to outflow is : 1 Kot = (Kit ∗ Ent ) n (Eq. 8.3) 1 This means that the amount of capital outflows Kot cannot exceed (Kit ∗ Ent ) n . Hypothesis 6 : domestic firms, foreign firms and investment decision process. We sup- pose that domestic firms invest their capital into foreign markets in function of the capital real return Rot of capital invested abroad :   Kot Ent .Γ∗t .It∗ Rot = (Eq. 9.1) Π∗t With (1−Γ∗t ) is international capital depreciation rate, It∗ is international real interest rate, Π∗t is international general price level. In growth rate terms, the equation (9.1) takes the following expression : rot = (Kot − EN t ) + Γ∗t + i∗t − Π∗t (Eq. 9.2) We suppose that domestic firms’ investment in foreign markets is elastic to capital real return as follow : φ Kot = (Rot ) (Eq. 9.3) With φ is the elasticity of capital outflows Kot to real return Rot . In growth rate terms, the equation (eq. 9.3) takes the following expression : kot = φrot (Eq. 9.4) We suppose also that foreign firms invest a part of their capital Kit in Morocco ; these capital inflows generate a real return Rit as follow : (Kit .Ent ) .Γt .it Rit = (Eq. 10.1) Pt 3. Capital outflows Kot are explained in Moroccan dirhams and capital inflows are Kit are explained in foreign currency. 210 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO With 1 − Γt is domestic depreciation rate, it is domestic interest rate and Pt is compen- sated general price level. In growth rate terms, the equation (eq. 10.1) takes the following expression : rit = (kit + ent ) + γt + it − pt (Eq. 10.2) Thus, we suppose that their investment in Morocco is elastic to capital real return as follow : ϕ Kit = (Rit ) (Eq. 10.3) With ϕ is the elasticity of capital inflows Kit to real return Rit . In growth rate terms, the equation (eq. 10.3) g the following expression : kit = ϕrit (Eq. 10.4) Under this hypothesis, Moroccan monetary authorities define their objective-function and take into consideration the objective-functions of all economic agents. Monetary authorities’ objective-function. We denote the monetary authorities’ objective- function WtM A and we define their objectives as follow : The first objective is to avoid misalignment of real effective exchange rate Ert from its ◦ ◦ equilibrium level Ert . As a result, Ert > Ert signifies that real effective exchange rate ◦ is depreciated and Ert < Ert signifies that real effective exchange rate is appreciated. Thus, monetary authorities try to conserve the equality E rt ◦ Ert = 1 or in growth rate terms ◦ ert − ert = 0. This means for the Central Bank minimizing the quantity 4 2 1 (ert − e◦rt ) (Exp. A) 2 The second objective is to stabilize the inflation rate pt with respect to target inflation rate p◦t , this require to conserve the equality pt − p◦t = 0. Therefore, the Central Bank minimizes the following quantity to avoid any situation of hyper-inflation pt > p◦t or deflation pt < p◦t : 2 1 (pt − p◦t ) (Exp. B) 2 The third objective for monetary authorities is to preserve the equality KitK∗E n nt = 1 or ot in growth rate the equality (kit + ent ) − nkot = 0. In other words, monetary authorities have to minimize the following quantity : 1 2 [(kit + ent ) − nkot ] (Exp. C) 2 From (Exp. A), (Exp. B) and (Exp. C) we can construct the monetary authorities’ objective-function : 4. We use the square to explain that any deviation of real effective exchange rate from its equilibrium is transformed to a loss for the Central Bank. We add also ? to eliminate square after derivation.  MODEL CONSTRUCTIONS 211 1 2 1 2 1 2 WtM A = α(ert − e◦rt ) + .β(pt − p◦t ) + ϑ[(kit + ent ) − nkot ] (Eq. 10) 2 2 2 With α is the weight granted by monetary authorities to real exchange rate misalign- ment, β is the weight granted by monetary authorities to inflation and ϑ is weight granted by monetary authorities to capital account restrictions(ϑ = 1 − α − β). Monetary authori- ties have to evaluate their objective-function WtM A under both fixed and flexible exchange rate regimes. Government objective-function. The Government (fiscal authorities) try to reduce the budget deficit by reducing the Compensation Fund. Consequently, the Compensation Fund Effect Dt is reduced to equalize Target Compensation Fund Effect Dt◦ . Thus, the govern- Dt ◦ ◦ ment tries to preserve the equality D ◦ = 1 or in growth rate (dt − dt = 0). When dt > dt , t so more public expenditures of compensation fund will increase budget deficit and when dt < d◦t , so less public expenditures of compensation fund will damage the purchasing power and disturb social harmony. As result, government have to minimize the following objective-function : 1 2 WtG = (dt − d◦t ) (Eq. 12) 2 Comparing budget deficit (Graphic 1) including compensation expenditures (green line) with budget deficit excluding compensation expenditures (red line), we remark clearly that the origin of Moroccan budget deficit is the compensation fund. Domestic firms’ objective-function. Domestic firms take into consideration the legal ca- ◦ pital to invest abroad Kot . They try to maximize their real return and preserve the equality Kot ◦ ◦ Kot = 1 or in growth rate kot − kot = 0. Thus, their objective is to minimize the following quantity : 1 ◦ 2 [kot − kot ] (Exp. A.1) 2 Moreover, domestic firms produce tradable and non-tradable goods and but command only non-tradables prices ΠN t . The welfare of domestic firms is also defined in terms of relative prices of non-tradable goods. That means domestic firms react by changing ΠN t . Indeed, domestic firms attempt to protect themselves by adjusting continuously the price of non-tradable goods ΠN t to the expected level Π◦tN . Consequently, they try to preserve the quality Π Nt Π◦ ◦ = 1 or in growth rate πN t − πN t = 0 and they have to minimize the Nt following quantity : 1 ◦ 2 (πN t − πN t) (Exp. B.1) 2 From (Exp. A.1) and (Exp. B.1), we the domestic firms’ objective-function becomes : 1 ◦ 2 1 ◦ 2 WtDF = µ (kot − kot ) + τ (πN t − πN t) (Eq. 13) 2 2 With µ is the weight granted by domestic firms to their investment in foreign markets and τ is the weight granted by domestic firms to non-tradables price level. 