US + Brazil Financial News (2/27)

The Treasury disclosed this morning the latest data on the stock and composition of its debt stock (DPMF). On average, over the last 12mo ending in Jan/18, the stock (including debt held at the central bank) grew 15.7%yoy or 12%yoy in real terms (deflated by IPCA inflation). The average interest rate paid on its debt was 9.4%, against an inflation of 3.2% in the period; i.e., a real rate of 6%. Meanwhile, data to be announced this week should show real GDP growth at 1.1%yoy and a running 12mo deficit on the Treasury accounts of 1.9-2% of GDP. Clearly, the numbers are inconsistent: A debt that grows at an annual rate of 12% in real terms; that pays a real interest rate of 6%; product of a fiscal system that produces running deficits of 2% of GDP; in a country that grows at an annual rate of 1.1%; is unsustainable. Of course, the data are backward looking. The hope is that, going forward, the system will adjust and the numbers re-giggle magically to produce declining ratios of Debt/GDP. It is a lot of wishful thinking, but then the market is adamantly hopeful about the outcome of the October elections. And, in the interim, truth-be-told, the Temer administration did a lot to pave the way to the holy-grail of “the new fiscal regime.”

It is important to focus on one aspect of the Treasury debt—admittedly, an arcane technical detail, but one with important implications for the future of policy.

Annex 5.3 of the Treasury’s report details the central bank’s actions in the secondary market for domestic Treasury debt. This includes operations to regulate liquidity and keep the interbank overnight rate close to the target rate of monetary policy. In other places this would be your run-of-the mill open-market-operations (think of the market for Fed funds) but, in Brazil, the so-called “operações compromissadas” have a twist. In contrast to most central banks, the Brazilian counterpart does not have a “below the line” valuation account to reflect changes in the local currency value of the dollar stock of international reserves. Because the exchange rate can be volatile, most central banks don’t think it is useful to reflect valuation changes in their balance sheets. Instead, “above the line” they show only changes that are due to real transactions in foreign exchange during the period under consideration; purchases or sales of foreign reserves. These, yes, valued at current exchange rates.

Because Brazil has a large volume of foreign reserves (19-20% of GDP) its practice of showing all valuation changes above the line produces large swings in the central bank’s balance sheet. The real problem is how they are accounted for. Positive results (i.e., an increase in valuation from BRL depreciation and/or net acquisition of reserves) are deposited in the Treasury’s account at the central bank. Negative results for the central bank (from BRL appreciation or net sales) are made whole by the Treasury issuing paper to the central bank. In theory, the amounts in the Treasury’s account at the central bank can be used only for debt abatement. Likewise, the Treasury debt outstanding with the central bank, only for open-market-operations. In practice, however, the transactions create a loophole for the central bank to fund the Treasury.

How? Because the Treasury never withdraws moneys from its account at the central bank. Instead, it lets the moneys accrue, and collects interest—at the average cost of Treasury debt held by the central bank. The catch is that this income is treated as any other income of the Treasury. It can be used to finance the deficit. OK, in symmetry, the Treasury pays the central bank interest on the stock of Treasury paper it holds. But considering what has happened over the last 12-15 years (just think of the “maxi” devaluations and the nearly $400bn purchase of reserves) on an accumulated basis, the quantities are vastly different. The central bank pays much more in interest than its receives from the Treasury. Moreover, because the quantities are vast, the stock of compromissadas is huge: 18.5% of GDP. And they trade in the secondary market. In Jan/18 the average daily trade volume was R$265bn or 7.5% of the outstanding stock of debt (DPMF). In total, on a net basis, the central bank repo-ed R$5,876.9bn in compromissadas. Some of the repos were overnight operations; the average maturity was 63 days. The point is that this total volume was 1.67 times the stock of the DPMF—i.e., the central bank, essentially, rolled-over more than once the entire stock of Treasury debt in a single month! Clearly, this is a dangerous way to conduct to monetary policy.

Which brings us back to progress during this administration. In November last year, the Senate Economic Commission (CAE) approved a proposal of the central bank to revamp this system. The project would create a valuation account and limit transfers between the Treasury and the central bank to real transactions with the stock of reserves. It also disciplines the sale of Treasury paper to the central bank. The idea is to vastly reduce the size of the Treasury account at central bank and make the flows of interest paid between the two institutions irrelevant from a fiscal or a quasi-fiscal perspective. This would be progress indeed. In 2014/15 during the Dilma regimes, the flow from the central bank to the Treasury was a main source of funding of the deficit—obscured from the public’s view and completely against the intent (if not the wording) of the fiscal responsibility law. Let’s hope that now that it will not deal with the Social Security Reform, the full Congress approves this proposal soon.

In addition to the debt and fiscal data discussed above, this week we get data on credit; on unemployment; on GDP in 2017; and on the trade account in Jan/18.