Financial liberalization, disaggregated capital flows and banking crisis: Evidence from developing countries

Economic Modelling 41 (2014) 124–132

Contents lists available at ScienceDirect

Economic Modelling journal homepage: www.elsevier.com/locate/ecmod

Financial liberalization, disaggregated capital flows and banking crisis: Evidence from developing countries☆ Helmi Hamdi a,⁎, Nabila Boukef Jlassi b a b

Aix-Marseille University, CERGAM (E.A 4225), France Laboratoire d'Economie d'Orléans (LEO), Université d'Orléans, Faculté de Droit, d'Economie et de Gestion, France

a r t i c l e

i n f o

Article history: Accepted 2 May 2014 Available online xxxx Keywords: Banking crises Financial liberalization Capital FLOWS Developing countries

a b s t r a c t The aim of this paper is to examine whether financial liberalization has triggered banking crises in some developing countries. We focus in particular on the role of capital flows as their volatilities threat economic stability and growth. In the empirical model, based on panel logit estimation, we use the two common financial liberalization indicators (defacto and dejure) for a panel of 58 developing countries observed during the period 1984–2007. Unlike the previous studies, this paper reveals that both indicators of financial liberalization did not trigger banking crises. However, the results show that foreign debt liabilities to total liabilities and foreign direct investment liabilities to total liabilities increase the likelihood of banking crises. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Financial liberalization is defined as the implementation of a set of measures aimed at eliminating the different restrictions and repression on the financial sector of a country that could hinder the wellfunctioning of its economy. These measures consist essentially on freeing interest rates, reducing credit control, eliminating barriers to entry, and removing restrictions on overseas financial transactions. For McKinnon (1973) and Shaw (1973), the weak growth and the lack of performance of developing economies during the sixties are due to the ineffectiveness of their financial markets which were fully controlled by the government. For them, developing economies were under “financial repression regime” in which the government intervenes in the monetary sphere to set interest rates and to fix the different tools of monetary policy. McKinnon (1973) and Shaw (1973) viewed the liberalization of interest rates and capital account as an efficient solution to eliminate directed credits and to remove control of interest rates and high reserve requirements. They consider the external financial liberalization as an important economic policy tool that enhances economic growth. McKinnon (1973) and Shaw (1973), also consider financial liberalization as a mainstay of economic reforms in developing countries (Balassa, 1989a, b).

☆ The authors wish to thank the anonymous referees for the useful comments to an earlier draft. They are also grateful to Professor Jean-Paul Pollin for his guidance and his useful comments and advices. ⁎ Corresponding author at: CERGAM. IAE Aix-en-Provence Clos Guiot Puyricard - CS 30063 13089 Aix en Provence Cedex 2. France. E-mail addresses: [email protected] (H. Hamdi), [email protected] (N.B. Jlassi).

http://dx.doi.org/10.1016/j.econmod.2014.05.010 0264-9993/© 2014 Elsevier B.V. All rights reserved.

They called policymakers of less developed economies to participate in the global financial integration to benefit the advantages of interconnected financial systems and to promote their banking and financial sectors. In the late eighties, financial liberalization became a strategy suggested by the International Monetary Fund and the World Bank under a framework called “Structural Adjustment Programs” (SAPs henceforth) to rescue fragile economies, notably those of developing countries (Hamdi et al., 2013). This framework suggests the easing of portfolio restrictions on banks, changing the ownership of banks, enhancing competition among financial institutions, integrating of domestic entities to international markets, as well as changing the monetary policy environment (Ucer, 1998). As a result, numerous countries adopted the SAPs and have progressively liberalized their economies with the aims of improving financial development and economic growth (Bekaert et al, 2005; Tornell et al, 2004). The debate on the impacts of financial liberalization on economic growth has received a great deal of attention by scholars and policymakers over the past three decades. However, the empirical studies have produced mixed and conflicting results. In fact, some authors (Levine, 2001; Mishkin, 2005; Prasad et al., 2003) showed that liberalization of capital flows can benefit both source and host countries by improving resource allocation, reducing financing costs, increasing competition and accelerating the development of domestic financial systems (IMF, 2012). On the other hand, several studies showed the adverse impact of financial liberalization and they demonstrated the potential role of liberalization on producing financial and economic crises (Caprio and Klingebiel, 1996; Demirguc-Kunt and Detragiache, 1998a, 1998b, 2000; Kaminsky and Reinhart, 1999; Mehrez and Kaufman, 2000).

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

The main purpose of this study is to investigate the consequences of financial liberalization indicators (dejure and defacto) on the likelihood of triggering a banking crisis in a large sample of developing countries. We employ a logit panel model to identify the factors that determine the occurrence of a bank crisis. Our methodology follows the previous studies, (Choudhry and de Haan, 2009; Demirguc-Kunt and Detragiache, 1998a, b; Joyce, 2010), but it differs in at least three points. First, we used more recent data which covers the period from 1984 to 2007. Second, we included in our sample more developing countries; 58 in all.1 Third, while recent studies have used only three indicators of foreign direct investments scaled by gross domestic product (GDP henceforth), we used in this paper six different ratios. Therefore, the paper focuses on the responses of foreign direct investment, portfolio flows, and other debt flows to financial liberalization and it examines the interaction between these indicators and the total foreign direct assets and liabilities. The main finding of this paper reveals that indicators of financial liberalization (dejure and defacto) did not trigger banking crises in our sample. The remainder of the paper is organized as follows: Section 2 presents a brief literature review, Section 3 describes the methodology and data, section four presents the empirical results and section five concludes. 2. Literature review Literature on financial liberalization is rich and abundant. Nevertheless, empirical evidence provided inconclusive results on the effects of liberalization on growth and financial stability. We can classify the literature in two categories. Authors of the first category demonstrated that financial liberalization improves financial development and contributes to higher longrun growth (King and Levine, 1993; Bekaert et al., 2004, Arteta et al. 2001, Edwards 2001, Mishkin, 2008). They showed how financial liberalization can play an important role in the development of financial institutions in emerging market economies and how both external and internal liberalizations tend to improve the financial infrastructure and bank governance (Schmukler, 2004a, b). For example, Mishkin (2008) claimed that liberalization can reduce the cost of capital, thereby encouraging investment which promotes growth. He showed that globalization of the financial system helps promote the development of better property rights and institutions that make the domestic financial sector works better in getting capital to productive uses (Mishkin, 2006). In another study, Prasad et al. (2003) and Kose et al. (2004) showed that investment in developing countries is constrained by an insufficient level of domestic saving. They showed that opening up an economy to capital flows will promote domestic savings, lowers the cost of capital and reduces the consumption volatility. Angkinand et al. (2010) showed that the removal of capital control can lead to more foreign direct investment (FDI), which will bring in new technology and management skill. Henry (2007) argues that liberalizations in emerging countries do increase investment activities and strengthen growth, but that these benefits are temporary. Bekaert et al. (2005) argued that stock market liberalizations do increase growth. Using industry level data, the study of Levchenko et al. (2009) also shows the positive effects of liberalizations on growth. Bonfiglioli (2008) studied the effects of financial integration on the productivity (TFP) and investment using a sample of 70 countries observed between 1975 and

