The article explores different models for predicting changes in short-term interest rates. Traditional models that only consider interest rate levels may not accurately capture how volatility changes over time. On the other hand, models that focus on past patterns of volatility may not fully account for the impact of new information. The researchers introduce a new model that considers both interest rate levels and unexpected information shocks. They find that the relationship between interest rate levels and volatility is not as strong as previously thought. Additionally, many existing models of interest rates may need to be improved to better predict changes in volatility.