HSBC Developed New Model to Predict Treasury Yields

The machine-learning tool forecasts a low probability of rising 10-year Treasury yields for investors over the next month.

Updated on Sept. 22, 2026 in Stock Markets

Isometric editorial illustration of a stack of bond certificates and balancing weights, representing quantitative financial modeling. — HSBC has introduced a machine-learning model, DUSTIN, to forecast 10-year U.S. Treasury yield directions based on market volatility and credit trends. AI Illustration. Upload story photo >

Live Poll

Would you trust a machine-learning model's projection to guide your personal investment decisions?

HSBC has introduced a machine-learning model called DUSTIN designed to project the direction of 10-year U.S. Treasury yields. The tool currently indicates a 28% probability that yields will rise over the next 21 trading days.

Why it matters

Understanding Treasury yield trends helps investors gauge broader interest rate environments, which influence the cost of borrowing for mortgages and other consumer credit. This model attempts to simplify complex market data into directional signals for monitoring those costs.

The DUSTIN model uses 17 variables, including momentum and volatility, to achieve a 65% accuracy rate on out-of-sample data. It currently assigns a 28% probability to rising 10-year Treasury yields.

The players

HSBC

A global banking institution that provides retail and commercial financial products, including savings accounts and credit services.

The details

The model processes inputs by converting raw market values into rankings to account for differing scales of data, such as credit performance and volatility. By analyzing these variables, the system identifies correlations where, for instance, a flattening yield curve suggests that 10-year Treasury yields are likely to decline. This quantitative approach is designed to provide clearer directional forecasts for treasury instruments.

The forecast window for 10-year Treasury yield direction covers the next 21 trading days.

Money Landscape

The use of machine learning to track Treasury yields follows the established pattern of using a flattening yield curve as a key indicator of market expectations. This approach helps ground yield projections within the broader historical range of interest rate shifts.

Changes in Treasury yields can influence the interest rates offered on consumer debt, such as fixed-rate mortgages and personal loans. You may wish to review your current debt obligations and consult with a financial professional about how yield trends could affect future borrowing costs.

The takeaway

While quantitative models offer new ways to track market signals, they remain projections rather than guarantees of economic shifts. Investors should monitor yield trends alongside their long-term financial goals and discuss any significant portfolio adjustments with a professional adviser.