1 We are interested to examine developing countries for several reasons. First, they are much exposed to external shocks as the level of their real income per capita is not sufficient to withstand a banking crisis. Second, banks' balance sheets of developing countries are mainly based on traditional activities and therefore, there is no diversification of risks. Consequently, they are vulnerable to any supply side shocks. Third, as Schmukler (2004) opined, deregulation, privatization, and advances in technology made foreign direct investment (FDI) and equity investments in emerging markets more attractive to firms and households in developed countries.

125

1999. The results for both dejure and defacto indicators suggest that financial integration has a positive direct effect on productivity, while it does not directly affect capital accumulation. In another recent study, Shehzad and De Haan (2009) investigated the impacts of financial liberalization on the likelihood of systematic and non-systematic banking crises for a large sample of developing and developed countries observed during the period 1973–2002. Using multivariate probit model, their results show that financial liberalization reduces the risk of systematic crises. Other studies conducted by Arteta et al. (2001), Edwards (2001), Bekaert et al. (2005), Alfaro et al. (2008), and Papaioannou (2009) focused on the role of institutions and show that liberalization in developing countries leads to larger capital inflows, and higher investment which in turn improve long-run economic growth. Authors of the second group argued that liberalization is a principal threat of economic stability and the main cause of banking crisis. The studies by Díaz-Alejandro (1985), Kaminsky and Reinhart (1999) and Kose et al.(2003) show that liberalization generate high macroeconomic volatility. Kaminsky and Reinhart (1999) presented evidence of a positive association between banking crises and financial liberalization for a panel of 20 countries observed during the period 1970–1995. They found that 18 out of 26 banking crises studied were preceded by a liberalization of the financial sector. They concluded that the probability of a banking crisis increases by 40% if a country liberalizes its domestic banking system. In another study, Demirgüç-Kunt and Detragiache (1999) studied a sample of 53 countries observed during the period 1980–1995 and their results revealed that banking crises were more likely to occur in a deregulated financial system. Weller (2001) surveys 26 emerging economies using monthly data into before and after financial liberalization. His results suggest strong evidence of increasing frequency and severity of financial crises in the period that follows the date of liberalization. Noy (2004) conducted a study to investigate the link between financial liberalization and banking crises for a sample of 61 non-OECD countries during the period 1975–1997. He employed an empirical model based on panel probit estimation and concluded that if liberalization is accompanied by insufficient prudential supervision of the banking sector, it will result in excessive risk taking by financial intermediaries and a subsequent crisis in the medium-run. He stated that more immediate danger is the loss of monopoly power that liberalization typically entails. Arteta and Eichengreen (2002) surveyed a sample of 75 emerging markets and developing countries during the period of 1975–1997. Their results suggest that capital account liberalization increases the likelihood of banking crises for countries that liberalize interest rate controls. Reinhart and Rogoff (2008) examined the determinants of banking crises for a large sample of countries over the period 1800–2008. They found that, since the early 19th century, there was a strong correlation between capital mobility and banking crises. The same study also showed that during the periods where capital mobility was interrupted,2 there was a remarkable decrease in banking crises. In a different type of analysis, Kaminsky (2008) examined the determinants of sudden stop of international capital flows in 26 emerging countries. She showed that a high level of financial integration increases the risk of sudden stop of capital flows, even in the absence of macroeconomic imbalances found in the host country. More recently, Joyce (2010) conducted a study to assess the effect of financial integration on the costs and duration of systemic banking crises for 20 emerging countries over the years 1976–2002. He showed that the nature of capital flows (in and out) plays a very important role on the stability of the banking sector of a country. He also found that an increase in foreign direct investment in a country tends to decrease the number and

2 For example, after the Second World War until the 1970s. According to Lane and Milesi-Ferretti (2005), the accumulation of larger stocks of gross foreign assets and liabilities has increased the magnitude of fluctuations in the value of cross-border holdings. Several other studies showed the close link between financial liberalization and banking crises.

126

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

duration of shocks, while foreign debt liabilities have the opposite effect. Furceri et al. (2011) studied the relationship between foreign capital inflows and the probability of occurrence of banking crises for a panel of 112 developed and developing countries observed during the period 1970–2007. The authors found that episode of important inflows of foreign capital significantly increase the probability of banking crises in the next two years. The authors found that the probability of occurrence of crises is even more important if capital inflows consist primarily of debt flows. According to the brief studies cited above, one may conclude that there is no strong consensus regarding the real effects of financial liberalization. Therefore, the main aim of this paper is to shed light on the relationship between financial stability and financial liberalization by focusing on several new aspects. Following the previous empirical research, we distinguish between the dejure and defacto financial liberalization and we will be particularly interested in the bidirectional aspect of financial liberalization, namely foreign assets (outflow) and foreign liabilities (inflow). We focus on the components of these two aspects to distinguish between foreign direct investment, portfolio investment and debt. 3. Methodology 3.1. Data and variables In this study, we use a Panel of 58 developing countries for a time period from 1984 to 2007. The list of these countries is presented in Annex 1. The main variable of the study is banking crises episodes which were drawn from the surveys of Caprio and Klingebiel (1996) Demirguc-Kunt and Detragiache (1997), recently updated by Caprio et al. (2005) and Laeven and Valencia (2012). The set of explanatory variables will be the same as Demirguc Kunt and Detragiache (2005) and they are categorized as follows: BC ¼ f ðBV; MV; FLVÞ: BC is the dependent variable that reflects the probability of occurrence of a bank crisis. It takes a value of 1 if a country experienced a crisis and zero for otherwise. BC is a function of three vectors: the first vector includes banking variables (BV); the second includes macroeconomic variables (MV) while the last vector includes indicators of financial liberalization (FLV). BV is a vector of 3 variables that reflects some characteristics of a country's banking sector. These variables are the ratio of broad money to the foreign exchange reserves of the central bank (M2/RES), which captures the vulnerability of the economy to sudden capital outflows triggered by a run on the currency (Büyükkarabacak and Valev 2008). Greater M2 to reserves ratio is expected to raise the likelihood of banking crises. We use the ratio of bank credit to the private sector scaled by GDP (CPS/GDP) which is an indicator of financial development of a country. We also use the growth of bank credit to the private sector (GC) which reflects the dynamic and evolution of lending activities in a country. MV is a set of macroeconomic variables which includes real gross domestic product per capita (GDPpc) which has been negatively linked to banking crises (Demirgüç-Kunt and Detragiache 1997). We include the rate of real GDP growth (GDPGR) which captures macroeconomic developments that affect the quality of bank assets. This ratio is expected to minimize the effects of financial crises. We will use Inflation rate (Inf.) which is measured by percentage change in the consumer price index. Finally, there would be a variable to measure the degree of openness of the economy. This variable is defined as export plus import scaled by GDP (Op.). The intuition for including this variable is straightforward. The source of all these variables is presented in Annex 2.

