What If Ireland Defaults? (8 page)

In this section we have discussed several alternative theories of financial contagion. Of the various theories, the pure contagion models are the least plausible. As Stiglitz (1999b) notes, while Brazil and Russia had few risk factors in common with Southeast Asian economies, both countries saw significant capital flight in the immediate aftermath of the Asian crisis. Similarly, Brazil suffered in the aftermath of the Russian crisis. In those cases, the effects arose from financial institutions and hedge funds with portfolio exposures to multiple emerging markets both within and outside of Asia, and especially from the financial constraints faced by those firms. More recently, disproportionate contractions in lending by banks in the crisis-affected countries helped spread crises to Eastern Europe and emerging markets.

Our discussion of the circumstances that precipitate contagion and spread of shocks to multiple economies has important policy implications for countries with significant international capital market linkages, including Ireland, which we discuss in the next section.

Contagion and Financial and Capital Market Liberalisation

Short-Run Exchange Rate Interventions

A standard response to the threat of contagion includes an international bailout package, the essential ingredient of which is a commitment of large amounts of financial support, some of which is used immediately for intervention to support the currency, and the rest is left to convince the market that more support will be provided, should the need occur. As Stiglitz (1999a) has commented, there are two things that are odd about these interventions, which often are ineffective (for example, in Russia in 1998, in East Asia in 1997 and in Argentina in 2001). First, why should a temporary intervention in the market have persistent effects? Moreover, if the crisis conveyed information about Mexico's fundamentals that are relevant to Argentina's situation, then even if the IMF intervention stabilised Mexico's exchange rate, it would not change market perceptions of the underlying weaknesses in Argentina's economy. Only if market participants were naïve enough just to look at the exchange rate (the outcome of market processes
and
intervention) would the intervention work.
²⁰
And secondly, why should an intervention in Mexico have any effect on Argentina? On the contrary,
if
the market thought that intervention was necessary but that intervention on behalf of Argentina was less likely than in the case of Mexico, an intervention in Mexico, even if successful in supporting the Mexican exchange rate, could have an adverse effect on Argentina.

There are two sets of models in which such temporary interventions might make sense. The first is in the presence of deep market irrationalities – where market participants are truly naïve and only look at exchange rates, not what brings them about; where they have simple beliefs about contagion – that a crisis in one country is like a communicable disease, and if we cure the symptoms in one country, it can affect its spread to others. The other is that there are multiple equilibria, and interventions help to move the economy from the ‘bad' equilibrium to the ‘good' one. A third explanation, which is a variant of the second explanation, is that markets are often prone to overshooting and interventions are an attempt to prevent that. Given the real consequences of overshooting discussed earlier, such interventions may make sense. Note that in each of these explanations market processes on their own are assumed to lead to sub-optimal outcomes. But the advocates of these interventions at the international financial institutions, which typically have placed strong confidence in the efficiency and stability of market processes, need to provide a clear delineation of the circumstances in which markets can be relied upon and those in which they cannot. Critics might argue that in the case of crises the market inefficiencies are so large that they simply can't be ignored, but they are likely present at other times as well (Greenwald and Stiglitz, 1987).

More broadly, however, the models that we have delineated in this paper provide a rationale for such exchange rate interventions. Markets with rational expectations but imperfect and asymmetric information are typically not efficient; even more so if markets are subject to irrational pessimism. Then the effects of such irrationalities (even if relatively small) can be large and persistent; markets may exhibit excessive volatility, and there can be
real
benefits to government efforts at stabilisation.

Optimal Financial Architecture

Stiglitz (2000, 2002, 2006, 2010b, 2010c) analyses the optimal design of international financial architecture given both the benefits of financial integration in achieving diversification and smoothing of negative consumption shocks and the costs of adverse spill-overs across markets due to financial contagion. Financial integration raises the overall risk of spill-overs of large negative shocks (Stiglitz, 2010c). Stiglitz (2010b) examines the trade-off between contagion and diversification associated with open capital markets in a risk-sharing context. He shows that risk-sharing arrangements can become a negative-sum game in the presence of bankruptcy costs and other commonly accepted financial market frictions. In the absence of such frictions, diversification achieved through risk-sharing arrangements benefits risk-averse investors and consumers. However, a number of plausible market frictions can set in motion a financial accelerator effect that leads the initial shock to gain magnitude and persist. With bankruptcy costs, full diversification may result in lower aggregate output (net of such costs), so much lower that it more than offsets the benefits from diversification. Capital market integration could increase, instead of lower, the likelihood of a financial crisis in a given economy. Even if risk sharing does not initially increase the likelihood of a crisis but only increases the probability of a near-crisis state, the resulting increase in borrowing costs accounts for trend reinforcement, which raises the odds of a crisis in the long run.
²¹

One analogy is with fuller integration of electricity grids, which saves on generating capacity but increases the risk of a broader systemic failure. In practice, well-designed electricity networks make use of circuit breakers. In international finance capital controls serve as such circuit breakers.

If well-designed capital controls could be incorporated to prevent contagion during crisis episodes without compromising the risk-sharing benefits of integration, integration would always be preferred. However, designing and implementing such a mechanism is very challenging in practice. Therefore, the choice of integration depends on the likelihood of a large shock (and ensuing systemic failure) relative to the level of country-specific risk and the costs associated with variability. Moreover, the types and severity of informational and other frictions present in different countries must be considered for a complete assessment of the trade-offs and benefits of capital market integration.

