Powers correlation analysis of non-stationary illiquid assets

TL;DR

This paper investigates the limitations of classical powers correlation in analyzing non-stationary illiquid assets, proposing robust tools to distinguish true long-term volatility effects from artifacts caused by changing zero returns probability.

Contribution

The paper introduces new powers correlation methods that are robust to non-stationarity and changes in zero returns probability, improving volatility analysis of illiquid assets.

Findings

Classical powers correlation can misinterpret volatility persistence.

Proposed tools effectively distinguish short-term from long-term volatility effects.

Empirical results suggest volatility effects are often only short-term.

Abstract

In this paper, the higher order dynamics of individual illiquid stocks are investigated. We show that considering the classical powers correlation could lead to a spurious assessment of the volatility persistency or long memory volatility effects, if the zero returns probability is non-constant over time. In other words, the classical tools are not able to distinguish between long-run volatility effects, such as IGARCH, and the case where the zero returns are not evenly distributed over time. As a consequence, tools that are robust to changes in the degree of illiquidity are proposed. Since a time-varying zero returns probability could potentially be accompanied by a non-constant unconditional variance, we then develop powers correlations that are also robust in such a case. In addition, note that the tools proposed in the paper offer a rigorous analysis of the short-run volatility effects, while the use of the classical power correlations lead to doubtful conclusions. The Monte Carlo experiments, and the study of the absolute value correlation of some stocks taken from the Chilean financial market, suggest that the volatility effects are only short-run in many cases.