Decoding the stock market dynamics in the banking sector: Short versus long-term insights

The severity of extreme fluctuations and crises within the global banking sector is escalating. Conventional models, operating on a single time scale, may misinterpret any shift as a change in the long-term trend, distorting market insights. To address this issue, the present paper introduces a hierarchical structure into the standard hidden Markov model, enabling the differentiation of short and long-term trends within the U.S. banking industry. Using NASDAQ Bank stock market index data from January 1, 2007, to July 31, 2023 at two different frequencies, we construct and evaluate different calibrations of the hierarchical hidden Markov model. Results reveal two long-term regimes: turbulent periods with high volatility, instability, and negative returns, and prevalent stable markets. Within each of them, two distinct states representing short-term trends are identified, exhibiting significant differences in persistence, likelihood, expected returns, and risk profiles. The results show that an investor should carefully differentiate between regimes on both hierarchies to make informed investment decisions.