In their analysis of the Asian financial crisis, Furman and Stiglitz (1998)
²²
and Stiglitz (2004) argue that while the adverse events affecting East Asian economies were at least to some extent exogenous (irrational investor perceptions, sudden changes in investor willingness to bear risk, interest rate increases in industrialised countries), the rapid liberalisation of capital flows and integration of domestic markets into global financial markets in the absence of a sound bank supervisory and regulatory framework contributed to the severity of the crisis. They find evidence that rapid growth in unhedged short-term debt exposures made East Asian markets vulnerable to sudden capital outflows and heightened the magnitude of the subsequent crisis. Moreover, financial integration limited the flexibility of the macroeconomic policy response because of the concern that interest rate reductions would exacerbate capital flight. In the aftermath of the Asian financial crisis and the Great Recession, the highly volatile, short-term, speculative nature of international capital movements has led many emerging market governments to reconsider the benefits of full liberalisation of capital flows (Calvo and Mendoza, 2000). Recently, the IMF has also argued that certain restrictions on cross-border capital flows may be desirable and included such restrictions in some of its recent programmes (for example, in Iceland).

Financial liberalisation refers to the opening of a country's financial system to banking institutions (and other financial institutions) from abroad. Research conducted before the crisis suggested that it provided one mechanism for the spread of a crisis from one country to others; as we have noted, the Great Recession reinforced these findings. One policy response is to question the single market principle, under which a bank that is regulated by one jurisdiction is allowed to operate freely in other jurisdictions. There is now a growing consensus that countries have to regulate all financial institutions operating within their jurisdiction (regardless of ownership) and that they should be organised as subsidiaries (not branches), to ensure that there is adequate capital within the country (United Nations, 2010).

Extensive work on crises and their propagation can be used to understand the history of financial crises, to draw inferences about the origins and spread of the recent financial crisis, and to devise policy frameworks to reduce the occurrence and magnitude of future crises. We have identified a number of mechanisms leading to crises and their contagion. Most of the plausible mechanisms require us to go beyond the standard macroeconomic frameworks based on rational agents with rational expectations operating in well-functioning financial markets. What is needed now is a comprehensive model that integrates various crisis transmission channels and provides a coherent set of policy recommendations both to reduce the magnitude and frequency of shocks, to stem contagion and to respond to the crises that nonetheless occur.

Endnotes

¹
The authors thank Charles Larkin, Brian Lucey and Constantin Gurdgiev for their helpful suggestions.

²
As we note below, this was a mistake, which is not uncommon in the presence of supervisory failures.

³
There were major institutional flaws in the design of NAMA which undermined its ability to fulfil its mission. These are not the subject of this paper.

⁴
We say typically because there are exceptions: in the Great Recession, though precipitated by the US banking crisis, the US appeared to be a safe haven, and its exchange rate appreciated. The subsequent low interest rates and depressed wages helped (at least temporarily) to buoy stock market prices, even though economic activity languished.

⁵
That this is not so in general—that markets with even large numbers of well-informed participants may look markedly different from those in which
all
are well informed—is one of the central messages of Salop and Stiglitz (1976). Grossman and Stiglitz (1980) showed that uninformed market participants could extract some, but not all, of the information from the prices generated by informed traders.

⁶
The essential insight was that with an overvalued exchange rate the country would generate a trade deficit, which foreign exchange reserves could only finance for a limited amount of time. Of course, if markets anticipated this, with rational expectations, the crisis would occur well before foreign exchange reserves were finally exhausted.

⁷
There is some evidence that normal trade adjustments, spurred in part by devaluation, were central to the resolution of the crisis; the bail-out, by temporarily leading to an exchange rate that was higher than it otherwise would have been, may in fact have impeded adjustment.

⁸
In these models, there is no way that market participants can anticipate when the economy might shift from one equilibrium to another.

⁹
See also the earlier work of Myers and Majluf (1984) and Greenwald, Stiglitz and Weiss (1984).

¹⁰
Either because managers are forced to bear some risk, as part of optimal incentive contracts, or because of bankruptcy costs. See Greenwald and Stiglitz (1990).

¹¹
Moreover, the value of firm equity can change rapidly, and there may be many claimants.

¹²
Similarly, Miller and Stiglitz (2010) use a model with collateral requirements to demonstrate how shocks can turn into crises in the presence of high leverage and overvalued assets.

¹³
More recent research has emphasised that individuals discount information that is inconsistent with their priors, and overweight information that is consistent. If a bubble is forming, they tend to weigh more heavily the information that is consistent with their beliefs. There can be equilibrium frictions, where they ‘rationally' believe that there is a bubble (see Hoff and Stiglitz, 2010).

¹⁴
Traditional economic theory – and economic policy – has taken ambiguous positions about these destabilising adjustments. It has been standard fare to worry about ‘overshooting'. Excessive exchange rate adjustments, it is thought, impede the adjustment of the market economy to the new (or ‘correct') equilibrium, and this provides justification for interventions to reduce the magnitude of the exchange rate adjustment. In some cases, there is evidence that such interventions actually impede the adjustment process. Indeed, one set of studies suggests that it was the normal foreign exchange adjustment mechanism which restored Mexico's growth, and that attempts to dampen the foreign exchange correction (driven by concerns about impact on foreign creditors) slowed down adjustment. In particular, if there had been larger foreign exchange adjustments accompanied by debt restructuring, the economy arguably would have recovered more quickly (Lederman, Menendez, Perry and Stiglitz, 2001, 2003).

¹⁵
Standard macro theories are of two minds about the role of wage and price rigidities. While the Hicksian IS-LM tradition focuses on wage and price rigidities, the Fisherian tradition revived by Greenwald and Stiglitz (1993a, 1993b, and the articles cited there) emphasises that with imperfectly indexed debt contracts wage flexibility may exacerbate economic downturns. In a model where both wages and prices are flexible, but imperfectly so, the economy can have sustained unemployment (see Solow and Stiglitz, 1968).