Regarding the variables of financial liberalization (FL), they are divided in 2 categories: dejure and defacto. The defacto measure is extracted from the database of Lane and Milesi-Ferretti (2007), updated in 2009. According to these authors, international financial integration (IFI) is measured by an index constructed through a dataset for the stocks of gross foreign assets and liabilities for 145 nations during the period of 1970–2004. It is as follows:

GDP

IFIit

¼

GFAi;t þ GFLi;t GDP i;t

where GFAi,t is the stock of gross foreign direct assets and GFLi,t is the gross stock of foreign direct investment liabilities. Following Demirguc-Kunt and Detragiache (1998a, b), Choudhry and de Haan (2009) Joyce (2010), among others, we use foreign assets and foreign liabilities scaled by nominal GDP, noted FA/GDP and FL/GDP respectively. The dejure index is the second indicator of financial liberalization. It is an index constructed by Chinn and Ito (2007) which includes four variables: variable indicating the presence of multiple exchange rates, variable indicating restrictions on current account transactions; variable indicating restrictions on capital account transactions; and variable indicating the requirement of the surrender of export proceeds. Before starting analyzing the empirical results, it is worth to have a look at the descriptive statistics of the different variables of the model. The results are displayed in Table 1. As we can see real economic growth of our sample is 3.42% which is moderate. However, growth rate of per capita GDP is relatively high (6.83%) with a minimum of 4.05% and maximum of 9.14%. Regarding inflation, the average rate is 2.49% which is reasonable. Credit growth appears to be weak as it stood at 1.55% on average. This indicator reveals the weak contribution of banks in the economic activities through financing channel and this can also explain the weak linkage between financial sector development and economic growth. All the variables of the model are lagged one year before the date of defacto financial liberalization to test the effect of the level of capital flows on the probability of occurrence of a banking crisis. 3.2. The model The econometric estimation is based on a panel logit regression approach3 as the dependent variable, which is the probability of occurrence of a banking crisis (BC), is assumed to be a binary choice variable: BC = 1 if a crisis occurs and 0 otherwise. The empirical estimation is based on a random sample of N developing countries, i = 1,…, N observed during T periods, where t = 1,…, T. Let yi,t⁎ denote whether country i experienced a banking crisis in year t. We suppose that yi,t⁎ is a linear function of banking and macroeconomic variables Xi,t, of two indicators of financial liberalization FLi,t (dejure and defacto), of a constant α and of a random error term ɛit.4 Hence, the linear function can be expressed as follows:



yi;t ¼ ∝ þ β1 FLi;t þ β2 Xi;t þ εi;t :

ð1Þ

3 With the use of panel logit regression we can see how changes in the different explanatory variables affect the probability of a bank crisis. In this paper, we preferred a random rather than a fixed logit model because the use of the later procedure requires the elimination, from the sample, of all countries that have not experienced crises. In our case this means the loss of a large quantity of information. 4 The errors are assumed to be independent and identically distributed across countries and over time. They are also assumed to have zero mean and unit variance. Further, they are supposed to have a logistic distribution.

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

127

Table 1 Statistics table.

BC GDPGR GDPpc M2/Reserves Inflation CPS/GDP CG Op. Dejure Defacto Liabilities/GDP (t − 1) Assets/GDP (t − 1) FDIL/Tot.Liabilities FDIA/Tot.Assets Debt L/Tot.Liabilities Debt A/Tot.Assets. FPEL/Tot.Liabilities FPEA/Tot.Assets Size Debt/GDP

Obs.

Mean

St. Dev.

Min

Max

1536 1445 1457 1372 1237 1344 1272 1407 1404 1426 1426 1425 1425 1077 1426 1425 1426 1425 1386 1358

.1295573 3.421713 6.835229 39.16153 2.495222 29.27446 1.551145 69.45314 −.2339293 1.582131 1.251492 .3308708 .225361 .0755085 .7502077 .5630739 .0244106 .0131464 13.3876 84.00239

.3359251 6.847834 1.074326 370.6495 1.522796 27.04324 34.1434 37.30908 1.414839 2.166722 2.08078 .3496893 .1633689 .096875 .1842804 .2029139 .0558812 .033851 5.053444 101.1026

0 −51.03086 4.056728 .000116 −3.296556 .6827951 −9.863692 12.34638 −1.85564 .1660583 .0495588 .0162082 .0004923 .0000193 .0680942 .0797269 0 0 2.975538 .9711362

1 106.2798 9.147261 7427.439 10.10283 165.7191 1201.817 280.361 2.45573 31.25265 30.95495 4.653005 .9251236 .6109056 1 .9959071 .462468 .3938915 43.47921 2092.92

CPS is the credit to private sector, CG is the Credit Growth, Op. is openness, FDIL is the Foreign Direct Investment Liabilities, FDIA is the Foreign Direct Investment Assets, FPEL is the Foreign Portfolio Equity Liabilities, FPEA is the Foreign Portfolio Equity assets, Debt L is the Foreign Debt liabilities, and Debt A is the Foreign Debt Assets.