212 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO Foreign firms’ objective-function. Foreign firms take into consideration the domestic fi- ◦ nancial requirement Kit . As a result, they look for a maximum real return Rit under the constraint of Kit . They try to maximize their real return and preserve the equality K ◦ ◦ = 1 it Kit ◦ or in growth rate terms kit − kit = 0. Thus, their objective is to minimize the following objective-function : 1 ◦ 2 WtF F = (kit − kit ) (Eq. 14) 2 Exchange rate regime choice. The economic problem for monetary authorities is to adopt the adequate exchange rate regime that minimizes their objective-function. We have four objective-functions : 1 2 1 2 1 2 WtM A = α(ert − e◦rt ) + .β(pt − p◦t ) + ϑ[(kit + ent ) − nkot ] (Eq. 11) 2 2 2 1 2 WG = (dt − d◦t ) (Eq. 12) 2 1 ◦ 2 1 2 WtDF = µ (kot − kot ) + τ (ΠN t − Π◦N t ) (Eq. 13) 2 2 1 ◦ 2 WtF F = (kit − kit ) (Eq. 14) 2 Model resolution First, we replace variables with their expressions as follows : 1 2 WtM A = α[ent + ΠT t − (1 − a) dt − ΠN t + adt − e◦rt ] 2 1 2 + .β[δ(ΠN t − adt )+ (1 − δ) (ΠT t + ent − (1 − a) dt ) − p◦t ] 2 1 2 + ϑ [kit + ent − n kot ] 2  2 1 δb c (1 − δ) WtG = (ΠN t − adt ) + [ΠT t + ent − (1 − a) dt ] − d◦t 2 aδ + (1 − a) (1 − δ) aδ + (1 − a) (1 − δ) 1 ◦ 2 WtDF = µ [φ[(Kot − EN t ) + Γ∗t + i∗t − ΠT t ] − kot ] 2 1 ◦ 2 + τ [ρ (ent + ΠT t − (1 − a) dt − ΠN t + adt − e◦rt ) + νmt − πN t] 2 1 ◦ 2 WtF F = [ϕ[(kit + ent ) + Γt + it − δ(ΠN t − adt )− (1 − δ) (ΠT t − (1 − a) dt + ent )] − kit ] 2  ESTIMATION OF MODEL PARAMETERS 213 Second, at the equilibrium the model implies that all variables equal the equilibrium values dt = d◦t , kot = kot ◦ , ΠN t = Π◦N t and kit = kit ◦ . The model takes the following form under the fixed exchange regime (EN t = 0) :  2   WtM A = 12 α[ΠT t − (1 − a) dt − ΠN t + adt − e◦rt ]  2 + 12 .β[δ(ΠN t − adt )+ (1 − δ) (ΠT t − (1 − a) dt ) − p◦t ]      2 + 1 ϑ [kit − n kot ]   h 2    i2 c(1−δ) δb WtG = 21 aδ+(1−a)(1−δ) (ΠN t − adt ) + aδ+(1−a)(1−δ) [ΠT t − (1 − a) dt ] − dt   WtDF = 1 µ [φ[Kot + Γ∗t + i∗t − ΠT t ] − kot ]2 +     2  1 ◦ 2 + 2 τ [ρ (ΠT t − (1 − a) dt − ΠN t + adt − ert ) + νmt − πN t ]      2 WtF F = 12 [ϕ[kit + Γt + it − δ(ΠN t − adt )− (1 − δ) (ΠT t − (1 − a) dt )] − kit ]  The model the following form under flexible exchange regime (ent 6= 0) :    WtM A = 21 α[ent + ΠT t − (1 − a) dt − ΠN t + adt − e◦rt ]2 +  + 12 .β[δ(ΠN t − adt )+ (1 − δ) (ΠT t − (1 − a) dt + ent ) − p◦t ]2 +      + 12 ϑ [(kit + ent ) − n kot ]2      h i2 c(1−δ) WtG = 12 aδ+(1−a)(1−δ) δb (ΠN t − adt ) + aδ+(1−a)(1−δ) [ΠT t − (1 − a) dt + ent ] − dt   WtDF = 1 µ [φ[(Kot − EN t ) + Γ∗t + i∗t − ΠT t ] − kot ]2 +     2 + 21 τ [ρ (ent + ΠT t − (1 − a) dt − ΠN t + adt − e◦rt ) + νmt − πN t ]2       WtF F = 21 [ϕ[(kit + ent ) + Γt + it − δ(ΠN t − adt ) − (1 − δ) (ΠT t − (1 − a) dt + ent )] − kit ]2  The first order conditions resulting from the model minimization under the fixed ex- change rate regimeand under the flexible exchange rate regime are solved using Matlab. 12.4 Estimation of model parameters Monetary authorities and domestic firms make a trade-off between two or three objec- tives and then they grant a weight for each of them as follow : Monetary authorities and domestic firms’ preferences α, β, ϑ, µ and τ . According to Bank Al-Maghreb, the prior objective of monetary policy is the price stabilizing, which is considered as a determinant factor to ensure the investment climate and the economic growth, and to protect the domestic purchasing power 5 . As well, to improve the external economic competitiveness, monetary authorities avoid real exchange rate misalignment. The capital account is introduced to quantify the capital flows liberalization’s effect in terms of competitiveness and inflation. As Latin American and Asian experiences learnt us, capital inflows to developing coun- tries raise the inflation and appreciate the real exchange rate. The theoretical mechanism 5. Bank Al-Maghreb, Department of Communication, Note d’information No 1, July 2006. 214 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO is explained as follow : Capital inflows to developing countries raise the foreign exchange reserves and expand the money supply. As a result, the inflation raises and the real ex- change rate appreciate under a fixed exchange rate regime, which undermines the external competitiveness. Consequently, the first objective is defined as inflation targeting. The second objective is to improve the external competitiveness (to avoid real exchange rate misalignment.) .And the third is to liberalize the capital account progressively (to get a more opened and integrated economy). Thus, we suppose that the weight of inflation-targeting is twofold the weight of real misalignment : β= 2α and that the weight of real misalignment is twofold the weight of capital liberalization : α= 2ϑ :   α + β + ϑ= 1  β= 2α  α= 2ϑ  The resolution of the system gives the following values α = 2/7, β= 4/7 and ϑ= 1/7. In Morocco, there is a demand of domestic firms (Also of households that we supposed included in domestic firms’ objective-function) to invest in foreign markets. For Moroccan domestic firms, the objective of non-tradable prices’ adjustment is not certainly the ma- jor problem comparing to importance of investment in foreign markets. Consequently, we can suppose that the importance to invest in foreign markets is twofold the importance to adjusted non-tradable prices :   1 2 µ = 2 ∗ (τ ) = 2 ∗ = 3 3 ◦ Equilibrium real effective exchange rate Ert . Using the Stock Flow Approach and the Auto Regressive Distributed Lag (ARDL) methodology, the equilibrium real effective ex- change rate in growth rate terms is between -0.03 and 0.2 (Graphic 2). We select the mean ◦ value of equilibrium real effective exchange rate Ert = −0.01917 over the period 1999- 2011 and then we change it in the interval [-0.03 ; 0.0] to test its effect on monetary autho- rities’ objective-function. Target inflation rate p◦t . In fact, the target level of inflation rate doesn’t vary over time. Generally, central banks fixe the percentage of 2% as targeted inflation rate : p◦t = 0.02. Proportion of compensation fund effect that reduces prices of non-tradables a and prices of tradables 1 − a. If a = 1, this means that Compensation Fund Effect is totally dedicated to reduce prices of non-tradable goods. And vice-versa, if a = 0 means that compensation Fund effect is totally dedicated to reduce price of tradable goods. We select the value of a = 0.5 for first simulation and we will vary it from 0 to 1 to test its effect on the exchange rate regime’ choice. Money supply mt . Money supply is measured by the large aggregate used by Moroccan central Bank "M3". Money supply growth (Graphic 3) varies between 0.04 and 0.18 with a downward tendency. Thus, we select the mean of recent values mt = 0.0509 and we change it to test its effect on monetary authorities’ objective-function.  