RDP 2004-03: Fear of Sudden Stops: Lessons from Australia and Chile 5. Concluding Remarks: Lessons for the Region at Large

In this paper we have distinguished between two inter-related dimensions of investors' confidence: country-trust and currency-trust. Doing so assists us to extract from the experiences of Australia and Chile lessons on how to improve the resilience of Latin America to external shocks. Chile is a good starting point because it has already resolved most other forms of instability, and hence it allows us to isolate the external problem more cleanly. It is then possible to discuss how additional constraints, such as extensive dollarisation or very weak monetary credibility, modify or limit the set of policies available.

There are several lessons for those economies that have only limited monetary credibility problems with residents. In building country-trust, the experience of Australia reinforces the obvious: a long history of sound institutions and non-opportunistic policies are important. Much of Australia's trust was generated by experiencing several substantial external shocks without defaulting. Clearly this was not the easy course of action at the time, and populist options – such as that of Premier Lang during the Great Depression – must have been tempting. But Australia did not behave myopically, and the rewards could not be clearer today. This contrasts with Argentina, a country with similar potential at the beginning of the 20th century. Another central pillar in Australia's external trust appears to be a solid and conservative banking system which learned the lessons of the banking crises at the end of the 19th century. These banks play a key role today in intermediating external resources into Australia, particularly when external conditions deteriorate.

Building currency-trust is also mostly a matter of common sense. It requires a good history of inflation and a clear framework governing monetary policy and the exchange rate. Inflation has been under control for most of Australia's history – at least relative to Latin American standards. Recently the clean float has contributed substantially to the external holding of Australian dollar denominated instruments, a trend that dates back only to the 1980s. Today, currency movements in the Australian dollar have a large exogenous component related to movements in commodity prices and the terms of trade. Therefore, not only is currency risk ‘exogenous’ to Australian policy-makers, but it is also highly correlated with the price of commodities, which is a widely marketed risk. In contrast, Cashin, Cespedes and Sahay (2002) fail to find a significant correlation between a similar index of commodity prices and the real exchange rate in Chile (at least over the 1980–2001 period). This is surprising given that, if anything, commodities make up an even larger share of Chile's exports.^[25] One explanation for the lack of correlation in Chile lies in domestic policies. In the early 1980s the exchange rate was fixed, while later it was subject to bands. The narrowing and widening of exchange rate bands during the turmoils of the end of the 1990s is another example of intervention. The Chilean peso risk then has a larger ‘endogenous’ component than Australian dollar risk. It is then subject to the usual concerns about moral hazard, time inconsistency, and so on. This makes it easier for Australians to find willing external buyers for Australian dollar risk. Not only is the risk ‘exogenous’ to Australian policy-makers, but it is also highly correlated with the price of commodities, for which substantial futures markets exist. In a sense, the problem is not that the Chilean peso has been a commodity currency, but that commodities have played a relatively small part in peso fluctuations. In recent years, Chile has begun moving to an Australian-style commodity-price-driven free float.

Sound and clear macroeconomic policies are not enough. Foreign investors need a liquid market in which they can take on currency risk with only limited exposure to other risks, such as credit risk. Countries with sound public finances can do this by developing a domestic public bond market in domestic currency (which could be indexed to the CPI, such as with Chile's Unidad de Fomento).^[26] Initially, as in Australia, these bonds will have high premia and the country must be willing to pay that cost. It is also likely that initially most of the holding will be done by domestic institutions which are less concerned with currency risk. Over time, if appropriate macro policies are implemented, as the country develops, currency-trust foreigners will be enticed by the high yields and will progressively hold a larger share of these bonds. The currency premia should then fall to ‘reasonable’ levels.

It is only a step from the adoption of sound monetary and exchange policies, and the development of a good set of benchmark bonds, to the development of a currency derivative market. Initially, such development may reduce foreign exposure to currency risk since foreigners may increase their participation in the local bond market but unload the currency risk. Assuming the country does develop currency-trust, this will only be a transitory cost. This cost is likely to be more than offset by the benefits of increased participation by foreigners in domestic financial markets and enabling domestic banks to eliminate currency-mismatch risk from their books. This should reduce the destabilising role played by banks during external shocks in Latin America.

There are several measures that can be undertaken to reduce external vulnerability while these institutions and markets are developing. In particular, external insurance can be designed to be indexed to contingencies that are highly correlated with external shocks, but are not under the direct control of the country and so subject to moral hazard, for example, commodity prices. Only this incomplete form of insurance is possible when there is a lack of currency-trust and country-trust. Complete insurance, for example, linked to domestic GDP or the local currency, can only be obtained when both kinds of trust exist. Over time, as trust is built, the mix of external insurance can be gradually shifted toward contingencies that include some endogenous factors, such as the local currency or GDP.

Similarly, the macroeconomic policy framework should not only be made as transparent and consistent as possible, but it should also be used to align private incentives with the aggregate risks that private decisions generate. Such objectives can be achieved, for example, by indexing specific macroeconomic policies to the same contingencies used to build external insurance. For example, faced with a temporary shortage of foreign currency, it is optimal for the central bank to inject international reserves into the domestic economy. However, if private-sector agents expect that this policy will be used in a crisis, and therefore anticipate the exchange rate will be stable, they will engage in excessive foreign currency borrowing. Hence, injections of international reserves during crisis periods should be accompanied by an expansion of the domestic money supply. By doing so, the central bank prevents the free insurance aspect of a strong exchange rate defence, but is still able to relax the international financial constraint faced by the private sector.

Most of these recipes also apply to dollarised economies, or economies where lack of monetary credibility with residents is widespread. This is true even though in such cases there is no hope of using monetary policy to provide adequate private incentives, or developing extensive derivative markets. These aspects of the plan need to be substituted in the transition by costlier measures such as taxes on certain capital inflows and large international liquidity ratio requirements for domestic banks. Similarly, the development of a domestic debt market in local currency is likely to be too expensive to develop quickly without being preceded by a clear effort to develop sound institutions and implement a transparent and credible macroeconomic policy program. Once these are implemented, there is no ready substitute for the passage of time. On the other hand, these economies have all the more reason to accelerate the development of external insurance mechanisms as described above, which may even be developed and fostered in domestic markets to improve the allocation of risks among residents.

Footnotes

For the period 1990–1999, for example the commodities included in the index amount to 58 per cent of total exports in Chile and 54 per cent in Australia. [25]

Note that the external insurance required is against shocks that depreciate the real exchange rate, not inflationary shocks per se. This is the reason UF instruments would work as well. [26]