High level chaos in the exchange and index markets

Many studies were inconclusive about the presence of chaos in financial markets due to test misspecification. Chaos tests present in the literature need noise-free time series, since any measurement error will induce the rejection of chaos. Moreover, chaos was merely tested on a low-level basis. This paper investigates the presence of a high-level noisy chaos in financial data; simulations were conclusive about the power of the test. When applied to six stock indexes and six exchange rates, the hypothesis of chaotic dynamics was rejected for all data.