8. Develop a global partnership for development
7.4 Financial crises and contagion
The world has seen banking crises in 69 countries since the late 1970s, and 87 currency crises since 1975. And the frequency of such crises has Figure 7.11 Net equity flows to developing countries, 2000–2006.
Source: World Bank WDI, 2007.
risen sharply over the last decade. After the recent series of meltdowns in Asia, Eastern Europe and Latin America, no observer can be surprised at the apparent instability of financial markets. The debate on the causes of these crashes will undoubtedly go on for a long time. Bad luck, in the form of exogenous shocks from abroad and from mother nature, and bad policy, in the shape of poor regulation and imprudent macro policies, doubtlessly carry some of the blame. But that cannot be the end of the story. The main message of this paper is that the potential for illiquidity was at the center of recent crises, and that short-term debt is a crucial ingredient of illiquidity. The empirical evidence is clear in that respect.
(Rodrik and Velasco, 1999: 22) As this quote makes clear, the incidence of financial crises in developing countries has certainly increased markedly since the 1980s. However, as illus- trated by Figure 7.12 below, this is by no means a new phenomenon, though it is certainly a growing one. This section will focus on, first, banking crises, and second, currency crises. This does not mean that crises in equity markets, for example, are unimportant, but their wider economic impact does not have the significance of banking and currency crises.
To recap some of the material covered in Chapter 1, the efficient market hypothesis states that market prices should fully and accurately reflect under- lying economic fundamentals, and will therefore only change when these fundamentals change, either directly or in relation to other financial assets.
However, if this were so, how are we to explain the fact that history is full of examples of speculative asset price bubbles – that inevitably burst, periodic bank runs and currency crises. It is difficult to argue that, in each case, the pre-crisis ‘boom’ accurately reflected fundamentals. Nor that the crisis was
Figure 7.12 Selected major financial crises, 1400–2000.
Source: World Bank, 2001.
triggered by a fundamental deterioration of these fundamentals, to which investors responded in a rational manner.11
The most important international financial crisis of recent decades is undoubtedly that which occurred in East and Southeast Asia between 1997 and 1999. Before distinguishing between different forms of crisis, Box 7.2 below sets out the chronology of the Asian crisis, where many of these differ- ent elements came together with devastating results.
Box 7.2 A chronology of the Asian crisis
July 2, 1997: In the face of huge speculative pressure, Thailand devalues the baht. The government requests technical assistance from the IMF.
July 8, 1997: Malaysia’s central bank intervenes to support the ringgit.
July 11, 1997: The Philippine peso is devalued. Indonesia widens the trading band for the rupiah.
July 18, 1997: The IMF releases US$1 bn from its ‘emergency funding mechanism’ for the first time to assist the Philippine authorities in sup- porting the peso.
August 5, 1997: Thailand agrees strict economic austerity conditions from the IMF in return for a $17 bn loan from the international lender and Asian nations.
August 14, 1997: Indonesia allows the rupiah to float freely. The cur- rency plummets in value.
Oct. 8, 1997: Indonesia requests assistance from the IMF and World Bank after the rupiah loses a third of its value.
Oct. 31, 1997: The IMF approves a $40 bn to a loan package for Indonesia. The government closes sixteen financially insolvent banks as part of a swingeing reform programme.
Nov. 17, 1997: The Bank of Korea abandons its effort to prop up the value of the won. It falls to a record low against the US dollar.
Nov. 21, 1997: South Korea requests IMF assistance. Critics call pro- posed terms of the agreement ‘humiliating’. A key component is large numbers of redundancies.
Dec. 3, 1997: The IMF approves a $57 bn package for South Korea, the largest in history.
Dec. 8, 1997: The Thai government announces the closure of fifty-six insolvent finance companies. Thirty thousand workers lose their jobs.
Dec. 18, 1997: Kim Dae Jung becomes South Korea’s first opposition to be elected president.
Dec. 23, 1997: The World Bank releases an emergency loan of
$3 bn to South Korea.
Dec. 24, 1997: Korea receives $10 bn in loans from the IMF and G7 to avoid a default on its short-term loan. It agrees to open its domestic financial markets to foreign institutions.
Jan. 7, 1998: Chief economist of the World Bank, Joseph Stiglitz openly questions the IMF ‘medicine’ of strict austerity.
