Banco Brascan has a culture of risk management. Back in 1996, the Brazilian bank was one of the first local firms to start implementing a market risk system. Its head of market risk, José Alberto Rebello Baranowski, was among the first in Brazil to gain internationally-recognised risk management accreditation. Earlier this year, one of Baranowski’s colleagues in the risk team, Renato da Silva Carvalho, followed suit – gaining certification from Garp and the Professional Risk Managers’ International Association (Prmia).
“We try to keep our methodology and technology as updated as possible,” says Baranowski. “It is necessary, given the sophistication of our activities,” he adds. The bank’s treasury operation is a specialist in arbitrage. Baranowski’s team uses an in-house parametric value-at-risk approach, underpinned by a RiskMetrics-type methodology. Risk estimates are based on volatility parameters of each factor and their correlation “Our system is adaptable. We can use up to 250 risk factors, selected from a total of 400 on a daily basis,” he says.
For example, one important risk factor that affects valuation – either explicitly as in the case of some derivatives, or more subtly in bonds, debentures and the like – is changes in the US dollar-real exchange rate. A parametric approach is not applicable to options because their returns are not normally distributed – a Monte Carlo simulation is used instead.
Beyond the market itself, one of the main hindrances to sophisticated risk management in Brazil can be the lack of data. Baranowski says this can present a problem, and assumptions must be made. “Typically, when we make assumptions it isn’t that much of a concern. For arbitrages, the position per se is not at risk, rather it’s the risk that a predicted price distortion does not occur.”
Research is a mainstay of Brascan’s risk team. In the aftermath of the abandonment of the real’s peg to the US dollar in January 1999, it pioneered a way of calculating a clean interest rate, when interest rate volatility swung violently and quoted interest rates were not always meaningful. More recently,
Baranowski and his team have turned their attention to US dollar-denominated Brazilian sovereign debt – so called C-bonds. “When purchasing this kind of debt, an exposure to Brazil risk specifically is what is often being sought,” Baranowski says.
What investors require, ideally, is a yield curve for Brazilian sovereign debt – something Baranowski describes as extremely difficult to accomplish. First, there are not bonds at every maturity point; also, liquidity can be patchy; and, finally, bonds are not uniform – some are fixed, some floating, and so on.
Undaunted by the challenge, the Brascan risk team developed a technique for tackling the thorny problem a few months ago. In essence, the team is now able to decompose the sovereign bond’s risk into systemic and country risk components, alongside a correlation dependency.
“We now have an understanding of how the US Treasury market can be used to hedge out systematic risk in Brazilian bonds,” Baranowski says. Shorting US Treasury futures of similar tenor can hedge bond positions; for shorter tenor bonds – eurodollar futures are used to hedge. Baranowski says: “We are always trying to learn and get new ideas. The Brazilian markets have many unique qualities – it’s an interesting place for quantitative work.”