We assume that a country i experiences a crisis if y⁎i,t N 0. Hence:  BC i;t ¼

1 0



if yi;t ≻0 otherwise

:

ð2Þ

in the total foreign direct investment and to give a better understanding of which ratio is the most likely to trigger a banking crisis. 4. Results 4.1. Determinants of banking crises

Consequently, the probability that a country i experiences a banking crisis in the year t can be expressed as follows:        P BC i;t ¼ 1 ¼ P yi;t ≻0 ¼ F β 1 FLi;t þ β2 Xi;t   exp β1 FLi;t þ β2 Xi;t   ¼ 1 þ exp β 1 FLi;t þ β2 Xi;t

ð3Þ

where F is the partition function and the probability of no crisis will happen is:     P BC i;t ¼ 0 ¼ 1−P BC i;t ¼ 1 ¼ 1− ¼

  exp β1 FLi;t þ β2 X i;t   1 þ exp β1 FLi;t þ β2 X i;t

1  : 1 þ exp β1 FLi;t þ β2 X i;t

ð4Þ

In this paper, we use a new method to estimate how foreign direct investment could generate a banking crisis. In some previous studies (i.e. Joyce, 2011), researchers have used three5 components of foreign direct liabilities scaled by the country's GDP. In this paper, we use six lagged indicators namely: Debt/Total Assets, Foreign Direct Investment Assets/Total Assets, Foreign Portfolio Equity Assets/Total Assets, Debt Liabilities/Total Liabilities, Foreign Direct investment Liabilities/Total Liabilities and Foreign Portfolio Equity Liabilities/Total Liabilities. The aim of this disaggregation is to measure the weight of each component

5 Foreign direct investment (FDIL/GDP), foreign portfolio equity liabilities (FPEL/GDP) and foreign debt investments (FDL/GDP).

Table 2 reports the estimations' results of the determinants of banking crises and the role of financial globalization on the triggering of these crises. As one can see, there are three different estimations. In the first one, we include the control variables commonly used in the empirical literature dealing with the issue of banking crises (Demirguc-Kunt and Detragiache, 1998a, b; Choudhry and de Haan, 2009; Bonfiglioli, 2008, Joyce, 2011). As we mentioned above, selected control variables are lagged for one period. The main results can be summarized as follows. First, the coefficient of economic growth, proxied by GDP growth (GDPGR), is negative and significant at the 1% level. This shows that banking crises get triggered when the economic performance of a country is weak. More rapid economic growth is associated with increasing incomes and a low probability of a banking crisis. Second, the coefficient of inflation is positive and significant at the 5% level. This indicates that a banking crisis is likely to occur in countries with high inflation rate. This result is in line with the literature that shows a positive effect of inflation on the likelihood of a banking crisis. Third, the variable M2/reserves is negative but not significant while the domestic credit growth (CG) and private sector credit to GDP (CPS/GDP) are both positive and significant at the 1% level. In terms of marginal effects, the results show that the coefficients of CG and CPS/GDP are 1.44% and 1.55 respectively. This confirms the effect of rapid domestic credit growth on the probability of occurrence of banking crises. Despite the huge finance and growth literature showing the positive relationship between greater credit levels and economic growth, it appears that rapid growth in bank credit to the private sector is associated with banking crises in developing countries. Our finding joint the previous studies (Demirgüc-Kunt and Detragiache 1997 Kaminsky and Reinhart, 1999 and Kaminsky and Reinhart, 1999) and follows the IMF (2004) estimates which found that about 75% of credit booms in emerging markets end in banking crises. In Eq. (2), we added two variables that reflect the liberalization process. The first variable is openness (Op.) and the second one is

128

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

Table 2 Determinants of banking crisis. 1

2

3

Variables

BC

Marginal effect

BC

Marginal effect

BC

Marginal effect

GDPGR (t − 1) GDPpc (t − 1) M2/Reserves (t − 1) Inflation (t − 1) CPS/GDP (t − 1) CG. (t − 1) Op. (t − 1) Defacto (t − 1) Liabilities/GDP (t − 1) Assets/GDP (t − 1) Constant

−0.106⁎⁎⁎ (0.0216) −0.0623 (0.161) −0.00160 (0.00167) 0.195⁎⁎ (0.0942) 0.0187⁎⁎⁎ (0.00576) 0.173⁎⁎⁎ (0.0658)

−0.00878⁎⁎⁎ (0.00200) −0.00517 (0.0133) −0.000132 (0.000140) 0.0161⁎⁎ (0.00776) 0.00155⁎⁎⁎ (0.000473) 0.0144⁎⁎ (0.00587)

−0.107⁎⁎⁎ (0.0226) −0.0267 (0.155) 0.0246⁎⁎⁎ (0.00813) 0.231⁎⁎ (0.0963) 0.0215⁎⁎⁎ (0.00588) 0.195⁎⁎ (0.0805) −0.00821⁎ (0.00447) −0.0691 (0.125)

−0.00904⁎⁎⁎ (0.00206) −0.00226 (0.0132) 0.00208⁎⁎⁎ (0.000719) 0.0195⁎⁎ (0.00815) 0.00182⁎⁎⁎ (0.000494) 0.0166⁎⁎ (0.00733) −0.000696⁎ (0.000378) −0.00586 (0.0106)

−0.107⁎⁎⁎ (0.0226) −0.0227 (0.166) 0.0245⁎⁎⁎ (0.00817) 0.230⁎⁎ (0.0971) 0.0215⁎⁎⁎ (0.00587) 0.196⁎⁎ (0.0805) −0.00813⁎ (0.00462)

−0.00903⁎⁎⁎ (0.00206) −0.00192 (0.0140) 0.00208⁎⁎⁎ (0.000723) 0.0195⁎⁎ (0.00822) 0.00182⁎⁎⁎ (0.000494) 0.0166⁎⁎ (0.00733) −0.000689⁎ (0.000390)

Observations Number of id Log lik Pseudo R2

−2.561⁎⁎ −1080 1000 58 −333.5 0.430

−0.0644 (0.142) −0.104 (0.527) −2.455⁎⁎ (1.153) 979 58 −321.1 0.451

−2.430⁎⁎ (1.096) 979 58 −321.1 0.451

−0.00546 (0.0121) −0.00884 (0.0447)

z-statistics are shown in parentheses. ⁎ Indicates 1% level of significance. ⁎⁎ Indicates 5% level of significance. ⁎⁎⁎ Indicates 10% level of significance.

