Applied Financial Economics, 2006, 16, 395–404 Behavioural and fundamental explanations of discounts on closed end funds: an empirical analysis George Emm. Halkos* and Theodore N. Krintas Department of Economics, University of Thessaly, Argonavton and Filellinon St., 38221 Volos, Greece This study extracts two factors related to the variability of Premium/ Discount: a behavioural and a fundamental. Evidence is provided to show that by using both factors one can achieve a better understanding of discounts as theories and the Closed End Funds Puzzle support it. I. Introduction The majority of financial and economic theory is based on the notion of rationality that characterizes the actions of individuals concerning all the available information in the decision making process. In other words the Efficient Market Hypothesis (hereafter EMH), which is actually defined from the above- mentioned sentence, has been the central proposition in finance for more than 30 years. It states that security prices in financial markets must equal fundamental values, either because investors are rational, or because arbitrage eliminates pricing anomalies (Shleifer, 2000). An alternative approach has been developed, attempting to better understand and explain how emotions and cognitive errors influence investors and the decision making process. Many researchers believe that the study of psychology and other social sciences can shed considerable light on the efficiency of financial markets and explain many stock anomalies. As an example, some believe that the out-performance of value investing comes from investors’ irrational overconfidence in exciting growth companies and from the fact that investors generate pleasure and pride from owning growth stocks. Researchers also believe that these human flaws are consistent, predictable and can be exploited for profit. Over the past four decades, investment decisions have been guided by efficient market theory. The theory is based on the notion that investors behave in a rational, predictable and unbiased manner. Empirical applications assume that investors aggres- sively correct stock prices to reflect all publicly available information. This study presents an empirical investigation of the determinants of the evolution from 1997 to 2004 of the discount on Greek closed end funds. This is achieved using a two-step procedure. First a dynamic factor analysis in financial time series is used and two factors are identified, one fundamental and one behavioural, which quantify the interplay of investor sentiment and arbitrage reaction. Then using these extracted factors, regression analysis is introduced in order to assess their effect on the premium/discount of Greek closed end funds. According to various diagnostic tests, the method- ology of Engle et al. (1987) has been utilized and maximum likelihood estimation of the ARCH-M model is used to assess the presence of ARCH effect in the data set at hand. In this way, apart from assessing the relative impact of the two factors on the discount, the study explores whether there were common patterns in premium discount as well as testing if the conditional variance influences signifi- cantly the behaviour of the premium discount at very high frequencies. *Corresponding author. E-mail: halkos@uth.gr Applied Financial Economics ISSN 0960–3107 print/ISSN 1466–4305 online ß 2006 Taylor & Francis 395 http://www.tandf.co.uk/journals DOI: 10.1080/09603100500400312