Jan. 8, 1998: The Indonesian rupiah plummets further after President Suharto unveils his state budget plan, which is seen as incompatible with the IMF reform program.
Jan. 12, 1998: Hong Kong-based Peregrine Investments – Asia’s larg- est private investment bank – files for liquidation.
Jan. 13, 1998: Students protest in the Indonesian capital against the IMF policies.
Jan. 15, 1998: President Suharto agrees to eliminate the country’s monopolies and state subsidies and signs a new deal with the IMF.
Jan. 28, 1998: International banks and South Korea agree to exchange $24 bn of short-term debt for longer-term loans.
Feb. 6, 1998: South Korean unions, government and businesses reach a landmark agreement to make redundancies legal.
March 9, 1998: The IMF delays a $3 bn instalment of loans to Indo- nesia, and criticises Suharto’s failure to implement agreed reforms.
March 23, 1998: Russian President Boris Yeltsin dismisses his entire cabinet.
April 8, 1998: Indonesia and the IMF reach a third agreement.
May 5, 1998: Students in Indonesia hold demonstrations across the country about fuel and food price rises.
May 12, 1998: Indonesian troops fire on a student protest, killing six and sparking riots.
May 21, 1998: President Suharto resigns after thirty-two years in power and is succeeded by Vice President Habibie.
May 22, 1998: The IMF postpones aid disbursement to Indonesia until political stability is restored.
May 27–28, 1998: A two-day general strike is held in South Korea to protest against 10,000 redundancies.
June 1, 1998: Russia’s stock market crashes and foreign exchange reserves fall to $14 bn.
June 25, 1998: Indonesia and the IMF announce a fourth agreement.
July 13, 1998: The IMF announces a package of $23 bn of emergency loans for Russia.
August 11, 1998: Trading on the Russian stock market is suspended.
August 17, 1998: Russia announces a devaluation of the rouble and defaults on debt payments. Latin American stock and bond markets plunge.
Sept. 23, 1998: The Fed orchestrates a $3.5 bn bailout to Long Term Capital Management (LTCM)
Sept. 29, 1998: The Fed cuts interest rates by a quarter point and again in October. World markets rally.
7.4.1 Banking crises
One area where it has long been accepted that crises can occur even without such a deterioration in fundamentals is in the banking sector. However, unlike the other forms of financial crisis covered in this chapter, banking crises can and do occur without any external financial influences. As we have seen previ- ously, banks are inherently susceptible to ‘runs’, where depositors with con- cerns over the solvency of the bank rush to withdraw their deposits. However, because banks have assets that can be liquidated instantly but use these assets to lend on a longer-term basis, a ‘run’ of this kind can lead to the closure of a bank, regardless of whether the bank really did have a solvency problem.12
There is also the possibility of ‘contagious’ bank runs, where the ‘run’
spreads from one bank to the rest of the system, threatening the viability of the entire banking sector – again, the long-term solvency of the system is largely irrelevant, as contagious bank runs can produce liquidity crises on a scale sufficient to undermine the entire banking sector regardless of the
‘soundness’ of the banks concerned. We have seen that banks play a pivotal intermediation role in the economy, and it is therefore not surprising that banking crises are associated with significant costs to the economy, as banks’
ability to perform this function is undermined.
The full cost to society of a banking crisis far exceeds the losses to the banks’ own balance sheets, however. For the World Bank (2001) there are three components of the true economic cost that need to be factored in:
1 The stock component is the ‘accumulated waste of economic resources that is revealed by the insolvency’. That is, instead of channelling savings to the most productive investments, banks have effectively wasted these funds by channelling them into uneconomic activities.
2 The public finance component can result as governments tend to assume at least some direct responsibility for the crisis, in order to bail out deposi- tors and others affected. In many instances, the government has little option but to intervene in this way, though there are obviously fiscal implications of doing so in terms of (a) expenditure cuts elsewhere, (b) tax increases, and/or (c) inflation arising from the printing of money to fund the intervention.
3 The flow component is a result of the reduced economic output with which banking crises are usually associated. This impact may be felt through any or all of the following channels:
• A sharp fall in investment and/or consumer spending due to a loss of confidence;
• A sharp fall in investment and/or consumer spending due to a lack of credit;
• A reduction in loans due to the loss of information on creditworthy investment opportunities;
• A failure in the centralised payment and settlement system.
