A EUROPEAN POLICY
This paper extends to test if the short and in the long run. Weak indica- the same short-run increase in cyclical tions are found that this may happen par- volatility arising from financial integration tially due to the anchoring of expectations is observed in this specific sample of “emerg-provided by the EU Accession, and to the ing markets". This work finds signs that, more robust institutional framework contrary to other emerging markets, this imposed by this process onto the countries in does not happen: for the future Member question. States, financial integration, similarly to the KEY WORDS: Enlargement, European outcome observed in mature market Union, financial liberalization, booms, 81 economies, reduces cyclical volatility both in busts, cycles, volatility.
Financial and capital flows' liberalization can play a fundamental role in increasing growth and welfare. Typically, emerging or developing economies seek foreign savings to solve the inter-temporal savings-investment problem. On the other hand, current account surplus countries seek opportunities to invest their savings. To the extent that capital flows from surplus to deficit countries are well intermediated and, therefore, put to the most productive use, they increase welfare. Liberalization can, however, also be dangerous, as has been witnessed in many past and recent financial, currency and banking crises. It can make countries more vulnerable to exogenous shocks. In particular, if serious macroeconomic imbalances exist in a recipient country, and if the financial sector is weak, be it in terms of risk management, prudential regulation and supervision, large capital flows can easily lead to serious financial, banking or currency crises. A number of recent crises, like those in East Asia, Mexico, Russia, Brazil and Turkey (described, for example, in IMF (2001)), and, to some extent, the Argentinean episode of late 2001, early 2002, have demonstrated the potential risks associated with financial and capital flows liberalization.
Central and Eastern Europe has a somewhat different experience, when compared to other emerging regions, concerning the financial liberalization process, as the process there seems to have been much less crisis-prone than in, for instance, Asia or Latin America. This maybe, at least partially, because the current high degree of external and financial liberalization in the Central Eastern European countries (CEECs), beyond questions of economic allocative efficiency, must be understood in terms of the process of Accession to the European Union. The EU integration process implies legally binding, sweeping liberalization measures-not only capital account liberalization, but investment by EU firms in the domestic financial services, and the maintenance of a competitive domestic environment, giving this financial liberalization process strong external incentives (and constraints). Those measures were implemented parallel to the development of a highly sophisticated regulatory and supervisory structure, again based on EU standards. This whole process happened also with the EU's technical and financial support, through specific programs-like the PHARE one, for these so-called Accession, and the TACIS, for the former Soviet Union ones- and direct assistance from EU institutions, like the European Commission, the European Parliament and the European Central Bank (also, on a very early stage of the transition process, the influence of the IMF in setting up policies and institutions in several countries in the region-an intervention widely considered to haven been successful-was important: see Hallerberg et al., 2002).
Additionally, EU membership seems to act as an anchor to market expectations (see Vinhas de Souza and Hölscher, 2001), limiting the possibilities of self- fulfilling financial crises and regional contagion (see Linne, 1999), which had the observed devastating effects in both Asia and Latin America (even a major event, like the Russian collapse of 1998, had very reduced regional side effects). Several regional episodes of financial systems' instability did happen (see Vinhas de Souza, 2002(a) and Vinhas de Souza, 2002(b)), but none with the prolonged negative consequences observed in other region (which was also due to the effective national policy actions undertaken after those episodes). This study's main aim is to expand the Kaminsky and Schmukler database (see Kaminsky and Schmukler, 2003), from now on indicated as K&S, to include the Accession and Acceding Countries from Eastern Europe (namely, for Bulgaria, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia and Slovenia). In their original work, K&S build an extensive database of external and financial liberalization, which includes both developed countries and countries from emerging regions (but not from Eastern Europe). With that, they create different indexes of liberalization (capital account, banking and stock markets: see Table I below) and using them individually and in an aggregate fashion, test for the effects and causality of this process on financial and real volatility, for the existence of differences between regions, and for the effects of the ordering of the liberalization process.