defacto liberalization and both are lagged of one year. The results of the model are similar to the previous one but M2/reserves becomes positive and significant at the 1% level. This shows that, following the liberalization process, when the financial sector holds considerable foreign currency liabilities it becomes more exposed to a banking and financial crises. The coefficient of openness (Op.) is negative and significant at the 10% level. This indicates that openness to external financial systems could reduce the probability of triggering a banking crisis. Regarding defacto liberalization, it has a negative sign but not significant. Turning now to the third equation; we replaced defacto liberalization by its two components which are foreign assets to GDP and foreign liabilities to GDP. Despite this disaggregation, the result is identical to model 2. 4.2. Defacto and dejure liberalization and banking crisis 4.2.1. Defacto liberalization and banking crisis In the section above we followed the previous empirical studies and we proxied defacto liberalization by an index of international financial integration (IFI). Empirical results show that defacto liberalization does not have any considerable role in triggering financial and banking crises. By dividing defacto into Liabilities/GDP and Assets/GDP ratios, we also found similar results. In this section we disaggregate these two measures by type of capital flows. Therefore, we will use the six ratios as explained above. The results are displayed in Table 3. As we have six indicators, we will separately introduce each ratio along with the set of explanatory variables. Unlike Joyce (2011), we keep the openness variable (OP) as it was significant and positive in Eqs. (2) and (3) of Table 2. In the first equation, we introduced the debt to total foreign assets ratio (debt/Tot.A) and we found that it does not trigger a banking crisis in developing countries. We also found that the results of the explanatory variables remained unchanged except for M2/Reserves which became positive and significant at the 5% level. Again, this shows the problem of holding large foreign currency liabilities. This scenario is the main reason behind the Asian financial crisis of 1997 in which the currency depreciation ravaged the financial sector of the so called “Tiger countries”. In Eq. (2) when we added foreign direct investment assets to total assets

(FDIA/Tot.A) we also found that this ratio does not trigger a crisis. However, inflation, credit growth and openness have become non-significant. Similar results were found while introducing foreign portfolio equity assets ratio (FPEA/Tot.A). In Eq. (4) when introducing debt to total foreign liabilities ratio we found that it could trigger a financial crisis and all the explanatory variables, except GDP per capita, are significant. In terms of marginal effects, a one percent increase in foreign debt entry raises the probability of banking crises in developing economies, by 17.5%. This shows that crises could happen in countries with a large debt ratio. The recent experience of PIIGS6 countries and Cyprus are the best witness of problem related to high debt. Similar results were found in with foreign direct investment liabilities to total foreign liabilities ratio but openness becomes insignificant. In terms of marginal effects, a one percent increase in foreign direct investment entry reduces the probability of banking crises in developing economies, by 22.3%. Finally, we found that foreign portfolio equity liabilities to total liabilities ratio is positive but not significant which means that it could not engender a crisis in these developing countries. In light of the results of the marginal effect of FDI and debt inflows, we can conclude that the probability that developing countries will experience a banking crisis mainly depends on these two types of capital inflows.

4.2.2. Dejure liberalization and banking crisis In this section, we aim at measuring whether or not dejure liberalization triggers banking crises in developing countries. To this end, we use only the dejure indicator and we introduce the different variables which are in liaison with the dejure liberalization. The results are displayed in Table 4. In the first equation, the dejure variable is negative and not significant. This shows that it does not trigger banking crises in the developing countries of our sample. On the other hand, all the explanatory variables, except GDP per capita, are significant. In the second equation, we introduced foreign liabilities to GDP ratio but the results remained unchanged and dejure does not produce a crisis. Again, when introducing the size of the government, measured by the

6

Portugal, Ireland, Italy, Greece and Spain.

Table 3 Banking crises and Defacto liberalization. 1 Variables

BC

BC

3 Marginal effect

BC

−0.104⁎⁎⁎ (0.0215) −0.000925 (0.0130) −0.0598 (0.150) 0.00173⁎⁎ (0.000758) 0.0213⁎⁎⁎ (0.00829) 0.0184⁎⁎ (0.00809) 0.236⁎⁎ (0.0949) 0.00194⁎⁎⁎ (0.000507) 0.0213⁎⁎⁎ (0.00577) 0.0162⁎⁎ (0.00714) 0.196⁎⁎

−0.00892⁎⁎⁎ (0.00197) −0.00511 (0.0128) 0.00182⁎⁎ (0.000728) 0.0202⁎⁎ (0.00806) 0.00182⁎⁎⁎

(0.0810) −0.000761⁎⁎ (0.000358) −0.00916⁎⁎ (0.00412) 0.0748 (0.0606) 2.101 (1.331)

−0.00822⁎⁎⁎ (0.00198)

4 Marginal effect

BC

BC

Marginal effect

−0.106⁎⁎⁎ −0.00916⁎⁎⁎ (0.0216) (0.00202) 0.0826 (0.163) 0.00717 (0.0142) 0.00203⁎⁎⁎ 0.0234⁎⁎⁎ (0.000720) (0.00799) 0.202⁎⁎ 0.0175⁎⁎ (0.0967) (0.00834) 0.0205⁎⁎⁎ 0.00178⁎⁎⁎

−0.0986⁎⁎⁎ (0.0216) 0.111 (0.162) 0.0213⁎⁎⁎ (0.00822) 0.173⁎ (0.0979) 0.0225⁎⁎⁎

−0.00818⁎⁎⁎ (0.00193) 0.00923 (0.0134) 0.00177⁎⁎ (0.000710) 0.0144⁎ (0.00815) 0.00187⁎⁎⁎

−0.100⁎⁎⁎ −0.00822⁎⁎⁎ (0.0216) (0.00191) 0.108 (0.160) 0.00884 (0.0131) 0.0211⁎⁎ 0.00173⁎⁎ (0.00820) (0.000704) 0.162⁎ 0.0133⁎ (0.0980) (0.00807) 0.0215⁎⁎⁎ 0.00177⁎⁎⁎

−0.104⁎⁎⁎ (0.0218) −0.0201 (0.157) 0.0248⁎⁎⁎ (0.00817) 0.239⁎⁎ (0.0960) 0.0214⁎⁎⁎

−0.00879⁎⁎⁎ (0.00199) −0.00171 (0.0133) 0.00210⁎⁎⁎ (0.000723) 0.0202⁎⁎ (0.00812) 0.00181⁎⁎⁎

(0.000482) 0.0168⁎⁎

(0.00577) 0.196⁎⁎

(0.000492) 0.0170⁎⁎

(0.00589) 0.200⁎⁎

(0.000490) 0.0166⁎⁎

(0.00588) 0.201⁎⁎

(0.000484) 0.0165⁎⁎

(0.00602) 0.195⁎⁎

(0.000506) 0.0165⁎⁎

(0.00742) −0.000783⁎⁎ (0.000348)

(0.0806) −0.00977⁎⁎ (0.00413)

(0.00752) −0.000848⁎⁎ (0.000356)

(0.0849) −0.00735⁎ (0.00422)

(0.00757) −0.000610⁎ (0.000352)

(0.0856) −0.00688 (0.00424)

(0.00755) −0.000565 (0.000349)

(0.0807) −0.00909⁎⁎ (0.00420)

