Symposium on Econometric Tools
面向《经济展望杂志》的广泛读者群，用通俗语言解释计量经济学工具的最新发展，以普通最小二乘法为例说明其原理和适用条件。
In the Broadway play Copenhagen, Niels Bohr and Warner Heisenberg impose the self-enforced requirement that they discuss quantum physics in a way that Bohr’s wife, Margrethe, can understand. If Broadway actors can attempt to explain the Heisenberg Uncertainty Principle to a lay audience, then surely econometricians and applied economists can explain recent developments in econometric tools to the diverse audience of economists that makes up the readership of this journal. As a starting point for offering a discussion of econometric tools to the broad audience of the Journal of Economic Perspectives, we recognized that ordinary least squares regression is the workhorse statistical method of economics, typically taught to every undergraduate economics major in a required econometrics or statistics course. The enduring popularity of ordinary least squares estimation is easy to understand. If the relationship between the dependent variable (sometimes called the outcome or response variable) and the explanatory variables (sometimes called the independent variables or regressors) is linear, and if the explanatory variables and the equation error are independent, then ordinary least squares using data from a random sample provides an unbiased estimate of the “true” regression line. 1 An unbiased estimator will not always yield the correct estimate—after all, no estimate is likely to be perfectly correct— but it will be correct on average.