ESTIMATION OF MODEL PARAMETERS 215 Capital depreciation rates Γt and Γ∗t . The domestic capital depreciation rate, we retain the value used for African countries 6 that is 0.06. For international capital depreciation rate, we retain the value estimated for total capital stock of the Euro zone 7 as the interna- tional capital depreciation rate 0.046. Among raisons (nominal exchange rate, interest rate, etc.) for which domestic firms look for more investment in foreign markets is that domestic capital depreciation rates is higher than international capital depreciation. However, these parameters are stable and their growth rate equal to zero : Γt = 0 and Γ∗t = 0. Interest rates it and i∗t . Moroccan Central Bank decreases the interest rate periodically. The domestic interest rate (Graphic 4) is reduced from 7% (in 1996) to 2.25% (in 2016). We retain the mean annual growth rate observed for the period 2009-2016 : it = −5.37% for first simulation and we fluctuate it afterward. For international interest rate, we take the U.S. interest rate (Graphic 5) as the World interest rate i∗t and we retain the mean annual growth rate of the same period above 2009-2016 : i∗t = −6.88%. This reflects that the world interest rate decreases rapidly than domestic interest rate. Capital account restrictions n. Restrictions on capital outflows are expressed as follow : (kit + ent ) kit n = = kot (kot − ent ) The value of elasticity n (Graphic 6) turns around 1.1 over the recent period of 2003- 2015. Elasticities of prices of non-tradable goods to real misalignment ρ and to money supply ν. The elasticities of the prices of non-tradable goods relative to the misalignment ρ and to domestic money supply ν are difficult to estimate. We retain the estimated values for the Tunisian 8 case that shows some similarities with respect to the Moroccan economy, thus we assume that ρ = 0.4 and ν = 0.6. The parameters b and c. The ratio πpN Nt t 1 = 1+b 1 may vary from 0 to 1. Thus, when 1+b →1 1 this means that pN t → πN t and b → 0 and when 1+b → 0 this means that πN t  pN t and (pT t +ent ) 1 1 b → ∞. Similarly, the ratio (π T t +ent ) = 1+c may vary from 0 to 1. Thus, when 1+c →1 1 this means that (πT t + ent ) → (pT t + ent ) and c → 0 and when 1+c → 0 this means that πT t + ent  pT t + ent and c → ∞. Because of the estimation difficulty to obtain these parameters, we start the simulation with the following assumed values : b = 31 and c = 14 . In order to test the parameters robustness, we vary their values in the interval of [0 ;3]. Economic openness δ : The degree of economic openness (1 − δ) is varying from 0 (fully closed) to 1 (fully open). For the Moroccan case, we can suggest that economic openness rate is between 50% and 80% (This corresponds to the sum of imports and exports divided by GDP). We start the first simulation with the value δ = 0.3 (Moroccan economy is opened at 70%). Elasticities of capital outflows φ and of capital outflows ϕ to real returns. The calculation of φ and ϕ is approximated as follow : ln(Total assets, millions of current US$) φ = ln(Primary income receipts, BoP, millions of current US$) 6. Rapport Économique sur l’Afrique 2010, élaboré par la Commission Economique pour l’Afrique, p. 134. 7. Monthly Bulletin of European Central Bank, p. 46 8. See S AMI (2006) 216 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO ln(Total liabilities, millions of current US$) ϕ= ln(Primary income payments, BoP, millions of current US$) (Source : Externatl wealth of nations) (Source :World bank data base) ϕ varies between 1.39 to 1.53 and turns around a mean of 1.45. φ varies between 1.50 and 2.74 and turns around a mean of 1.83 (Graphic 7). We remark that the value of φ tends to the value of ϕ from 1985 to 2011. We select the following values : ϕ= 1.45 and φ= 1.83 for first simulation. ϕ varies between 1.39 to 1.53 and turns around a mean of 1.45. φ varies between 1.50 and 2.74 and turns around a mean of 1.83 (Graphic 7). We remark that the value of φ tends to the value of ϕ from 1985 to 2011. We select the following values : ϕ= 1.45 and φ= 1.83 for first simulation. Price of tradable goods ΠT t . We consider that European union’ inflation rate is tradable goods’ inflation, because of the high dependence of Morocco to Europe. The inflation in EU (Graphic 8) is very low over the recent period of 2013-2015. We select the value of 1%, so Π∗t = ΠT t = 0.01. 12.5 Simulations and discussions The recapitulation of initiated and estimated values and interval of its variations are presented in Table 1. First, we perform the first simulation basing on the selected values to know which exchange rate regime is actually the optimal choice for monetary autho- rities. Second, we vary each parameter according to the interval of variation in order to compare monetary authorities’ objective-function under fixed and flexible exchange rate regime under the current conditions. Simulation based on selected values The simulation of monetary authorities’ objective-function under fixed exchange rate regime and under flexible exchange rate regime (Table 2) - based on the selected values- shows that monetary authorities’ losses under the fixed exchange rate (0.006855) are re- latively higher than losses under the flexible exchange rate (0.006280). The exchange rate flexibilization reduces monetary authorities’ losses by 9.16%. The desegregation of monetary authorities’ losses (Table 3) shows that under the fixed exchange rate regime 70.20% of monetary authorities’ losses results from real