One underlying hypotheses of this work is that the existing regulatory and institutional framework in Eastern Europe, plus a more sustainable set of macro policies, played an important role in enabling liberalization to largely deliver the welfare enhancing outcomes that it is supposed to. Such an “anchoring" role of the European Union in the CEECs, through the process of EU membership, and through the effective imposition of international standards of financial supervision and regulation, may indicate that, beyond multilateral organizations like the IMF or the OECD, a greater, pro-active regional stabilizing role in emerging markets by regional actors, for instance, the United States, or by some regional sub-grouping, like Mercosur, may also be welfare enhancing for other “emerging" regions.
2. CAPITAL ACCOUNT
The achieving of capital account liberalization happened rather swiftly in most of the countries in our sample: by the mid 1990s, all bar Bulgaria and Romania had been declared Article VIII compliant (for those two countries, this happened in 1998: see Table II below).
One of the main driving forces behind this was the process of European Integration, for which external liberalization is a pre-requisite: in the early to mid-1990s, all the countries had signed Association Agreements with the European Union (frequently preceded by trade liberalization agreements with the EU, also called “Europe" trade agreements, usually with years given to the countries to prepare for their full implementation) and formally applied for EU membership. Another additional factor supporting liberalization was IMF and OECD membership: four of the larger countries in our sample became OECD members during the second half of the 1990s. Another factor to be considered, is the endogenous decision process to liberalize in a sustainable fashion.
3. BANKING SECTOR
Financial integration, in the form of the opening up the banking sector to foreign banks, is seen as being positive, on a micro level, as foreign banks are usually better capitalized and more efficient than their domestic counterparts (of course, the domestic banking sector eventually catches-up: for an indication of this process at the ACs, see, among others, Tomova et al., 2003). Also from a macroeconomic perspective, financial integration maybe positive for the Eastern European countries, both for long run growth and, as there are indications that foreign banks do not contract either their credit supply nor their deposit base, in helping to smooth the cycle (see de Haas and Lelyveld, 2003: they find some indication that this is linked to the better capitalization base and prudential ratios, as better capitalized domestic banks behave similarly to foreign banks). Given the bank-centered nature of virtually all the financial systems of the future Member States, this is particularly important for them. In most of the member states, the initial stage of the creation of the two-tier banking system, modeled on the Western European “universal bank" system, was characterized by rather liberal licensing practices and limited supervision policies (aimed at the fast creation of a de novo commercial, private banking sector: see Fleming et al., 1996, Balyozov, 1999, Enoch et al., 2002, Sörg et al., 2003). This caused a mushrooming of new banks in those countries in the early 1990s. Parallel to this, a series of banking crises, of varied proportions, affected most of those de novo banking systems, due to this lax institutional framework, inherited fragilities from the command economy period (the political need to support state-owned, inefficient industries, with the consequent accumulation of bad loans and also the financing of budget deficits), macroeconomic instability, risky expansion and investment strategies and also sheer inexperience, both from the investors and from regulators. Progressively, the re-capitalization, privatization and internationalization of the banking system (mostly into the hands of EU financial conglomerates), coupled with the implementation of a more robust, EU-modeled institutional framework, did away with most of those problems.
Two of the worst cases where the set of Baltic banking crises and the Bulgarian episode, which are described in more detail below. Other smaller banking crises happened in Estonia in 1994 and 1998, and in Latvia in 1994. Caprio and Klingebiel, 2003, report smaller episodes of “financial sector distress" in the Czech Republic (94-95), Hungary (93), Poland (91-93), Romania (98-00), Slovakia (97) and Slovenia (92-94). The initial proliferation of banks was, quite naturally, followed by a process of consolidation and strengthening-parallel to the privatization of the remnant state-owned components of the financial system- of the banking sector in most of those economies (in Bulgaria, from 81 banks in 1992 to 35 in 2001, in the Czech Republic from 55 in 1995 to 38 in 2001, Estonia, from 42 in 1992 to 7 currently, while Hungary had 33 banks in 2002, showing only a very slight decrease from the early 1990s, Latvia from 56 in 1994 to 23, Lithuania from 27 in 1993 to 13, in Poland from 81 in 1995 to 71 in 2001, in Romania from 45 in 1998 to 41 in 2001, in Slovakia from 22 in 2000 to 19 in 2001, and in Slovenia, where the number fell from 25 to 21 during 2001 alone).
This consolidation process was frequently led by foreign companies, which
now hold the majority of the assets of the banking system in virtually all of
them-contrary to the situation in the current EU Member States-bar Slovenia.
