Predicting Bubbles: A Two-Step Guide Under Fluctuating Markets

Economists have been grappling with the challenge of predicting bubbles in financial markets. One recent study took a fresh approach by dividing the market into four regimes and assuming that volatility can change over time. This four-regime bubble model focuses on identifying the start, peak, collapse, and recovery of market bubbles. The researchers proposed a two-step method to improve the accuracy of these predictions. First, they estimated when the bubble starts to grow rapidly, when it collapses, and when the market returns to normal, all while assuming a constant level of market volatility. Next, they collected the leftover data, which they call residuals, and used a more advanced method called weighted least squares (WLS) to refine their estimates. Through computer simulations, they found that this two-step process significantly improved the accuracy of predicting bubble breaks compared to the traditional ordinary least squares (OLS) method in some cases.