Time-series Properties of Earnings: The Case of Georgian Stock Exchange

Erekle Pirveli and Jochen Zimmermann

The Georgian Stock Exchange (GeSE) is one of the smallest and most illiquid capital markets worldwide. Inefficient regulation and high transaction costs feature high in describing the embryonic stage of Georgia's stock market. This paper, by supplying first-hand empirical evidence on time-series properties of the GeSE firms’ earnings, attempts to consider and explain the limited stock market properties from earnings quality perspective. Statistical tests predicate on the primarily collected financial information of 83 Joint Stock Companies registered at the GeSE. The work covers the years of from 2005 to 2013, constituting around 500 firm-year observations. The major finding suggests that reported earnings at the GeSE are poorly persistent and predictable, making it difficult for investors to assess firm value. This is a pervasive phenomenon: Investors are practically unable to estimate firm finances up to a three year horizon. It is further revealed that from the earnings components such as cash-flows from operations and total accruals, the former operates with higher persistence and predictability. Financially pernicious years (2008-2009)’ negative effect on the properties of earnings’ components, as opposed to earnings themselves, is also detected. The findings are robust to time and industry fixed effects.

Key Words: Time-series properties of earnings; the Georgian Stock Exchange; earnings quality determinants; pooled OLS regression

JEL Codes: G11, G14, M41, O16

DOI : http://dx.doi.org/10.21102/jbpr.2015.07.101.04