This process now has a component of regional expansion of the Eastern
European banks themselves, or, more precisely in most cases, the regional expansion of Western banks via some of their locally-owned subsidiaries (see Sörg et
al., 2003, ibid). The share of banking assets to GDP, nevertheless, is still far below
the Euro area average (which stood at around 265% of GDP by end 2001), compared with 47% in Bulgaria, 136% in the Czech Republic, 72% in Estonia and
Latvia, 32% in Lithuania, 63% in Poland, 60% in Hungary, 30% in Romania,
96% in Slovakia and 94% in Slovenia (data also for 2001). Another peculiar feature of the banking system in the region is that foreign currency lending -usually
euro-denominated-to residents is very high, especially in the Baltic republics:
with 80% of total loans in Estonia, 56% in Latvia and 61% in Lithuania. Also, the
Baltic countries have substantial shares of deposits by non-residents, with over
10% in Estonia and Lithuania and close to 5% in Latvia (Latvia, with its close
trading ties to Russia, has a particular strategy of selling itself as a stable financial services center to CIS depositors: see IMF, 2003(b), ibid). The supervision system has also substantially improved, and, following recent international-and
EU- best practice, is now centered in independent universal supervisory agencies
in the most advanced of those countries (Reininger et al., 2002, ibid., estimate
that the formal regulatory environment for the Czech Republic, Hungary and
Poland is actually above the EU, and that its actual enforcement level is at its
average;Liive, 2003, gives a description of the Estonian experience that culminated in the
creation of the EFSA -Estonian Financial Supervisory Authority- in January
3.1 BANKING CRISES IN EASTERN EUROPE
The Baltic bank crises were, to different degrees, linked to liquidity difficulties related tolerations with Russia (in the November 1992 Estonian case, by the freezing of assets held by some Estonian banks in their former Moscow headquarters, while the Latvian and Lithuanian episodes of, respectively, March and December 1995, were caused by the drying-up of lucrative trade-financing opportunities with Russia, whose export commodities, at that time, were still below world price levels) and regulatory tightening (Latvia, Lithuania), compounded by the elimination of credit opportunities with the implementation of the Estonian and Lithuanian CBAs (Currency Board Arrangements). In Lithuania, as in Bulgaria, the financing of the budget deficit also played a role. In the Estonian and Latvian cases, around 40% of the assets of the banking system where compromised, in the Lithuanian and Bulgarian cases, around a third. The Bulgarian 1996-1997 crisis eliminated a third of its banking sector, and led the country to hyperinflation (reaching over 2000% in March 1997, see Yotzov, 2002). Its roots lie in the political instability that preceded it (which, on its turn, led to inadequate real sector reform, with state-owned, loss making enterprises being financed via the budget deficit or through arrears with the, at the time, still mostly state-owned part banking sector: those arrears were, in turn, partially monetized by the Bulgarian National Bank -BNB- and the largest state bank, the State Savings Bank -SSB). Periodic foreign exchange crises (March 1994, February 1997) and bank runs (late1995, late 1996, early 1997) were part of this picture. The implementation of tighter supervisory procedures during 1996 (giving the BNB the power to close insolvent banks), and a tightening of policy actually led to more bank runs. A caretaker government in February 1997 (before a newly elected government took power in May) paved the way to longer lasting reform and the implementation of the CBA, with its tighter budget constraints towards both the government and the banking sector. This reform process happened with the support from multilateral institutionsamely, (namely the IMF).
4. STOCK MARKETS
The existence of stock markets is assumed to be beneficial for economic performance. In principle, it provides a way for companies to raise capital at lower costs than through simple banking intermediation, and because it is not as restricted a source of capital as internal financing. Also, it is assumed that the existence of alternative modes of finance may reduce the likelihood of credit crunches caused by problems with the banking sector (see Greenspan, 2000). Additionally, the existence of external ownership is (or was, given the recent problems with market-based governance in the US and the EU, and the shift towards a more regulated environment) assumed to provide better governance for the management of firms. The majority of economic analyses seem to support the position that a diversified financing mix is positive for economic growth and stability. As described in the previous section, all the financial sectors in the Member States are bank-centered, with stock markets playing marginal roles in most of them (and, in some, a very marginal role: in Bulgaria, Slovakia and Romania, their average market capitalization in GDP terms is below 5%: see Figure I below).
