ࡱ> iq`\ bjbjqPqP 8T::s  ^j D  jjj8&| E:U< DDDDDDD$HhTK+DA +D lE DDr T  "$jQe $E0Es zK=.K  4K ! F+D+DkdE d&D &   THE DETERMINANTS OF FINANCIAL EUROIZATION IN A POST-TRANSITION COUNTRY: DO THRESHOLD EFFECTS MATTER? This paper investigates long run and short run determinants of financial euroization (FE) using both linear and threshold models. We model deposit euroization (DE) and credit euroization (CE) in Croatia, a post-transition country recording very high and persistent unofficial FE. Results suggest that only portfolio view is important for explaining DE and CE. Market failure view does not seem to matter for FE in Croatia. Both nominal and real exchange rate changes have a strong effect on FE in the long run, the former is more important for DE, while the latter for CE. In the short and long run CE is also determined by matching behaviour of banks foreign currency position. Both DE and CE respond to changes in inflation and exchange rate volatility. Threshold cointegration confirms that FE determination is subjected to significant threshold effects while error correction models suggest FE adjustment is very slow and asymmetric, partly due to very strong FE persistence. Key words: financial euroization, transition, cointegration, threshold JEL Classification: C22, E44; F31; F41 1. INTRODUCTION Origins of unofficial financial euroization (FE) in most post-transition European countries can be traced back as far as to 1980s. In spite of economic and political stability, strong growth performance and increased central bank credibility witnessed in the last 15 years, the incidence of FE continued to persist in most of post-transition European countries. Exchange rate depreciations experienced in some countries during the economic downturn in 2009 showed just how dangerous it is to underestimate the pervasiveness of FE. To that end, EBRD in its latest Transition report for 2010 cites FE as one of the major weaknesses of post-transition European countries that needs to be tackled in order to achieve stable and sustainable growth in the future (EBRD, 2010). Policy makers in post-transition economies acceding to EU did not insist on financial system de-euroization due to the assumption that the Euro adoption is imminent, which in turn implied that FE was a temporary phenomenon. Moreover, since banking systems of post-transition countries consisted mostly of banks owned by parent banks from the Eurozone, FE was considered as a natural extension of the financial integration process. However, after the global economic crisis shook the economies of the region it has become increasingly obvious that the adverse effects of FE as well as the limits it imposes on policymakers were perhaps somewhat underestimated. Large exchange rate depreciation in euroized countries can not only affect the stability of the banking system, but can also deepen or even trigger recession, instead of helping the economy to recover (Cook, 2004). In such a situation, it is clear that more research is needed in order to understand the FE phenomenon in the region and to undertake appropriate policy actions. We analyse data for Croatia, one of the post-transition countries characterised by very high and persistent FE. We use empirical analysis in order to answer the following key questions. First, what drives deposit and credit euroization and are any of the drivers susceptible to policy measures? Second, is FE more elastic to changes in nominal or real effective exchange rate (REER)? Third, are deposit and credit euroization characterised by threshold effects? Besides having direct implications for economic policy, the results of this analysis also contribute to the body of literature in several respects. First, to the best of our knowledge, all existing empirical studies on substituting domestic currency with Euro use linear models, although it is well documented that euroization (dollarization) increases sharply during episodes of macroeconomic instability and that it remains stubbornly high even long after successful stabilizations thereby implying either asymmetric or threshold effects. In the same vein, it is less likely that the level of FE will change if nominal or real effective exchange rate changes are small or negative (i.e. if the exchange rate appreciates). The opposite is true for large and positive changes of the nominal or real effective exchange rate. Several studies like Neanidis and Savva (2009) and Rennhack and Nozaki (2006) incorporate an index of asymmetric exchange rate movement in a linear econometric framework, which could be considered as an implicit confirmation of existence of a bias toward local-currency depreciation. Moreover, threshold effects may also matter for responses of FE to exchange rate volatility and inflation changes. Therefore, in this study, we estimate both linear and threshold models of short run FE behaviour in order to determine whether any threshold effects exist, and in which cases linear models of FE are appropriate. Second, by using separate models we test the effects of nominal and real effective exchange rate changes on FE. Although the portfolio and market failure approach to explaining FE emphasize REER as FE determinant, empirical papers usually control either for nominal exchange rate or for REER changes and by doing so they neglect to compare which exchange rate has a stronger effect on FE. We feel that nominal exchange rate changes might have a stronger effect on FE because they are more easily observed and thus agents may react stronger than to REER changes. Moreover, threshold cointegration models used in this paper assume the adjustment is discrete, or in other words, all agents adjust long run discrepancies in the level of euroization at the same time. The observability of nominal exchange rate facilitates the discrete adjustment, thus making the nominal exchange rate more suited for threshold cointegration model when compared to REER. Finally, unlike other empirical studies on FE for post-transition European countries, our analysis includes the period of the latest economic downturn that coincided with an intensive FE upswing. Hence, inclusion of the newest observations may add relevant information content to the empirical model and consequently provide new insights into FE behavior. The remainder of the paper is organized as follows. Section two reviews the empirical literature on FE giving special attention to the results for European post-transition countries and Croatia. The third section presents a brief overview of FE in Croatia. The fourth section discusses the applied methodology, data and results of the empirical analysis, while section five concludes the paper. 2. LITERATURE REVIEW The term euroization (or dollarization) refers to a variety of cases in which a domestic currency is to a certain degree substituted by a foreign currency. While official euroization implies legally approved foreign currency usage for all money functions, unofficial euroization presents a situation in which both domestic and foreign currency are used simultaneously, although a foreign currency is not a legal tender currency. Unofficial euroization comes in different forms: currency substitution (foreign currency pertains medium of exchange function), asset substitution (foreign currency pertains store of value and unit of account function) and real euroization (foreign currency used for denomination of prices and wages). Instead of asset substitution, during the last decade the literature coined the term financial euroization (FE), further divided into deposit and credit euroization. Deposit euroization (DE) is the propensity of households, enterprises and even governments to hold deposits in foreign currency, while credit euroization (CE) is a propensity of commercial banks to approve loans either in foreign currency or indexed to foreign currency. Early literature on dollarization in Latin America focuses primarily on currency substitution and its repercussion for conducting monetary policy. Currency substitution is thus explained through negative correlation between demand for local currency and the rate of inflation (Savastano, 1996). Levy Yeyati (2006) outlines three main paradigms in modern theoretical analysis of FE. A portfolio view explains FE as an outcome of decisions on an optimal (minimum variance) portfolio given the real returns on different currencies. Returns of local currency assets are uncertain due to domestic inflation while returns on foreign currency assets are uncertain due to real exchange rate risk (Ize and Levy Yeyati, 2003; Basso, Calvo-Gonzales and Jurgilas, 2007). Since this setting assumes that the uncovered interest rate parity holds, the emphasis is placed on variances. Namely, if the variability of domestic inflation increases relative to the real devaluation rate, the local currency becomes less attractive and the FE ratio rises. The market failure view explains FE as a result of optimal decisions by risk neutral agents in the presence of default risk (enhanced by moral hazard/asymmetric information). If the central bank of an economy characterized by large and sudden exchange rate depreciations in the past, insists on a stable exchange rate, moral hazard problems will arise and borrowers and lenders will not fully internalize risks of borrowing (or lending) in foreign currency. Consequently, they benefit from lower cost (and lower risk) of borrowing and lending in foreign currency, which in turn increases the level of FE. The institutional view analyses the presence of FE in relation to the credibility of central bank and government policies, i.e. it explains FE as a result of domestic market and legal imperfections. Weak institutional framework and low credibility of economic policies increase the government quest to build up confidence by using the exchange rate anchor. Since such dollar- or euro-friendly regulations represent a commitment mechanism by which the government or monetary authorities borrow credibility and make the cost of a potential depreciation prohibitively high, FE becomes the collateral cost of low institutional credibility (Nicol, Honohan, Ize, 2005; Rajan and Tokatlidis, 2005). Several empirical papers focus on FE in post-transition European countries; all of them use linear models to control for the behavior of FE except a study by Heimonen (2001). For the case of Estonia, Heimonen (2001) estimates possible portfolio shifts between two substitute currencies, Euros and dollars. He identifies the dynamics of such a substitution in both short and long term using threshold cointegration. However, his research does not focus on determinants of financial euroization, nor it considers the substitution between the domestic and foreign currency. By examining foreign currency shortages in dollarized economies, Rajan and Tokatlidis (2005) claim that liability dollarization is a response to institutional weaknesses and lack of monetary credibility. Moreover, their results indicate that higher inflation in the past leads to dollarization strengthening while the opposite is not true when inflation falls suggesting CE is persistent. Neanidis and Savva (2009) estimate a panel of 11 Central and Eastern European (CEE) and Commonwealth of Independent States (CIS) countries in order to model short run determinants of DE and CE. In the short run, negative effects of depreciation and monetary contraction are particularly pronounced in countries with high euroization. CE is in the short run driven by DE, while both types of euroization are influenced by interest rate differentials. Basso, Calvo-Gonzales and Jurgilas (2007) use an unbalanced panel of 24 CEE and CIS countries in order to explain FE. The results indicate that access to foreign funds as well as higher interest rate differentials increase CE, although they decrease DE. Moreover, the tradeoff between inflation and REER volatility is found to be a significant factor in explaining FE, thereby proving the validity of the portfolio approach. Similar results are confirmed also by Rosenberg and Tirpk (2008) who find that CE increases together with rising interest rate differentials, foreign funding and openness. Other authors agree that higher degree of trade openness contributes to CE. Luca and Petrova (2008) use aggregate data in a panel of 21 CEE and CIS countries in order to model CE. Empirical results suggest CE is positively related to interest rate differentials, domestic monetary volatility and DE, while it is negatively related to exchange rate volatility. Moreover, export openness seems to promote CE, as export oriented firms use natural hedges. Authors also show that the presence of a deep forward foreign currency market decreases the level of CE. As explored by Guscina (2008), financial development, openness and falling inflation promote government debt de-euroization and therefore tend to decrease CE. Stix (2010) uses a unique household survey data for Croatia, Slovenia and Slovakia in order to understand factors driving DE persistence. Probit estimations show that expectations of inflation and exchange rates do not affect DE, while trust in the banking system determines the choice between foreign currency cash and foreign currency deposits. On the other hand, age and level of income and education of savers are positively correlated with DE. Brown, Kirschenmann and Ongena (2009) analyze CE in Bulgaria using a unique data set on individual loan contracts of a Bulgarian bank to small and medium size businesses. The results show that Bulgarian firms are more likely to request a foreign currency denominated loan if they have foreign exchange revenues and if the requested loan amount is larger and the maturity longer. Higher inflation volatility and interest rate differential also promote demand for foreign currency denominated loans. At the same time, banks seem hesitant to lend long-term in local currency and are eager to match the currency structure of their assets and liabilities. Moreover, exploring a large sample of small firms from a number of emerging market economies, Brown, Ongena and Ye_in (2009) confirm the results of Brown, Kirschenmann and Ongena (2009) but find that CE is more strongly related to foreign currency revenues than to interest rate differentials. Kokenyne, Ley and Veyrune (2010) separately model CE and DE in Croatia in order to determine whether prudential regulation has an effect on the level of FE. They find that both CE and DE show a moderate response to prudential measures in Croatia. Besides prudential measures, CE also reacts to changes in REER and inflation, while DE seems to be characterized by a high degree of persistence. 