Sugar Giant Turns to Algorithms for Price Predictions

(Bloomberg) -- The world’s largest sugar producer is turning to computers to predict where prices are going next.

Sao Paulo-based Raizen Energia SA has hired QuantumBlack, a data and analytics company controlled by McKinsey & Co., to develop an algorithm capable of predicting price moves for sugar, according to people with knowledge of the matter.

The algorithm should be ready to be used by the company in about a year, one of the people said. In addition to broader market data, the analytics will also include historical data from Raizen’s 860,000 hectares (2.1 million acres) of sugar cane, including yield patterns, to help predict changes in supply, the person said. Raizen is a joint venture between Cosan SA and Royal Dutch Shell Plc.

Raizen declined to comment on the move. McKinsey said company policy prohibits client disclosures.

Algorithms have been widely used by investors and hedge funds to operate in agricultural markets. In recent years, consultancies and trading houses have been seeking to improve their pricing capabilities with the use of algorithms, but this is the first time in Brazil’s sugar market that a giant producer is leading a project of its own.

The move comes as Raizen believes it has the potential to improve its sugar forecasts, given its scale, the person said. The company is aiming to use the analytical tools as part of its risk-management strategy and doesn’t have plans to act as a high-speed or algorithmic trader, the people said. QuantumBlack will provide price forecasts for various scenarios and also the rationale behind them.

Raizen runs 26 cane mills in Brazil, producing about 4.3 million metric tons of sugar per year. That’s equal to about 15 percent of the nation’s exports. The company has a raw sugar joint venture named Raw with Singapore-based Wilmar International Ltd. to trade the sweetener sourced from Brazil. The contract with QuantumBlack doesn’t include Raw, one of the people said.

©2019 Bloomberg L.P.