All of these countries had (re-)established stock markets by the mid-90s (see Table III above). About half of the future Member States used them to drive the initial process of re-privatization, either via mass issues of voucher certificates for residents (the most famous case of this strategy was the Czech Republic), or via IPOs (Initial Public Offerings) re-privatization processes, to lock-in domestic and foreign strategic investors (see Claessens at al., 2000). In the voucher-driven privatization, the initial large number of investors and traded stocks in those stock markets was soon concentrated in a rather limited number of institutional investors-domestic and foreign- and “blue chip" stocks. In the IPO-driven markets, the number of stocks and investors actually tended to increase with time, albeit from a rather concentrated base.
Even in the largest ones, nevertheless, market capitalization, as a GDP share, was and remains rather low (see Figure I below), and far below the EU average (around 72% of GDP). Only in the Czech Republic, Estonia, Hungary and Slovenia the average market capitalization is above a 20% GDP share, while in Romania is below 1% in several years. Also, the average market turnover is equally below the one observed in comparable EU economies. Similarly to what is observed in the banking sector, the initial regulatory environment was deliberately lax, and the regulators were plagued by much the same problems of inexperience and limited number of staff and resources.
This does not mean that domestic agents in those countries lack access to the financial services supposed to be provided by stock markets: the very process of opening up, the increase in cross-border trade in financial services, the harmonization of rules for capital trading with the EU (including the ongoing efforts of the Lamfalussy Committee towards a single European market for securities: according to the current proposal, small and medium size firms would be able to use a simplified prospectus valid throughout the EU and choose the country of its approval), plus the development of information technology, all imply that is not actually necessary-nor economically optimal, given economies of scale-for each individual country to have its own separate stock market. One must also recall that the current national stock markets in the mature developed economies are themselves the result of process of consolidation-and closing-of smaller regional stock markets (as was observed in Bulgaria in the early 1990s), which still today coexist with larger, dominant national stock exchanges even in some mature markets, like Germany and the US. Nevertheless, the observed tendency of domestic larger companies, with presumed better growth prospects, to list abroad (see Table IV below), due to the obvious cost and liquidity advantages of the larger international stock markets, does seems, on balance, to deprive those stock markets of liquidity (see Claessens at al., 2003). On the other hand, nonresidents seem to play a major role in most of those markets (accounting for 77% of the capitalization in Estonia, 70% in Hungary and half of the free-float capitalization in Lithuania).
All the specific questions described above concerning the way those stock exchanges were founded and their later developments, plus their relative smallness and shallowness, affect the dynamics of their stock market indexes (SMI), and are clearly reflected by them (as one may see in Figure II, below). This, coupled with the rather limited duration of the series, may affect their adequacy as proxies of financial cycles.
Source: Datastream, modified by the authors. The price indexes here were converted to US Dollars and re-based to a common reference period were they equal 100, May of 1998. The country codings are as described in the Annexes.
5. ESTIMATED INDEXES
The construction of the index for this new sample of countries was the core of this work. A comprehensive effort was done to crosscheck the information collected from papers and publications with national sources. Below we present the estimated monthly index, for the period January 1990 to June 2003 (see Figure III). The base data for its construction was collected from IMF and EBRD publications, and then exhaustively verified both with national sources and with works written about the individual countries and the region. This is an index that falls with liberalization, where maximum liberalization equals one and minimum three (in this sense, one could actually see it as an index of financial repression). As an additional robustness check, the year-end value of the index here constructed was regressed on the combined EBRD's yearly indexes of banking sector reform and non-banking financial sector reform. The results from a panel regression with the index constructed here on the LHS and the EBRD index on the RHS yield a coefficient of .60, and correlations among the individual country- specific index series range from -0.91 to -0.35.
As one may see from Figure III above, the process of integration and liberalization was almost continuous throughout the 1990s and early 2000s. The spikes in the “Full Liberalization Index" in the early 1990s do not indicate reversals: the merely reflect the entry into the sample of the newly independent Baltic republics. As former members of the Soviet Union, they “enter" the world as highly closed economies, but those countries introduced liberalization reforms almost immediately from the start. After this, a slight increasing trend, that does reflect a mild liberalization reversal, is observed, starting mid-1994 and lasting until early 1997, from when a continuous liberalization trend is observed.
