Some Econometric Recipes for High-Frequency Data Cooking
According to Ragnar Frisch (1933), econometrics involves the mutual penetration of quantitative economic theory and statistical observation. Today, econometric research rarely has all the ingredients spelled out in Frisch's definition. Most econometric models are either very structural, contrived, and unfortunately typically rejected or else they are statistical summaries of stylized data features and unfortunately difficult to link unambiguously to a single theoretical paradigm. Because my role on the panel is to discuss econometric issues, let me try to spell out what econometricians have to offer to the field of high-frequency data analysis. It is fair to ask, What do we learn from high-frequency data that we do not from traditional time series? More