Clearly, the larger the initial capital shortfall of the banks, the greater will be the final economic impact. In developing countries alone, the World Bank (2001) estimates that the combined cost of banking crises from 1980 to 2000 was more than $1 trillion.
Figure 7.13 below details the estimated fiscal cost of banking crises in individual countries over the past two decades. As we can see, even the small- est impact – in Sweden – amounted to 3% of GDP, while the largest – in Indonesia – wiped out almost 50% of national wealth. For all countries, the average impact of a banking crisis was a reduction in GDP of 17.5% and, as the reference to Sweden makes clear, such crises are not restricted to develop- ing countries. Rather, as discussed in previous chapters on the domestic financial system and financial liberalisation, countries that deregulate their banking systems without first establishing a robust and effective framework of prudential regulation and supervision greatly increase the likelihood of a banking crisis occurring.
In many ways it could be argued that having a banking crisis is something of a ‘rite of passage’ that must be undergone if a liberalised financial system is to develop successfully, though the cost of this transition would seem to most reasonable people to be extremely high – possibly higher than any benefits that liberalisation might ultimately bring.
Although banking crises can and do happen in both developed and developing countries, it is also the case that concerns about the solvency of banks are associated with situations of weak prudential regulation and supervision, which are more likely to pertain in developing countries. In addition to this factor, the World Bank (op. cit.) gives the following reasons why banking crises are more likely to occur in developing than developed countries:
Figure 7.13 Total fiscal costs of banking crises, 1980–2000.
Source: World Bank, 2001.
1 Inadequate prudential regulation and supervision of the banking sector does not prevent banks taking on excessive risks, as they will tend to do otherwise.13
2 They experience more severe information problems, which lead banks to make suboptimal lending decisions.
3 Developing country banks have less diversified investment portfolios, reflecting the more concentrated economies in which they operate. This leaves them more vulnerable to economic shocks.
4 The higher volatility of financial assets (equities, bonds, exchange rates etc.) in developing countries leads to higher banking sector fragility, as developing country banks are largely invested in their own domestic markets.
5 Financial markets in developing countries are dominated by banks. As a result, companies tend to be highly leveraged and therefore more vulner- able to economic shocks. Lower access to longer-term capital also increases the vulnerability of firms to shocks.
6 State ownership of banking assets is linked to banking sector vulnerability.
7 There is poor sequencing of financial liberalisation.
Clearly, the costs of banking crises are such that every effort should be made to prevent them. In this regard, robust regulation and supervision is crucial, as is the legal and commercial framework in which banks operate. However, because of the features of developing countries listed above, there remains a higher probability of banking crises in developing than in developed markets even if regulation and supervision were perfect. The volatility and concentra- tion of the financial system of many developing economies increases the probability of genuine insolvencies, whilst the often poor information avail- able also increases the possibility of unfounded fears of insolvency sparking bank runs. We have seen how banking crises can occur independently of any external financial involvement. However, it is also the case that domestic banking crises can also be precipitated by sharp movements of externally driven financial variables, and that these movements may be driven more by perception than reality. This mixture of reality and perception is most readily seen in a second important aspect of financial crises: currency crises.
7.4.2 Currency crises
As described by Dani Rodrik at the start of Section 7.4, there has been a sharply increased incidence of currency crises, particularly though not exclusively in developing and emerging economies, since the collapse of the system of fixed exchange rates in the early 1970s.
If we recall that the key theoretical pillar in support of the switch to fixed exchange rates was that ‘rational arbitrageurs’ would ensure that exchange rates remained close to their equilibrium/fundamental/‘fair’ values (Friedman, 1953), then this would suggest that currency crises are caused by a
sharp deterioration in the economic fundamentals of the country concerned.
Indeed, as we shall see, early attempts to model and explain currency crises broadly adopted this position, only for the weight of evidence to lead econo- mists to question whether this perspective could accurately describe these events.