misalign- ment, 9.20% from inflation deviation and 20.60% from capital account disequilibrium. Under the flexible exchange rate regime 51.99% of monetary authorities’ losses becomes from real misalignment, 9.46% from inflation targeting and 38.55% from capital account. The monetary authorities’ objective function desegregation shows that the choice of the flexible exchange rate regime reduces losses relative to real misalignment by -25.94%, but, increases relatively the losses relative to inflation by +2.76% and consequently the losses relative to capital account by +87.16%. Accordingly, the exchange rate regime flexibilization allows monetary authorities to avoid the real exchange rate appreciation and then to preserve the external economic com- petitiveness. The mechanism is that under the flexible exchange rate regime, the exchange  SIMULATIONS AND DISCUSSIONS 217 rate adjusts in nominal terms rather than in real terms. The exchange rate regime flexibiliza- tion affects moderately the inflation because of the compensation fund effect adjustment. However, the nominal exchange rate adjustment is not without risks. As determined in the model under the flexible exchange rate regime, the nominal exchange rate 9 grows by ent =1.26%. It is interpreted as a nominal depreciation. Consequently, under the unchanged capital account restrictions (n = 1.1), capital in- flows grow rapidly (Kit = 17.37% under fixed exchange rate regime and Kit = 20.97% under flexible exchange rate regime) than capital outflows (Kot = 3.01% under fixed ex- ◦ change rate regime and Kot = 3.48% under flexible exchange rate regime). This is why the losses relative to capital account under the flexible exchange rate becomes higher. Varying capital account restrictions We vary the parameter n (Table 4) to show how more capital account liberalization re- duces monetary authorities’ losses under flexible exchange rate regime and raises them un- der fixed exchange rate regime. Thus, the flexible exchange rate choice remains the optimal choice as long as the parameter of capital account restrictions is between 0.90 < n < 1.25. The minimal monetary authorities’ losses (0.006173) is reached under the flexible ex- change rate regime when n = 1.04 (Graphic 9). Consequently, allowing more capital to outflow with respect to capital inflows (decreasing n from 1.10 to 1.04) is compatible with the choice of the flexible exchange rate regime. In cases of massive capital flight (n≤0.90) and massive capital influx (n≥1.25), the fixed exchange rate becomes an optimal choice because of high monetary authorities’ losses. The desegregation of monetary authorities’ objective function under flexible ex- change rate regime shows the losses’ structure (Graphic 10). The monetary authorities’ losses under flexible exchange rate regime results from capital account disequilibrium in case of capital flight (high capital account deficit), and from real misalignment in case of capital influx (high real exchange rate appreciation). Varying monetary authorities’ preferences The parameters α, β and ϑ play a crucial role in the exchange rate regime choice. The results change by classing the order of importance of parameters according to six possi- bilities (Table 5). If monetary authorities consider the real exchange rate misalignment as the most (least) important objective or when they consider the capital account equilibrium as the least (most) important objective, the flexible (fixed) exchange rate come to be the optimal exchange rate regime. 9. Is the equilibrium nominal exchange rate resulted from model resolution and it is expressed in function of parameters. 218 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO Domestic firms’ preferences Varying domestic firms’ preferences has no effect on monetary authorities’ objective- function under both fixed and flexible exchange rate. Regardless the order of importance of domestic firms’ preferences, the flexible exchange rate remains the optimal choice of monetary authorities. Economic openness rate Increasing the economic openness rate from 50% to 80% decreases the monetary au- thorities’ losses under both fixed and flexible exchange rate regime ; however, the flexible exchange rate regime remains the optimal choice (Table 6). Compensation fund repartition The compensation fund repartition between tradables and non tradables (Table 7) has a significant effect on monetary authorities’ choices. As long as a < 0.9, the flexible exchange rate regime remains the optimal choice for Moroccan monetary authorities. At the repartition a = 0.9, monetary authorities’ losses under the fixed exchange rate regime equal to losses under the flexible exchange rate. Exceeding the limit a > 0.9, the price of non tradables decreases consequently, and because the price of tradables is stable, the inflation decreases with respect the target level (2%), which raises monetary authorities’ losses under the flexible exchange rate. As a result the fixed exchange rate regime becomes the optimal choice. The minimal monetary authorities’ losses are reached when a = 0.8 under the flexible exchange rate regime choice (Graphic 11). Consequently, the optimal repartition of compensation fund under the flexible exchange rate regime is 80% for non- tradable prices and 20% for tradable prices. Increasing b (c) decreases monetary authorities’ losses under both fixed and flexible exchange rate regime (Tables 8 and 9). However, if b ≥ 34 (c  3), the fixed exchange rate regime becomes the optimal choice. In other words, when non-tradable prices (tradable prices) are reduced by more than 50% (by more than 75%), the inflation declines to a lower rate as regards the target level, which causes more monetary authorities’ losses under the flexible exchange rate compared to losses under the fixed exchange rate. The desegregation of monetary authorities’ objective function shows that increasing pa- rameters b and c reduces monetary authorities’ losses relative to real misalignment and to capital account, but raises losses relative to the inflation under the flexible exchange rate regime (Graphics 12 and 13). Accordingly, under the fixed exchange rate regime, increa- sing parameters