Infinite-Variance Error Structure in Finance and Economics

Abstract

Many macroeconomic and financial data exhibit large outliers and high volatility so that their returns are usually modeled to follow an infinite-variance stable process. Extreme behaviors in such data tend to exist especially for emerging markets due to frequent existence of high economic turmoil. A relatively new area of research studies that model the financial returns as infinite-variance stable errors exists for emerging markets as well as for industrialized countries. This study aims to briefly introduce the reader the concept of infinite-variance stable distributions, discuss some existing studies on unit root and co-integration tests that assume infinite-variance stable error structure, and then to point out the potential lines of research while showing the significance of this relatively new concept.