Brazil: Foreign Exchange Reserves, the BNDES and Public Debt

The accumulation of foreign exchange reserves and loans to the BNDES increase public spending and shorten the maturity of the public debt.

One cannot deny that there has been an improvement in Brazil’s public debt indicators. This stands out even more clearly when compared to the situation in developed countries, which increased their debts significantly to mitigate the recession and bail out their financial systems.  However, two economic policy actions have been having a negative effect on Brazilian public debt: the accumulation of foreign currency reserves resulting from the Central Bank’s (CB) sterilized interventions and the National Treasury’s (NT) loans to financial institutions, notably the National Bank for Economic and Social Development (BNDES).

Both actions increase gross debt without affecting net public debt (which deducts the public sector’s main assets). In the case of foreign exchange interventions, the public sector purchases an asset –  foreign currency reserves –  as a counterpart of the liability issued – the public debt. In the second case, the NT’s loans to the BNDES constitute the assets acquired by the public sector.

Although they do not initially increase net debt, the two operations are very onerous over time. In the case of the accumulation of foreign currency reserves, the latter yield the low dollar rates of interest paid in international markets. The International Reserve Management Report –  available at the BC’s website –  informed that the reserves yielded a scant 0.83% in 2009. In the case of the BNDES, on the other hand, the interest paid on the NT’s loans is about the same as the long-term interest rate (TJLP), the heavily subsidized rate charged by the BNDES on its operations (6% a year). To finance these two initiatives, the government issues debt, at a rate of interest that is currently equal to or higher than the Selic rate (10.75% a year). The result may be seen in Graph 1. Although the Selic rate has fallen considerably since 2008, the public debt’s implicit interest rate has remained practically unchanged. This is because the operations cited above raise the public debt’s cost substantially, as they involve exchanging low-yielding assets for highly onerous liabilities.

As well as increasing the public debt’s cost, the accumulation of foreign currency reserves, also leads to a deterioration in the profile of the Federal Bonded Debt (made up of government bonds issued in Brazil). In order to buy dollars, the CB issues money. The CB itself then has to remove the money it has issued from circulation, otherwise the Selic rate would fall, preventing it from fulfilling the target established by the COPOM (Monetary Policy Committee).  This operation, which involves “soaking up” the liquidity generated by the purchase of reserves, is known as sterilization. In Brazil, sterilization operations have been undertaken via repo issues, rather than by selling longer-dated government bonds.

Graph 2 shows that there has been a clear increase in the debt’s average maturity since 2005. This is very positive, as it diminishes rollover risk and encourages the much needed lengthening of durations in our fixed income market. The Federal Bonded Debt’s average maturity is currently more than 41 months.

However, if short-term repos are included (as they should be), the average maturity falls to a much lower 33 months. It is true that the increase in banks’ reserve requirements, announced at the beginning and the end of 2010, tends to mitigate this decline in average durations because banks exchange repos for reserve requirements at the CB. But even after taking this effect into account, if the CB continues to perform its onerous sterilized interventions, the public debt’s average duration will continue to decline, unless the NT and CB begin to issue longer-dated debt instead of using repos.  

The analyses of the Federal Bonded Debt available at the TN’s website typically ignore repo operations. This is curious, given that they currently represent more than 25% of Federal Bonded Debt (having already accounted for more than 35% in the recent past). The graphs presented there, showing a decline in the proportion of debt maturing within 12 months, or the composition of the public debt by indexators, do not provide an accurate picture of the overall Federal Bonded Debt’s situation.

In sum, new sterilized interventions and new TN loans should undergo a cost-benefit analysis. As well as the increase in gross debt, these costs should include the increasing cost of the public debt and the deterioration in the Federal Bonded Debt’s profile.  

Graph 1


Graph 2