b and c raises the compensation funds effect which reduces non-tradables prices and tradable prices, and then reduces the inflation to equalize the target level. Ho- wever, under the flexible exchange rate regime, the increase of b and c raises the exchange rate in nominal terms (nominal depreciation) which raises severely the inflation. The result indicates that the choice of the flexible exchange rate regime is not without risk, especially the risk relative to inflation, and that the compensation fund is counter effect under the flexible exchange rate. Non-tradable prices’ elasticities Varying non-tradable prices’ elasticities to the real misalignment ρ and to money supply ν doesn’t affect largely the results (Tables 10 and 11). The flexible exchange rate regime remains the optimal choice for monetary authorities. Capital flows’ elasticities Increasing the elasticity of capital outflows to real return φ reduces monetary autho- rities’ losses under both fixed and flexible exchange rate regimes (Table 12). Thus, if  SIMULATIONS AND DISCUSSIONS 219 φ≥1.73, the flexible exchange rate regime becomes the optimal choice. Moreover, increa- sing the elasticity of capital inflows to real return ϕ reduces also monetary authorities’ losses under fixed and flexible exchange rate regimes (Table 13). However, if ϕ≥1.55, the fixed exchange rate regime becomes the optimal choice. These results indicate that dimini- shing returns of capital outflows (more capital outflows with fewer returns) is compatible with the flexible exchange rate regime and that the diminishing returns of capital inflows (more capital inflows with fewer returns) are compatible with the fixed exchange rate re- gime. The desegregation of monetary authorities’ objective function under the fixed exchange rate regime shows that increasing φ (Graphic 14) doesn’t affect monetary authorities’ losses relative to real misalignment and to inflation deviation, but, it reduces losses relative to capital account resulting from the increase of capital outflows. Therefore, the desegrega- tion under the flexible exchange rate regime show that the raise of capital outflows reduces consequently losses relative to real misalignment (because of real depreciation) and relati- vely losses relative to inflation deviation, and raises losses relative to capital account. The desegregation of monetary authorities’ objective function under the fixed exchange rate regime shows that increasing ϕ (Graphic 15) reduces consequently losses relative to real misalignment and relatively losses relative to inflation deviation. In contrary, increa- sing ϕ under the flexible exchange raises the losses relatives to real misalignment and inflation deviation, and decreases losses relatives to capital account. Equilibrium real exchange rate ◦ Decreasing the equilibrium real exchange rate Ert (real appreciation) reduces monetary authorities’ losses under both fixed and flexible exchange rate regimes. The results (Table 14) show that the flexible exchange rate regime is compatible with the real appreciation case, which corresponds to the Moroccan case. When the economy is involved in real depreciation case, the fixed exchange rate regime becomes the optimal choice. Money supply growth Increasing money supply growth mt raises monetary authorities’ losses under both fixed and flexible exchange rate regimes (Table 15). When the money supply grows by more than 9%, the fixed exchange rate regime becomes the optimal choice. High domestic money supply raises the prices of non-tradable goods and then the inflation rate under the flexible exchange rate with respect to the fixed exchange rate. Interest rates The flexible exchange rate is the optimal choice as long as the world interest rate i∗t is growing within the interval −7.38% <i∗t < −2.88% (Table 16) and as the domestic interest rate it is growing within the interval −15.37% <it < −4.37% (Table 17). Accordingly, once the gap between the domestic interest rate and the world interest rate exceeds the interval ] − 8.49%, 2.51%[ under the domestic interest rate variation and the interval] − 2.01%, 3.49%[ under the world interest variation, the fixed exchange rate becomes the optimal exchange rate regime (Graphics 16 and 17). The excess corresponds to capital fight situation when it i∗t and to capital influx when it i∗t . Finally, increasing tradable goods’ inflationΠT t raises monetary authorities’ losses un- der both fixed and flexible exchange rate regime ; however, the flexible exchange rate re- gime remains the optimal choice (Table 18). Thus, the flexible exchange rate regime allows the economy to adjust to the increase in the tradable goods’ inflation by nominal exchange rate rather than the real exchange rate. 220 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO 12.6 Concluding remarks In a context of financial liberalization and economic integration, developing countries try to adopt their liberalization policy approaches taking into account the recent expe- riences. In this context, Morocco conducts a gradual liberalization of trade flows and of capital flows to maximize its advantages from international capital markets. Moroccan monetary authorities initiated a total liberalization of capital inflows and a partial liberali- zation of capital outflows. This situation contributes to the real exchange rate appreciation and raises the question of risks under the fixed exchange rate regime This paper examines the Moroccan authorities’ opportunity to move towards the flexible exchange rate regime under a gradual policy of capital account liberalization. Many factors plead for the exchange rate flexibilization such as economic theory predictions (triangle of incompatibilities), recent developing countries’ experiences and crises (Latin American and East Asian’ experiences) and institutional recommendations (International Monetary fund’ consultations). Moreover, The Moroccan Central Bank announced its decision to reform the exchange rate regime. Based on the findings of E ZZAHID et M AOUHOUB (2014), we continue to develop a new theoretical game model adapted to the Moroccan economic conditions. Thus, we consider a new model with four economic agents, namely