Comments on Part IV

Adrián Armas, Eduardo Levy Yeyati, and Alain Ize
Published Date:
July 2006
  • ShareShare
Show Summary Details
Augusto de la Torre

The two very good chapters in this section deal with significantly different aspects of the debate on financial dollarization and de-dollarization. Chapter 10, by Holland and Mulder, discusses the potentially beneficial role that inflation-indexed debt can play for the government’s debt management in particular and for the development of the local currency debt market more generally. Chapter 11, by Hardy and Pazarbasioglu, by contrast, examines key issues in episodes of forced or involuntary de-dollarization. I will not attempt to discuss the substance of each of these chapters. To do so might prove unmanageable within these short comments and, in any case, would fail to do justice to the two chapters, given the differences in topic and approach. Instead, I will discuss a single question – can CPI-indexed debt significantly help induce the voluntary de-dollarization of financial contracts? My comments are hopefully complementary to the chapter by Holland and Mulder, which does not really address this question directly.

In the Latin region, financial contracts that are indexed to the CPI are typically denominated in a unit of account which is itself linked to the CPI in a backward-looking fashion but with a relatively short lag. For convenience and to honour the popularity of Chilean nomenclature, I will henceforth refer to that unit as UF. Now, I will argue that the question of whether UF-denominated debt contracts can help promote de-dollarization is not as simple as it might appear to be. The evidence that UF-based financial contracting was historically crucial in preventing financial dollarization does not imply that the UF can help in reversing dollarization once it has taken root.1 For example, the use of a stake to support a young tree can indeed prevent it from growing crooked, but this does not imply that the same support would straighten a crooked tree once mature.

That hysteresis is a feature of financial dollarization should not be surprising if initial conditions and path dependence are taken seriously. To illustrate this point, consider the case of the Chilean UF. The UF was instrumental in preventing financial dollarization in Chile because it was accepted by the population as a store of value for savings and this in turn was a function of initial conditions and path dependence. Initial conditions included not just an active promotion of the UF by the government, but also a high and highly volatile inflation, a high pass-through (due to widespread backward-looking wage and price indexation) and, importantly, severe legal and regulatory impediments to using of the dollar for financial contracting at home (see Herrera and Valdés, 2005). Given this initial context, path dependence enabled the UF to become rooted in financial markets. Why? Because the associated changes in financial contracting practices became self-reinforcing as a consequence of the increasing returns and network externalities inherent in institutional change – that is, as a consequence of the large set-up costs of new UF-consistent institutions, the subsequent lowering of uncertainty and transaction and information costs and the associated spill-overs for financial contracting. One can thus surmise that if financial dollarization could not be prevented and instead became locked-in in other countries, it must have been due to a complex interaction of different initial conditions and a different process of path dependence compared to those that obtained in Chile. This reasoning highlights a crucial but often downplayed point in the de-dollarization debate – that an isolated instrument such as the UF, which was helpful in averting dollarization under a given institutional matrix, may simply not stick in another institutional milieu.

The burden of the proof thus falls on the proponents of the UF as a de-dollarization tool. They would need to build a solid case – a task that in my mind has not yet been done – of why savers/investors would find the UF more appealing than the dollar in the circumstances that typically characterize the financially dollarized Latin countries of today. These circumstances – which are in sharp contrast to those present in Chile when the UF was introduced – include a low rate of inflation, a low pass-through, the well-entrenched position of the dollar in financial contracting, the absence of legal and regulatory impediments to dollar-based financial intermediation2 and the high degree of integration with international financial markets. I am already convinced that a deep market for UF-denominated long-duration debt would have many positive benefits for financial development, provided that savers use the UF intensively. What is much less clear to me is why investors would do so in the mentioned circumstances and if they already rely on the dollar.

The challenge before us is, therefore, to better understand the conditions under which, given the mentioned circumstances and considering institutional path dependence, the UF could help de-dollarize financial contracts. To frame the discussion of this question more pointedly, I would like to submit four propositions developed in detail in a paper that I recently wrote with Sergio Schmukler (de la Torre and Schmukler, 2004).

Proposition 1: the currency of denomination (dollar, peso, or UF) and the duration (short or long) of a financial contract are devices used by rational investors to cope with systemic risks (i.e., macroeconomic volatility and weak contractual institutions) in emerging economies. In particular, dollarized contracts, UF contracts and short-duration peso contracts are substitute ways of hedging price risk (i.e., inflation and interest rate volatility).

Proposition 2: there is an inherent trade-off between exposure to price risk, on the one hand, and exposure to price-induced default risk, on the other. Thus, by investing in a short-duration (peso or UF) contract or in a dollarized contract, the investor hedges against unexpected movements in inflation or the real interest rate at the expense of taking on an open exposure to the risk that the debtor might default if hit by an unexpected rise in the real interest rate (in the case of a short-duration contract) or by an unexpected and large real devaluation (in the case of a dollarized contract).

Proposition 3: as systemic risks rise, the equilibrium outcome settles in favour of investor protection against price risk (through dollarization or short duration) at the expense of exposure to price-induced default risk, instead of the other way around. One reason for this is the fact that the contract is a legal claim that gives the investor an option to litigate and recover some value in the event of default.