3. FINANCIAL EUROIZATION IN CROATIA The Croatian financial system is characterized by a very high and persistent level of euroization. Judging from data on DE and CE collected both by Luca and Petrova (2008) and Levy Yeyati (2006), Croatia is (together with Armenia and Georgia) by far the most financially euroized country in Europe. Moreover, once Croatia joins the EU (which may be expected during 2013 or 2014), it will become an EU member state with the most pervasive FE. The main features of FE in Croatia encompass using Euro as a unit of account and store of value, very high level of both DE and CE, less pronounced currency mismatches on commercial banks balance sheets and very pronounced currency mismatches on balance sheets of non-tradable business sectors and households. Similar features can also be identified in other Western Balkan countries (Serbia, Macedonia, Kosovo and Albania), some CIS countries (Armenia and Georgia) and to a smaller degree in two Baltic countries, Latvia and Lithuania. On the contrary, in many Central and Eastern European countries DE is much less pronounced, credits are indexed also to CHF and JPY, and commercial banks exposure to currency mismatches is much more pronounced. Euroization of the Croatian financial system first appeared in 1980s during the period when Croatia was a part of former Yugoslavia. It emerged due to hyperinflation, several sharp devaluations of the domestic currency as well as due to a general instability of the domestic banking system. The emergence of euroization was supported also by a significant inflow of workers remittances and high tourism revenues, both in foreign currency. In the first few years of 1990s, the Croatian economy was stricken by war and by a transition recession that further destabilized the financial system. Economical stagnation in a combination with hyperinflation amounting to 1,616 percent annually in 1993, resulted in an erosion of economic policies credibility and caused additional distrust in the domestic currency. In spite of the successful implementation of the Stabilization program in October 1993 and low inflation rates in later years, general confidence in the domestic currency has not been established. The domestic currency has never fully assumed its unit of account function, while store of value function was mostly reserved for Deutsche mark and later Euro (Vizek, 2006). Consequently, both deposit and credit euroization remained a permanent characteristic of the financial system, although their level varied somewhat over the years. In the period from 1994 to 2000, DE was particularly high, since the ratio of foreign currency deposits in total deposits revolved from 85 to 90 percent. This was the consequence of macroeconomic instability exhibited during 1993, compounded by a banking crisis, brief economic downturn and increasing depreciation pressures on nominal exchange rate experienced in 1998-1999 period. [Insert Figure 1 here] After 2000, DE leveled off and gradually descended to 65 percent recorded during 2006-2008 period. A prolonged period of macroeconomic stability and increasing monetary policy credibility, higher interest rates on domestic currency deposits and appreciation of the domestic currency contributed most to such movements. Along with deposit euroization, credit euroization also trended downwards as evident in the data gathered since 1999. Starting from 2000, when the ratio of foreign currency denominated loans amounted to 85 percent of total loans, the level of CE started to subside to reach its bottom during 2008 at around 60 percent. After the global financial crisis erupted and macroeconomic instability increased, both deposit and credit euroization downward trends reversed, suggesting that credibility of domestic currency and monetary authorities was never truly established. EMPIRICAL ANALYSIS 4.1. METHODOLOGY The aim of the econometric analysis is to model long run and short run behaviour of DE and CE. Thereby, we use both linear (Johansen cointegration and vector error correction model) and threshold approaches (Engle-Granger threshold cointegration and threshold error correction model). Each aspect of financial euroization (DE and CE) is modelled with two models with the same endogenous variables, except for a variable that proxies the exchange rate. As explained in the introduction, we use two exchange rate proxies: REER and nominal exchange rate of domestic currency to euro. The following general-to-specific modelling strategy is employed: first we estimate the model for both DE and CE with the largest number of potential determinants and test for Johansen cointegration in order to obtain a long run equilibrium relationship between euroization and their determinants. If no cointegration is found, we reduce the model by one variable and test again for cointegration. Once a linear long run relationship is found, we test for threshold cointegration. If threshold cointegration is established, we continue with the estimation of the threshold error correction model of DE and CE. If no threshold cointegration is found, we estimate a vector error correction model of DE and CE in order to model the short run behaviour of financial euroization. We continue with the exercise by reducing further the number of exogenous variables until we find significant threshold cointegration relationship and derive a threshold error correction model. Results of this exercise will show that Johansen cointegration is detectable in relatively large models for which no threshold effects are discovered. On the other hand, threshold effects are present in relatively parsimonious models of DE and CE. For Johansen cointegration test, we use finite sample corrections of trace and max statistics (Cheung and Lai, 1993) instead of original test statistics (Johansen, 1988; Johansen, 1991) mainly because trace and max statistics tend to detect cointegration more often in finite samples when models include large number of variables and lags. In order to detect threshold cointegration in the behaviour of FE in Croatia, we use a threshold cointegration method developed by Enders and Siklos (2001) by extending generalized threshold autoregressive (TAR) and momentum-TAR (M-TAR) tests for unit roots. The model used for testing threshold cointegration has the following specification:  EMBED Equation.3  (1) where  EMBED Equation.3  is the disturbance term of Engle-Granger long run model, while  EMBED Equation.3 and  EMBED Equation.3  are the Heaviside indicator functions for the TAR and the M-TAR model respectively, such that:  EMBED Equation.3  (2) in the TAR case, and  EMBED Equation.3  (3) in the M-TAR case.  EMBED Equation.3  and  EMBED Equation.3  are the values of the threshold and  EMBED Equation.3  is a sequence of IID random variables with mean zero and a constant variance. Threshold  EMBED Equation.3  is either set to zero or is determined endogenously using an algorithm developed by Chan (1993). In each of the four cases, depending on the type of asymmetry under consideration EMBED Equation.3  and on the value of the threshold, we estimate the regression equation (1) and test the null hypotheses  EMBED Equation.3  and EMBED Equation.3 , which are direct tests of cointegration existence. The empirical F statistics obtained from the latter test is compared to the critical values tabulated by Wade, Gilbert and Dibooglu (2004). Finally, we test the null hypothesis  EMBED Equation.3  using the Wald test in order to determine whether the cointegration relationship is characterised by threshold effects. 4.2. DATA Data for the DE model are available in monthly frequency from January 1997 to May 2010. As a proxy for DE, we use the share of foreign currency deposits in total deposits. Apart from the standard variables used in this kind of empirical exercise as possible DE determinants like nominal exchange rate, REER, exchange rate volatility (calculated as a standard deviation of daily exchange rates in a given month), CPI inflation, M1 monetary aggregate, the share of M4 in M1 monetary aggregate (a variable representing financial system development), we also use three variables suggested by the main paradigms in theoretical analysis of FE. The first variable is past rate of inflation suggested by the currency substitution view (Savastano, 1996), where past inflation in the contemporaneous period is calculated as an average inflation during 24 months (in case of DE models) or 8 quarters (in case of CE models) prior to the contemporaneous period. The second variable is the minimum variance portfolio (MVP) suggested by the portfolio view (Ize and Levy Yeyati, 2003) that measures the Euro share of the MVP calculated from historical inflation and real depreciation rates. The third variable represents the market failure view (Levy Yeyati, 2006) that measures the correlation between probability of default and REER. The more procyclical REER is, the stronger is the euroization bias. Levy Yeyati (2006) proxied the probability of default by real GDP growth, but since we need observations in monthly frequency, we use the industrial production growth that is in Croatia closely correlated to GDP changes. Data for the CE model are available in quarterly frequency from the third quarter of 1999 to the second quarter of 2010. As a proxy for CE, we use the share of foreign currency loans and loans indexed to foreign currency in total loans. Besides using all endogenous variables already specified above as potential DE determinants, we also use total bank foreign debt and the share of foreign currency deposits in total deposits, as additional CE determinants representing commercial banks foreign currency funding sources and consequent matching behaviour of banks foreign currency position. All variables are seasonally adjusted using X12ARIMA method and are in logarithms. ADF unit root test results suggest all variables are stationary in first differences. A more detailed description of variables and their sources can be found in the Appendix. 4.3. RESULTS Johansen cointegration tests suggested the following variables are cointegrated with DE: exchange rate (nominal or REER), exchange rate volatility, minimum variance portfolio, monetary aggregate M1 and inflation. The long run model of CE includes: exchange rate (nominal or REER), exchange rate volatility, minimum variance portfolio, bank foreign debt, deposit euroization, the share of M4 over M1 aggregate and average past inflation. Unrestricted and restricted long run coefficients, together with related adjustment parameters are reported in Table 1 for DE model and in Table 2 for CE model. Results of individual zero restrictions on cointegrating parameters suggest none of the variables displayed in Table 1 can be excluded from the long run equilibrium relationship. Since we found one cointegration vector, we need only one restriction per model to be able to identify long run coefficients and adjustment parameters. However, we introduce a greater number of restrictions in order to add more information content to the system. In the long run DE model with REER as a proxy for the exchange rate, we restrict the long run coefficient for REER to be equal to -1. Besides, we impose weak exogeneity of DE and exchange rate volatility. Concerning the DE model with the nominal exchange rate, we do not manage to restrict the coefficient for the nominal exchange rate. Since the estimated long run elasticity of DE to changes in nominal exchange rate and REER amounts to 2.4 and 1 respectively, we can conclude that the depreciation of the nominal exchange rate has a much stronger effect on the long run behaviour of DE. Furthermore, in the DE model with the nominal exchange rate, we impose the unit elasticity restriction on the coefficient for inflation together with weak exogeneity of exchange rate volatility and monetary aggregate M1. [Insert Table 1 here] Long run coefficients also suggest that exchange rate volatility and inflation contribute to a rise in DE, while monetary expansion has an effect in the opposite direction. Results