monetary authorities, government, fo- reign firms, and domestic firms. We explore the optimal exchange rate regime under these conditions such as the presence of the compensation fund effect and of the government strategy to supress progressively that fund, the foreign firms and the international capital returns, the domestic firms and the presence of some restrictions on capital outflows, etc. Starting with a first simulation based on actual economic parameters, the results show that Moroccan monetary authorities’ losses under a choice of flexible exchange rate regime is lower than losses under a choice of fixed exchange rate regime. The decomposition of the objective-function’s losses shows that the main source of risks is the real exchange rate misalignment and the capital account disequilibrium respectively. Consequently, the choice of flexible exchange rate regime reduces significantly monetary authorities’ losses. However, losses relative capital account equilibrium ; under flexible exchange rate, become relatively higher (table 3). To succeed their strategy of exchange rate flexibilization, Moroccan monetary authori- ties have to liberalize more capital outflows with respect to capital inflows. According to our results, decreasing the parameter n from 1.10 to 1.04 allows monetary authorities to reduce their losses under the flexible exchange rate regime. Thus, Moreover, it’s true that the compensation fund allows price stabilizing under the fixed exchange rate. However, its repartition between non-tradable prices and tradable prices and its extent is of the crucial importance. The optimal repartition according to the results is to allocate 80% of the compensation fund to non-tradable prices and 20% to tradable prices, because as Balassa-Samuelson effect stipulates for developing countries, non-tradable prices grows rapidly that tradable prices. As well, increasing b and c reduces the inflation with respect the target level under the fixed exchange rate regime ; but, it produces a counter effect under the flexible exchange rate regime. Consequently, reducing the compensation fund is recommended under the choice of exchange rate flexibilization.  RÉFÉRENCES 221 Références AGÉNOR, P.-R. (1991). Credibility and exchange rate management in developing countries. AGÉNOR, P.-R. (1994). Credibility and exchange rate management in developing countries. Journal of Development Economics, 45(1), 1-16. A IZENMAN, J. (1994). Monetary and real shocks, productive capacity and exchange rate regimes. Economica, 407-434. C HIN, D. M., M ILLER, P. J. et al. (1995). Fixed vs. floating exchange rates : a dynamic general equilibrium analysis. Federal Reserve Bank of Minneapolis. C ORBO, V. & H ERNANDEZ, L. (1996). Macroeconomic adjustment to capital inflows : lessons from recent Latin American and East Asian experience. The World Bank Research Observer, 11(1), 61-85. D E V ITA, G. & K YAW, K. S. (2011). Does the choice of exchange rate regime affect the economic growth of developing countries ? The Journal of Developing Areas, 135- 153. D EVEREUX, M. B. & E NGEL, C. (2001). The optimal choice of exchange rate regime : Price-setting rules and internationalized production. In Topics in Empirical Interna- tional Economics : A Festschrift in Honor of Robert E. Lipsey (p. 163-194). Univer- sity of Chicago Press. E ZZAHID, E. & M AOUHOUB, B. (2014). Capital account liberalization and exchange rate flexibility : Scenarios for the Moroccan case. Economics Discussion Papers. F RIEDMAN, M. (1953). The case for flexible exchange rates. G HOSH, A. R. (1996). Does the exchange rate regime matter for inflation and growth ? 2. H OFFMANN, M. (2007). Fixed versus flexible exchange rates : Evidence from developing countries. Economica, 74(295), 425-449. L EVY-Y EYATI, E. & S TURZENEGGER, F. (2003). To float or to fix : evidence on the impact of exchange rate regimes on growth. American economic review, 93(4), 1173-1193. M UNDELL, R. A. (1961). A theory of optimum currency areas. The American economic review, 51(4), 657-665. M UNDELL, R. A. (1963). Capital mobility and stabilization policy under fixed and flexible exchange rates. Canadian Journal of Economics and Political Science/Revue cana- dienne de economiques et science politique, 29(4), 475-485. S AMI, B. A. M. (2006). capital account liberalization and exchange rate regime choice, what scope for flexibility in Tunisia ? Z HANG, Z. (2001). Choosing an exchange rate regime during economic transition : the case of China. China Economic Review, 12(2-3), 203-226. 222 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO ANNEXES 1. The base model The variables used in the base model are the general price level p, the targeted general price level p∗ , the price of tradable goods Pt , the price of non-tradable goods PN , the no- minal exchange rate En , the real exchange rate Er , the equilibrium real exchange rate Er∗ , and the domestic money supply M. The structural parameters of the model are the weight granted by monetary authorities to competitivenessα, the weight granted by monetary au- thorities to inflation β= 1−α, the openness degree 0 < 1−µ< 1, the elasticity of prices of non-tradable goods relative to the misalignment of real exchange rate ρ and elasticity of prices of non-tradable goods relative to domestic money supply of real exchange rate υ. The general price level p is function of prices of tradable goods pt that are determined internationally (supposed to be stable) and of prices of non-tradable goods pN that are determined within the economy. The mechanism is described as a Gobb- Douglas function : µ 1−µ P = (PN ) × (En × PT ) (Eq. 1) In growth rate terms, equation (Eq. 1) takes the following expression : p = µ.pN + (1 − µ).en (Eq. 1.1) The prices of non- tradable goods pN are determined by two factors. The first is the deviation of the real exchange rate from its equilibrium level. Thus, a real depreciation increases the external competitiveness of the economy. As a result, we observe an increase in the demand of tradable goods (i.e. increase in exports). Therefore, domestic firms shift resources from the non-tradable goods sector to the tradable goods sector. This transfer