Fear of Sudden Stops: Lessons From Australia and Chile¹

1. This paper is a revised version of the paper prepared for the Financial Dedollarization: Policy Options conference at the Inter‐American Development Bank, 1–2 December 2003. The views expressed in this paper are those of the authors and should not be attributed to the Central Bank of Chile, the Reserve Bank of Australia or any other institution with which they are affiliated. We thank Guy Debelle, Luci Ellis, Darren Flood, Ilan Goldfajn, Simon Guttmann, Paulo Mauro, Tony Richards and Alejandro Werner for their comments. Erwin Hansen provided excellent research assistance.

Abstract

Latin American economies are exposed to substantial external vulnerability. Domestic imbalances and terms of trade shocks are often exacerbated by sharp shifts in the net supply of external capital (sudden stops). At times, these sudden stops can be the main shock. In this paper we explore ways of overcoming external vulnerability, drawing lessons from a detailed comparison of the response of Chile and Australia to recent external shocks and from Australia’s historical experience. We argue that in order to understand sudden stops and the mechanisms to smooth them it is useful to identify and then distinguish between two inter‐related dimensions of investors’ confidence: country‐trust and currency‐trust. Lack of country‐trust is the fundamental problem behind sudden stops. Lack of currency‐trust in turn weakens a country’s ability to deal with sudden stops and real external shocks. We discuss steps aimed to improve these two dimensions of investors’ confidence in the medium run, and policies to reduce the impact of country‐trust and currency‐trust weaknesses in the short run.

1. This paper is a revised version of the paper prepared for the Financial Dedollarization: Policy Options conference at the Inter‐American Development Bank, 1–2 December 2003. The views expressed in this paper are those of the authors and should not be attributed to the Central Bank of Chile, the Reserve Bank of Australia or any other institution with which they are affiliated. We thank Guy Debelle, Luci Ellis, Darren Flood, Ilan Goldfajn, Simon Guttmann, Paulo Mauro, Tony Richards and Alejandro Werner for their comments. Erwin Hansen provided excellent research assistance.

Key Words:

• Sudden stops
• Australia
• Chile

JEL Classification:

• E44
• E52
• F31
• F33

Notes

1. This paper is a revised version of the paper prepared for the Financial Dedollarization: Policy Options conference at the Inter‐American Development Bank, 1–2 December 2003. The views expressed in this paper are those of the authors and should not be attributed to the Central Bank of Chile, the Reserve Bank of Australia or any other institution with which they are affiliated. We thank Guy Debelle, Luci Ellis, Darren Flood, Ilan Goldfajn, Simon Guttmann, Paulo Mauro, Tony Richards and Alejandro Werner for their comments. Erwin Hansen provided excellent research assistance.

2. The RBA did intervene, notably in mid 1998 when there was seemingly large amounts of speculative activity, but was certainly willing to allow the currency to depreciate.

3. We use the IMF effective exchange rates for comparability. The depreciation of the Australian dollar based on the RBA’s Trade Weighted Index, which uses different weights, over this period was less than 5%.

4. Unlike output and terms of trade we look at deviations from the period 1990–1997 because 1990 marks the return of capital inflows to emerging markets.

5. The rising share of foreign assets is not a result of the depreciation of the peso. This is evident from the line in Figure 7 that plots the share of foreign assets adjusting for changes in the dollar/peso exchange rate.

6. See Caballero (2002) for a more extensive discussion of this point.

7. For example, when there is an exogenous output gap of 2%, pass‐through is below 0.115.

8. There is extensive theoretical literature on this issue. See for example Krugman (1999a,b), Aghion et al. (2001) and Cespedes et al. (2000). Empirical results are less abundant and categorical, see Bleakley and Cowan (2002).

9. Note that these measures of mismatch overstate the sensitivity of Australia’s indebtedness to currency movements as Australia has a large foreign‐currency asset position on derivative contracts, as detailed in Section 2.4

10. Source: ‘Informe Estadisticas de la Deuda Publica’, Ministerio de Hacienda Chile.

11. Similar results are found by Bleakley and Cowan (2002), and Cowan (2002).

12. For comparison purposes we carry out a similar set of regressions for Argentina. We find that the currency ‘matching’ that we see in Chile does not take place in Argentina. Not only is average dollar debt considerably higher in Argentinean firms, but it is distributed equally between tradable and non‐tradable sectors.

13. Close to two‐thirds of turnover in Australian is cross‐border. None of Chile’s turnover is cross‐border (BIS, 2002).

14. This measure of foreign‐currency ‘exposure’ ignores the elasticity of export revenue to exchange rate fluctuations and substantial foreign equity asset holdings of A$228 billion. Indeed, accounting for foreign equity assets, in 2001 Australia had a net foreign‐currency asset position of A$149 billion.

15. Strictly, currency‐trust could play an important insurance role even under a fixed parity since it could allow agents to modify out‐of‐equilibria scenarios and therefore prevent some perverse outcomes. However, this would still require extensive contracting with foreigners in local currency.

16. See Caballero (2003) and Caballero and Panageas (2003).

17. At the turn of the century, GDP per capita was around 90% of the levels in the US and the UK. Prior to the severe 1890s recession in Australia, GDP per capita was around 40% greater than the US (Maddison, 2003).

18. The checklist consisted of inflation, the nominal exchange rate, interest rates, the balance of payments, monetary aggregates and the general state of the economy.

19. Contributing policy changes include an increase in the legislated minimum wage, the relaxation of price controls, and the effective depreciation resulting from the devaluation of sterling in September 1949.

20. The British Colonial Stock Act of 1900 enabled many trust funds to purchase Dominion bonds, giving the Dominions an even greater advantage over many domestic borrowers. Australian governments were able to issue long‐dated debt from early on. In 1913 their fixed‐maturity debt had an average maturity of just under 18 years, while 6% of debt had been issued as perpetuities.

21. An alternative interpretation is offered by Bordo et al. (2003). They argue that the development of the domestic market for government bonds accelerated in those periods in which international markets were closed.

22. The pick‐up in swaps toward the end of the 1990s is exaggerated somewhat by the RBA’s use of swaps to conduct open market operations in the face of declining liquidity in government securities markets.

23. The conversion of domestic government debt to lower interest rates was voluntary but a large propaganda campaign led to 97% of borrowers taking up the new loans with longer maturities and interest rates reduced by 22.5%. The 3% of ‘dissenters’ ended up having their debt compulsorily converted.

24. The Loan Council, formed in 1923, facilitated the coordination of government borrowing. The Council was formalised later in the 1920s with government debt amalgamated and state debt explicitly guaranteed by the Commonwealth.

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

26. 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.