for monetary aggregate M1 are not surprising given the fact that the increase in money supply promotes local currency deposits. Those findings are confirmed also by earlier studies. [Insert Table 2 here] Along with the long run DE models, we also try imposing similar restrictions for CE models. The unit elasticity restriction on the REER coefficient in the CE model is rejected with Chi-square equal to 6.01 and the underlying p-value 0.01. We impose weak exogeneity of the exchange rate volatility and the ratio of monetary aggregate M4 over M1. In the CE model with the nominal exchange rate, we restrict the long run coefficient for the nominal exchange rate to -1 and impose weak exogeneity of exchange rate volatility, bank foreign debt and M4/M1. As is the case for DE, we find a positive relationship between CE and both nominal and real effective exchange rate, suggesting that CE also increases after depreciation. However, opposite to DE, CE seems to react more strongly to changes in REER when compared to the nominal exchange rate. Growing exchange rate volatility, bank foreign debt, DE and average past inflation contribute to a rise in CE, while financial system development (represented by M4/M1) reduces it. As a next step, we construct vector error correction models for the four cases examined where each case consists of the error correction term derived from the belonging restricted cointegration vector. To save space, we present only error correction models for deposit and credit euroization. All error correction models for DE and CE satisfy diagnostic tests and are presented in Table 3. One can notice that disequilibria in short run CE models adjust very slowly, while there seems to be no adjustment in the case of DE. Block exclusion restrictions on all lags of individual right-hand side variables suggest that in the short run DE only responds to changes in inflation and minimum variance portfolio. Moreover, results suggest that persistence is important both for DE and CE short run behaviour. In the CE model with incorporated nominal exchange rate, changes in DE, bank foreign debt and average past inflation all have an effect on euroization in the short run. To sum up the results of linear models, portfolio view seems to be important for explaining both DE and CE in the long run, thus confirming the results of Basso, Calvo-Gonzales and Jurgilas (2007). Market failure view does not seem to matter, while currency substitution view matters only for explaining the behaviour of CE. Besides variables suggested by the theory, DE in the long run increases as a result of consumer price inflation and nominal and real effective exchange rate depreciation. DE is more elastic to changes in nominal exchange rate when compared to REER changes, while the opposite is true for CE. CE also shifts upwards after an increase in bank foreign debt and DE, suggesting that bank currency matching behaviour is very important for the long run CE determination in Croatia. Contrary to results of Luca and Petrova (2008) and Kokenyne, Ley and Veyrune (2010), we find that greater exchange rate volatility promotes both DE and CE. Results also suggest that monetary expansion promotes local currency deposits, thereby confirming the results of Neanidis and Savva (2009). The adjustment is quite slow and characterised by a high degree of persistence, thereby supporting Kokenyne, Ley and Veyrune (2010) findings. The emergence of persistence as FE driver could however be a result of choice of DE and CE proxies. Namely, the backward looking feature of variables used as proxies (stocks of foreign exchange denominated deposits/credits to total deposits/credits) might give rise to persistence tendencies and therefore the importance of persistence must be somewhat downplayed. Moreover, in the short run both CE and DE react to changes in inflation, but they do not react to changes in, neither nominal, nor real effective exchange rate. DE also responds to short run changes in MVP, while CE responds to short run changes in banks foreign sources of financing, thus supporting Brown, Kirschenmann and Ongena (2009), Brown, Ongena and Ye_in (2009) and Basso, Calvo-Gonzales and Jurgilas (2007) findings. [Insert Table 3 here] As a next step, we test the detected four Johansen cointegration relationships to threshold cointegration. Comparing the test statistics for  EMBED Equation.3  test with the critical values tabulated in Wade, Gilbert and Dibooglu (2004), we conclude there is no evidence of threshold cointegration. Thus, we conclude that Johansen cointegration and VECM models of DE and CE are well specified and continue with testing the threshold cointegration on more parsimonious models. In case of CE, we detected threshold cointegration in quintvariate cases, while in case of DE threshold cointegration is present only in bivariate cases. Table 4 summarises the test results for threshold CE models with five variables. Four models with CE as the dependent variable satisfy all diagnostic tests and reject the null of no cointegration and the F test for symmetric adjustment at least at the 5% level. The threshold values are positive, similar and very close to zero for all cases that demonstrate an asymmetric adjustment of CE to changes in bank foreign debt, exchange rate volatility, either inflation or average past inflation and either nominal or real effective exchange rate. We proceed by creating threshold error correction models of CE for four cases exhibiting threshold cointegration. [Insert Table 4 here] Threshold error correction with estimated coefficients and associated p-values of the constant term and the adjustment parameters, the block exclusion restriction test for lagged changes of the dependent and the explanatory variables with associated F statistics and p-values, together with diagnostic tests are presented in Table 5. The first thing one notices is that the dependent variable is not weakly exogenous. Moreover, CE adjusts the discrepancies in both regimes, with the adjustment being faster when the disequilibrium is above the threshold (true in three out of four cases). Adjustment parameters range from -0.266 (in the case of the real effective exchange rate) to -0.150 (in the case of the nominal exchange rate), but even with the largest adjustment parameter (-0.266) it takes approximately 4 periods or one year to correct the discrepancies. The observed slow adjustment might be explained by persistence, presence of which is indicated by the results of Granger causality tests for lagged dependant variable. Since there is persistence in the adjustment process, fundamentals take a longer time to react which in turn prevents the adjustment to unfold more swiftly or completely. However, as in the case with VECM, persistence might also be a result of a stock definition of the dependant variable. Concerning the other CE determinants included in threshold models, in the short run only the nominal exchange rate has an affect on CE, although this effect is marginally significant. [Insert Table 5 here] Table 6 summarizes test results of three bivariate DE models for which threshold effects were found. These models suggest DE reacts in the long ran to nominal exchange rate, REER and exchange rate volatility. Threshold values in all three models are negative and very close to zero. As a final step, we construct threshold error correction models for three cases exhibiting threshold cointegration with the results displayed in Table 7. DE responds to discrepancies from the equilibrium in all three models. In the model with REER, DE adjusts discrepancies in both regimes, but the adjustment is faster when the discrepancies are below the threshold. In the case with the nominal exchange rate, the adjustment occurs in one regime, when discrepancies are above the threshold. Significant adjustment parameters are very small and range from -0.016 to 0.044 implying it even takes up to two years to correct the long run disequilibria. [Insert Table 6 here] [Insert Table 7 here] 5. CONCLUDING REMARKS In order to address FE with the appropriate policy measures, it is very important to understand its determinants. Namely, if policymakers aiming at de-euroization do not understand the process that led to equilibrium FE, with FE sometimes being an optimal outcome in an economy characterised by a set of frictions, policy measures aimed at de-euroization could result in a failure and may even have detrimental effects on the stability of the financial system. Results of the empirical analysis presented in this paper suggest that FE is an equilibrium outcome of several factors. Models nested in linear framework suggest that the exchange rate, exchange rate volatility, inflation, and the trade off between inflation and real depreciation (suggested by the portfolio view of FE) are important for explaining the long run behaviour of both DE and CE. In addition, CE is in the long run affected by banks behaviour aimed at matching their foreign currency positions, which in effect implies that banks limit their exchange rate risk by shifting it to borrowers and by lending in foreign currency. In the short run, DE and CE are characterised by a high degree of persistence that seems to prevent the disequilibrium adjustment from taking place. Exchange rate and exchange rate volatility changes do not affect DE and CE in the short run, but inflation changes do, as well as banks behaviour aimed at matching foreign currency positions. Threshold effects that could only be detected in more parsimonious models of DE and CE suggest that the adjustment of disequilibrium from previous periods is discrete and asymmetric. In most cases, it is faster in regimes characterised by the FE level being lower than what would have been suggested by the fundamentals, and slower in regimes with the FE level being above equilibrium values. This finding might provide an explanation for the observed FE behaviour not only in Croatia, but elsewhere in the region. Namely, it took over a decade of stabile economic performance for FE to very gradually subside, but after the unfavourable change of economic conditions in 2008 it swiftly shot up to its original higher levels. As far as the policy recommendations are concerned, tackling FE in Croatia is quite challenging and characterised by a high degree of uncertainty. The main dilemma faced by policy makers is whether to go ahead with an across-the-board de-euroization of the financial system (ideally a symmetric one in order to avoid a build-up of currency mismatches) or to work towards a swift Euro adoption once Croatia joins the EU. However, the latter option is highly uncertain not only due to political and international economic factors, but also because Croatia is likely to have serious problems with fulfilling the Maastricht criteria related to fiscal deficit and public debt. If however, policy makers choose the Euro adoption, than they should accept euroization as a fact and manage its risks, primarily using regulation. In the opposite case, if Euro adoption ceases to be a viable exit option, policy makers will have to address very high and persistent FE. In order to reduce FE, Croatia and other post-transition European countries with high DE and CE first have to reform macroeconomic policies and institutions in order to increase their credibility. Zettelmeyer, Nagy and Jeffrey (2010) suggest that de-dollarisation in Latin American countries only began to fall after countries introduced credible macroeconomic policies based on floating exchange rate and inflation targeting regimes. Importance of credible macroeconomic policies is confirmed also in post-transition Europe where two countries with the oldest and most established floating exchange rate and inflation targeting regimes (Poland and the Czech Republic) have the lowest level of FE. Although credibility of macroeconomic policies and institutions was not controlled for in our empirical analysis, it is likely that it has to rise if policymakers want to reduce FE in Croatia. Fiscal adjustment will also be necessary not just in order to increase fiscal credibility and meet the Maastricht debt and deficit criteria, but also to make inflation targets credible over the longer term. Given that the empirical analysis results suggest exchange rate developments and exchange rate volatility influence deposit and credit euroization in the long run, it is essential that derivatives markets allowing borrowers to hedge against currency risk are developed. These markets would allow borrowers to hedge open foreign currency positions at affordable prices, thus reducing the overall sensitivity to exchange rate changes that promotes FE. The consequent elimination of the exchange rate risk should especially matter for lowering borrowing in foreign currency generated by the corporate sector. Besides developing derivatives markets, Croatian authorities should also consider developing further local currency money and promoting local currency bond markets. At the moment, local currency money market is relatively shallow, while most of the corporate and government bonds issued at the domestic capital market are indexed to foreign currency. REFERENCES Basso HS, Calvo-Gonzales O, Jurgilas M (2007): Financial Dollarization: The Role of Banks and Interest Rates. ECB Working Paper, no. 748. Brown M, Ongena S, Ye_in P (2009): Foreign currency borrowing by small firms. Working paper Swiss National Bank, no. 2009-2. Brown M, Kirschenmann K, Ongena S (2009): Foreign currency loans - demand or supply driven?. European Banking Center Discussion Paper, no. 2009-21. Caner M, Hansen BE (2001): Threshold Autoregression with a Unit Root. Econometrica, 69(6):1565-1596. Chan KS (1993): Consistency and Limiting Distribution of the Least Squares Estimator of a Threshold Autoregressive Model. The Annals of Statistics, 21(1):520-533. Cheung YW, Lai KS (1993): Finite-sample Sizes of Johansen's Likelihood Ratio Tests for Cointegration. Oxford Bulletin of Economics and Statistics, 55(3):313-329. Cook D (2004): Monetary policy in emerging markets: Can liability dollarization explain contractionary devaluations. Journal of Monetary Economics,  HYPERLINK "http://www.sciencedirect.com/science?