Researchers from Ateneo de Manila University have created advanced artificial intelligence (AI) tools capable of predicting money market interest rates, providing valuable insights for policymakers and business leaders in managing financial risks and economic strategies.
The team employed two deep learning models—Multi-layer Perceptrons (MLP) and Vanilla Generative Adversarial Networks (VGAN)—to forecast changes in the Philippine Benchmark Valuation (BVAL) rates. These rates, which reflect borrowing costs and savings rewards, are influenced by factors like inflation, supply-demand dynamics, and central bank policies. Accurate forecasting of these rates is critical for sound economic and financial decision-making.
“Interest rates are among the most important macroeconomic factors considered by both government and private entities when making investment and policy decisions. A reliable forecast is a requisite to sound management of exposure to different types of risk,” the researchers noted in their study.
Breaking down the AI models
MLP, a type of artificial neural network, processes data through interconnected layers, making it suitable for detecting complex patterns in financial datasets. It excelled in providing accurate predictions using fewer variables and simpler setups. On the other hand, VGAN—a system comprising two opposing networks, a generator and a discriminator—proved effective in analyzing more intricate datasets, particularly under complex scenarios.
Both models successfully anticipated trends in one-, three-, six-month, and one-year BVAL rates during pre-pandemic and pandemic periods. The researchers integrated 16 key economic indicators, such as inflation, exchange rates, and credit default swaps, into their analysis, further enhancing the models’ reliability.