of resources lowers the production of non-tradable goods (decrease of supply), which leads to an increase in their prices. A real exchange rate appreciation produces exactly the opposite effects. The deviation of the real exchange rate from its equilibrium level (misalignment) is measured by EErr∗ and the elasticity of prices of non-tradable goods relative to the misalign-  ρ ment is denoted by ρ> 0. So the first factor is denoted by EErr∗ where the real exchange rate Er is given by Er = En × ppNt . The second factor affecting the price of non-tradable goods is domestic money supply, denoted by M. The elasticity of the price of non-tradable goods with respect to M is ν> 0. Consequently, the price of non- tradable goods pN is : #ρ En ppNt " pN = .Mv (Eq. 2) Er ∗ In growth rate terms, the equation (Eq. 2) takes the following expression : pN = ρ. [En − pN − Er ∗ ] +v.m (Eq. 2.1) Monetary authorities are confronted to a trade-off between competitiveness and price stability. Competitiveness is defined by the deviation of the real exchange rate from its  ANNEXES 223 equilibrium level (or target level er ∗ ) and the price stability is defined by the square of the deviation of inflation from its target level p∗ . The objective-function of monetary authori- ties is given as a loss-function : 1 2 WMA = −α. [er − er ∗ ] + β.[p − p∗ ] (Eq. 3) 2 Where α and β are monetary authorities preferences for real exchange rate misalign- ment and inflation targeting. Replacing er and p by their expressions gives the following equation : 1 2 WMA = −α. [En − pN − Er ∗ ] + β.[µ.pN +(1−µ).en − p∗ ] (Eq. 3.1) 2 The welfare of domestic firms depends on the relative price of non-tradable goods. Indeed, domestic firms attempt to protect themselves continuously by adjusting the price of non-tradable goods pN to changes in the expected price of non-tradable goods PeN . The objective-function of domestic firms is given as follows : 1 2 WDF = .[pN − peN ] (Eq. 4) 2 Replacing peN by its expression gives the following equation : 1 2 WDF = .[pN − ρ. [Ene − pN − Er ∗ ] − v.m] (Eq. 4.1) 2 Under a fixed exchange rate regime, we have en = 0. Consequently p =µ.pN and pN = ρ. [(−pN ) − Er ∗ ] +v.m. The objective-functions of the two agents become : 1 2 WMA fixe = −α. [−pN − Er ∗ ] + β.[p−p∗ .] (Eq. 5) 2 1 2 WDF fixe = .[pN − ρ. [−pN − Er ∗ ] − ν.m] (Eq. 6) 2 The model resolution gives us the expression of the objective-function of the monetary authorities depending on the model’s parameters. 2 ρ.Er ∗ − ν.m ν.m−ρ.Er ∗    MA ∗ 1 ∗ W fixe = −α. − Er + ×β. µ. −p (Eq. 7) 1+ρ 2 1+ρ Under a flexible exchange rate regime, En may change and thus En is different from 0. Then, the objective-function of the two agents will be as in the general form developed earlier. The inflation p is replaced by its value µ.pN + (1 − µ) .En in WMA . Thus, under a flexible exchange rate the objective functions of the monetary authorities and domestic firms are obtained as follows. 1 2 WMA flexible = −α. [En − pN − Er∗ ] + β.[µ.pN + (1 − µ) .En − p∗ ] (Eq. 8) 2 224 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO 1 2 WDF flexible = .[pN − ρ. [En − pN − Er∗ ] − ν.m] (Eq. 9) 2 The model resolution gives us the objective-function of the monetary authorities under the exchange rate flexibility choice : α + βµνm+βσ − βp∗ µ − βνm−βEr ∗ +2βµ − βEr ∗ µ2   WMA flexible = − α. β(1 − 2µ + µ2 − µρ + ρ)  2 (Eq. 10) 1 α + β. 2 β(1−µ) 1. Graphics Graphic 1 : budget deficit with and without compensation fund, Source : Rapport on compensation fund, Moroccan 2014 finance bill Graphic 2 : Equilibrium real effective exchange rate, author’s estimation .03 .02 .01 .00 -.01 -.02 -.03 -.04 80 82 84 86 88 90 92 94 96 98 00 02 04 06 08 10  ANNEXES 225 Graphic 3 : Aggregate M3 in growth rate, source : Bank-Al-Maghrib 0,20 0,18 0,16 0,14 0,12 0,10 0,08 0,06 0,04 0,02 0,00 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 Graphic 4 : Moroccan interest rate, source : www.tradinfeconomics.com, Bank Al-Maghrib Graphic 5 : US. Interest rate % (LR), source : World Bank Data base 8,00 7,00 6,00 5,00 4,00 3,00 2,00 1,00 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Graphic 6 : FDI net inflows and net outflows (BoP, current USD) 226 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO Graphic 7 : Elasticities of capital outflows φ and of capital outflows ϕ to real returns 3,00 2,50 2,00 1,50 φ 1,00 ϕ 0,50 0,00 Graphic 8 : Inflation, consumer prices (annual %), source : WB database 12,00 10,00 8,00 6,00 4,00 2,00 - 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 (2,00) European Union Morocco Graphic 9 : capital account restrictions and monetary authorities’ losses 0,012 0,011 0,01 0,009 0,008 Fixed Flexible 0,007 0,006 0,005 0,004 0,85 0,9 0,95 1 1,05 1,1 1,15 1,2 1,25 1,3 1,35 1,4  ANNEXES 227 Graphic 10 : The structure of monetary authorities’ losses under the flexible exchange rate regime 0,0160 0,0140 0,0120 0,0100 RM 0,0080 IT 0,0060 KA 0,0040 0,0020 0,0000 0,8 0,85 0,9 0,95 1 1,05 1,1 1,15 1,2 1,25 1,3 1,35 1,4 1,45 Graphic 11 : Compensation fund repartition and monetary authorities’ losses 0,025 0,02 0,015 Fixed Flexible 0,01 0,005 0 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 Graphic 12 : Varying parameter b and desegregation of monetary authorities’ losses 0,006 0,005 RM/FIXED 0,004 RM/FLEXIBLE 0,003 KA/FIXED KA/FLEXIBLE 0,002 IT/FIXED IT/FLEXIBLE 0,001 0 0 1/3 2/3 1 4/3 5/3 2 7/3 8/3 3 Graphic 13 : Varying parameter c and desegregation of monetary authorities’ losses 0,006 0,005 RM/FIXED 0,004 RM/FLEXIBLE 0,003 KA/FIXED KA/FLEXIBLE 0,002 IT/FIXED IT/FLEXIBLE 0,001 0 0 1/3 2/3 1 4/3 5/3 2 7/3 8/3 3 228 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO Graphic 14 : Varying parameter φ and desegregation of monetary authorities’ losses 0,006 0,005 RM/FIXED 0,004 RM/FLEXIBLE 0,003 KA/FIXED KA/FLEXIBLE 0,002 IT/FIXED IT/FLEXIBLE 0,001 0 1,63 1,68 1,73 1,78 1,83 1,88 1,93 1,98 Graphic 15 : Varying parameter ϕ and desegregation of monetary authorities’ losses 0,01 0,009 0,008 0,007 RM/FIXED 0,006 RM/FLEXIBLE 0,005 KA/FIXED KA/FLEXIBLE 0,004 IT/FIXED 0,003 IT/FLEXIBLE 0,002 0,001 0 1,30 1,35 1,40 1,45 1,50 1,55 1,60 Graphic 16 : the interest rate gap it − i∗t and monetary authorities’ gap M A/f lexible M A/f ixex Wt − Wt under domestic interest rate variation 0,002 0,001 0 -0,0949-0,0849-0,0749-0,0649-0,0549-0,0449-0,0349-0,0249-0,0149-0,0049 0,0051 0,0151 0,0251 0,0351 -0,001 -0,002 -0,003 -0,004 Graphic 17 : the interest rate gap i∗t − it and monetary authorities’ gap M A/f lexible M A/f ixex Wt − Wt under world interest rate variation 0,0025 0,0020 0,0015 0,0010 0,0005 - -0,0351 -0,0201 -0,0151 -0,0051 0,0049 0,0149 0,0249 0,0349 0,0449 -0,0005 -0,0010 -0,0015  ANNEXES 229 Table 1: Estimated parameters and interval of variation Baseline Parameters Interval of variation simulation { } { } { } { } { } none none none Table 2: Monetary authorities’ losses under fixed and flexible regimes Choices Monetary authorities’ losses Fixed Exchange rate regime 0.006855 Flexible Exchange rate regime 0.006280 Table 3: desegregation of monetary authorities’ losses under fixed and flexible regimes Choices objective function’s Losses Percentage Total components Real misalignment RM 0.004812 70.20% Inflation targeting IT 0.000631 9.20% 0.006855 Fixed Exchange rate regime Capital account KA 0.001412 20.60% Real misalignment RM 0.003265 51.99% Inflation targeting IT 0.000594 9.46% 0.006280 Flexible Exchange rate regime Capital account KA 0.002421 38.55% 230 CAPITAL ACCOUNT LIBERALIZATION IN MOROCCO Table 4: Monetary authorities’ losses and capital account restrictions 0.80 0.85 0.90 0.95 1 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.45 Capital capital flight capital influx flows Fixed 0.007043 0.007011 0.006979 0.006948 0.006917 0.006886 0.006855 0.006825 0.006796 0.006766 0.006737 0.006708 0.006680 0.006652 Flexible 0.015576 0.010262 0.007811 0.006700 0.006261 0.006175 0.006280 0.006487 0.006747 0.007031 0.007323 0.007614 0.007897 0.008170 Table 5: Monetary authorities’ losses values and preferences Order of importance Fixed 0.008368 0.005861 0.010615 0.006855 0.011350 0.012604 Flexible 0.010251 0.006648 0.012041 0.006280 0.009737 0.011520 Table 6: Objective function values and economic openness rate 1-s 0.50 0.55 0.60 0.65 0.70 0.75 0.80 Fixed 0.009102 0.008417 0.007817 0.007298 0.006855 0.006485 0.006181 Flexible 0.007569 0.007190 0.006849 0.006546 0.006280 0.006052 0.005860 Table 7: Monetary authorities’ losses and compensation fund repartition 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Fixed 0.007090 0.007054 0.007014 0.006969 0.006916 0.006855 0.006784 0.006699 0.006596 0.006468 0.006306 Flexible 0.006844 0.006791 0.006699 0.006593 0.006456 0.006280 0.006059 0.005805 0.005635 0.006468 0.022710 Table 8: Monetary authorities’ losses and parameter b b 0 1/3 2/3 1 4/3 5/3 2 7/3 8/3 3 Fixed 0.007367 0.006855 0.006439 0.006094 0.005805 0.005559 0.005348 0.005165 0.005006 0.004865 Flexible 0.006537 0.006280 0.006096 0.005960 0.005856 0.005777 0.005714 0.005663 0.005623 0.005589 Table 9: Monetary authorities’ losses and parameter c c 0 1/4 1/3 2/3 1 4/3 5/3 2 7/3 8/3 3 Fixed 0.007079 0.006855 0.006795 0.006600 0.006459 0.006351 0.006267 0.006199 0.006143 0.006069 0.006056 Flexible 0.006822 0.006280 0.006205 0.006045 0.005973 0.005932 0.005906 0.005888 0.005874 0.005864 0.005863 Table 10: Monetary authorities’ losses and elasticity of non-tradable goods’ prices to real misalignment 0.25 0.30 0.35 0.4 0.45 0.50 0.55 Fixed 0.006858 0.006857 0.006856 0.0068559 0.0068553 0.0068547 0.0068541 Flexible 0.006236 0.006252 0.006267 0.006280 0.006292 0.006304 0.006314  ANNEXES 231 Table 11: Monetary authorities’ losses and elasticity of non-tradable goods’ prices to money supply 0.45 0.50 0.55 0.6 0.65 0.70 0.75 Fixed 0.006757 0.006789 0.006821 0.006855 0.006891 0.006927 0.006964 Flexible 0.006148 0.006181 0.006225 0.006280 0.006347 0.006426 0.006516 Table 12: Monetary authorities’ losses and elasticity of capital outflows to real return 1.63 1.68 1.73 1.78 1.83 1.88 1.93 1.98 Fixed 0.007526 0.007307 0.007129 0.006980 0.006855 0.006749 0.006658 0.006579 Flexible 0.007916 0.007454 0.007026 0.006635 0.006280 0.005964 0.005689 0.005455 Table 13: Monetary authorities’ losses and elasticity of capital inflows to real return 1.30 1.35 1.40 1.45 1.50 1.55 1.60 Fixed 0.012200 0.009687 0.008022 0.006855 0.006004 0.005362 0.004863 Flexible 0.008455 0.007526 0.006814 0.006280 0.005895 0.005633 0.005475 Table 14: Monetary authorities’ losses and Equilibrium real exchange rate - 0.03417 -0.02917 -0.02417 - 0.01917 -0.01417 -0.00917 -0.00417 0.0000 +0.00417 +0.00917 +0.01417 Fixed 0.006172 0.006395 0.006623 0.006855 0.007093 0.007315 0.007583 0.007793 0.008006 0.008266 0.008532 Flexible 0.005600 0.005731 0.005958 0.006280 0.006698 0.007212 0.007822 0.008404 0.009052 0.009917 0.010878 Table 15: Monetary authorities’ losses and domestic money supply 0.0109 0.0209 0.0309 0.0409 0.0509 0.0609 0.0709 0.0809 0.0909 0.1009 Fixed 0.006578 0.006638 0.006777 0.006777 0.006855 0.006940 0.007030 0.007127 0.007229 0.007338 Flexible 0.006218 0.006136 0.006120 0.006168 0.006280 0.006457 0.006698 0.007004 0.007374 0.007809 Table 16: Monetary authorities’ losses and world interest rate -0.0888 -0.0738 -0.0688 -0.0588 -0.0488 -0.0388 -0.0288 -0.0188 -0.0088 0.0000 +0.0088 +0.0188 +0.0288 Fixed 0.007880 0.007333 0.006855 0.006447 0.006109 0.005840 0.005640 0.0055101 0.005449 0.005453 0.0055109 0.005641 0.005842 Flexible 0.009298 0.007570 0.006280 0.005428 0.005014 0.005037 0.005498 0.006397 0.007734 0.009272 0.011150 0.013695 0.016678 Table 17: Monetary authorities’ losses and domestic interest rate -0.0537 -0.1637 -0.1537 -0.1437 -0.1337 -0.1237 -0.1137 -0.1037 -0.0937 -0.0837 -0.0737 -0.0637 -0.0437 -0.0337 -0.0237 -0.0137 -0.0037 0.0037 0.0137 0 0.049478 0.043853 0.003656 0.002582 0.038578 0.033653 0.029079 0.024854 0.020979 0.017545 0.014279 0.011455 0.006855 0.005081 0.001858 0.001483 0.001433 0.001432 0.001666 0.00898 Fixed 0.037662 0.026806 0.009372 0.007549 0.006280 0.050737 0.043922 0.005566 0.005407 0.005802 0.006752 0.008257 0.008954 0.009728 0.031957 0.014682 0.018169 0.011749 Flexible 0.02221 0.0054 Table 18: Monetary authorities’ losses and prices of tradable goods 0.005 0.010 0.015 0.020 0.025 0.030 Fixed 0.006264 0.006855 0.007486 0.008154 0.008862 0.009609 Flexible 0.005781 0.006280 0.006855 0.007507 0.008234 0.009038