Robust zero autocorrelation tests for financial time series

Abstract: Financial theory claims that returns from a broad class of financial instruments should be unpredictable and calls this '(financial) market efficiency.' Translated into mathematically manageable terms, these theories suggest that the price process must be a martingale difference. For its part, statistical theory has generated a good amount of evidence to detect this type of dynamics. However, practice shows that these statistical devices fail when confronted with financial data. Part of the explanation lies in the influence that the so-called 'stylized facts of financial returns' have on testing procedures. In this talk we will explore this dissonance between theory and practice by developing new autocorrelation tests that work appropriately on financial time series. We show that it is not difficult to 'harden' existing procedures and illustrate how the new robust tests do their job quite well.

Semblance: Actuary, doctor in Mathematical Sciences from UNAM, he is a full-time professor in the Department of Physics and Mathematics at the Universidad Iberoamericana CDMX. He has collaborated with different educational institutions, such as the Faculty of Sciences of the UNAM, the Center for Research in Mathematics (CIMAT), in Guanajuato, the Center for Economic Research and Teaching (CIDE), and the Department of Economics of the UC3M, in Spain. Currently level 1 of the SNI, he is primarily interested in time series statistics and econometric theory. His research has two distinct lines. On the one hand, the development of theoretical devices for the best analysis and modeling of the dynamics of time series. On the other, the application of different econometric techniques to the study of social phenomena such as economic growth and the relationship between the economy and crime.