(0.00734) −0.000771⁎⁎ (0.000354)

Observations Number of Id Log lik Pseudo R2

−2.336⁎⁎ (1.021) 979 58 −320.1 0.452

Marginal effect

−8.472 (7.353)

−0.736 (0.633) 2.103⁎⁎ (0.947)

0.175⁎⁎ (0.0779)

−2.713⁎⁎ (1.080)

−0.223⁎⁎ (0.0871) 0.711 (2.733) −2.553⁎⁎ (1.074) 979 58 −321.2 0.451

0.0602 (0.232)

0.179 (0.115)

FDIL/Tot.L (−1)

−3.127⁎⁎⁎ (1.166) 979 58 −320.5 0.452

6

Marginal effect

Debt/Tot.L (t − 1)

FPEL/Tot.L (t − 1) Constant

5

BC

−2.928⁎⁎⁎ (1.061) 962 58 −319.1 0.454

−5.031⁎⁎⁎ (1.523) 979 58 −318.6 0.455

−2.800⁎⁎⁎ (1.051) 979 58 −317.7 0.457

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

−0.0994⁎⁎⁎ (0.0217) GDP.cp (t − 1) −0.0112 (0.158) M2/Reserves (t − 1) 0.0209⁎⁎ (0.00875) Inf. (t − 1) 0.222⁎⁎ (0.0972) CPS/GDP (t − 1) 0.0234⁎⁎⁎ (0.00616) CG. (t − 1) 0.196⁎⁎ (0.0803) Op. (t − 1) −0.00921⁎⁎ (0.00430) Debt/Tot.A (t − 1) 0.905 (0.748) FDIL/Tot. A (t − 1) FPEL/Tot.A (t − 1) GDPGR (t − 1)

2 Marginal effect

z-statistics are shown in parentheses. ⁎ Indicates 1% level of significance. ⁎⁎ Indicates 5% level of significance. ⁎⁎⁎ Indicates 10% level of significance.

129

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

902 57 −277.9 .52458753

−0.00972⁎⁎⁎ (0.00225) −0.0107 (0.0134) 0.00188⁎⁎⁎ (0.000668) 0.0217⁎⁎ (0.00871) 0.00206⁎⁎⁎ (0.000489) 0.0248⁎⁎ (0.0110) −0.000801⁎⁎ (0.000373) 0.000605 (0.00822) −0.0315 (0.0205) 0.00303 (0.00226) 0.000434 (0.000316) −0.129⁎⁎⁎ (0.0270) −0.142 (0.179) 0.0251⁎⁎⁎ (0.00839) 0.288⁎⁎ (0.115) 0.0274⁎⁎⁎ (0.00655) 0.330⁎⁎ (0.137) −0.0107⁎⁎ (0.00499) 0.00806 (0.109) −0.419 (0.272) 0.0403 (0.0303) 0.00578 (0.00422) −2.452⁎ (1.319) −0.0102⁎⁎⁎ (0.00230) −0.0104 (0.0138) 0.00203⁎⁎⁎ (0.000702) 0.0245⁎⁎⁎ (0.00892) 0.00200⁎⁎⁎ (0.000510) 0.0170⁎⁎ (0.00792) −0.000754⁎ (0.000386) −0.00218 (0.00844) −0.0125 (0.0125) 0.00320 (0.00232) −2.411⁎ (1.300)

931 57 −288.0 .50741142 954 58 −302.5 .4826219

5. Conclusion

−2.030 (1.245)

−0.133⁎⁎⁎ (0.0262) −0.136 (0.180) 0.0265⁎⁎⁎ (0.00858) 0.319⁎⁎⁎ (0.114) 0.0262⁎⁎⁎ (0.00662) 0.222⁎⁎ (0.0949) −0.00984⁎ (0.00505) −0.0285 (0.110) −0.164 (0.162) 0.0417 (0.0304) −0.130⁎⁎⁎ (0.0257) −0.123 (0.179) 0.0259⁎⁎⁎ (0.00860) 0.295⁎⁎⁎ (0.110) 0.0250⁎⁎⁎ (0.00655) 0.210⁎⁎ (0.0907) −0.00773 (0.00489) −0.0996 (0.110) −0.107 (0.157) −0.00976⁎⁎⁎ (0.00215) −0.00685 (0.0135) 0.00202⁎⁎⁎ (0.000722) 0.0231⁎⁎⁎ (0.00886) 0.00201⁎⁎⁎ (0.000522) 0.0166⁎⁎ (0.00783) −0.000700⁎ (0.000368) −0.00894 (0.00853) −0.123⁎⁎⁎ (0.0237) −0.0866 (0.171) 0.0256⁎⁎⁎ (0.00861) 0.292⁎⁎⁎ (0.110) 0.0254⁎⁎⁎ (0.00655) 0.210⁎⁎ (0.0914) −0.00886⁎ (0.00465) −0.113 (0.109)

The aim of this paper is to examine the consequences of financial liberalization on the economic and financial stability of developing countries. Precisely, we tried to understand whether the adoption of financial liberalization was followed by banking crises or not. For this purpose we used data of 58 developing countries observed during the period 1984–2007. The empirical analysis is based on panel logit model and the econometric procedure was performed in two steps. In the first step, we used the dejure and defacto measures of capital account liberalization to analyze the determinants of banking crises. Empirical results show that dejure liberalization does not trigger a banking crisis in our sample. Regarding defacto liberalization, results also show that it does not have any considerable role in triggering financial and banking crises. By dividing defacto into Liabilities/GDP and Assets/GDP ratios, we also found similar results. In the second step, we split the stock of assets and liabilities into six components to measure the weight of each indicator in the total foreign assets and liability separately. The empirical results show that foreign debt liabilities to total liabilities and foreign direct investment liabilities to total liabilities generate banking crises in our sample of developing countries.

954 58 −302.7 .48221632 Observations Number of Id Log lik Pseudo R2

z-statistics are shown in parentheses. ⁎ Indicates 1% level of significance. ⁎⁎ Indicates 5% level of significance. ⁎⁎⁎ Indicates 10% level of significance.