_ob=PublicationURL&_tockey=%23TOC%235937%232004%23999489993%23518765%23FLA%23&_cdi=5937&_pubType=J&view=c&_auth=y&_acct=C000050661&_version=1&_urlVersion=0&_userid=4758817&md5=85243f957bfac944fc0f469440109e1f" 51(6):1155-1181. EBRD (2010): Transition Report 2010. Enders W, Siklos PL (2001): Cointegration and Threshold Adjustment. Journal of Business & Economic Statistics, 19(2):166-176. Guscina A (2008): Impact of Macroeconomic, Political, and Institutional Factors on the Structure of Government Debt in Emerging Market Countries. IMF Working Papers, no. 205. Heimonen K (2001): Substituting a Substitute Currency: The Case of Estonia. BOFIT Discussion Paper, no. 11. Ize A, Levy Yeyati E (2003): Financial dollarization. Journal of International Economics, 59:323347. Johansen S (1988): Statistical Analysis of Cointegration Vectors. Journal of Economic Dynamics and Control, 12(2-3):231-254. Johansen S (1991): Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregression. Econometrica, 59(6):1551-1580. Kokenyne A, Ley J, Veyrune R (2010): Dedollarization. IMF Working Papers, no. 188. Levy Yeyati E (2006): Financial Dollarization: Evaluating the Consequences. Economic Policy, 21(45): 61-118. Luca A, Petrova I (2008): What Drives Credit Dollarization in Transition Economies?. Journal of Banking and Finance, 32(5):858869. Neanidis KC, Savva CS (2009): Financial dollarization: Short-run determinants in transition economies. Journal of Banking and Finance, 33(10):1860-1873. Nicol GD, Honohan P, Ize A (2005): Dollarization of the banking system: Causes and consequence. Journal of Banking & Finance, 29(7):1697-1727. Rajan R, Tokatlidis I (2005): Dollar Shortages ad Crises. International Journal of Central Banking, no. 2: 177-220. Rennhack R, Nozaki M (2006): Financial dollarization in Latin America. IMF Working Paper, no. 7. Rosenberg C, Tirpk M (2008): Determinants of Foreign Currency Borrowing in the New Member States of the EU. IMF Working Paper, no. 173. Savastano M (1996): Dollarization in Latin America: Recent evidence and some policy issues. IMF Working Paper, no. 4. Stix H (2010): Euroization: what factors drive its persistence? Household data evidence for Croatia, Slovenia and Slovakia. Applied Economics, 42(1-16). Tong H (1983): Threshold Models in Non-linear Time Series Analysis. New York Springer Verlag, no.21. Vizek M (2006): Econometric Analysis of Monetary Transmission Channels in Croatia. Privredna Kretanja i Ekonomska Politika (Economic Policy and Economic Trends) 109: 2861. Wane, A, Gilbert S, Dibooglu S (2004): Critical Values of the Empirical F-Distribution for Threshold Autoregressive and Momentum Threshold Autoregressive Models. Department of Economics, Southern Illinois University, Discussion Paper 13. Zettelmeyer J, Nagy PM, Jeffrey S (2010): Addressing Private Sector Currency Mismatches in Emerging Europe. EBRD Working Paper, no. 115. TABLES Table 1. Long run coefficients and short run adjustment factors (deposit euroization) ModelVariable in cointegration vectorBefore imposing restrictionsAfter imposing restrictions#Testing restrictionTesting weak exogeneity=0 Chi^2(1)=0 Chi^2(1)Deposit euroization with real effective exchange rateDeposit euroization1.000.0221.0000.0017.3 [0.00]**1.54 [0.21]Exchange rate volatility-0.076-1.18-0.0740.0022.3 [0.00]**1.37 [0.24]REER-0.9340.093-1.0000.1137.1 [0.007]**14.9 [0.00]**M10.141-0.1650.146-0.2289.5 [0.002]**5.2 [0.022]*Inflation-0.5210.042-0.5460.0416.6 [0.01]*15.0 [0.00]**Minimum variance portfolio0.584-0.6540.539-0.77332.9 [0.00]**24.5 [0.00]**Constant2.362-2.51-5.8 [0.016]*-Deposit euroization with nominal exchange rateDeposit euroization1.000.0241.0000.0255.6 [0.02]*3.6 [0.05]Exchange rate volatility-0.145-0.200-0.1510.0024.6 [0.00]**0.07 [0.78]Nominal exchange rate##-2.4280.070-2.4090.07219.5 [0.00]**24.0 [0.00]**M10.238-0.0850.2080.0010.5 [0.00]**2.3 [0.13]Inflation-1.0810.026-1.0000.0258.2 [0.004]**8.5 [0.004]**Minimum variance portfolio0.512-0.3420.539-0.28611.3 [0.00]**8.1 [0.004]**Constant4.078-3.261-17.2 [0.00]**-Note:  - cointegrating coefficient;   adjustment coefficient; p-values presented in brackets; ** - null hypothesis rejected at 1 percent significance level; * - null hypothesis rejected at 5 percent significance level. # In the case of the model with REER, restrictions are accepted with Chi^2(3) = 3.23 [0.36], while in case of the model with the nominal exchange rate the restrictions are accepted with Chi^2(3) = 3.0531 [0.3835]; ## The restriction that the long run coefficient of the nominal exchange rate in a restricted model is equal to -1 is rejected with Chi^2(5) = 14.8 [0.01]. Source: Calculation of the authors. Table 2. Long run coefficients and short run adjustment factors (credit euroization) ModelVariable in cointegration vectorBefore imposing restrictionsAfter imposing restrictions#Testing restrictionTesting weak exogeneity=0 Chi^2(1)=0 Chi^2(1)Credit euroization with real effective exchange rateCredit euroization1.0000.0591.0000.0593.9 [0.048]*7.7 [0.006]**REER##-3.6970.038-3.6690.0375.8 [0.02]*4.9 [0.03]*Bank foreign debt-0.269-0.181-0.261-0.1887.7 [0.006]**4.3 [0.04]*Exchange rate volatility-0.0531.764-0.0350.0005.1 [0.02]*3.3 [0.07]Minimum variance portfolio 1.174-0.2691.187-0.31520.1 [0.00]**6.8 [0.009]**M4/M10.5990.0220.6040.0001.9 [0.17]0.3 [0.61]Constant6.880-6.770-2.8 [0.09]-Credit euroization with nominal exchange rateCredit euroization1.000-0.0981.000-0.1616.4 [0.01]*17.8 [0.00]**Exchange rate-3.5230.014-1.0000.0326.3 [0.01]*1.1 [0.30]Bank foreign debt-0.4320.074-0.3400.00013.2 [0.00]**0.7 [0.41]Deposit euroization-3.5010.010-2.6380.03116.8 [0.00]**0.2 [0.7]Average past inflation-0.252-0.0060.515-0.0080.1 [0.75]17.6 [0.00]**Minimum variance portfolio-1.1030.371-0.8060.51833.3 [0.00]**9.5 [0.00]**Exchange rate volatility-0.0320.375-0.0080.0002.9 [0.09]0.1 [0.74]M4/M11.057-0.1100.0820.0002.2 [0.14]5.1 [0.02]*Constant2.463-0.929-2.4 [0.13]-Note:  - cointegrating coefficient;   adjustment coefficient; p-values presented in brackets; ** - null hypothesis rejected at 1 percent significance level; * - null hypothesis rejected at 5 percent significance level; # In the case of the model with REER, restrictions are accepted with Chi^2(2) = 3.36 [0.19], while in the case of the model with the nominal exchange rate the restrictions are accepted with Chi^2(4) = 8.26 [0.08]; ## The restriction that the long run coefficient of real effective exchange rate in a restricted model is equal to -1 is rejected with Chi^2(1) = 6.01 [0.01]. Source: Calculation of the authors. Table 3. Vector error correction models Dependent variable Deposit euroizationt Deposit euroizationt Credit euroizationt Credit euroizationtConstant0.049 [0.45]0.047 [0.41]0.004 [0.25]-0.0005 [0.89]ECT_reert-10.02 [0.44]--0.052 [0.03]*-ECT_ex t-1-0.015 [0.42]--0.145 [0.00]**A1(L)Exchange_rate_volatilityt-11.49 [0.14]1.332 [0.21]0.001 [0.78]0.003 [0.42]A2(L)REERt-11.27 [0.24]--0.435 [0.08]-A2(L)Exchange_ratet-1-0.875 [0.57]-0.158 [0.57]A3(L)M_1t-11.03 [0.42]0.930 [0.52]--A4(L)Inflationt-12.03 [0.03]*1.82 [0.06]--A5(L)Minimum_variance_portfoliot-12.08 [0.028]*2.31 [0.014]*0.020 [0.65]0.047 [0.21] A6(L)Deposit_euroizationt-13.05 [0.002]**3.73 [0.0002]**-0.419 [0.002]**A7(L)M_4/M_1t-1--0.115 [0.28]0.087 [0.30]A8(L)Bank_foreign_debtt-1--0.103 [0.06]0.181 [0.00]**A9(L)Average_past_inflationt-1---4.198 [0.002]**A10(L)Credit_euroizationt-1--0.479 [0.002]**-0.007 [0.96]Number of lags121211Number of observations1611614343sigma0.00270.00280.00660.0050R^20.690.670.590.77AR test1.06 [0.39]1.22 [0.30]2.36 [0.09]1.33 [0.28]ARCH test0.31 [0.94]1.76 [0.41]0.30 [0.83]1.20 [0.33]RESET test2.13 [0.15]0.33 [0.56]0.62 [0.44]1.13 [0.30]Note: p-values presented in brackets; ** - null hypothesis rejected at 1 percent significance level; * - null hypothesis rejected at 5 percent significance level; ECT_reer error correction term from the restricted cointegration vector for deposit euroization with REER as an endogenous variable; ECT_ex error correction term from the restricted cointegration vector for deposit euroization with nominal exchange rate as an endogenous variable; lag length chosen according to SBIC - Schwartz Bayesian Information Criterion; EMBED Equation.3 is the first order polynomial in the lag operator L; statistics corresponding to  EMBED Equation.3  refers to F statistics and associated p-value of block exclusion restriction on all lags of an individual variable. Source: Calculation of the authors. Table 4. Threshold cointegration (quintvariate cases) Dependent variableCredit euroizationEndogenous variablesNominal exchange rate, exchange rate volatility, average past inflation, constantNominal exchange rate, exchange rate volatility, bank foreign debt, constantNominal exchange rate, exchange rate volatility, inflation, bank foreign debtREER, exchange rate volatility, average past inflation, constantM-TARM-TARM-TARM-TARThreshold =0.00540.00320.00670.0046 EMBED Equation.3 =-0.205-0.179-0.227-0.246 EMBED Equation.3 =-0.491-0.536-0.520-0.469Tmax-1.70-1.63-1.89-2.02 EMBED Equation.3 11.496**14.451***12.127**10.530** EMBED Equation.3 5.81 [0.02]**9.02 [0.00]***5.74 [0.02]**4.00 [0.045]**AR test0.24 [0.87]0.87 [0.47]0.28 [0.84]0.12 [0.95]Number of lags4444Number of observations43434343Note:  EMBED Equation.3  and  EMBED Equation.3  denote the adjustment parameters; p-values presented in brackets; lag length chosen according to SBIC - Schwartz Bayesian Information Criterion; *** - null hypothesis rejected at 1 percent significance level; ** - null hypothesis rejected at 5 percent significance level; * - null hypothesis rejected at 10 percent significance level. Source: calculation of the authors. Table 5. Threshold EC model (quintvariate cases) summary of estimation results Dependent variableCredit euroizationEndogenous variablesNominal exchange rate, exchange rate volatility, average past inflation, constantNominal exchange rate, exchange rate volatility, bank foreign debt, constantNominal exchange rate, exchange rate volatility, inflation, bank foreign debtREER, exchange rate volatility, average past inflation, constantM-TARM-TARM-TARM-TARThreshold =0.00540.00320.00670.0046Constant-0.001 [0.73]-0.0004 [0.81]-0.0006 [0.87]-0.0013 [0.71] EMBED Equation.3 -0.262 [0.01]**-0.150 [0.088]*-0.248 [0.01]**-0.266 [0.007]*** EMBED Equation.3 -0.177 [0.03]**-0.178 [0.021]**-0.181 [0.03]**-0.177 [0.03]**A1(L)REERt-1#---2.27 [0.12]A1(L)Exchange_ratet-11.80 [0.18]2.57 [0.09]*1.19 [0.32]-A2(L)Exchange_rate_volatilityt-11.88 [0.17]0.36 [0.70]0.72 [0.50]2.47 [0.10]A3(L)Average_past_inflationt-10.28 [0.60]--0.26 [0.62]A4(L)Inflationt-1--0.61 [0.55]-A5(L)Bank_foreign_debtt-1-0.78 [0.47]1.30 [0.29]-A6(L)Credit_euroizationt-123.11 [0.00]***20.39 [0.00]***22.95 [0.00]***24.80 [0.00]***R2 0.680.600.750.71Number of lags of explanatory variables2222Number of observations43434343AR test0.60 [0.62]0.80 [0.50]0.60 [0.62]0.85 [0.48]ARCH test0.19 [0.90]0.22 [0.88]0.51 [0.68]0.17 [0.91]Note: #numbers represent F statistics and the corresponding p-values of the Granger causality test for the respective variable; p-values are presented in brackets; lag length chosen according to SBIC - Schwartz Bayesian Information Criterion; *** - null hypothesis rejected at 1 percent significance level; ** - null hypothesis rejected at 5 percent significance level; * - null hypothesis rejected at 10 percent significance level. Source: calculation of the authors. Table 6. Threshold cointegration for DE (bivariate cases) summary of estimation results Dependent variableDeposit euroizationExplanatory variableNominal exchange rateREERExchange rate volatilityType of testM-TARM-TARTARThreshold =-0.0030-0.0029-0.1048 EMBED Equation.3 =-0.017-0.016-0.048 EMBED Equation.3 =0.0430.038-0.161Tmax3.052.78-1.25 EMBED Equation.3 8.133**7.061*7.970** EMBED Equation.3 14.51 [0.00]***12.39 [0.00]***3.94 [0.047]**AR test0.68 [0.69]0.59 [0.76]1.51 [0.17]Number of lags441Number of observations161161161 Note:  EMBED Equation.3  and  EMBED Equation.3  denote the adjustment parameters; W denotes the Wald test; p-values presented in brackets; lag length chosen according to SBIC - Schwartz Bayesian Information Criterion; *** - null hypothesis rejected at 1 percent significance level; ** - null hypothesis rejected at 5 percent significance level; * - null hypothesis rejected at 10 percent significance level. Source: calculation of the authors. Table 7. Threshold EC model for DE (bivariate cases) summary of estimation results Dependent variableDeposit euroizationExplanatory variableNominal exchange