Lies, Damned Lies, and Statistics? Examples from Finance and Economics
CEJEME 5: 231-248 (2013)
18 Pages Posted: 30 Dec 2015
Date Written: December 30, 2015
Abstract
As the saying goes, lies can be classified into the following, from bad to worse: lies, damned lies, and statistics (the worst of all). But is this saying really true? Many puzzles in finance and economics have accumulated over the years. Seemingly reasonable economic theories have been badly rejected by the data, giving rise to suspicions that something is wrong, and not just with the theories. What if the data contained some hidden features that, if neglected, could be misleading the analysis? What if these features required new statistical tools to bring them out and handle them? With a few examples collected from my research with colleagues, this paper will show how the conclusions could be quite different from what was originally thought.
Keywords: Flexible density specification, option pricing, term structure of interest rates, expectation hypothesis, nonlinear long-memory, macroeconomic dynamics.
JEL Classification: C10, C5, C32, G1, E3, E6
Suggested Citation: Suggested Citation