−2.350⁎⁎ (1.165)

Annex 1. List of countries

GDPGR (t − 1) GDP.pc (t − 1) M2/Reserves (t − 1) Inf. (t − 1) CPS/GDP (t − 1) CG. (t − 1) Op. (t − 1) L.F.dejure (t − 1) Liabilities/GDP (t − 1) Size Public debt/GDP (t − 1) Constant

1

government final consumption expenditure as a share of GDP, we do not find any change. Finally, in the last equation, we added the public debt of the country as a share of GDP. It is well known that larger government budget deficits are expected to increase the probability of crises. In the estimation, we find that the public debt ratio is positive and significant at the 10% level and the signs of the other explanatory variables remain unchanged. It appears that when the budget deficit increases, developing countries are incapable to support the high costs of insolvent banks. Moreover, governments facing severe fiscal imbalances are more likely to use the financial sector as an off-budget source of funding for government objectives, by pressuring banks to direct loans to favored borrowers. According to Keefer (2001), since securing repayment of loan obligations from these borrowers is typically a difficult proposition for banks, these pressures can translate into solvency difficulties for the financial system.

−0.0103⁎⁎⁎ (0.00231) −0.00974 (0.0141) 0.00204⁎⁎⁎ (0.000721) 0.0233⁎⁎⁎ (0.00886) 0.00198⁎⁎⁎ (0.000521) 0.0166⁎⁎ (0.00778) −0.000612 (0.000386) −0.00788 (0.00865) −0.00848 (0.0124)

3

BC Variables

Table 4 Banking crises and dejure liberalization.

Marginal effect 2

BC Marginal effect BC

4

BC Marginal effect

Marginal effect

130

Algeria Argentina Armenia Azerbaijan Bangladesh Belarus Bolivia Brazil Bulgaria Burkina Faso Cameroon China Colombia Congo. Dem. Rep. Congo. Rep. Dominican Republic Ecuador Egypt. Arab Rep. El Salvador Ghana

Guinea Guinea-Bissau Guyana Haiti Indonesia Jamaica Jordan Kenya Latvia Liberia Lithuania Madagascar Malaysia Mali Mexico Morocco Mozambique Nicaragua Niger Nigeria

Panama Peru Philippines Romania Russian Federation Senegal Sri Lanka Tanzania Thailand Togo Tunisia Turkey Uganda Ukraine Venezuela. RB Vietnam Yemen. Rep. Zambia

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

131

Annex 2. Definitions and sources of variables in the model Variables

Definition & source

BC GDP GDP.pc M2/Reserves Inf. CPS/GDP C.G Op.

Dummy variable equals 1 if a country i experienced a systemic banking crisis at date t and 0 otherwise. Source: Laeven and Valencia (2012). GDP growth (annual %). Source: World Development Indicators (WDI) GDP per capita (constant 2000 US$). Source: World Development Indicators (WDI). Money and quasi money (M2) to total reserves ratio. Source: World Development Indicators (WDI). Inflation, consumer prices (annual %). Source: World Development Indicators (WDI) Domestic credit to private sector (% of GDP). Source: World Development Indicators (WDI). Claims on private sector (annual growth as %). Source: IFS Trade (% of GDP) = the sum of exports and imports of goods and services measured as a share of gross domestic product. Source: World Development Indicators (WDI). Dejure liberalization. Source: Chinn and Itô (2008) actualisée en (2011) Defacto liberalization = (stock of external assets + stock of external Liabilities/GDP). Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of external Liabilities/GDP. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of external assets/GDP. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of direct investment Liabilities/Total Liabilities. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of direct investment assets/Total Assets. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of debt liabilities/Total Liabilities. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of debt assets/Total Assets. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of portfolio equity liabilities/Total Liabilities. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 Stock of portfolio equity assets/Total Assets. Source: Lane and Milesi-Ferretti (2007) updated to cover the period 1970–2011 General government final consumption expenditure (% of GDP). Sources: World Development Indicators (WDI) Gross general government debt to GDP. Source: Abbas et al. (2010)

Dejure Defacto Liabilities/GDP (t − 1) Assets/GDP (t − 1) FDIL/Tot.Liabilities FDIA/Tot.Assets. Debt.L./Tot.Liabilities Debt.A./Tot.Assets. PEQL/Tot.Liabilities PEQA/Tot.Assets Size Public Debt/GDP

References Abbas, S.M. Ali, Belhocine, N., El-Ganainy, A., Horton, M., 2010. A Historical Public Debt Database, IMF Working Paper WP/10/245. , Washington, DC. Alfaro, L., Kalemli-Ozcan, S., Volosovych, V., May 2008. Why Doesn't Capital Flow from Rich to Poor Countries? An Empirical Investigation. The Review of Economics and Statistics, MIT Press 90 (2), 347–368. Angkinand, A.P., Sawangngoenyuang, W. And, Wihlborg, C., 2010. Financial liberalization and banking crises: a cross-country analysis. Int. Rev. Finance 10, 263–292. Arteta, C., Eichengreen, B., 2002. Banking crises in emerging markets: presumptions and evidence. In: Blejer, M., Skreb, M. (Eds.), Financial Policies in Emerging Markets. MIT Press, Cambridge, MA. Arteta, C., Eichengreen, B., Wyplosz, C., 2001. When Does Capital Account Liberalization Help More than it Hurts? NBER Working Paper no. (8414). Balassa, B., 1989a. Financial liberalization in developing countries. Background Paper for the World Development Report. The World Bank September 1989 WPS 55. Balassa, B., 1989b. Financial liberalization in developing countries. Background paper for the World Development Report. The World Bank September 1989 WPS, 55. Bekaert, G., Harvey, C.R., Lundblad, C., 2004. Growth volatility and financial liberalization. , Unpublished working paper, Duke University. Bekaert, G., Harvey, C.R., Lundblad, C., 2005. Does financial liberalization spur growth? J. Financ. Econ. 77 (1), 3–56. Bonfiglioli, Alessandrar, 2008. Financial integration, productivity and capital accumulation. Journal of International Economics, Elsevier, vol. 76(2), pp. 337–355 (December). Büyükkarabacak, B., Valev, N., 2008. Credit Expansions and Banking Crises: The Roles of Household and Firm Credit. , Georgia State University, ASSA Meetings of the North American Economics and Finance Association. Caprio, G., Klingebiel, D., 1996. Bank insolvencies: cross-country experiences. World Bank Policy Research Paper, 1620. Caprio, G., Klingebiel, D., Laeven, L., Noguera, G., 2005. Banking crisis database. In: Honohan, Patrick, Laeven, Luc (Eds.), In Systemic Financial Crises: Containment and Resolution. Cambridge University Press, New York, NY. Chinn, M., Itô, H., 2007. A new measure of financial openness. La Follette School Working Paper No. 2007–033. Chinn, M., Itô, H., 2008. A new measure of financial openness. J. Comp. Policy Anal. 307–320. Choudhry, T.S., de Haan, J., 2009. Financial reform and banking crises. CESifo Working Paper Series, 2870. CESifo Group Munich. Demirguc Kunt, A., Detragiache, E., 2005. Cross-country empirical studies of systemic bank distress: a survey. IMF Working Papers 05/96. International Monetary Fund. Demirguc-Kunt, A., Detragiache, E., 1998a. Financial liberalization and financial fragility. World Bank, Working Paper(N. 1917). Demirguc-Kunt, A., Detragiache, E., 1997. The Determinants of Banking Crises - Evidence from Developing and Developed Countries, IMF Working Papers 97/106, International Monetary Fund. Demirgüç-Kunt, A., Detragiache, E., 1828. The Determinants of Banking Crises: Evidence from Industrial and Developing Countries. Policy Research Working Paper, World Bank. , Washington, D.C.1997. Demirguc-Kunt, A., Detragiache, E., 1998b. The determinants of banking crises in developing and developed countries. IMF Staff. Pap. 45 (1), 81–109.