rateREERExchange rate volatilityType of testM-TARM-TARTARThreshold-0.0030-0.0029-0.1048Constant0.00007 [0.78]0.0002 [0.53]-0.0003 [0.43] EMBED Equation.3 -0.016 [0.03]**-0.020 [0.007]***-0.007 [0.28] EMBED Equation.3 0.024 [0.18]0.044 [0.006]***-0.007 [0.37]A1(L)REERt-1#-0.99 [0.47]-A1(L)Exchange_ratet-10.83 [0.62]--A2(L)Exchange_rate_volatilityt-1--1.13 [0.35]A4(L)Deposit_euroizationt-13.52 [0.0002]***3.60 [0.0001]***5.42 [0.0000]***R2 0.400.440.36Number of lags of explanatory variables121210Number of observations161161161AR test1.20 [0.31]2.04 [0.057]*0.65 [0.72]ARCH test0.39 [0.91]0.45 [0.87]0.63 [0.73] #numbers represent F statistics and the corresponding p-values of the Granger causality test for the respective variable; p-values are presented in brackets; lag length chosen according to SBIC - Schwartz Bayesian Information Criterion; *** - null hypothesis rejected at 1 percent significance level; ** - null hypothesis rejected at 5 percent significance level; * - null hypothesis rejected at 10 percent significance level. Source: calculation of the authors. APPENDIX Data description and sources VariableSourceDescriptionDeposit euroizationCroatian National BankShare of foreign currency deposits in total deposits.Credit euroizationCroatian National BankShare of foreign currency loans and loans indexed to a foreign currency in total loans.Real effective exchange rateCroatian National BankThe index of the real effective exchange rate is a weighted geometric average of the index of bilateral exchange rates of the kuna adjusted for the relevant relative price indices. Producer price index is used as a deflator, (2005=100). Exchange rateCroatian National BankAverage periods nominal HRK/EUR exchange rate.Exchange rate volatilityCroatian National BankMonthly average of daily exchange rate volatility given by a ratio of standard deviation and average daily exchange rates in four months prior to the observed period.M1Croatian National BankNarrow money.M4/M1Croatian National BankRatio of M4 over M1 monetary aggregate.InflationCroatian National BankCroatian Consumer Price Index, (2005=100).Average past inflationCroatian National BankIn each contemporaneous period calculated as an average of the Croatian Consumer Price Index (2005=100) two years prior to the contemporaneous period.Minimum variance portfolioCroatian National BankMinimum variance portfolio calculated as:  EMBED Equation.DSMT4  where cpi stands for Croatian Consumer Price Index and reer for Croatian real effective exchange rate.Bank foreign debtCroatian National BankBanks gross external debt.REER cycleCroatian National Bank and Central Bureau of StatisticsCorrelation between real effective exchange rate (reer) and the business cycle represented by industrial production (ip) calculated as:  EMBED Equation.DSMT4 .  Table 1. Summary statistics for all variables ObservationsMeanStandard deviationSkewnessExcess kurtosisMinimumMaximumDeposit euroization1610.790.08-0.58145-1.084970.650.88Exchange rate1617.370.19-0.560380.028276.907.70REER161101.955.67-0.26418-0.4337088.89112.41Exchange rate volatility1610.730.511.747673.438910.062.77Inflation16196.2811.750.07354-0.8727874.94116.21M116131041.914673.30.09799-1.4145811274.9555832.98M4/M11614.100.300.567490.454453.314.90Average past inflation16192.9811.72-0.01836-0.8685671.97114.86Minimum variance portfolio1610.200.170.36107-0.00834-0.270.63REER cycle161-0.100.270.11923-0.16704-0.770.56Credit euroization430.7530.077-0.35290-0.820920.6080.86Bank foreign debt436446.333201.34-0.23897-1.678142035.2210724.16 Table 2. Critical values of F statistics used for testing TAR and M-TAR threshold cointegration (null hypothesis  EMBED Equation.3 = EMBED Equation.3 =0) with %&38JLhiln  ) . w  [ k m ǻ|hRfhC]mH sH hC]mH sH hRfhRf]mH sH hRf]mH sH hRfhK]mH sH hK]mH sH hRfhR\]mH sH hRfhR\mH sH h)tmH sH hCmH sH  hQ0JhQhCfh ChAhh$/jklmnC D z78F $7$8$H$a$gdf0Z$a$gd;M$a$gd$a$gdK$a$gdG$ $7$8$H$a$gdRf$a$gdN [      D c B C D   F H U {qmie^Z^Z^Z^eZhv h hvh>HxhVhjh 0JUh{h?cmH sH h;h{mH sH hG$hG$PJhN) hG$hG$hr hG$hrhG$hR\hG$mH sH hRfhR\mH sH hRfhR\]mH sH hRf]mH sH hRfhRf]mH sH hRfB*phhRfhRfB*ph U t v $-./0 '.06@AFUZbijȼȼĴĪĦhTU7jh0JUh1RrhC hh&h@hKh\ h hQ hIhQh)th>Hx h hNhNhvhQ h YhQA !+-Jpxyz)+>]_')5678BLlv̼ظرЛ h6h h6h7R h6hk!Hhk!Hh> hgh=n hgh hgh>hDh>Hxh*h}}h`^hnqdhhVhhAmhh hhNhTU7hQh@h 4FHJPz QXZ_aeEF`ټxmh=nh6mH sH h=nhIQmH sH h>HxhgmH sH jh~0JUmH sH hC h>Hxh[K h>HxhIQ h6hIQ h6hg h6hh=s h6hsh7R h6hD h6hB/xhQ h6h  h6hn u h6hk!H h6h*^b}~;=goz{  DEF}hJ`hg h=nh h=nh h!> h=nh}}hI h=nhgh]sQh)tho h=nh=nhf0Z h=nhIQh!omH sH h)tmH sH h=nhgmH sH h=nhgdmH sH hf0ZmH sH h=nhIQmH sH .FNQW`os%5<DHaeijr ѻƚ珄yncnhqh<mH sH hqh+{mH sH hqh^LmH sH h+{hmH sH h+{h^LmH sH h+{hsmH sH h+{hH mH sH h+{h)tmH sH h+{h]sQmH sH h+{h mH sH h+{h!omH sH h+{h}}mH sH h+{hmH sH  h+{h h+{h2$ !%&?HObf#ABRVlvwx| ( , 9 V ƺƺƺƺƨxxoxxh<]mH sH hfR]mH sH h#]mH sH h=s]mH sH hN]mH sH h+{hS\mH sH hS\mH sH h1~\mH sH h+{h+{\mH sH hf0Z\mH sH hqhR2\mH sH hqhf0Z\mH sH hqh<\mH sH hqh+{\mH sH +V ^ _ ` a !!?!!!!!!!""" ","2"R"T"i"}"""""""""~#### $$º²ͪ͢{wsokh.-h%Xh ihMgh h$rKhxPh)phy h}}h}}h hb4&h?ch B*phh<mH sH hrmH sH hM*mH sH hhwIhhwImH sH hhwIhqmH sH h]mH sH hShk!H\mH sH hN]mH sH h ]mH sH *F` a !!""&&*-//1Z:=C $7$8$H$a$gd $7$8$H$a$gd#0$a$gd+{ $7$8$H$a$gd[K $7$8$H$a$gdMg $7$8$H$a$gdS$a$gdhwI$a$gd $a$gdq$a$gdG$$a$gdj|$V$^$n$$$$$$$$$$$$$%*%4%I%N%P%X%Z%`%b%%%%%&&&$&w&&&&&&&&&ⷜ┿✿}sish~OJQJ^JhtOJQJ^J h}}h}}hxPhKhMgmH sH hC[mH sH hh]LhmH sH h~$mH sH h'ZmH sH hSmH sH hh]LhKmH sH h mH sH hCmH sH hAmH sH hh]LhxPmH sH hh]Lh$rKmH sH h.-h$rK)&&&)'?'X'Y''''''''''''''(.(0(1(B(O(Z([(\(b(c(m(}((((ù⋁ìzszlele^Z^elezeh h h h h> h hh]L h hK h h h OJQJ^Jh h OJQJ^JhMgOJQJ^Jh ?OJQJ^Jh h3:OJQJ^JhOJQJ^Jh{:?OJQJ^JhthtmHsHh!>OJQJ^JhtOJQJ^JhCOJQJ^Jh~OJQJ^J#((((()E)F)G)T)b)g)))))))*********N+Q+R+T+d+e++++ѩzslses^s hMghyc hMgh hMgh6 hMghschMgh3:OJQJ^JhOJQJ^Jh h3:OJQJ^Jh h>B*ph# h h h h3:h:OJQJ^Jh~h`WbhhJ` h h>h.fOJQJ^Jh'ZOJQJ^Jh h@OJQJ^J h h@#++++++++++++,,,, ,$,0,1,W,Z,b,j,k,l,x,z,,,,,,,,,,-̊~rk hMghhMgh3:]mH sH hMghK]mH sH h]mH sH hMghK]mH sH hMgh=s]mH sH hS]mH sH hMghgd]mH sH hMgh ]mH sH hMgh6]mH sH hMgh6] hMgh hS hMgh6 hMghsc hMghL?%--1-2-y-----]mH sH h+{hpD]mH sH h+{h{:?]mH sH h+{h<]mH sH h+{hC[]mH sH h+{h)p]mH sH h+{hL]mH sH h+{h K]mH sH h+{hgd]mH sH h)phgd]mH sH 111111Y2p2X3d3e3t3v33333333333333344647484¶vmdXh(8hj]mH sH h#0]mH sH h?r]mH sH h9=]mH sH h]mH sH hh)p]mH sH h(8hn_q]mH sH hn_q]mH sH h(8h>]mH sH h(8hC[]mH sH h(8h~]mH sH h]]mH sH hRG]mH sH h>yO]mH sH  h>yO] h+{hgd h+{hIh+{8494?4H4I475=5Z5f5i5~555555555 6 66666$6)626465666>6ɷɫɢɋxmbmbmWhMhMmH sH hMhZTmH sH hMhfmH sH hmH sH h(8hfmH sH h(8hZTmH sH h(8hf]mH sH hM]mH sH h(8h(8]mH sH h]mH sH hP]mH sH h(8hJa]mH sH h(8hU]mH sH h?r]mH sH h(8hj]mH sH h]mH sH >6?666666667777777778 848=8>8B8T8[8\8ȿȶxmx`SKhmH sH h(8h(8PJmH sH h(8h KPJmH sH h(8h mH sH h(8h KmH sH hMh]mH sH hMh KmH sH hMh mH sH h]]mH sH hU]mH sH hM]mH sH h]mH sH hMhM]mH sH  hMhMOJQJ^JmH sH hmH sH hMhMmH sH hMmH sH \8u8x888888888P9R9999(:/:6:Y:Z:^:_:e:i:k:m:::~rir]iT]Ih(8h)qmH sH hn_q]mH sH h(8h)q]mH sH hb]mH sH h(8h~]mH sH hMhPPJmH sH h&PJmH sH h]PJmH sH hCaAPJmH sH hbJPJmH sH h#0PJmH sH h(8h PJmH sH hn_qmH sH hn_qhn_qmH sH hmH sH h#0mH sH hmH sH h(8h(8mH sH :::;;;;;>;;;;;;<%<&</<3<9<A<H<X<^<g<p<<<<<<<+=Q=U=f=v=z=ױxxphp]hh'ZhNmH sH h{:?mH sH hNmH sH h'Zh'ZmH sH hmH sH h'Zh(8mH sH h(8mH sH h(8h[OmH sH hmH sH h~@mH sH h)pmH sH h(8h-AmH sH h(8hhqmH sH hhqmH sH h(8hmH sH h(8h)qmH sH h(8h+A mH sH $z===== >>>%>)>*>5>6>[>_>>>>>>@@@@hAAAAAVBWBXB[B`Bڴ}qe\eqh$]mH sH hmh$]mH sH hmhm]mH sH hmh%{]mH sH hmhM]mH sH hmh~]mH sH h(8h]mH sH hhmH sH h]mH sH h'Zh(8mH sH h'ZhA$mH sH h%fmH sH hmH sH h'Zh'ZmH sH hNmH sH h'ZhNmH sH !`BeBnBoBC!CCACKCWCbCcCgCCCCCCCCCCCCD2D3DHDSDTDUDDDDļ{skck[k[k[k[hB{mH sH hmH sH hdmH sH hWLfmH sH hlMmH sH hLmH sH h;mH sH hx9hQmH sH hQmH sH hQ hmhhwI hmhb4&h?ch KmH sH hmhmmH sH h$\mH sH hmhm\mH sH h%{]mH sH h]mH sH h)p]mH sH #=C>CbCcC H HKMLMNNNNZR[RnRoRR$a$gd'$ & F h^`a$gd'$a$gdB{$a$gdb4&$a$gdU0$a$gdQ$a$gdq$a$gdBO$a$gdMN$a$gdhwI $7$8$H$a$gdb4&DDDEEEJEOEPEUEEQFF G GGG4GOGkGGGGGGGGGHH H H Hؽ{rj_Th\ h&cmH sH hYh-amH sH hM@mH sH hM@\mH sH h'\mH sH hYhY\mH sH hqhY\mH sH hqh'mH sH hqhWLfmH sH hqhYmH sH hYhYmH sH hC mH sH hWLfmH sH hYmH sH h?*mH sH hB{mH sH hA$mH sH hdmH sH  HHHH&H7H=HEHOHUHaH|HHHHHHHHHIII$I,I-I.IYI^I`IbIIIIIIIIJJJ*J/J7JCJDJߏ߳yߨh\ hRtmH sH h\ h'mH sH h'mH sH !jh\ hRt0JUmH sH h\ hpbmH sH h\ hzmH sH h\ h?*mH sH h\ h mH sH h\ hB{mH sH h\ h;mH sH h\ hQmH sH h\ hlMmH sH -DJHJZJ[JpJJJJJJJJJJJJJJJJJJKKK KKKKKKKKLL$L%L&L*L:LILaL~LLLɾԳꏳ߄yh\ hMgmH sH h\ h]mH sH h\ h%mH sH h\ hB{mH sH h\ hyhImH sH h\ hymH sH h\ hgmH sH h\ h;mH sH h\ hImH sH h\ hpbmH sH h\ hQmH sH h\ hmH sH ,LLLLLLLLLLLLLLLLLLLLLLLM MM#MJMKMLMMMNNNNNNNȽ~vrnjbhB{mH sH hB{hvhU0h;mH sH hsmH sH hs h\ hsh\ h'ewmH sH h\ h\9mH sH h\ hmH sH h\ hMmmH sH h\ hQmH sH h\ hpbmH sH hqmH sH h\ hsmH sH hC mH sH h\ h;mH sH  h\ hMg%NNEOIOOOOOP!PUPPPPPPQQLQNQQQQGRLRYRZR[RnRtRRRRRRSSESMSOShSiSSSST2TEThTTTTTTTŽ𽹵hiEh?*h.h h[h.8hG>hhKh2`QhMgh'hMh'mH sH hP{mH sH hsdhsdmH sH h(mH sH h1RrmH sH hsdmH sH hB{mH sH hMgmH sH 6RRTWYYZ[[9\Y\w\x\g]]]] ^`gdp!$^`a$gdp!$a$gd%$$a$gdOe$a$gdiE$a$gdOe$a$gd?$a$gdAb$a$gdv$a$gd|$a$gd.$a$gdlMTTTTTT8UjUUUZV[V\VcVxVVV)W*WOWPWSW_WWWWWWWW"X#XxXXXXXYYYYMYNYYYYYYYYYY#ZCZLZSZlZZZhOJQJ^JhhOJQJ^JhvOJQJ^Jhv hhh hh=|h=|hshMgh6Nh[h|h2`Qh h.hG>h.8h&hEh :ZZZZZZZZZ [[ [#[$[,[0[4[5[6[@[N[T[U[^[e[z[[[[[[[[[\\)\ǽdzǜǒǜǜǜLjtmiea]aehAbhRh?*h? hh?jh:$h?0JU h:$h?hAbOJQJ^JhKOJQJ^Jhh?OJQJ^JhQUOJQJ^JhvOJQJ^JhOeOJQJ^Jh?OJQJ^Jh&hEhhOJQJ^Jh.OJQJ^JhOJQJ^JhvOJQJ^J$)\7\8\X\Y\Z\m\n\o\p\s\t\u\x\~\\\\\\\\\\\\\\\\\\\\\]̿߻}}k^jhh?EHU#j@N hh?CJUVaJh?*h%$j%hh%$EHU%jHP h%$CJUVaJmHsH hh%$jhh%$UhKhp!jhiEhiEEHU%jWP hiECJUVaJmHsHjhh?Uh? hh? h*SJh?hR#]]]]e]f]g]h]i]|]}]~]]]]]]]]]]]]]]]]]]] ^ ^ ^ ^^^%^γƢ΢΢ΏƢpcjShh?EHU#jWN hh?CJUVaJj8hhp!EHU%jP hp!CJUVaJmHsHhRj% hhiEEHU%jP hiECJUVaJmHsHhp!hiE hh?jhh?Uj hh?EHU#jN hh?CJUVaJ$%^&^'^(^L^M^N^a^b^c^d^v^y^^^^^^^^^^^^^^^___6_E_J____ʯΞΚΖreaa]hKhyjhhOeEHU#jN hhOeCJUVaJ hhOejhhOeUh%$h?hOejhRh__EHU%jP h__CJUVaJnH tH jhRUhR hh?jhh?Ujrhh?EHU#jwN hh?CJUVaJ#__________```````"`#`6`7`8`9`U`ɶ}r[IAhKB*ph"j!hwhOeB*EHUph,jN hwhOeB*CJUVaJphhwhvB*ph"jhwhOeB*EHUph,jPN hwhOeB*CJUVaJphjhwhOeB*UphhwhOeB*phhKhOe hhOehyjhhOeUjh~h~EHU%jsP h~CJUVaJmHsHU`c`l`m`n`o`~```````````a!a"a5a6a7a8a9a;a>aMaVaZa|aaaļĢᗇlZ"jm#hwh-MB*EHUph4jP hwh-MB*CJUVaJnH phtH jhwh-MB*Uphhwh-MB*phjhw0JB*Uphhwh60B*phhwB*phhwhwB*ph"jhwhOe0JB*UphhwhOeB*phhKB*phhwhKB*ph aaaaaaaaaaaaaabb'b(bjbbbbXccccd;dB*phh|oB*phh;MhA B*ph h;Mh h;MhK h;MhhNuhhhBI6hLhBhhU'h?chv hh?hwhOeB*ph"jhwh-M0JB*Uph&]aaaah\k]kjkkk"m#mMprrrr0t1t$a$gdc $7$8$H$a$gdc$a$gdc$a$gdLH $7$8$H$a$gdjx$a$gd > $h^ha$gdB$a$gdv$a$gdOe$a$gdwddWe[ereseeeeeeeeeeeeeeeef.f;fAfSfhfsfwf~fffffffffffffffٺ|tph5vhbB*phhB*phh]3B*phh;Mh]3B*phhKhwh]3hC h|o h;MhA hh;gt h;Mh;MhNuB*phh >B*phh;MB*phh;Mh;MB*phh;MhA B*phh|oB*phh5vB*ph+ffggg*g/g5g6g7g:g;gB*phhB*phhBI6B*phhB*phhhU'B*phh5vB*phhP{hv h5vh5vh]3h3ZB*phhbB*phh;Mh]3B*phh]3B*phhB*phh5vh3Z)yiiiiijl?lllllllmmmm m!m"m#mymmmmmmmmmmmwnnno~oֿηjhc0JU h)hchch]3h@h >B*phh3ZB*phjh0JB*UphhB*phhB*phhB*phh@h@B*phh@hBB*phhB:~oooooooooo*p,pJpKpLpMp\p_ppppppppqq rr8rWrdrrrrrrrrrrrrss+s.s5ssst%t/t0t1tFtGtHt[t\tt.u@uTuXuuuuΪh)hcCJaJh?rhcCJaJ hjxhc hhch@hcCJaJh?rh?rCJaJjh)hc0JUhch?r h)hcC1tGtHtHxIx||>jkЋef|}%& $dha$gdc$dh7$8$H$a$gdc $7$8$H$a$gdc$a$gdc$a$gdcuupvvvvvvvvvwwwwwwxxFxGxHxWxxxynyyyyyyyyyyEzIzZzzzzzzzz {,{7{X{a{l{t{w{{{{{{{||||||||||}}lhhhc]mH sH h(8hc]mH sH  h0thch?r hV4Uhc hhc h)hchcJlF_ÀĀ(br҄ԄЅ҅÷榛֦}rg\jh)hcUhhhcCJaJhjhcCJaJhhhcB* CJaJph h)hchhmH sH h(8hcmH sH hcmH sH hc\mH sH h:hc\mH sH  h:hchThc]mH sH h(8hc]mH sH hhhchc]mH sH hmhc]mH sH $҅+EJKQֆ׆ILpqtćƇՇևއ :<ÈLjЈԈ܈al46;@JjkȾϺϺϳϺϺϬϬϡhhhcCJaJ hf0hc hE hchhjhc0JU h;fhchcjh)hcUj%h)hcEHUjP h)hcUV h)hc< $JKl͋΋ϋЋ #-09\̌͌Όό (3CGehl !+4Cv)4MQdef{|}߾涫haJhcCJaJhcCJaJh67hcCJaJh67 h1hch)hcB*ph h)hchchhhcCJaJhhcCJaJCO_sݑ0125=>lϒҒKZ_p˓̓ "$%&<=RSTkw˔ 67ڗQǿǴǴǴhXgh. B*phhXgh. B*ph# h6HHB*ph# h. B*ph# h8/h. B*ph# h. hcCJaJh67hcCJaJh67 h)hchcB&<=STjk8 FG $7$8$H$a$gd8 $h^ha$gd8$a$gd8 $7$8$H$a$gd. $a$gd. $a$gdc$a$gdcĘȘ ͚ӚjnZ\xҝӝX`aekmnBFѾx$jhh. 0JU\mH sH hh. \mH sH hU>\mH sH h6HH\mH sH h. \mH sH ht]h. \mH sH ht]h. B*ph# h6HHB*ph# h. B*ph# hXgh. B*ph# h. B*phhXgh. B*phh6HHB*ph+џޟ0134)1BiʡΡҡӡݡYZǢˢТۢ"BC^h !#9>_ûûûûûִδִδδΪ֟֟ hph. hd#jh. 0JU hzh. h. B*ph# hzh. B*ph# hrNh~`h. hh. h. \mH sH hh. \mH sH hrN\mH sH >ĥǥ-7èĨ*;DFGNX @ptªЪҪԪޫvk_hbh86CJaJhWh8CJaJh86CJaJh8CJaJhbh8CJaJhbh8CJ]aJmHsH"hbh86CJ]aJmHsHh8CJaJmHsHhbh8CJaJmHsHhh8CJaJ hh8h)>hchd#h&sGh. hh. hrN%GҪԪde WX}~$-DM a$gd8$a$gd8 $7$8$H$a$gd86EQY]^de߬  #Ono!>@ADEIJWX˿˴{oh8B*CJaJph# h_+h8jh_+h8CJUaJh_+h86CJaJh_+h8CJaJh2h86CJaJh2h8CJaJhWrh86CJaJhWrh8CJaJhh8CJaJhh86CJaJhh8CJaJh8CJaJ*Xew|}~¯ ǹr`NBr7rh{@h8CJaJh8CJaJmHsH"hzh86CJ]aJmHsH"h{@h86CJ]aJmH sH h{@h8CJaJmH sH h8CJaJmH sH h8CJaJh{@h8CJaJh{@h86CJaJh{@h80JCJaJh{@h80J6CJaJhh8CJaJh&wh8CJaJmHsHhLHh8CJ]aJmHsHhLHh8CJaJmHsH7NOPr±ñ ߽qqqaqRqJh8CJaJhd1h8CJaJmHsHh\h86CJaJmHsHh\h8CJaJmHsHh8CJaJmHsHhh8CJaJhh80JCJaJhh80J6CJaJh80J6CJaJ+h{@h8CJOJQJ\^JaJmH sH h8CJaJmH sH h{@h8CJaJmH sH "h{@h86CJ]aJmH sH  Pop|Ʋز޲ҳާҘҘҘҘ҆Ҙ{o`oQQh3h8CJaJmHsHh3h8B*CJaJph# h8B*CJaJph# h3h8CJaJ"hzh86CJ]aJmHsHhzh8CJaJmHsHh8B*CJaJph# hBh8CJaJmHsHh\h86CJaJmHsHh8CJaJmHsHh\h8CJaJmHsHh8CJaJh\h8CJaJQR׳سqrwxٵڵcdڶ۶tuڷ۷ $7$8$H$a$gdc$a$gd8 $7$8$H$a$gd80?IQRųdzʳ˳ճ׳س>?]fjkpqĵxiiiWii"hsGh86CJ]aJmHsHhsGh8CJaJmHsHh8B*CJaJph# hLUh8B*CJaJph# hLUh8CJ]aJmHsH"hLUh86CJ]aJmHsHhLUh8CJaJmHsHhh8B*CJaJph# "h3h86CJ]aJmHsHh3h8CJaJmHsHh8CJaJmHsHqrxѴӴ !;=ekuwxеصEGX^acdԳŝuufhh8CJaJmHsHhph86CJaJh86CJaJh8CJaJhph8CJaJhh8CJaJhk+h8CJaJ"hk+h86CJ]aJmHsHhk+h8CJaJmHsHhk+h8B*CJaJph# hMLB*CJaJph# *hVh8B*CJaJph# #dѶڶ۶6Whjstuy{}˷ѷԷٷڷ۷.