Noteworthy here is the fact that virtually none of the obvious candidates for a reversal of liberalization (the 1997 Asian Crisis, the collapse of the Czech monetary arrangement in 1997, the collapse of the Bulgarian monetary arrangement in 1996/97, the 1998 Russian Crisis, the 1999-2001 oil price shocks-as all those economies are highly dependent of imported energy sources) seems to have driven these mild liberalization reversals. Comparing the Full Index constructed here with the one constructed by K&S, for similar time samples, one may observe that the ACs start substantially below the average level of other emerging markets- i.e., they are more liberalized, but both the “entry" of the initially less liberalized former Soviet republics, plus continuous liberalization efforts in the emerging market K&S set reverse this situation. A similar liberalization reversal trend in both the ACs and the merging market set is observed from early 1994, but it is actually slightly stronger on the ACs sample, until its reversal in 1996. By the end of our sample, the ACs are clearly below the final value for the emerging set in K&S's sample. This sort of remarkably fast pattern of the ACs' “leapfroging" towards best international practice is also observed in several types of institutional frameworks, like, for instance, monetary policy institutions and instruments (see Vinhas de Souza and Hölscher, 2001): a process that virtually took decades for Western central banks was compressed in a half a dozen years in the Future Member States. Nevertheless, by the end of the sample, both emerging and ACs are still above the level of mature, developed economies. Analyzing the individual components of the index (see Figure V), one may see that, abstracting again from the initial spikes in the index, which are, as explained above, caused by the addition of new countries to the sample, the 1994/1997 reversal of liberalization was essentially driven by the Financial Sector liberalization component. As will become clear with the country specific analysis below, this was related, in most cases, to-and here it must be stressed that those were rather limited reversals-to the banking crises that plagued several countries in our sample in the early to mid 1990s.
Comparing now the individual components of the Full Index constructed here with the ones from K&S, again for emerging and mature economies, it becomes clear that the reversals observed in Figure IV were driven by different sources in the emerging set (increase in capital account restrictions) and ACs set (financial sector): see Figure VI. All the indexes for mature economies are, again as one would expect, substantially lower.
One could, in principle, aggregate the countries in our sample in three different groups: rapid liberalizers (the ones that followed a “big bang" early approach, without major reversals: Bulgaria, Estonia, Latvia, Lithuania), consistent liberalizers (the ones that followed a more delayed path, but also without major roll backs: the Czech Republic, Hungary, Poland) and cautious liberalizers (the ones whose liberalization path was either openly inconsistent or downright mistrustful: Romania, Slovakia, Slovenia).
5.1 COUNTRY-BY-COUNTRY LIBERALIZATION PATH.
In Bulgaria, virtually no sign of a liberalization reversal is observed, even during the substantial stress experienced by the country during the banks runs of 1996/97 and the ultimate collapse of the floating regime in 1997 (beyond ad hoc restrictive measures adopted by the banks themselves). As in most of the countries in my sample, the stock market is the last one to liberalize, but does so in a faster fashion. Nevertheless, this is in most cases a data quasi-artifact that arises from the later (re-)constitution of the stock exchange itself. In the Czech Republic, a limited reversal of the financial sector liberalization is observed from late1995 to late 1997, namely, via the imposition of limits on banks' short-term open positions towards on-residents, as a way to limit the exposure of the financial sector to the inflows brought about by the hard peg and the potential gains with interest rate differentials. After the peg was replaced by the current float regime, this restriction is duly removed. In Estonia, again, virtually no sign of a liberalization reversal is observed, even during the bank runs of the early 1990s, the unwinding of the 1997 bubble, nor during the 1998 Russian crisis. Again, the stock market is the last one to liberalize, but one more time, this arises from the later constitution of the stock exchange. In Hungary, also no signs of any liberalization reversal are observed. Hungary was an early reformer, introducing some liberalization measures already during the late 1980s, but the profile of its reform path is much more discounted through time, as compared, for instance, with the Baltic countries. In Latvia, a rather limited reversal of the financial sector liberalization is observed from mid 1996all the way to early 2003: resulting from the 1996 banking crisis, specific aggregate lending limits to regions (i.e., limits on exposure to non-OECD countries, bar the other Baltic republics) are imposed. In Lithuania, a limited