Brazil’s Central Bank Lost $1.8bn on Swaps, Made $4.8bn on Reserves

Central Bank

Every Brazilian paper this week reported that the Brazil central bank swap book lost just over nine billion reais in June. Almost none reported that the same release showed reserves earning nearly twenty-five billion, and the two numbers are linked.

A currency swap at the Brazilian central bank is a synthetic dollar sale. The bank offers the market protection against a rising dollar, paying out the exchange-rate move and receiving the domestic interest rate.

Its international reserves, meanwhile, are held in dollars. When the real weakens, those reserves are worth more reais, and the swap book bleeds.

One instrument cannot lose without the other gaining. That is not a coincidence to be discovered, it is the definition of a hedge.

What the Brazil central bank swap numbers really show

Put two months of the bank’s own series side by side and the point becomes clear. In June the swap position lost just over nine billion reais while reserves management earned almost twenty-five billion.

In January the swap position gained more than forty-two billion reais. Reserves management lost more than seventy-one billion, so the signs flip together, both times.

Six months apart, two opposite currency moves, one steady structure. A reader shown only the swap line in January would think the bank brilliant, and in June that it was inept.

It was neither. The bank says so itself, in the same note every month, that it seeks no profit from swaps or reserve management.

The January figures make the case more strongly still. That month the losses on reserves ran to more than seventy billion reais, dwarfing the swap gain, and drew almost no comment at all.

Its stated aims are to supply hedges to the market during volatility and to keep a cash cushion for a crisis. Those are insurance goals, and insurance is not supposed to make money.

The number nobody prints

Now do the arithmetic the release invites and nobody performs. Reserves earned nearly twenty-five billion reais gross, and the net result, meaning return minus funding cost, was a little over six billion.

Subtract one from the other and you have the cost of carrying the reserves. That is above eighteen and a half billion reais in a single month, or about three and six-tenths billion dollars.

Run the same subtraction on January and it comes to nearly thirteen billion reais. That line is negative every month, in good currency weather and bad, because reserves are funded with domestic debt at a policy rate of fourteen and a quarter percent.

This is the true price of Brazil’s insurance policy. It is the only figure in the release that never changes sign, and the swap result is theatre by comparison.

Whether the premium is worth paying is a fair question, and a large one for a country with a heavy debt load. It is simply not the question the monthly swap headline is answering.

The bank has also been shrinking the book. It bought dollar futures in May for the first time in a decade, a reverse swap that runs against a position measured in the tens of billions of dollars.

Why the papers disagree with each other

If you saw two different June losses quoted this week, both were right. The cash measure recorded a shade over nine billion reais and the accrual measure a shade over eight.

The gap of roughly a billion reais is settlement timing, because the cash effect lands the next business day. Accrual counts what happened in the month, whenever the money moves.

The month-on-month swing is the other thing worth holding. May produced a gain of nearly one and four-tenths billion reais, so the turn to June is a move of more than ten and a half billion, roughly two billion dollars, in thirty days.

Nothing about the bank’s policy changed in those thirty days. The currency did, and this instrument is built to register exactly that.

A separate line, the result of foreign-exchange operations, showed a loss near two billion reais. It is a small number beside the others and it drew no headlines at all.

For a foreign investor the practical lesson is narrow but useful. None of this is a signal about Brazilian solvency, and the bank publishes its entire swap position and maturity schedule daily, in a public dataset running back to 2011.

Anyone can check the book rather than the headline. The gap in knowledge here is not the central bank’s fault, it belongs to whoever declined to open the second page of the release.

Did the Brazil central bank swap position really lose money?

Yes, just over nine billion reais in June on the cash measure. The same release showed reserves management earning nearly twenty-five billion.

Why do the two lines move in opposite directions?

Reserves are held in dollars and the swap book is a synthetic dollar sale. A weaker real lifts one and hurts the other.

What does it cost Brazil to hold reserves?

Gross return minus net result implies a funding cost above eighteen billion reais in June alone. That line is negative every month.

More: Brazil news in English, every day from The Rio Times.

Frequently Asked Questions

Why did Brazilian newspapers report different figures for the central bank's swap loss in June?

Two different accounting methods were used to calculate the loss, producing figures that were R$1.015bn apart. This is why various publications printed different numbers from the same official release.

How do the central bank's currency swaps and international reserves relate to each other?

The two instruments move in opposite directions by design. When the real weakens, the swap book loses money while the dollar reserves gain value in reais.

In June, for example, the swaps lost R$9.277bn while reserves earned R$24.893bn in the same release.

What is the actual cost of holding Brazil's international reserves?

Holding the reserves cost R$18.629bn in June alone. That reduced the gross reserve gain of R$24.893bn to a net figure of R$6.264bn after costs.

The central bank keeps these reserves not to make a profit but as emergency backing for its currency swap operations.