ƷƧƷ}t}i^hrw*hcCJaJhoh8CJaJh86CJaJhq8h86CJaJhq8h8CJaJh8CJaJhh8CJaJhJ,h86CJaJmHsHhJ,h8CJaJmHsHh8CJaJmHsHhh8CJaJmHsHh!Ovh86CJaJmHsHh!Ovh8CJaJmHsH.U\{*`vw}"$ºǴ|t|i|^ZSLHLAHAHA h)hchc hKhc hohch8h.9 h.9 CJaJh.9 6CJ]aJhMLCJaJh.9 CJaJhcCJaJhrw*hc6CJ]aJhrw*hcCJaJhrw*hcB*CJaJph# %hrw*hcB*CJaJmHph# sH"hc6B*CJaJmHph# sH"h1~6B*CJaJmHph# sH(hrw*hc6B*CJaJmHph# sHwxºκJܻ $$Ifa$gdc$a$gdc $7$8$H$a$gd8 $7$8$H$a$gdc]gdcºFH&4WX`lwտ޿[\d:>@Btv'(h)hcH*OJQJ^JhchcOJQJ^JnHtH hhcOJQJ^JnHtH hqhc h)hcOJQJ^JnHtHhthcCJaJh)hcCJaJhqhcCJH*aJhcCJaJhqhcCJaJ3ܻ޻kdY($$IfTl4VֈL B &&&&nn0 44 lap<T޻(*ȼԼ "Vdp~Ffg/ $IfgdcFf{+ $$Ifa$gdcʽؽ  &-3:>HLUFf7FfC3 $$Ifa$gdcUVWahnu{ž̾ھ߾FfBFf> $$Ifa$gdcFf:AUZ`flpx|ǿ̿ӿFfyJ $IfgdcFfF $$Ifa$gdcӿԿտ #$%(.5;@ENRYZ[elFfUFf#R $$Ifa$gdcFfUNlry  $IfgdcFf]FfY $$Ifa$gdcJnoJZr $$Ifa$gdc $7$8$H$a$gdc$a$gdc$a$gdcFfia :;<=Jmnouv FH!"-.klwx|·vvvvvv *h)hcCJaJhthcCJaJh)hcCJH*aJh)hcCJaJ h)hchchcCJaJh@hcCJaJ h@hcOJQJ^JnHtHh)hcH*OJQJ^JhcOJQJ^JnHtH h)hcOJQJ^JnHtH.kdc$$Ifl4Vֈ <O!&&&&nn0 44 laWp<Zh Ffj $IfgdcFff $$Ifa$gdc"&./0BIPW^blpxyzFfPrFfrn $$Ifa$gdc !'-39=DHOPQZ`Ff}Ff z $$Ifa$gdcFf.vCDNOtuvw"#-.ijtu &NO[\`f  34Խh)hcCJH*aJhhcCJaJhthcCJaJ *h)hcCJaJh)hcCJaJh)hch|$1CJaJhcCJaJhqhcCJaJA`bhjnuwx Ff $IfgdcFfց $$Ifa$gdc#'./0BIOV\ajnuvwFfbFf $$Ifa$gdc '.4;AFOS\]^w~FfFf $$Ifa$gdcFf@~ FfFf $$Ifa$gdc \.\ $7$8$H$a$gdc$a$gdcFf^ $$Ifa$gdc 456LPLMNO*,XZ 2FLXZbdjlxz˿˿˿˿˰˰˰˰˰˰˿˰˰˰˰˿˰˰˰˰˰˰h@hcCJaJmH sH h@hcCJH*aJh@hcCJaJ h)hchch)hcCJH*aJhcCJaJh)hcCJaJEE99999 $$Ifa$gdckd$$IfTlTrt "0#  t0644 laT&46N9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdcNXfjx $$Ifa$gdcE99999 $$Ifa$gdckd$$IfTlTrt "0#  t0644 laT,6D9kd7$$IfTlTrt "0#  t0644 laT $$Ifa$gdc$*,BZ\^dflrRVXhjlrtʻhDhcCJH*aJh6hcCJH*aJh@hcCJaJmH sH h@hcCJaJh8zhcCJH*aJhcCJOJQJ^JaJhcCJH*aJhcCJaJ@DP^jx $$Ifa$gdcE99999 $$Ifa$gdckdǨ$$IfTlTrt "0#  t0644 laT ,9kdW$$IfTlTrt "0#  t0644 laT $$Ifa$gdc,:>JXZt9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdct~ $$Ifa$gdcE99999 $$Ifa$gdckdw$$IfTlTrt "0#  t0644 laT Zdv9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdcv $$Ifa$gdc",BE99999 $$Ifa$gdckd$$IfTlTrt "0#  t0644 laTBFRfh9kd'$$IfTlTrt "0#  t0644 laT $$Ifa$gdc9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdc@BDJLx~*.JLPRVXZ\^bEenoӿӴ꩝ꎴ{%j=P hcCJUVaJmHsHjh)hcCJUaJhhcCJ]aJhhcCJaJh)hcCJaJhM<hcCJaJhcCJH*aJh@hcCJaJh8zhcCJH*aJhcCJaJhcCJOJQJ^JaJ. ,> $$Ifa$gdc>@E99999 $$Ifa$gdckdG$$IfTlTrt "0#  t0644 laT9kd׭$$IfTlTrt "0#  t0644 laT $$Ifa$gdc,.LR9kdg$$IfTlTrt "0#  t0644 laT $$Ifa$gdcRX\`b9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdc9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdc9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdc 9kd$$IfTlTrt "0#  t0644 laT $$Ifa$gdc ',38?DK $$Ifa$gdc$$7$8$H$Ifa$gdc KLV[bE4(( $$Ifa$gdc$$7$8$H$Ifa$gdckd7$$IfTlTrt "0#  t0644 laTbgnsz $$Ifa$gdcE4(( $$Ifa$gdc$$7$8$H$Ifa$gdckdDZ$$IfTlTrt "0#  t0644 laT $$Ifa$gdcE911$a$gdc $7$8$H$a$gdckdW$$IfTlTrt "0#  t0644 laT9:MNOP )G[\]IMԹԡ̒̒̒̒̒̒q`!j8h>JhcCJEHUaJ#jN h>JhcCJUVaJjh>JhcCJUaJh>JhcCJaJhc h)hc!jĵhZ>hcCJEHUaJ%j

hcCJEHUaJ# 3FG\Ekd$$Iflp0#) t 6`\0644 lap $$\&`#$/Ifa$gdc$$\&`#$/If^a$gdcgdcI'kd'$$Ifl4prE S#) t 6`\0644 lap $$\&`#$/Ifa$gdc$$\&`#$/Ifa$gdc:'''$$\&`#$/Ifa$gdckdڹ$$Ifl4prE S#) t 6`\0644 lap (kd$$IflrE S#) t 6`\0644 lap $$\&`#$/Ifa$gdc$$\&`#$/Ifa$gdc$$\&`#$/7$8$H$Ifa$gdcB*$$\&`#$/7$8$H$Ifa$gdckdV$$IflrE S#) t 6`\0644 laXYlmnop!:]dexy°”‚q””iW#jN h>JhcCJUVaJhcCJaJ!jOh>JhcCJEHUaJ#jN h>JhcCJUVaJhn`hcCJaJ!jOh>JhcCJEHUaJ#j,N h>JhcCJUVaJh>JhcCJaJjh>JhcCJUaJ!jh>JhcCJEHUaJ#j3 N h>JhcCJUVaJ$+29:/kd$$IflrE S#) t 6`\0644 la$$\&`#$/Ifa$gdc:?EKQW$$\&`#$/Ifa$gdc$$\&`#$/7$8$H$Ifa$gdcWXpB*$$\&`#$/7$8$H$Ifa$gdckd$$IflrE S#) t 6`\0644 lapy/kd$$IflrE S#) t 6`\0644 la$$\&`#$/Ifa$gdc$$\&`#$/Ifa$gdc$$\&`#$/7$8$H$Ifa$gdc B*$$\&`#$/7$8$H$Ifa$gdckd$$IflrE S#) t 6`\0644 la !$$\&`#$/Ifa$gdc!"1B*$$\&`#$/7$8$H$Ifa$gdckdO$$IflrE S#) t 6`\0644 la13579:/kd$$IflrE S#) t 6`\0644 la$$\&`#$/Ifa$gdc:QTWZ]$$\&`#$/Ifa$gdc]^ ]B::.: $7$8$H$a$gdc$a$gdckd$$IflrE S#) t 6`\0644 layz{') ]f89PQde±ԩԩԩԃ||ԩԩԩԩԩj#jN h>JhcCJUVaJ h)hchchhcCJ]aJhhcCJaJh@hcCJaJhcCJaJ!jGh>JhcCJEHUaJ#j3 N h>JhcCJUVaJh>JhcCJaJjh>JhcCJUaJ!j)h>JhcCJEHUaJ']p8Lkdf$$Iflw01 $  t 6`0M%44 lap $$&`#$/Ifa$gdc$$&`#$/If^a$gdc'$$&`#$/Ifa$gdckd$$Ifl4wr1  u$ llll t 6`0M%44 lap2$$&`#$/Ifa$gdc'$&`#$/Ifgdckd$$Ifl4wr1  u$ llll t 6`0M%44 lap2 $$&`#$/Ifa$gdc )kd|$$Iflr1  u$ llll t 6`0M%44 lap2 "*19@HO$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc OP)kd>$$Iflr1  u$ llll t 6`0M%44 lap2efg $NPRXZtvzrtv|~ 8:<B±ԩԩzԩԩԩԩh6hcCJH*aJh8zhcCJH*aJhcCJOJQJ^JaJhcCJH*aJhcCJaJ!jh>JhcCJEHUaJ#j3 N h>JhcCJUVaJh>JhcCJaJjh>JhcCJUaJ!jh>JhcCJEHUaJ0Phox$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc )kd$$Iflr1  u$ llll t 6`0M%44 lap2$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc  ()$&`#$/Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2(,04>L$$&`#$/Ifa$gdcLN|)$&`#$/Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2|$$&`#$/Ifa$gdc)$&`#$/Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2(2@JXbp$$&`#$/Ifa$gdcpr)$&`#$/IfgdckdE$$Iflr1  u$ llll t 6`0M%44 lap2$$&`#$/Ifa$gdc)$&`#$/Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2$26$$&`#$/Ifa$gdc68n)$&`#$/Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2BDfldh78Z\oDhiopӾӯxmaӃYRNhc h)hchcmH sH hhcCJ]aJhhcCJaJh@hcCJaJh>JhcCJaJmH sH h>JhcCJH*aJmH sH hcCJH*aJmH sH hu hcCJaJmH sH h>JhcCJH*aJhcCJH*aJh>JhcCJaJh8zhcCJH*aJhcCJaJhcCJOJQJ^JaJnr|$$&`#$/Ifa$gdc)kd$$Iflr1  u$ llll t 6`0M%44 lap2 ,@L`$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc `b)kdM$$Iflr1  u$ llll t 6`0M%44 lap2bjt~$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc)$&`#$/7$8$H$Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2$$&`#$/Ifa$gdc)$&`#$/Ifgdckd$$Iflr1  u$ llll t 6`0M%44 lap2$$&`#$/Ifa$gdc)kd$$Iflr1  u$ llll t 6`0M%44 lap2',38?DKPW$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc WX)kdU$$Iflr1  u$ llll t 6`0M%44 lap2Xbgnsz$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc Dh) $7$8$H$a$gdckd$$Iflr1  u$ llll t 6`0M%44 lap2hi5kd$$Iflp0 t 6`044 lap $$&`#$/Ifa$gdc$$&`#$/If^a$gdcgdc $7$8$H$a$gdcp5Axy.➍{j_hn`hcCJaJ!jMh>JhcCJEHUaJ#j,N h>JhcCJUVaJ!jh>JhcCJEHUaJ#j3 N h>JhcCJUVaJ!jKh>JhcCJEHUaJ#jN h>JhcCJUVaJjh>JhcCJUaJh>JhcCJaJhcCJaJhc h)hc#)45B9kd_$$Iflp\L  t 6`044 lap $$&`#$/Ifa$gdcBHNRS_g9kd$$Iflp\L  t 6`044 lap $$&`#$/Ifa$gdcgowx9kd$$Ifl\L  t 6`044 lap $$&`#$/Ifa$gdcx$$&`#$/Ifa$gdc$$&`#$/7$8$H$Ifa$gdcT:%%$$&`#$/Ifa$gdc$$&`#$/7$8$H$Ifa$gdckdi$$Ifl\L  t 6`044 la?%$$&`#$/7$8$H$Ifa$gdckd$$Ifl\L  t 6`044 la$$&`#$/Ifa$gdc?%$$&`#$/7$8$H$Ifa$gdckd$$Ifl\L  t 6`044 la$$&`#$/Ifa$gdc ?kd$$Ifl\L  t 6`044 la$$&`#$/Ifa$gdc./012a#'<UWx·¯¯¯¯¯¯•ra¯¯¯!j"h>JhcCJEHUaJ#j3 N h>JhcCJUVaJ!jh>JhcCJEHUaJ#jN h>JhcCJUVaJhMLCJaJhcCJaJhn`hcCJaJh>JhcCJaJjh>JhcCJUaJ!jFh>JhcCJEHUaJ#jN h>JhcCJUVaJ$28BHRWa$$&`#$/Ifa$gdc$$&`#$/7$8$H$Ifa$gdcabjovT:%%$$&`#$/Ifa$gdc$$&`#$/7$8$H$Ifa$gdckd$$Ifl\L  t 6`044 lav{?kd?$$Ifl\L  t 6`044 la$$&`#$/Ifa$gdc$$&`#$/Ifa$gdc$$&`#$/7$8$H$Ifa$gdcT:%%$$&`#$/Ifa$gdc$$&`#$/7$8$H$Ifa$gdckd$$Ifl\L  t 6`044 la?7777$a$gdckdm$$Ifl\L  t 6`044 la$$&`#$/Ifa$gdcgdc$a$gdcx&:QUp|df˦˔˦q`WhcCJH*aJ!jEh>JhcCJEHUaJ#j3 N h>JhcCJUVaJ!jh>JhcCJEHUaJ#jN h>JhcCJUVaJjh>JhcCJUaJhc h)hcht?hcCJaJh>JhcCJaJhhcCJ]aJhhcCJaJhcCJaJh@hcCJaJ"%&;QVdo_kdA$$Iflw0 l  t 6`044 lap $$&`#$/Ifa$gdcop}N9999$$&`#$/Ifa$gdckd$$Iflw\ hl  t 6`044 lap N<''$$&`#$/Ifa$gdc$&`#$/Ifgdckdy$$Iflw\ hl  t 6`044 lap 9dh$&`#$/7$8$H$Ifgdckd+$$Ifl\ hl  t 6`044 lap $$&`#$/Ifa$gdc?kd$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc"0FTb$$&`#$/Ifa$gdcdh$&`#$/7$8$H$IfgdcbdT:%%$$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdckd$$Ifl\ hl  t 6`044 la?kdd$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc.028:TVZ~ DF Ӳӥyh>JhcCJaJmH sH h>JhcCJH*aJmH sH h$CJH*aJmH sH hcCJH*aJmH sH h>JhcCJH*aJh6hcCJH*aJhcCJH*aJh>JhcCJaJh8zhcCJH*aJhcCJOJQJ^JaJhcCJaJ+ (,.-kd $$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc$&`#$/Ifgdc.\ftx|$$&`#$/Ifa$gdc$&`#$/Ifgdc|~TB---$$&`#$/Ifa$gdc$&`#$/Ifgdckd$$Ifl\ hl  t 6`044 la?%dh$&`#$/7$8$H$IfgdckdS$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc %1BC?kd$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdcCGLQV$$&`#$/Ifa$gdcdh$&`#$/7$8$H$IfgdcVWT=(($$&`#$/Ifa$gdc$&`#$/7$8$H$Ifgdckd$$Ifl\ hl  t 6`044 la?($&`#$/7$8$H$IfgdckdB$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc?%dh$&`#$/7$8$H$Ifgdckd$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc?kd$$Ifl\ hl  t 6`044 la$$&`#$/Ifa$gdc $$&`#$/Ifa$gdcdh$&`#$/7$8$H$Ifgdc THHHHHHH $7$8$H$a$gdckd1$$Ifl\ hl  t 6`044 la". $$Ifa$gdc$a$gdc $7$8$H$a$gdc.CZ[H6E\!5ag~˼혍~sd혍혍혍hqphcB*CJaJphhqphcCJaJh;hcB*CJaJphhDhcCJaJh;hcCJaJ hUXhc h#rhc hKhchch>JhcCJaJmH sH h>JhcCJaJhhcCJ]aJhhcCJaJhcCJaJh@hcCJaJ'./CZqhhh $Ifgdckd$$IflFN #  t06    44 laqhhh $IfgdckdQ$$IflFN #  t06    44 la1H6qhhW$$7$8$H$Ifa$gdc $Ifgdckd$$IflFN #  t06    44 la67E\qhhh $IfgdckdG$$IflFN #  t06    44 lacqhhh $Ifgdckd$$IflFN #  t06    44 lacdg~qhhh $Ifgdckd=$$IflFN #  t06    44 laqhhh $Ifgdckd$$IflFN #  t06    44 la7MNBCZ[\] -NOò{h%jP hcCJUVaJmHsH jhcB*CJUaJphhcB*CJaJphhVthcB*CJaJphhVthcCJaJ!jhGhcCJEHUaJ%jՑP hcCJUVaJmHsHjhcCJUaJhDhcCJaJh;hcCJaJhcCJaJ(qhhh $Ifgdckd3$$IflFN #  t06    44 la 7Nqhhh $Ifgdckd$$IflFN #  t06    44 laB^qhhhhh $Ifgdckd)$$IflFN #  t06    44 la qhhh $Ifgdckd8 $$IflFN #  t06    44 la Oqhhhh $Ifgdckd $$IflFN #  t06    44 la% >bq!Ll RSTZ[\]ʾyyyyyyyyyyyyqjjc hA(hc h"hch)hc5)hchcB*CJOJQJ^JaJphhchcCJOJQJaJhc h#rhch;hcCJaJhJhcCJaJhcB*CJaJphhVthcB*CJaJph jhcB*CJUaJph*j. h&7hcB*CJEHUaJph&%&38KTdqii`TTTTT $$Ifa$gdc $Ifgdc$a$gdckd$$IflFN #  t06    44 la dlt $$Ifa$gdctukd$$IfTlִH n&!&X\ t0'    44 laTu $$Ifa$gdc $Ifgdckd$$IfTlִH n&!&X\ t0'    44 laT $$Ifa$gdc $Ifgdckd$$IfTlִH n&!&X\ t0'    44 laT  $$Ifa$gdc $Ifgdc !kd$$IfTlִH n&!&X\ t0'    44 laT!:>CHPX]b $$Ifa$gdc $Ifgdcbckd$$IfTlִH n&!&X\ t0'    44 laTcmqw} $$Ifa$gdc $Ifgdckd$$IfTlִH n&!&X\ t0'    44 laT $$Ifa$gdc $Ifgdckd$$IfTlִH n&!&X\ t0'    44 laT $$Ifa$gdc $Ifgdckd$$IfTlִH n&!&X\ t0'    44 laT!'-6?EL $$Ifa$gdc $IfgdcLMkd$$IfTlִH n&!&X\ t0'    44 laTMhlqv~ $$Ifa$gdc $Ifgdckdr$$IfTlִH n&!&X\ t0'    44 laT $$Ifa$gdc $Ifgdckd`$$IfTlִH n&!&X\ t0'    44 laT $$Ifa$gdc $Ifgdc kdN$$IfTlִH n&!&X\ t0'    44 laT '/8AIR $$Ifa$gdc $IfgdcRSkd<$$IfTlִH n&!&X\ t0'    44 laTST.>NOPVLkdg $$IfTl4F_ t06    44 laT $$Ifa$gdc $7$8$H$a$gdc$a$gdc]p @cɸ۞yjy^y^y^yShM<hcCJaJh<%hc5CJaJh<%hcCJaJmHsHh<%hcCJaJ h"hcU!jHh>JhcCJEHUaJ#j3 N h>JhcCJUVaJhcCJaJ!j*h>JhcCJEHUaJ#jN h>JhcCJUVaJh>JhcCJaJjh>JhcCJUaJ hA(hchcthe unknown threshold Number of observations2-variable case5-variable case1 lag4 lags4 lags0.900.950.990.900.950.990.900.950.99TAR506.357.5410.295.836.989.568.109.4612.541005.956.999.395.666.668.978.509.7312.431505.946.989.295.786.768.938.7710.0212.732006.037.059.355.896.889.059.1710.4013.062506.147.119.386.077.089.329.3410.5713.205006.417.399.666.387409.639.7811.0513.67M-TAR507.228.4911.556.547.7610.508.9010.2613.441006.978.1510.676.617.7310.149.5210.8113.721506.757.8710.406.547.629.969.6610.9913.742006.627.7210.046.467.519.859.7511.0513.832506.617.7610.156.417.449.719.8411.1213.825006.527.559.936.427.479.7210.0511.3114.07Source: Wane, Gilbert and Dibooglu (2004). Table 3. Threshold cointegration ModelDeposit euroization with real effective exchange rateCredit euroization with nominal exchange rateTARM-TARTARThreshold = 0.0081Threshold = 0.0055Threshold = -0.0126 EMBED Equation.3 =-0.15-0.0054-0.297 EMBED Equation.3 =-0.059-0.13-0.535Tmax-1.12-0.067-1.84 EMBED Equation.3 5.505.515.16 EMBED Equation.3 0.67 [0.44]1.03 [0.18]0.85 [0.36]AR test1.22 [0.29]1.10 [0.36]0.61 [0.61]Note:  EMBED Equation.3  and  EMBED Equation.3  denote the adjustment parameters; W denotes the Wald test ; p-values presented in brackets. Source: calculation of the authors.  Authors are grateful to two anonymous referees for their comments and suggestions. All remaining errors are ours.  One must however note that due to high level of public debt, Euro adoption can not be considered as an exit option for those new EU member states and candidate countries whose public debt level has approached or surpassed the Maastricht criteria. Moreover, the authorities in these countries focused rather on restricting borrowing in foreign currencies, while saving in foreign currencies was generally very limited.  The notable exception is Heimonen (2001) who focuses on substitution of a substitute currency, while we are interested in financial euroization.  At the time, Deutsche mark was used to substitute the domestic currency, not Euro.  More details about TAR and M-TAR models can be found in Tong (1983), Caner and Hansen (2001), and Enders and Siklos (2001).  For the tests, we used the larger of the t values and F statistics that were later denoted by Tmax and  EMBED Equation.3  both in the text and in the corresponding tables.  The complete table with critical values tabulated by Wade, Gilbert and Dibooglu (2004) is provided in the Appendix.  The null hypothesis assumes linearity, while the alternative assumes threshold behaviour. Test statistics is denoted by W both in the text and in the corresponding tables.  Unit root test results are not presented in the paper to save space, but can be obtained upon request from the authors.  Due to space considerations, Johansen test results are not presented in the paper, but can be obtained upon the request from the authors.  Please note that the restrictions relate to long run coefficients written in vector notation.  Restriction that the long run coefficient of the nominal exchange rate in a restricted model is equal to -1 is rejected with Chi-square test statistic equal to 14.8 [0.01].  