1. Introduction
There are many different views as to the objectives of debt management but in the
case of Brazil the paramount ob jective of debt management should be that of reducing the
country’s fiscal vulnerability.1 This calls for funding at low cost but also for minimizing
the risk of large interest payments due to unexpected changes in interest rates and/or in
the exchange rate.2 Risk minimization is accomplished, as shown by Goldfa jn (1998), by
choosing debt instruments which both ensure a low volatility of returns and provide a hedge
against fluctuations in the primary budget, in the interest payments and in the value of the
other liabilities.
In this paper we present a simple model where debt management helps to stabilize the
debt ratio and thus reduces the probability of a debt crisis. Reducing the uncertainty of the
debt ratio, for any expected cost of debt service, is valuable in that it lowers the probability
that the fiscal adjustment may fail because of a bad shock to the budget.
The optimal debt composition is derived by looking at the relative impact of the risk and
cost of alternative debt instruments on the probability of missing the stabilization target.
This allows to price risk against the expected cost of debt service and thus to find the
optimal combination along the trade off between cost and risk minimization.
The optimal debt structure is thus a function of the expected return differentials between
debt instruments, of the conditional variance of debt returns and of their covariances with
output growth, inflation, exchange-rate depreciation and the Selic rate. We estimate the
relevant covariances with three alternative methods. The first approach exploits the daily
survey of expectations on GDP growth, inflation, the exchange rate and the Selic rate. The
second method relies on a small structural model of the Brazilian economy estimated on
monthly data for the period 1999:03-2003:07. The one-year ahead unanticipated components
of the Selic rate, the exchange rate, inflation and output growth are estimated as the 12-
month cumulated impulse responses of these variables to shocks to inflation, the output gap
and the EMBI spread. The third approach approximates the one-year ahead unanticipated
components of the relevant variables using the residuals of forecasting regressions run on
quarterly data for the period 1995:3-2003:1.
The empirical evidence suggests that a large share of the Brazilian debt should be
indexed to the price level. Price indexation should be preferred to Selic-rate indexation
while the share of dollar denominated (and indexed) bonds should be further reduced from
the current high level.3 These policy prescriptions appear robust to alternative methods
of estimating the optimal debt structure. The share of fixed-rate bonds should also be
increased. Fixed-rate debt avoids large interest payments when the Selic rate rises during
a crisis or reacts to negative supply shocks and thus when debt stabilization is endangered
by slow output growth. Because of their short maturity, below two years, fixed-rate bonds
1See Missale (1999) for a review of the literature on the objectives of debt management.
2see, Garcia (2002).
3For a similar conclusion in favor of price indexation see Bevilaqua and Garcia (2000). Goldfajn (1998)
also hardly finds an explanation for the high share of foreign-denominated debt.