Demirgüç-Kunt, A., 1999. Detragiache. , E. Financial liberalization and financial fragility, World Bank Policy Research Working Paper No 1917. Demirguc-Kunt, A., Detragiache, E., 2000. Monitoring banking sector fragility: a multivariate logit approach. World Bank Econ. Rev. 287–307. Diaz-Alejandro, C.F., 1985. Good-bye financial repression, hello financial crash. Journal of Development Economics (19). Edwards, S., 2001. Capital Flows and Economic Performance: Are Emerging Economies Different? NBER Working Paper no. 8076. Furceri, D., Guichard, S., Rusticelli, E., 2011. The Effect of Episodes of Large Capital Inflows on Domestic Credit. OECD Economics Department Working Paper 864. Paris France. Hamdi, H., Hakimi, A., Djelassi, M., 2013. Did financial liberalization lead to bank fragility? Evidence form Tunisia. Int. J. Bus. Financ. V7 (5). Henry, P.B., 2007. Capital account liberalization: Theory, evidence, and speculation. Journal of Economic Literature 45 (4), 887/935. IMF, 2012. Liberalizing Capital Flows and Managing Outflows, Background Paper. International Monetary Fund, WDC (March 16, 2012). Joyce, J., 2010. Financial Globalization and Banking Crises In Emerging Markets. Open Econ. Rev. 1–21. Joyce, J., 2011. Financial Globalization and Banking Crises in Emerging Markets. Open Economies Review, Springer vol. 22 (5), 875–895. Kaminsky, Graciela, 2008. Crises and sudden stops: evidence from international bond and syndicated loan markets. NBER Working Paper, 14249. Kaminsky, Graciela L., Reinhart, Carmen M., 1999. The twin crises: the causes of banking and balance of payments problems. Am. Econ. Rev. Vol. 89 (No.3), 473–500. Keefer, P., 2001. When do special interests run rampant? Disentangling the role in banking crises of elections, incomplete information, and checks and balances, Policy Research Working Paper Series 2543, The World Bank. King, R.G., Levine, R., 1993. Finance and growth: Schumpeter might be right. Quarterly Journal of Economics 108, 717–738. Kose, A., Prasad, E., Rogoff, K., Wei, S.-J., 2004. Financial globalization, growth and volatility in developing countries. CEPR Discussion Papers 4772, C.E.P.R. Discussion Papers. Kose, M.A., Prasad, E.S., Terrones, M., 2003. Financial Integration and Macroeconomic Volatility. IMF Staff Papers 50, 119–142. Laeven, L., Valencia, F., 2012. Systemic banking crises database: an update. IMF Working Papers 12/163. International Monetary Fund. Lane, Philip R., Milesi-Ferretti, Gian Maria, 2007. The external wealth of nations mark ii: revised and extended estimates of foreign assets and liabilities, 1970–2004. J. Int. Econ. 73 (2), 223–250. Levchenko, Andrei, Rancière, A., Romain, Mathias, Thoenig, July 2009. Growth and risk at the industry level: The real effects of financial liberalization. Journal of Development Economics, Elsevier vol. 89 (2), 210–222. Levine, Ross, 2001. International financial liberalization and economic growth. Rev. Int. Econ. 9 (4), 688–702. McKinnon, R.I., 1973. Money and Capital in Economic Development, The Brookings Institution, Washington D.C. Mehrez, G., Kaufman, D., 2000. Transparency, liberalization and banking crises. World Bank Policy Research Working Paper No. 2286, 2286. Mishkin, Frederic S., 2005. Is financial globalization benefical? NBER Working Papers, 11891. National Bureau of Economic Research, Inc. Mishkin, F.S., 2006. Globalization — a force for good? BIS Review 96/2006. Mishkin, Frederic S., 2008. The next great globalization: how disadvantaged nations can harness their financial systems to get rich. Princeton University Press (January 23, 2008).

132

H. Hamdi, N.B. Jlassi / Economic Modelling 41 (2014) 124–132

Noy, I, September 2004. Financial liberalization, prudential supervision, and the onset of banking crises. Emerging Markets Review, Elsevier. vol. 5(3), pages 341–359. Papaioannou, E., March 2009. What drives international financial flows? Politics, institutions and other determinants. Journal of Development Economics vol. 88 (2), 269–281. Prasad, Eswar, Rogoff, Kenneth, Wei, Shang-Jin, Ayhan Kose, M., 2003. Effects of financial globalization on developing countries: some empirical evidence. Research Department, International Monetary Fund. Occasional Paper No. 220. Reinhart, Carmen, Rogoff, Kenneth, 2008. Banking crisis: an equal opportunity menace. NBER paper, 14587. Schmukler, S.L., 2004a. Benefits and risks of financial globalization: challenges for developing countries. Globalization, Growth, and Poverty, World Bank Policy Research Report.

Schmukler, Sergio L., 2004b. Financial globalization: gain and pain for developing countries. Economic Review. Federal Reserve Bank of Atlanta, pp. 39–66 (issue Q 2). Shaw, E.S., 1973. Financial Deepening in Economic Development. , Oxford University Press, New York. Tornell, A., Westermann, F., Martinez, L., 2004. The positive link between financial liberalization, growth, and crises. National Bureau of Economic Research Working Paper Series No. 10293. Ucer, M.E., 1998. Notes on financial liberalization. Macroeconomic Management: New Methods and Current Policy Issues, Organized by African Economic Research Consortium (AERC), Nairobi, April 20–30 1998. Weller, C., 2001. Financial Crises and Financial Liberalization: Exceptional Circumstances or Structural Weaknesses.". Journal of Development Studies 38 (1), 98–126.