Foreign exchange market data analysis reveals statistical features that predict price movement acceleration

Area of Science:

• Quantitative Finance
• Financial Econometrics
• Market Dynamics

Background:

• Accessible financial data enable the inference of market dynamics and model creation.
• Recent years have seen the uncovering of several stylized facts in financial markets.

Purpose of the Study:

• To analyze foreign exchange data to uncover a statistical financial law.
• To characterize a new stylized fact of financial markets and its deviation from traditional models.

Main Methods:

• Extensive analysis of foreign exchange data.
• Statistical analysis to identify power-law relationships and volatility patterns.
• Comparison of real data with theoretical simulations from the Black-Scholes model.

Main Results:

• A "breaking-acceleration effect" was identified: volatility increases significantly when prices exceed historical highs or lows.
• The probability of breaking historical price extremes follows a power law in both real and simulated data.
• Real data showed a lower probability of breaking resistance than predicted by the Black-Scholes model, but higher volatility when extremes were breached.

Conclusions:

• The study reveals a "resistance effect" where markets exceed historical extremes less often than predicted by the Black-Scholes model.
• When market extremes are breached, volatility is substantially higher than expected.
• These findings suggest that Markovian models do not fully capture the complexities of market dynamics.