The first major theoretical work in the field was Krugman (1979). Adapting earlier work by Salant and Henderson (1978),14 Krugman developed the first standard analysis of the process of speculative attack. Krugman demon- strated how inconsistencies between domestic economic considerations and the maintenance of an external exchange rate peg could provoke a specula- tive attack, resulting in the collapse of the peg. For example, expansionary domestic monetary policy results in balance of payments problems and the central bank finances this shortfall through the use of its foreign exchange reserves. At some point, however, reserves fall below a particular level, trig- gering a speculative attack. The remaining reserves are then exhausted and the currency is forced to devalue, thereby eliminating the inconsistency of expansionary domestic policy and the maintenance of a fixed currency peg.
Krugman stresses that it is not necessary for reserves to have been com- pletely exhausted for an attack to be successful. Where policy inconsistencies are apparent, an attack will take place at the earliest point that it could be successful: market actors are all attempting to get out just before the attack, and are aware that others are attempting to do the same, as soon as an attack would be sufficient to wipe out the remaining reserves it will therefore be initiated.
This early canonical model places the responsibility for the success of speculative attacks on currency pegs squarely with the governments in ques- tion. From this perspective, the market is merely making a rational move in response to inconsistent economic policies, and is therefore responding to, rather than participating in, real-world events. The speculative attacks that have occurred since 1979, however, led many to question the relevance of this
‘first-generation’ model. It was pointed out that, commonly, the simple case of expansionary domestic policy did not accord with the facts. More complex theories arose in the 1980s which attempted to align theory more closely with reality. These approaches saw the role of markets as far greater than simply reacting to fundamentals; with the markets viewed as perhaps creating, rather than simply responding to, circumstances.
This second generation of models investigated the possibility of the exist- ence of ‘multiple equilibria’, wherein the perceptions of market actors are crucial. Multiple equilibria may arise where the sustainable level of a currency depends on whether or not a speculative attack occurs. So, whilst current fundamentals are taken into account, so are future fundamentals; however, these differ depending on whether or not a speculative attack occurs. There- fore, it may be that the currency peg is compatible with current and future fundamentals in the absence of a speculative attack, but incompatible with current and future fundamentals after an attack has taken place.
Obstfeld (1986) offers theoretical and empirical evidence in support of the multiple equilibria hypothesis, arguing that a ‘grey area’ exists where the sustainability of a currency peg is determined by the attitude of market act- ors. He stresses that where clear incompatibilities in policy exist then the processes identified by Krugman (1979) may be relevant, but that this model does not adequately describe the diversity of real-world situations, and, cru- cially, underestimates the role that the markets can play in shaping outcomes.
The distinction between these two approaches is very important. The argument rests upon the extent to which markets are rational and objective processors and users of information, and/or the extent to which they act in a self-fulfilling or collectively irrational way. That is, do markets move in response to economic fundamentals – as orthodox theory suggests – or are market movements more driven by market sentiment, which may be decoupled from underlying economic fundamentals but can then come to shape them?
This also relates to the next issue to consider in this section, which is the increasing prevalence of financial crisis contagion.
7.4.3 Financial crisis contagion
It is now generally well accepted that contagion – or spillover effects – from financial crises are not fully explained by fundamental factors alone. The debate on this subject has been conducted since the ERM crises of 1992/3 and intensified after first the Latin American ‘tequila crisis’ of the mid-1990s and, more recently, the Asian crisis of 1997/8. At issue have been the explan- ations for the temporal and geographical clustering of financial crises.
On the one hand some have argued that, whilst crises are indeed clustered in both a geographical and temporal sense, there are sound economic reasons for this, with markets discriminating in the countries that come under specu- lative attack; that is, only countries with weak fundamentals are vulnerable.
The opposing view has held that the primary causes of the observed clustering lie in the behaviour of international investors, rather than in the fundamentals of the countries concerned. One reason why this debate is significant is for its implications for the justification, or not, for international assistance for crisis-affected countries. If the former camp is right, and coun- tries primarily come under speculative attack because of weak and/or declin- ing economic fundamentals, then international assistance to bail out such countries may not be justified. Alternatively, if the primary cause lies in the changing behaviour of international investors, then the countries affected cannot be said to be culpable and thus may be deserving of international assistance (Eichengreen et al., 1996).
Some commentators working in this area have suggested that although contagion can be observed in the aftermath of a number of financial crises, this does not, of itself, indicate shared causes or uniform transmission mech- anisms. Just as each financial crisis is, to some extent, a unique event, so contagious episodes may also be relatively heterogeneous in terms of causes