Tables with critical values can be found in Table 2 of the appendix. The results of the three models whose statistics are the closest to establishing threshold cointegration are presented in Table 3 of the Appendix  Public debt at the end of 2010 amounted to 45 percent of GDP. Projections based on intertemporal budget constraint suggest that if no reforms of public spending are introduced, public debt could easily reach 60 percent of GDP by the end 2017, thus precluding the adoption of the common currency.  The eroded credibility of macroeconomic policies is obvious knowing that Croatia has a tightly managed float exchange rate system (that could also be defined as a quasi currency board), sustained throughout years because of fear of floating (Vizek, 2007).     V]defkpuzMkd $$IfTl4\_'  t0644 laT $$Ifa$gdczq$$7$8$H$Ifa$gdckkd$$$IfTl) t0644 laTFfX" $$Ifa$gdc  !&,2Ff(Ff% $$Ifa$gdc237<AFKPUZ`fgkpuzFf1Ff. $$Ifa$gdcFf+kkd6$$IfTl) t0644 laTFf4 $$Ifa$gdc  #(.39?@DINTY^chntFf:Ff7 $$Ifa$gdctuy~FfCFf@ $$Ifa$gdcFf= ?@ag$$7$8$H$Ifa$gdcgdc$a$gdcFfF $$Ifa$gdc=?@FGa-./0HI\]^_ԳԐm\TTThcCJaJ!jPh>JhcCJEHUaJ#j,N h>JhcCJUVaJ!jkMh>JhcCJEHUaJ#j3 N h>JhcCJUVaJ!jJh>JhcCJEHUaJ#jN h>JhcCJUVaJjh>JhcCJUaJh>JhcCJaJ h"hchch09hcCJaJh"'hcCJaJ nbQQQ$$7$8$H$Ifa$gdc $$Ifa$gdckd!I$$Ifl4VF @Z%z 2 t0%6    44 la[OOOOO>$$7$8$H$Ifa$gdc $$Ifa$gdckdI$$Ifl4V\ '@Z%z    t0%644 la28@GJkd7J$$Ifl4V\ '@Z%z    t0%644 la$$7$8$H$Ifa$gdcGHahnu\KKKK$$7$8$H$Ifa$gdckdL$$IflV\ '@Z%z    t0%644 lauv{\KKKK$$7$8$H$Ifa$gdckdO$$IflV\ '@Z%z    t0%644 la\KKKK$$7$8$H$Ifa$gdckdP$$IflV\ '@Z%z    t0%644 la\KKKK$$7$8$H$Ifa$gdckdR$$IflV\ '@Z%z    t0%644 la'(;<=>CDWXYZbf}{jbbbbWh>JhlMCJaJhcCJaJ!jYh>JhcCJEHUaJ#j3 N h>JhcCJUVaJ!jVh>JhcCJEHUaJ#jN h>JhcCJUVaJhLHhcCJaJh<%CJaJjh>JhcCJUaJ!jSh>JhcCJEHUaJ#jN h>JhcCJUVaJh>JhcCJaJ   \KKKKKKK$$7$8$H$Ifa$gdckdU$$IflV\ '@Z%z    t0%644 la !O\WWOG?$a$gdf0Z$a$gd!$a$gdAgdckdjV$$IflV\ '@Z%z    t0%644 laNOPHR59ȯzsshWLDhG]DCJaJh:$hG]DCJaJ!jh:$hG]D0JCJUaJhBOhG]DmHsH hBOhG]DjhBOhG]D0JUhG]D\mH sH hvnhG]D\mH sH h>HxhG]D\mH sH  h>HxhG]Djh>HxhG]D0JUh!hG]D\mH sH  h!hG]Djh!hG]D0JUh hG]DmHsHhG]DjhG]D0JUhthlMCJaJZ39H!KM N P Q S T V W Y Z $a$gd. gdc$a$gdc $7$8$H$a$gdc$a$gd:gdw$a$gd!>$a$gd?gd?*9Z[rLgh234Ͳͣ͊wpi^hhG]DmHsH h-MhG]D hwhG]DhG]DB*phhwhG]DB*phhG]DmH sH !j.[hq]hG]DEHUmH sH jP hG]DUVmHsHjhq]hG]DUmH sH hq]hG]D6mH sH hq]hG]DmH sH hhG]D6 hhG]DhG]DjhG]D0JUh:$hG]DCJaJ$89:GHI  !"#KL  I L M N vnnnnnnnnjh'\BhG]DmH sH !jhE/hG]D0JUmH sH hE/hG]DmH sH hxhG]DmHsH hh bhG]D h)hG]DjhG]D0JU h.hG]DhG]D h9[JhG]Djh9[JhG]D0JUh9[JhG]DCJaJmHsHh9[JhG]DCJaJ!jh9[JhG]D0JCJUaJ(N O Q R T U W X [ \ hthlMCJaJh'\Bjh'\BU Z [ \ gdc6&P 1h:pCf. A!"#$% %Dd b  c $A? ?3"`?2oߥT!9lͮKD `!CߥT!9lͮ$@-@CxڝUOA~3ӥe)ٶhE0RI I 6ʁHQo%v 6Ѥ*b '?D㥇z7у#on]:͛YOIJttmjgXJ'[$78PS@9c*hӬ0WE\K+GtAF}m s2S{\|}-]D,hEk" d+r?-wƾ : m`:#ʚ|.4A]k4O*Xոiǭc99֐\|C~a5B&P5'Mߡdî.Vtr0$33I.:.n ^j;^Ͱ@H$s~εo{22a#męk2hcD#7QGM Aj25wBƝ@]N0~`Vς*hhn&: q!\^gSl`NR鄒D 5ppĂ{=Q>,.&c(9)|ODԹQZlX(.;( B(6` ȕC0Г?F6czi `! >0Г?F6cz@H|xuQ1o`}9mN  I%"JM,H*،[ 7(q<L,7tΞ??:"`CU1q|;5 P!s*8AvZ[T3ZftҌ\9c|9b/iW 78h/=i² |JMʴZq.3vaQwYKn[fV\A8)r~]&n^ |,fOFk r^?ѷ{zwnwtkg^} Օizwx}|.ٯn5#xX˻EƷ ݋[#Ө1J SwH fԐhͲ97pW(V9Yߍiji갉m"̫px>M&FogWsM $~hDd hb  c $A? ?3"`?2^t$M M[&z:A `!2t$M M[&z@H|xcdd`` @c112BYL%bpubDd @hb  c $A? ?3"`?2^$Ѵ@}0:U `!2$Ѵ@}0@ |xcdd`` @c112BYL%bpuArF~h2cxRYU)ȃʤ|>forə%6=):;b~-Bm:WQba$53016Lf[fU+*Qd^cY+rpwSo3DgYQ>ĹŹd^tǙ%gg.iIm^jc$ϻΣDπ;rGR1K cxJ}w>D0UMev֋ ¶:~4+>xlxxfI, qAN1YߐFԭj{`8u2/SU՝P*~R6i>>7<:p!;z2e(sU)@η0Lכ9U&@JN&2D=>&KqQؓF)$[z^? t_:\ A( f`]za/yq'ѐ7&% vFc>Yzj#&EmDpi}㿺]^݆MKn)R~06ᩂh'w+pWf мDd b  c $A? ?3"`?2e3_tK%?W  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~      !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~Root Entry} Fs%@5Data ]WordDocument|8TObjectPool0g$s%_1352535639,F0g$ $Ole CompObjiObjInfo "%&'()*+,/23456789:=BGLQTUVWX[`cfilopqtwxyz} FMathType 5.0 Equation MathType EFEquation.DSMT49qu4DSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!/ED/APG_APAPAE%B_AC_A %!AHA_D_E_E_A  Dm t Equation Native P_1352536136w F $ $Ole CompObj i==I jt r 1 m t"-1 ++1"-I jt ()r 2 m t"-1 ++e t ,            j==1,2 FMathType 5.0 Equation MathType EFEquation.DSMT49quDSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!/ED/APG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  m t FMicrosoft Equation 3.0 DS EqObjInfo Equation Native _1319993152F $ $Ole CompObjfObjInfoEquation Native :_1319993352c'F $ $uation Equation.39q5hV I 1t FMicrosoft Equation 3.0 DS Equation Equation.39q58 I 2tOle CompObjfObjInfoEquation Native :_1352535732F $ $Ole  CompObj!iObjInfo# FMathType 5.0 Equation MathType EFEquation.DSMT49quDSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!/ED/APG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  I 1t ==Equation Native $:_1352535811OF $ $Ole -CompObj .i1     if       m t"-1 e"t 1 0     if       m t"-1 <<t 1 {   FMathType 5.0 Equation MathType EFEquation.DSMT49qObjInfo!0Equation Native 1F_1325325143@$F $ $Ole ;u*DSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!/ED/APG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  I 2t ==1     if       Dm t"-1 e"t 2 0     if       Dm t"-1 <<t 2 {   FMicrosoft Equation 3.0 DS Equation Equation.39qL  1CompObj#%<fObjInfo&>Equation Native ?6_1325325175;)F $ $Ole @CompObj(*AfObjInfo+CEquation Native D6 FMicrosoft Equation 3.0 DS Equation Equation.39qLHD  2 FMicrosoft Equation 3.0 DS Equation Equation.39q_1352460819JE.F $ $Ole ECompObj-/FfObjInfo0H& 1  t FMicrosoft Equation 3.0 DS Equation Equation.39q5 Sz Equation Native I6_13199934903F $ $Ole JCompObj24KfObjInfo5MEquation Native N)_13531514878F $ $Ole OCompObj79PiObjInfo:REquation Native SD_13199954721"=F $ $ FMathType 5.0 Equation MathType EFEquation.DSMT49qw(DSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!/ED/APG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  I 1t  or  I 2t () FMicrosoft Equation 3.0 DS Equation Equation.39q5#Ԏ  i =0 FMicrosoft Equation 3.0 DS EqOle YCompObj<>ZfObjInfo?\Equation Native ]?_1319995528BF $ $Ole ^CompObjAC_fObjInfoDauation Equation.39q59x*  1 = 2 =0 FMicrosoft Equation 3.0 DS Equation Equation.39qEquation Native bU_1352463077GF $ $Ole dCompObjFHefObjInfoIgEquation Native hL_1351410967LF $ $Ole j0  1 = 2 FMathType 5.0 Equation MathType EFEquation.DSMT49qDSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!'_!/!/G_CompObjKMkiObjInfoNmEquation Native n_1352707133m6QF $ $AAPAE%B_AC_A %!AHA_D_E_E_A  F FMathType 5.0 Equation MathType EFEquation.DSMT49q@DSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!ED/AP     "! bc#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`ajdegfhklmnopqrstuvwxyz{|}~A| `!93_tK%?W` h`\xVkA3fGU(-6BmS Mhm)6E/%HfZ@ăx՛QC@ԓ 7lt5'ٙ7fYnL ‡RT*tT<ԜeA 4tdڠ'q|>`Tgya4^L_ynQ.dۆ: U@/t=2iLN y}.4cncz Ag؆Ec@_1]&C=4fo*w Maن e^Cqdҵ_nl1s oʴߘ7N%BĚk\gsI|M0H?wG{ dc[`11u;AEUg9ٯ$9ij[5u-u'ľA4YcYK֛k:뾔}m qD\QSF..Mh>rBP<  \Uw'_ʳ\ a4)ůΞx1vŵZq07ěb\J kE}dۅa]X*%ȃP,Ȗ~3c/`J`L޹#*ƠPn?3q2_>%*˫mp!Q2S,xDF`[G2C%u~r`?8SG`_$ŸJC[橸q MTY)~V7F''f6] 2}oDd Tb  c $A? ?3"`?2iT} HaQ]E `!=T} HaQ] RXJ xcdd``> @c112BYL%bpu @c112BYL%bpu @c112BYL%bpudKF&&\ s:  02/L0Dd b   c $A ? ?3"`? 2?Ǯ* goq' `!Ǯ* goq':@`!xcdd`` @c112BYL%bpu 18\ _pUsi#.F5 %\P[.a Cr;F&&\y @ ]` {> 1@`2`V/6=Dd Lb   c $A ? ?3"`? 2zKޔN:=Q7c `![zKޔN:=Q7 h)xڝSoQmAF6M޻@CFo$ЭhO=s&/z2){,iLݷy3͛yB>G$B4&."ħ+(fЗXΚoyK1[EPGΦ l(!OHtVhDL8=iuDx:1;~-v1` 50 rg-n 8GHӀc1,^5얱f[N81̈Cjݔ=pV[5O|uGpG83Tw: ;! B0uaiUqEB,p4;8O5)(Siq>grY͊YysQ^λuFz4#n[z]Y*s<TKs ja42<U)ZL l,n;P49Nk'μ[2h@Oބ$\re8fz_8zV4 I<Dd hb   c $A ? ?3"`? 29Ewved?)b! `!Z9Ewved?)B@@|(xcdd``Vdd``baV d,FYzP1n:&&! KA?H1Z00< UXRY7S?&,e`abM-VK-WMcpXpU eF\ XcBܤ\pr,$s@FBU3DwBelB&{A‚R2 \+4=8|Ĥ\Y\P 2C b> 1,˅p2TDd Tb   c $A ? ?3"`? 2%Y?mqz]! `!r%Y?mqr  XJ@xcdd``dd``baV d,FYzP1n:&v! KA?H1Z ǀqC0&dT20 KXB2sSRs~.b*\iYt9WcHA%<?}2¨Wa%w3@penR~g\pb#! #l+a|1&P s=i`b[Ǎ8v0o(121)W2ԡ1 =v@Dd Tb  c $A? ?3"`? 2Haѕ/_JSYf# `!^Haѕ/_JSY  8XJ,xuQJA};w4 *``UETbD3bv ;%X ;+0;-3 T/a EbÝUIBf ^LL|ÈU]M z xfS{&N2yO{_ى dWc#Sdr#-[*HG>?;*a;UĹǺ*ǘJ>as(|։%*tS dx",ZfBv`.Llo:ol?"}_Dd b  c $A? ?3"`?2HY@SNCY,Q % `!HY@SNCY,Q *`R!xMQ/Q.#m!nrB"sWJmbѤV'?UUI.&73c88jG%y~_zlÂ7/X77a}-ܓD 8Y !m7Or/ƸVr\Qgj N;~l2#يG/>J\l2`5_|*ײ\)VRinC'c; \4-elWvaś"&SA}霑:kmD0bj󽂹d.m,fg;YpSnZR^0BMw>UX+\q--h >F7 M|7:k+~ t1,_wB7"fۍSfD>F0Q6;EYmw`$$If!vh55555n5n#v#v#v#vn:V l4V0 ++,5555n9/ /  / / ap<T$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4V &&0 ++,555R5S5R5S5n9/ / / apPTMkd)$$IfTl4VִL J B &&&R&S&R&Snn &&0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / apPT/kd-$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kd1$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kdt5$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kdP9$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kd,=$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kdA$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / apPT/kdD$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kdH$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / apPT/kdL$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / apPT/kdP$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / apPT/kdTT$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / apPT/kd"X$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / apPT/kd[$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT$$If!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / /  / apPT/kd_$$IfTl4TִL J B &&&R&S&R&S&n&n0     44 lapPT\$$IfW!vh55555n5n#v#v#v#vn:V l4V0 ++,5555n9/ /  / / aWp<$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4V &&0 ++,555R5S5R5S5n9/ / / aWpPEkdd$$Ifl4Vִ <O!&&&R&S&R&Snn &&0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / aWpP'kdh$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kdl$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kdp$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kdyt$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kdWx$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kd5|$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpPOle rCompObjPRsiObjInfoSuEquation Native vG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  A i (L) FMathType 5.0 Equation MathType EFEquation.DSMT49q@DSMT5WinAllBasicCodePages_1352707132rVF $ $Ole {CompObjUW|iObjInfoX~Equation Native _1309080487[F "$ "$Ole CompObjZ\fTimes New RomanSymbolCourier NewMT Extra!ED/APG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  A i (L) FMicrosoft Equation 3.0 DS Equation Equation.39qObjInfo]Equation Native 6_1309356823h`F "$ "$Ole ;   1 FMicrosoft Equation 3.0 DS Equation Equation.39q6   2 FMicrosoft Equation 3.0 DS EqCompObj_afObjInfobEquation Native 6_1309080876Y^eF "$ "$Ole CompObjdffObjInfogEquation Native fuation Equation.39q;J  ( 1 = 2 =0) FMicrosoft Equation 3.0 DS Equation Equation.39q_1309081069jF "$ "$Ole CompObjikfObjInfolEquation Native ^_1352634837 ToF "$ "$Ole CompObjnpi;B  W( 1 = 2 ) FMathType 5.0 Equation MathType EFEquation.DSMT49q|DSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!'_!/!/G_$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / / aWpP'kd$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / aWpP'kd$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kdχ$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kd$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / / aWpP'kd$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / aWpP'kdi$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / aWpP'kd9$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / aWpP'kd $$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / / aWpP'kdٞ$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$IfW!vh555R5S5R5S5n5n#v#v#vR#vS#vR#vS#vn:V l4T0 +,555R5S5R5S5n9/ / / /  / aWpP'kd$$Ifl4Tִ <O!&&&R&S&R&S&n&n0     44 laWpP$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aT$$If!vh5 5555#v #v#v:VlT t6,5 5aTDd 0,b  c $A? ?3"`?2' KgΘǚ4+`! KgΘǚ4kxڝROQ-E)%=.$ AO^8PMq*yY 2%[# @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJW$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55Dd ,Tb  c $A? ?3"`?2i%9Cm`!:eDE 8`!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55bDd Th b  c $A? ?3"`?2\⤳X`!\⤳X @|Nxcdd``dd``baV d,FYzP1n:&&.! KA?H1} UXRY7S?&,e`abM-VK-WMcpXggF\ 9AZ5+a F%>?e`JaaT 0X'27)?a?WG!_;[X hrm/0-ۣsAc `ƥ(ePyĤ\Y\˰dP"CD|b@33X} $$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55bDd hb  c $A? ?3"`?2S=N/)|3`!S=N/)|3@e |Nxcdd``Ned``baV d,FYzP1n:&&! KA?H1 ㆪaM,,He` @201d++&18\ _˹ӼXt9F=vVJ?g># dQ%o V {+ssZ9|=pd[O(%P=Y2It2L:8e"{@i#RpeqIj.\E.Y`1$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55$$If!vh5)5555#v)#v#v:Vl t 6`\6,5)55Dd Tb  c $A? ?3"`?2h ùulഝD m`!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJWDd ,Tb  c $A? ?3"`?2i%9Cm`!:eDE `!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5 5#v #v:Vlw t 6`M%,5 5p $$If!vh5 5l5l5l5l#v #vl:Vl4w t 6`M%+,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl4w t 6`M%+,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2Dd Tb  c $A? ?3"`?2h ùulഝD D`!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJW$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2Dd ,Tb  c $A? ?3"`?2i%9Cm`!:eDE $`!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh5 5l5l5l5l#v #vl:Vl t 6`M%,5 5lp2$$If!vh55 #v#v :Vlp t 6`,55 p $$If!vh5555#v#v#v:Vlp t 6`,555p $$If!vh5555#v#v#v:Vlp t 6`,555p $$If!vh5555#v#v#v:Vl t 6`,555p Dd Tb  c $A? ?3"`?2h ùulഝD `!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJW$$If!vh5555#v#v#v:Vl t 6`,555Dd ,Tb  c $A? ?3"`?2i%9Cm`!:eDED `!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5555#v#v#v:Vl t 6`,555$$If!vh5555#v#v#v:Vl t 6`,555bDd Th b  c $A? ?3"`?2\⤳X `!