AI Case Study
Bank of America enhances its published currency research with machine learning
Bank of America is the first among three biggest U.S. banks to provide machine learning based currency research for the foreign exchange (FX) market. In fear of market turmoil, following political issues in Italy in the beginning of summer 2017, the bank began using machine learning programs to advise clients on what to sell and buy. To predict how the euro-dollar currency pair could possibly perform, BoA trained its system using both supervised and unsupervised learning on data, such as government spending and consumer confidence. Its inference was that in the aftermath of the Italian election the euro would likely weaken.
"The bank’s currency strategists are using machine-learning programs -- which enable computers to comb through vast amounts of data to draw inferences and make predictions on their own -- to tell clients what to buy and sell. They began producing AI-based research last month, as political turmoil in Italy roiled financial markets and sparked fears of another existential crisis in Europe.
“It’s hard to learn from the historical data because of the nature of the FX market, so we try to really push the frontier” with alternative data and machine learning, said Alice Leng, the currency strategist who authored Bank of America’s AI-based research.
Quant funds have used machine learning for years. But at a time when Wall Street research is increasingly commoditized, it’s not hard to see why Bank of America is trying to capitalize on one of the hottest buzzwords in finance.
"The bank’s models concluded that in the aftermath of the Italian election, in which euro-skeptic parties swept into power, the common currency would likely weaken. However, fears of a deep and sustained selloff against the dollar, like the one witnessed during the European debt crisis, were overblown."
"The team used both supervised learning, when the machine receives training in how to process information, and unsupervised learning, when no classification guidelines are given."
"For the team’s first study, Bank of America’s machine-learning algorithms sifted through fundamental and survey data, such as government spending and consumer confidence, to determine how the euro-dollar currency pair might perform."