reversal of the financial sector liberalization is observed from early 1998, also resulting from the experienced banking crisis: reserve requirements on deposits on foreign accounts by non-resident are introduced; In Poland, no signs of any liberalization reversal are observed. Similarly to Hungary, the profile of its reform path is much more discounted through time; In Romania, no signs of any liberalization reversal are observed, but the reform path is a decidedly slow and cautious one: at the end of the sample, it has the highest (i.e., less liberalized) score for the “Full Index" of all countries in the sample: 1.60 (see Table V). In Slovakia, no signs of any liberalization reversal are observed. Here, the reform path is characterized by a broad stagnation since the Czechoslovak partition till 1998/1999, when, after a change in the political leadership, reforms are re-started, reaching after that levels similar to the other “Vise grad" countries in a rather quick fashion. In Slovenia, one of the most consistently cautious Member States concerning the advantages of integration and liberalization, reversals are indeed observed in all three indexes, since early 1995in the capital account and financial sector components, and from early 1997 in the stock market one. Since early 1999, with the entry in effect of the EU Association Agreement, across-the-board further (re)liberalization measures have been introduced.
6. FINANCIAL CYCLES AND LIBERALIZATION
The financial cycle coding which is used by K&S defines cycles as a at least twelve month-long strictly downwards (upwards) movement, followed by a equally upwards (downwards) 12-month movement from the through (peak) of a stock market index, measured in USD, as they should reflect returns from the point of view of an international investor. As described in the stock market section of this work, one must be warned that there are specific factors in the countries in our sample that may affect the effectiveness of a stock market index as an adequate proxy of financial cycles, at least for the sample here considered. Beyond that, these series have a rather limited time extension (our sample covers the 01:1990-06:2003 period). Adapting K&S criteria to the limited time dimension of our sample, we use a less stringent definition of “cycle", the same algorithm as above but with a 3-month window for the cycle (Edwards et al., 2003, use a 6-month window). With this we get 118 observations for all countries in our sample. Of these 118 cycles, 61 are upward, with an average of 7.51 months duration, and 57 are downward, with an average of 8.20 months of duration.
The main aim of this paper was to extend the index developed by Kaminsky and Schmukler, 2003, for a specific sample of countries, namely, the previously centrally planned economies from Central and Eastern Europe, and to perform a similar analysis on them. Our results do lend some support to the basic assumption of this study: in spite of all the limitations of the time series used (their shortness, the fact that they were buffeted by several country-specific and common shocks), a re-estimation of K&S's core regressions strongly supports the notion that financial liberalization does generate benefits both in the short and in the long run, measured via the extension of the amplitude of upward cycles and its reduction for downward cycles of stock market indexes. Importantly, these results diverge from K&S, as in their work “emerging markets" experience a relative short run increase in the amplitude of downward cycles. Another noteworthy feature is that only minor liberalization reversals, led by the financial sector component, were observed in the aggregate index. Also, those reversals do not seem to be driven by “contagion" from shocks in other emerging markets (like the Asian or Russian crisis), but reflect country-specific shocks. When considering the individual components of the index separately, again signs of minor reversals in financial sector liberalization are observed, related to temporary reactions to the several banking crisis observed in the region. Concerning the importance of institutions and of the EU Accession, this paper's initial assumption was that the mostly positive results above would come about due to the anchoring of expectation provided by the perspective of entry into the EU already by mid-2004 (or 2007, in the case of Bulgaria and Romania) for the countries here analyzed, and by the imposition of a more robust macro and institutional framework by the requirements of the Accession process itself. Signs of this are not found in the K&S regressions, perhaps because the liberalization index itself captures the effects of the EU Accession process.
Finally, using a different framework than K&S's to assess the affects of liberalization on financial, real and nominal volatility, most of the econometric results seem to support the previous ones, but they seem to indicate that the capital account liberalization is the element that most consistently and significantly reduces volatility. On this final section, the majority the econometric results seem to support some specific role for the EU Enlargement process in reducing volatility.