\⤳X @|Nxcdd``dd``baV d,FYzP1n:&&.! KA?H1} UXRY7S?&,e`abM-VK-WMcpXggF\ 9AZ5+a F%>?e`JaaT 0X'27)?a?WG!_;[X hrm/0-ۣsAc `ƥ(ePyĤ\Y\˰dP"CD|b@33X} $$If!vh5555#v#v#v:Vl t 6`,555bDd hb  c $A? ?3"`?2S=N/)|3 `!S=N/)|3@e |Nxcdd``Ned``baV d,FYzP1n:&&! KA?H1 ㆪaM,,He` @201d++&18\ _˹ӼXt9F=vVJ?g># dQ%o V {+ssZ9|=pd[O(%P=Y2It2L:8e"{@i#RpeqIj.\E.Y`1$$If!vh5555#v#v#v:Vl t 6`,555$$If!vh5555#v#v#v:Vl t 6`,555$$If!vh5555#v#v#v:Vl t 6`,555$$If!vh5555#v#v#v:Vl t 6`,555Dd Tb  c $A? ?3"`?2h ùulഝDH `!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJWDd ,Tb  c $A? ?3"`?2i%9Cm`!:eDEf `!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5 5#v #v:Vlw t 6`,5 5p $$If!vh5 555#v #v#v#v:Vlw t 6`,5 555p $$If!vh5 555#v #v#v#v:Vlw t 6`,5 555p $$If!vh5 555#v #v#v#v:Vl t 6`,5 555p $$If!vh5 555#v #v#v#v:Vl t 6`,5 555Dd Tb  c $A? ?3"`?2h ùulഝD `!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJW$$If!vh5 555#v #v#v#v:Vl t 6`,5 555Dd ,Tb  c $A? ?3"`?2i%9Cm`!:eDE `!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t ObjInfoqEquation Native ,_1352634894tF "$ "$Ole AAPAE%B_AC_AE*_HA@AHA*_D_E_E_A   var(Dcpi)+cov(Dcpi,Dreer) var(Dcpi)+var(Dreer)+2cov(Dcpi,Dreer)CompObjsuiObjInfovEquation Native _1352536287yF "$ "$ FMathType 5.0 Equation MathType EFEquation.DSMT49qDSMT5WinAllBasicCodePagesTimes New RomanSymbolCourier NewMT Extra!'_!/!/G_AAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  correl[12-month average(Dip),12-month average(Dreer)] FMathType 5.0 Equation MathType EFEquation.DSMT49quDSMT5WinAllBasicCodePagesOle CompObjxziObjInfo{Equation Native Times New RomanSymbolCourier NewMT Extra!/ED/APG_APAPAE%B_AC_AE*_HA@AHA*_D_E_E_A  FOh+'0   0< \ h t 0FINANCIAL EURIZATION IN A TRANSITION COUNTRYmvizek6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555$$If!vh5 555#v #v#v#v:Vl t 6`,5 555y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55Dd 0b " c $A? ?3"`?!2X9BN`!X9BNF`kxڭTkQ6[4OBѤ&M4%%u.$ݰz xA{ zx٫H]*4WhM޼Λy gAC d$w|]Ģad B78.rPRת60گ#K9++!.2T)UOajZV-:,8bAnhA⍶@rb.1wFy h53CM9+/}K,E` F?Ф] DOFM4 <%bW=am 1e\N~_x쳴|k0~&H]+|> ,9^y$$If!vh5 55#v #v#v:Vl t6,5 55y$$If!vh5 55#v #v#v:Vl t6,5 55oDd ,b  c $A? ?3"`?2Fb d '!r  `!Fb d '!`@+[xڵSOoQ@Ÿ lJn HMt),YIx0$xГojbL6LęDz4=f̼y \9a x8cd2-vϿ"J7 2M7HR%V~+zR "$]kŏ~t9Cj[T=CWp$wk57:<%R0MNFSCTԘ֊'O*0 ʈ#\p,6R)ɶ} ˾SOÀwZr^"LU5'"i g eP%o)|fG0:Бe|}>#kyxx/XҚrjW yB\v>!OxhjI!Ηܣ@.D7ǕJtnEU`] `puV0!6Yʆ557Pl%g+ h-@tK'Ut@%h R$ $Љb:G~•EQ} NϿW7?ޝ 04-7bQua3! τ a8=.V=y$$If!vh5 55#v #v#v:Vl t6,5 55$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555T$$If!vh5X5\555555#vX#v\#v#v#v#v#v#v:Vl t',5X5\555555TDd Tb $ c $A? ?3"`?#2h ùulഝD n`!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJWDd ,Tb % c $A? ?3"`?$2i%9Cm`!:eDE `!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh555#v#v#v:Vl4 t6+,555T$$If!vh5555#v#v#v:Vl4 t6+,555T$$If!v h555555555 5 #v#v#v#v#v :Vl4 t6+, 55555 T-kd}!$$IfTl4 _ ' U# t06((((44 laTQ$$If!vh5)#v):Vl t6,5)T$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd$$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd'$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd*$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd-$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd0$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd3$$IfTl _ ' U# t06((((44 laTQ$$If!vh5)#v):Vl t6,5)T$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd'7$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd&:$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd%=$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd$@$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd#C$$IfTl _ ' U# t06((((44 laT$$If!v h555555555 5 #v#v#v#v#v :Vl t6, 55555 T*kd"F$$IfTl _ ' U# t06((((44 laT$$If!vh5z 525 #vz #v2#v :Vl4V t%6+,5z 525 $$If!vh5z 5 5 5 #vz #v #v :Vl4V t%6+,5z 5 5 $$If!vh5z 5 5 5 #vz #v #v :Vl4V t%6+,5z 5 5 Dd Tb & c $A? ?3"`?%2h ùulഝD K`!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJW$$If!vh5z 5 5 5 #vz #v #v :VlV t%6,5z 5 5 Dd ,Tb ' c $A? ?3"`?&2i%9Cm`!:eDE M`!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~$$If!vh5z 5 5 5 #vz #v #v :VlV t%6,5z 5 5 $$If!vh5z 5 5 5 #vz #v #v :VlV t%6,5z 5 5 bDd Th b ( c $A? ?3"`?'2\⤳XP`!\⤳X @|Nxcdd``dd``baV d,FYzP1n:&&.! KA?H1} UXRY7S?&,e`abM-VK-WMcpXggF\ 9AZ5+a F%>?e`JaaT 0X'27)?a?WG!_;[X hrm/0-ۣsAc `ƥ(ePyĤ\Y\˰dP"CD|b@33X} $$If!vh5z 5 5 5 #vz #v #v :VlV t%6,5z 5 5 bDd hb ) c $A? ?3"`?(2S=N/)|3S`!S=N/)|3@e |Nxcdd``Ned``baV d,FYzP1n:&&! KA?H1 ㆪaM,,He` @201d++&18\ _˹ӼXt9F=vVJ?g># dQ%o V {+ssZ9|=pd[O(%P=Y2It2L:8e"{@i#RpeqIj.\E.Y`1$$If!vh5z 5 5 5 #vz #v #v :VlV t%6,5z 5 5 $$If!vh5z 5 5 5 #vz #v #v :VlV t%6,5z 5 5 Dd Tb * c $A? ?3"`?)2h ùulഝD 5W`!< ùulഝܒ HXJ xcdd``> @c112BYL%bpu;vPHPcdbR ,.IeԡRYv] `sJWDd ,Tb + c $A? ?3"`?*2i%9Cm`!:eDE SY`!=%9Cm`!:eD XJ xcdd``> @c112BYL%bpuؽF2 Ma`ǕQy FE\Ps}v&"=0yĤ\Y\d.P"CD|bb#3XL~Dd b  c $A? ?3"`?2OՀ>K4Cr[ `!OՀ>K4C*RxMQ=OA}$w&O-LhSv穧 466-=+qv=ٙ73o%:R8c>bK4| .`QB3A' gm'B:bd|\QdK|lFNeVB1Cf$?O hA4`~qAB D7KU\N9nZ,UܦszTV.jv( p [4VAC IyOVm36'Lx7R6!]Kgyד$'kY;c[ئb4HF?Нf (Ke忶؏p:ZBEMT\cd|(*,.3:;RW^ckpwyz !&-/0KMRY^egh &),/23;@GLSX_dklv{X|}*/=HIV\bfgs{%-.FLV\fkuv~%9:Oejx%,7>EF^dkq|#(49EVW[`ejk  !"#$%&'()*+,-. &/6BCWn'(E\JKYpwx{34Kb,Vr +c 9:GL_hx '-45NRW\dlqvw15;AJSY`a| 03;CLU]fgh(?O_`agnuvw| #(-27=CDHMRW\afkqwx| $)/49?DJPQUZ_ejoty $%PQrx !)*CIQXYry %*12`k DJY2\^_abdeghjkn000 @@@ NormalCJ_HaJmH sH tH\@\ hj Heading 2$<@& 56CJOJQJ\]^JaJZA@Z !Default Paragraph Font, Char CharRi@R  Table Normal4 l4a (k(No ListjOj  Char Char Char Char dCJOJQJaJmH sH tH 6U@6  Hyperlink >*B*phO 9 Char Char Char Char Char1 Char Char1 Char Char Char Char dCJOJQJaJmH sH tH >@"> M Footnote TextCJaJ@&@1@ MFootnote ReferenceH*j@Cj M Table Grid7:V0PZ@RP M Plain Text CJOJ QJ ^J aJmH sH tH jObj q Char Char Char Char dCJOJQJaJmH sH tH tOrt   Char Char Char Char Char dCJOJQJaJmH sH tH B'@B CfComment ReferenceCJaJ<< Cf Comment TextCJaJ@j@ CfComment Subject5\HH Cf Balloon TextCJOJQJ^JaJLOL j|m Char dCJOJQJaJmH sH tH O +N3 Char Char Char Char Char Char Char Char Char1 Char dCJOJQJaJmH sH tH @B@@ Rt Body Text$a$ mHsHtH 0a@0 hj HTML Cite6]D?QVWWa7cbffg{am t4X^mFp r u mTjklmnCDz 7 8 F`a"%'')Z2U9V9z9{9!>">cCdCDDDDrHsHHHHHJMOOPQQQRqRRRSSSSWWWW^tauaaa:c;cefhhhhHjIj_j`j`nan0r1r z!z7z8z|GH^_BCYZv01GH‘Ιϙ#$67@A45QR67ǥȥ:;9;BCݭ&'()+-/1?LMǮȮɮ $*17;EJSTUX^ekrvįͯίϯ !"+138:GIJyװް %,29?DMR[\]`fmsx}űɱӱԱձ &'(179?AOQRƴǴ"C`}µϵе $*04=AKLMT[ahnrz~ĶȶжѶҶ ;AHNUZcgqrsy·ƷͷϷз$+/7<EFGU\bios{¸Ƹ͸θϸ׸ .5<BIMTYbcdϹֹܹ  $+/789BHJPRV]_`ؼټ+BXnox~ǽɽнؽڽ۽*/6<CIPV]^lqxzξӾھ AFOT]cjpwxǿȿٿۿݿ$*34TVXZ`jk !&'/4;@GLSX_`jov{00000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000 0 0 0 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 00000 0 0 0 00 00 0 0 0 0 0 0 0 0 0 0 0 00 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 0 0 0 00 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 00 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 00 0 0 0 0 0 0 0 00 0 0 00000 0 0 0 0 0 0 00 00 00 00 0 0 00 0 00 0 0 0 0 00 0 00 0 0 00 00 00 00 0 0 00 0 00 0 0 0 0 00 0 00 0 0 00 00 0 0 0 0 00 00 0 0 0 0 00 00 00 00 0 0 00 00 0 00 0 0 0 0 00 00 0 0 0 0 00 00 0 0 0 0 0 00 0 0 0 0 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 00 0 0 00 00 00 00 0 0 00 00 00 00 0 00000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 00 0 0 00 00 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 00000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 00 0 0 00 00 00 00 0 0 00 00 00 00 0 0 0 0 0 00 0 0 00 00 00 0 0 0 00 00 00 00 0 0 00 0 0 00 0 0 0 0 00 0 0 0 0 00 00 0 0 0 00 00 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 00 0 0 00 00 00 00 0 00000 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 0 0 00 00 00 0 0 0 0 0 0 0 0 0 0 0 000000000000000000000000000000 0 0 0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 0 0 0 00 00 0 0 00 00 00 0 0 0 0 0 0 0 00 0 0 0 0 0 0 00 0 0 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 0 0 00 00 00 0 0000000000000000000000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 000 0 0 0 0 0 0 0 00 0 000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0      !"#$%&'()*+,-./0123456789:;<=>?@ABCD 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 000 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000 0 0 0 0 0 0 0 0 0 00 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00 00 00 0 00@0@0@0@0@0@0@0@0@0@0@0@0@0@0@000r @0I00@0I00@0I00@0I00I00‘ΙϙnK00 K00 K00K00K00K00K00O?K00  U F V $&(+-/184>6\8:z=`BD HDJLNTZ)\]%^_U`adfyi~oul҅X qd.º4yeBp.x]9N \  #3<DM[ciF=CR]1t&Gܻ޻Uӿl`~ ND,tvB>RKb:Wp!1:]] OP(L|p6n`bWXhBgxavob.|CV .6c dtu !bcLM RSVz2tGu Z \      !"$%&'()*+,-./012456789:;=>?@ABCEFGHIJKLNOPQRSTUVWXYZ\]^_`abdefghjklmnopqrstuvwxyz{|}~[ qRRRRRRRSSSSSSSSSSS T!T#T)T=T?TeTyT{TTTTUUUV3V5V:VNVPV9WMWOWzzzx|Mac,.lxdxz.BD,.FZ\Vnp*>@Ymo8LNThjm:::::::::::::::X::::::::::::::::::::::::::::u :8@0(  B S  ? OLE_LINK1 OLE_LINK2 OLE_LINK5 OLE_LINK6 OLE_LINK3 OLE_LINK4 _Hlk277066135***nn` /T / /T /T / /~ /Ԃ / /T / /ԃ / /Ԅ /T / / / /, /l  /쨎 !/ "/, #/l $/, %/ &/l '/쥎 (/, )/ */, +/, ,/짎 -/l .//w ?/w @/x A/Tx B/x C/x D/y E/Ty F/y G/y H/z I/Tz J/z K/z L/{ M/T{ N/{ O/{ P/| Q/T| R/| S/| T/} U/T} V/} W/} X/, Y/l Z/ [/쐵 \/, ]/l ^/ _/쑵 `/, a/l b/ c/쒵 d/, e/l f/ g/쓵 h/, i/l j/ k/씵 l/, 11   YYu(u(2222O4O4 8 888r9r9?:Y:e::::P;P;<<<<<8=C=C=~>>>?@ABCDEGFHIJKLMNOPQRSTUVWXYZ[\]^_88 ""``|(|(2222W4W48888y9y9F:`:l::::W;W;<<<<=>=L=L=>>>CQCQ]]OuOuϐϐ WW--ppORR//ddNNn  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEGFHIJKLMNOPQRSTUVWXYZ[\]^_9*urn:schemas-microsoft-com:office:smarttagsState8'*urn:schemas-microsoft-com:office:smarttagsCityB`*urn:schemas-microsoft-com:office:smarttagscountry-region="*urn:schemas-microsoft-com:office:smarttags PlaceName= *urn:schemas-microsoft-com:office:smarttags PlaceType9_*urn:schemas-microsoft-com:office:smarttagsplace  `_`_`_`_`___````_`_`_`_`_```_`_`_```````_``__``_`_`__``_`'_`_`"_ `__'"_" '_'_'_`__'_`_"_ _'__` G!J!T!Z!i!n!|!!|''''''''''T(\(~((2*7*<*F*v+~+++?-D-R-Z-////11Z2^222"4.43494 7777A7M7R7X7777777::0:6:W WZZ[[[[\\G]M]rrrrquxuuuuuuu^vfvkvpvvvvvvvyyyyyyyyyyzz{"{ HS}-37<}5<RU^d̤פJRYcdk}ȥХԥץۥ!'KUZf"ç;@DN%ǩΩשߩ<AS\bj իݫ*5AKWam۽zx{_g?G'iq__aabbdeghjknhla l SSSS%T(Tkk"rGIǣ1P5ƥȥh 9fc8CKǴϴټ~ƽн׽*./5 (}rw hpQY__aabbdeghjkn33333333333333333333333333sHHqRRSSSSSSUU1HMJyRез`ټ*4T4pl Normal.dotmvizek2Microsoft Office Word@vA@ѓ@5@56#՜.+,D՜.+,\ hp  Eizgxt -FINANCIAL EURIZATION IN A TRANSITION COUNTRY Titlel 8@ _PID_HLINKSA$ `-http://www.sciencedirect.com/science?_ob=PublicationURL&_tockey=%23TOC%235937%232004%23999489993%23518765%23FLA%23&_cdi=5937&_pubType=J&view=c&_auth=y&_acct=C000050661&_version=1&_urlVersion=0&_userid=4758817&md5=85243f957bfac944fc0f469440109e1f   FMicrosoft Office Word Document MSWordDocWord.Document.89qrqd|z0K.FF^k,n(\pxb,+ :ghq(%Qx2`__aabbdeghjkn__aabbdeghjkn; We4?9*hC̢eW?,^`o(.h\h^h`\o(..0^`0o(...L0L^L`0o(....   ^ `o( ..... l l ^l `o( ...... 0`0^0``o(....... `^``o(........ PP^P`o(.........^`o(.^`o(.808^8`0o(..808^8`0o(... ^`o( .... ^`o( ..... `^``o( ...... `^``o(....... pp^p`o(........h^`OJQJo(hHh^`OJ QJ ^J o(hHohpp^p`OJ QJ o(hHh@ @ ^@ `OJQJo(hHh^`OJ QJ ^J o(hHoh^`OJ QJ o(hHh^`OJQJo(hHh^`OJ QJ ^J o(hHohPP^P`OJ QJ o(hHhh^h`o(.^`o(..0^`0o(...0^`0o(....   ^ `o( ..... @ @ ^@ `o( ...... `^``o(....... x`x^x``o(........ HH^H`o(.........; We4eW*hC         JEYM)_ C{ o #EY M#J$<`\)}o@)y<kJ-z.U7o8NgHNgHJ-@ M`\)z.U_ tVy<k o }o7o8@]EwtVC{sTmh&#AD+Ng?rn$6D3c,p~ (=AfoIt <%U0V_ BZTuQR1~ % C   \ u L} 3B \ . z  .9 +A P W 2v y M+.=bJRw'7 ?FL[z\.;QRW3ZlqUr2,E@LPW`7QQf"78z(H)Kl~&4K&3C]gr0_dIrx|3:EK;fgs_gd kerzBse&:)X$t*?BWchq~1 1e s  0= H K 7h #!#!"" (" z""#>#I%#;#%$A$$ $jV${[$g$z$% %Q%y&b4&yU&e&x&"'hU'Y't'4(A(%(:+(S(V( W()&).)>)A)D)_N)vv)}x)*?*+6+7+k+o+12,J,U,f,;-.-3-RI-pM-Y ./m/8/#0.O0f0u1|$1[*1B1d1p1q17!2Z+2R2Z|2~2K33lL3]3<5 66z66/6BI67$7t37TU7V7Va7 ~7(8.8E8 u8u 909Q9\9"c9s9:FD:J: ;m;; +;C;\;b;-j;h<<<@,<-<M<Y<c<0=9=G>U> >G>Z>Tk>q>q>r>|>?.?l?1?{:?Q?jk?t?@@)@M@y@AA-ACaAyOPA8PYP\P}P ?Q2`Q]sQ}QR =RBIRW`R,S'SNSZSU2UV4U)>ULUQUdU3iUsUtVWW4W:WD!X%X)X-,X06XHXaX YYbYY Z)Z*Zf0ZHxyy]yzTz _z/|zP{/{ {+{[;{ d{)i{u{|| |=|C|_|<_|}.}f9}]}n}}}~f~~k4~G~D&7OJ7u4^*I)tx34DVnKBC=n,j@~ H"Wcwfq5vUvBx__bgtX!Sg|*~@Cbm::]_ K{c -Zq36>Jei#Y9DZrsX W/(,L?MP`b!osy m;YIjD!(5ItZdoCSpZc4>FtTabIk{C-:G~`4!x,yIycfroJDr.rG0#.7tvyi4:<aJpq@ -'Zp[ sBU\\{' S2\_8 EHUy2D.E!QpbXgG$NBBDVw<_0 %N?OQYhq{@DGSatQ0=Oe *-19@Q_&cJf1ByW"@L JNWqzY,'2#rNu5s!>yfRGTp'.2gj {~ ,|{3K%2?c"(4{9cOUn 4D&)>oAjPo W#AR j|.8iEn %1?34*Jm~"js &WrQvn/!^#DgrK"|o&so"#,Jy~/RPU~;BV4xXhmsjx(MOC[(rv~? w5CS{~q8v|7^dmk>CRs$]]`^i$:Xk+pDIVlg#<*_+1x""r? ATtt*IQ___@$7RfGjq&):Nj/6G=>GSW%*3BOadX4:@Ab!d#(KyJZ&h[\_mty)+,(0CE^~$_DIvs "8LHbi@019CPw[5 VP^gbcJMs(0IObCoc@fR ;27}X N)67;@U%{:1+%2!_voxz(;J`-ae%M#B{ KLY|I68o p!;=Arn!Yta}9qD#_\g2pR !gOoV]Zp IvzL#oM$z 25-MPPJakW+x360#3{;QDQ^d(<_f!L5K&Pjo x|A 7Gq y N<qRtfz68T\|{,F.M>gXp2Yݭ&'()+-/1?LMǮȮɮ $*17ESTUX^ekrͯίϯ !"+138:GIJyװް %,29?M[\]`fmsxűӱԱձ &'(179?AOQRǴ"C`}µϵе$*0=KLMT[ahnzĶжѶҶ;AHNUcqrsy·ͷϷз$+7EFGU\bio{¸͸θϸ .5<BITbcdϹֹܹ  +789BHJPR]_`ټ+BXnoxǽɽؽڽ۽*6CP]^lxzξھ AO]jwxǿȿٿۿݿ$34TVXZjk !&'/;GS_`jv4GZ[p] 18?FMNSY_ekl)56EGIKMNehknqr qL (6ETcd|(*,.:;R^kwyz !-/0KMYegh &),/23;GS_klv}*/HIV\bfgs{%-.FVfuv~%9:Oej%7EF^k|#4EVW[`ejk,&/6BCWn'(E\JKYpwx{34Kb, +c 9:GL_hx '-45NRW\dlqvw15;AJSY`a| 03;CLU]fg'(?O_`agnuvw| #(-27=CDHMRW\afkqwx| $)/49?DJPQUZ_ejoty $%Qrx)*CIQXYry %12~>n5a0 U!05aaJ1 0Uy0@\d<78yz8888XXXXXXm@@@ @X@|@@@@@@@ @Unknown g: Times New RomanTimes New Roman5Symbol3& : ArialO AdvEPSTIMMS MinchoSCMR12Times New RomanyTimesNewRomanPS-ItalicMTTimes New Roman_M Times-RomanTimes New Roman3Times5& zaTahoma?5 : Courier New;Wingdings"1CFCFCF 6#x6#x!4dtt2qHX ?2,FINANCIAL EURIZATION IN A TRANSITION COUNTRYmvizekmvizek