Proposition 4: for a given level of systemic risk and investor appetite for systemic risk, efforts to repress financial dollarization may only displace (rather than reduce) risk. In other words, the repression of the currency mismatch may, all other things equal, only lead to a compensating increase in another mismatch – i.e., the duration mismatch implied in short-duration peso debt contracts – or to the shifting of dollar contracts from the local market to the offshore market.

Given these propositions, let us go back to the question of whether the UF could really draw investor interest in economies with low inflation and where financial dollarization is already well established. Assume, first, that the relevant macroeconomic environment is one where inflation and the interest rate are more volatile than the real exchange rate. Per the seminal article of Ize and Levy Yeyati (2003), financial portfolios in that environment would be tilted in favour of the dollar. Assume, now, that the government attempts to introduce the UF, all else equal. Would investors shift voluntarily from dollar-based to UF-based financial contracts? The answer is most likely no. Why? Because the UF offers no obvious advantage to investors as a risk-coping device compared to the dollar. Take a long-duration UF-denominated contract (i.e., a long-term UF contract with a fixed interest rate). It would be almost as good as the dollar in protecting the investor against inflation risk but, unlike the dollar, it would not enable the investor to hedge against real interest-rate risk. Take, now, a short-duration UF contract (i.e., a short-term UF contract, or a long-term UF contract where the interest rate adjusts frequently, say, every month). It would be almost as good as the dollar in protecting the investor against inflation and real interest-rate risks. But, unlike the dollar, the short-duration UF would not protect the investor against real interest rate-induced default risk. Moreover, in an environment where inflation and the real interest rates are more volatile than the real exchange rate, the UF would be an inferior risk-coping device compared not only to the dollar but also to a short-duration peso contract.3

It follows from the previous reasoning that isolated gimmicks are unlikely to lead to de-dollarization. Financial dollarization is a symptom, a response to weaknesses in macroeconomic and institutional fundamentals that give rise to high systemic risks. Attacking the symptom without addressing the underlying causes can backfire. Hence, the sufficient condition for financial de-dollarization is also a fundamental condition, namely, the permanent reduction in systemic risk. On the macroeconomic policy front, such a reduction requires structural improvements in the solvency of the government. Progress towards more resilient fiscal solvency would in turn underpin a robust framework of independent monetary policy featuring inflation targeting-cum-exchange flexibility. As a result, the volatility of inflation and the real interest rate would be durably reduced relative to the volatility of the real exchange rate, thereby creating the type of credible nominal stability that would foster the use of the local currency as a store of value for savings, with the consequent de-dollarization of financial contracts.

The de-dollarization effect of improvements in fiscal and monetary fundamentals could be, no doubt, enhanced by complementary policies in other fronts, including actions to strengthen the prudential and contractual frameworks as well as the introduction of CPI-indexed long-duration debt contracts. The UF alone would accomplish little, but the UF in an environment of improving macroeconomic and institutional fundamentals could have a significant impact. In that context, appetite for CPI-indexed debt could be boosted through a government debt management strategy that would seek to develop a reliable yield curve in long-duration CPI-indexed sovereign bonds. That could, of itself, lead to improvements in fiscal viability – by narrowing the scope for time-inconsistent government misbehaviour and by reducing the government’s exposure to risk for a given cost. However, the clincher for the establishment of the UF has to come ultimately from strong investor demand. This would clearly materialize to the extent that the reformed pension systems of the Latin American region mature, thereby giving rise to a growing market for annuities, with annuity providers (typically life insurance companies) increasingly needing long-duration CPI-indexed assets to hedge their growing long-duration CPI-indexed liabilities.


    AridaP.E.Bacha and A.Lara-Resende (2005) ‘Credit Interest and Jurisdictional Uncertainly: Conjectures on the Case of Brazil’ in F.GiavazziI.Goldfajn and S.Herrera (eds). Inflation Targeting Debt and the Brazilian Experience 1999 to 2003 (Cambridge, MA: MIT Press).

    de la TorreA. and S.Schmukler (2004) ‘Coping with Risk Through Mismatches: Domestic and International Financial Contracts for Emerging Economies’ Working Paper No. 3212 (Washington, DC: World Bank).

    HerreraL. O. and R.Valdés (2005) ‘Dedollarization Indexation and Nominalization: The Chilean Experience’ The Journal of Policy ReformVol. 8No. 4 pp. 281-312.

    IzeA. and E.Levy Yeyati (2003) ‘Financial Dollarization’ Journal of International EconomicsVol. 59 (March) pp. 323-47.


Financial dollarization was also prevented in some cases – notably Brazil – by the practice of indexing debt contracts to the local short-term interest rate, combined with legal and regulatory limitations on the use of the dollar for private financial contracts at home (Arida, Bacha and Lara-Resende, 2005).

In fact, the use of the dollar is not only not restricted in the financially dollarized countries; it is, as noted by the papers by Ize and Levy Yeyati, rather promoted by the macro policies and the regulatory environment. I see this pro-dollarization bias as a natural by-product of path dependence.

As the experiences of Colombia and Ecuador during the 1990s show, high real interest-rate volatility is public enemy number one of the UF. Such volatility caused the budding UF to disappear in Ecuador in the mid-1990s and it caused the UF to be replaced by indexation to the nominal interest rate in the case of Colombia in the second half of the 1990s.

    Other Resources Citing This Publication