Minister Mantega’s suggestion to revise the concept of gross debt should be much broadened so that credibility on the Brazilian fiscal accounts could be regained.
Minister Mantega sent, last month, a letter to IMF`s managing director Christine Lagarde, requesting a methodological revision of the “…IMF measurement of the Brazilian Gross General Government Debt”. What seems to motivate the minister is that the requested revision should remove from the gross debt part of the National Treasury’s (NT) bonds held by the Brazilian Central Bank (BCB), those that are not being used as collateral to repos. The change would reduce gross debt from 68.0% to 58.7% of GDP (data from December 2012), bringing it below 60.0%, usually considered as the threshold for over-indebtedness.
Regardless of the minister’s motivation, the BCB has been using the new methodology to compute the Brazilian public gross debt for many years. The new methodology has merits. For example, this year, the NT has been redeeming its debt in net terms. The Brazilian constitution does not allow the NT to hold deposit accounts in any bank, only at the BCB. This characteristic creates a heavy burden to monetary policy, as a redemption of public bonds creates an increase in the monetary base of the same amount. To keep the basic interest rate (Selic) at the target, set by the monetary policy committee (COPOM), the BCB then has to mop up the excess of liquidity generated by the bond redemption. That means, by the end of the process, that the securities issued by Treasury were replaced by repos from the BCB, using government bonds as collateral. Hence, the public debt should remain the same, only with a change of composition. And, indeed, using BCB’s new concept of gross debt, it barely changed: it went from 58.7% of GDP, in December of 2012, to 59.3 % of GDP, in June of 2013. Yet, when the traditional concept, adopted by the IMF is used, it fell from 68.0% to 64.8% of GDP. This fall occurred because the bonds used as collateral for the repos were already counted as part of the gross debt. Thus, the redemption of Treasury securities, although fully replaced by repos, reduced the gross debt, according to the traditional concept.
The discussion brought up by the minister is relevant and deserves attention. However, if the minister really wants to improve Brazil’s fiscal accounting, the discussion should be much broader and extended to many more issues than the mere revision of the gross debt measurement. After all, over the last years, the minister and the NT secretary have promoted a number of measures that caused regrettable loss of credibility of the Brazilian fiscal numbers. It is long due a serious discussion of how to improve our discredited fiscal accounting.
There are important points to deal with in this debate. The first is about the primary balance measurement. As it is well known, the primary deficit is defined as government expenses, excluding interest payments, minus government revenues. Interest expenses depend on the interest rate and the size of the public debt. Over the last years, the Treasury has lent hundreds of billions of reais to government-owned banks, especially to BNDES, at rates lower than the Treasury’s cost of capital, so that these banks could loan to firms and households, at heavily subsidized rates. An underestimation of the subsidy given by loans to BNDES is the difference between Selic (the NT cost of capital; today at 8.5%) and TJLP (long-term interest rate, a rate set by the government, used for BNDES loans; today at 5%), 3.5 percentage points. Except for a small part, these subsidies do not appear on the fiscal statistics as primary expenses, but as an interest expense1. This is conceptually wrong; granting subsidies so that BNDES may lend at low interest rates to firms is a primary expense and, just like any other, should be part of the budget. The current methodology masks this expense in the interest account, jeopardizing the cost-benefit analysis of this public policy, which has much contributed to the increase in the Brazilian public debt. The Treasury loans to public banks should also be included in the budget.
Recognizing the interest-rate subsidies as primary expenses would bring the additional benefit of discouraging the accounting tricks that have been undermining our fiscal accounts’ credibility. It became a terrible habit of the current administration to transfer subsided resources from the Treasury to state-owned companies, so that they may generate fictional profits and transfer artificially high dividends back to the NT, thereby unduly inflating the primary surplus. If the subsidy’s total cost were properly recognized as a primary expense, there would not be room for this kind of accounting gimmick.
Another relevant topic is the revision of the net debt measurement. Roughly speaking, the net debt is the gross debt subtracted of some public sector’s financial assets. Among these, there are the loans that the NT made to government-owned banks, over 400 billion reais. It is true that payments owed to the Treasury from its loans to public banks are not linked to potential default losses on the loans made by such banks to firms and households. However, if the default losses are big enough to require a recapitalization of any public bank (which has already happened many times before), the bill will footed by the NT. In other words, the value of R$1 lent by the Treasury to a public bank is different from the value of R$1 in international reserves, and that should be recognized in net debt’s measurement.
Finally, it would be very beneficial to create a fiscal council, such as the Chile did recently, and many other countries have done in the past in different formats. The Chilean one is composed by renowned specialists from outside the government, and its mandate is to implement the measurement of the structural surplus, i.e., the fiscal surplus adjusted by the business cycle. Brazil’s Central Bank has announced that this concept might guide henceforward its monetary policy decisions. There is much more reason for fiscal policy to do the same. If so, the council would be in charge of improving, clearing and consolidating the structural surplus measurement. It would not set the fiscal policy, an inalienable attribution of democratically elected governments. But it should determine if the fiscal policy stance is or is not compatible with governments announced goals for public deficit and debt2. Of course, such council should also deter the very damaging proliferation of accounting tricks that we have witnessed in the last few years. Certainly, it would be a great step to rehabilitate the battered credibility of our fiscal statistics and improve fiscal policy.
1For example, the interest rate of loans given by BNDES under the Programa de Sustentação do Investimento (Program to Sustain Investment, PSI), to finance purchases of machinery and equipments, is 3% per year, while the TJLP is 5% and the Selic is 8,5%. The difference between PSI’s interest rate and TJLP is considered a primary expense – interest equalization – but the same does not apply to the difference between Selic and TJLP, which is considered an interest expense.
2The Lei de Responsabilidade Fiscal (Fiscal Accountability Act), mentions the creation of a fiscal management council. Therefore, this debate is even more timely, as its implementation requires Congress to enact a law.