计量经济学5大门派，风云诡谲舞林争霸

计量经济学派其实是存在的，大家对整个世界的认知是有不同视角的。计量经济学派是从他们的研究方法划分的，有些比较喜欢从理论角度建模，有些比较倾向于从数据角度建模，而有些比较倾向于搞随机试验设计来建模。这些计量经济学派活跃者都在欧美国家，可以看出中国计量经济学背后的理论逻辑依然比较落后，毕竟大家善于用软件做些数据实证。

考尔斯委员会计量经济学派：the Cowles Commission approach

The thrust of the Cowles approach was a specific, probabilistic framework in estimating simultaneous equations to model an economy. Its ultimate goal in doing so was to gain policy insight. The Cowles approach structured its models from a priori economic theory. One of its main contributions was in exposing the bias of ordinary least squares regression in identifying coefficient estimates.

Consequently, Cowles researchers developed new methods such as the indirect least squares, instrumental variable methods, full information maximum likelihood method, and limited information maximum likelihood method. All of these methods used theoretical, a priori restrictions. According to an article by Carl F. Christ, the Cowles approach was grounded on the following assumptions:

1, simultaneous economic behavior; 2, linear or logarithmic equations and disturbances; 3, systematic, observable variables without error; 4, discrete variable changes as opposed to continuous; 5, a prior determination of exogeneity and endogeneity; 6, the existence of a reduced form; 7, independence of the explanatory variables; 8, a priori identified structural equations; 9, normally distributed disturbances with zero means, finite and constant covariances, a nonsingular covariance matrix, and serial independence; 10, a dynamically stable system of equations.

网站链接：http://cowles.yale.edu/people/research-faculty-staff

向量自回归计量经济学派：the Vector autoregression approach

Vector autoregression (VAR) is a stochastic process model used to capture the linear interdependencies among multiple time series. VAR models generalize the univariate autoregressive model (AR model) by allowing for more than one evolving variable. All variables in a VAR enter the model in the same way: each variable has an equation explaining its evolution based on its own lagged values, the lagged values of the other model variables, and an error term.

VAR modeling does not require as much knowledge about the forces influencing a variable as do structural models with simultaneous equations: The only prior knowledge required is a list of variables which can be hypothesized to affect each other intertemporally.

伦敦经济学院计量经济学派：the LSE approach to econometrics - originated with Denis Sargan now associated with David Hendry (and his general-to-specific modeling). Also associated this approach is the work on integrated and cointegrated systems originating on the work of Engle and Granger and Johansen and Juselius(Juselius 1999)

The LSE approach to econometrics, named for the London School of Economics, involves viewing econometric models as reductions from some unknown data generation process (DGP). A complex DGP is typically modelled as the starting point and this complexity allows information in the data from the real world but absent in the theory to be drawn upon. The complexity is then reduced by the econometrician by a series of restrictions which are tested.

One particular functional form, the error-correction model, is often arrived at when modelling time series. Denis Sargan and David Forbes Hendry (with his general-to-specific modeling) were key figures in the development of the approach and the one way the approach has been extended is through the work on integrated and cointegrated systems by Robert F. Engle, Clive Granger, and Søren Johansen. Another commonly used functional form is distributed lag or autoregressive distributed lag.

应用校准建模的计量经济学派：the use of calibration - Finn Kydland and Edward Prescott（真实周期理论学派）

试验和双倍差分法的计量经济学派：the experimentalist or difference in differences approach - Joshua Angrist and Jörn-Steffen Pischke.

The experimentalist approach to econometrics is a way of doing econometrics that, according to Angrist and Krueger (1999): … puts front and center the problem of identifying causal effects from specific events or situations. These events or situations are thought of as natural experiments that generate exogenous variations in variables that would otherwise be endogenous in the behavioral relationship of interest.

An example from the economic study of education can be used to illustrate the approach. Here we might be interested in the effect of effect of an additional year of education (say X) on earnings (say Y). Those working with an experimentalist approach to econometrics would argue that such a question is problematic to answer because, and this is using their terminology, education is not randomly assigned. That is those with different education levels would tend to also have different levels of other variables.

And these other variable, many of which would be unobserved (such as innate ability), also affect earnings. This renders the causal effect of extra years of schooling difficult to identify. The experimentalist approach looks for an instrumental variable that is correlated with X but uncorrelated with the unobservables.

《END》

写在后面：各位圈友，一个等待数日的好消息，是计量经济圈应圈友提议，09月04日创建了“计量经济圈的圈子”知识分享社群，如果你对计量感兴趣，并且考虑加入咱们这个计量圈子来受益彼此，那看看这篇介绍文章和操作步骤哦（戳这里）。进去之后一定要看“群公告”，不然接收不了群信息。