Bubbles, Can We Spot Them? Crashes, Can We Predict Them?

Johansen and Sornette proposes that the crash has fundamentally an endogenous origin and exogenous shocks only serve as triggering factors. This endogenous force is shown in price as power law log-periodicity (PLLP) signature prior to a crash. We estimate the highly nonlinear model developed by them using a hybrid approach which combines scatter search, genetic adaptor and tabu search. The model is applied to two property data sets (Hong Kong Office Price Index and Seoul Hosing Price Index) and one property related stock price (Korea General Construction Stock Price Index). The fitting of the original model to these data sets was unsuccessful, due to the lack of the power law. We hence fit the data using a modified model, and the results are encouraging when crash-date prediction is the aim