Econometric Theory/Assumptions of Classical Linear Regression Model

The estimators that we create through linear regression give us a relationship between the variables. However, performing a regression does not automatically give us a reliable relationship between the variables. In order to create reliable relationships, we must know the properties of the estimators  \hat{\alpha}, \hat{\beta} and show that some basic assumptions about the data are true.one must understand that having a good dataset is of enormous importance for applied economic research.

UnbiasednessEdit

Under the following four assumptions, OLS is unbiased. This means that:
E(\hat \alpha) = \alpha
E(\hat \beta) = \beta

LinearityEdit

The model must be linear in the parameters.
The parameters are the coefficients on the independent variables, like \alpha and \beta. These must be linear, so having \beta^2 or e^\beta would violate this assumption.

Sample VariationEdit

The x_is cannot all have the same value.

Random SamplingEdit

The x_i values must be randomly selected. In other words, there is no correlation between two different x values: Cov(x_i, x_j) = 0 for i \neq j.

Zero Conditional MeanEdit

The mean of the error terms, given a specific value of the independent variable x_i, is zero. E(\epsilon_i | X_i) = 0.

Efficiency of OLSEdit

Given the following two assumptions, OLS is the Best Linear Unbiased Estimator (BLUE). This means that out of all possible linear unbiased estimators, OLS gives the precise estimates of \alpha and \beta.

With the third assumption, OLS is the Best Unbiased Estimator (BUE), so it even beats non-linear estimators. Also given this assumption, \hat \alpha is distributed according to the Student's t-distribution about \alpha, and \hat \beta is distributed in such a way about \beta.

No HeteroskedasticityEdit

The variance of the Error terms are constant.  Var(u_i|x_i) = \sigma^2. This means that the variance of the error term u_i does not depend on the value of x_i. If this is the case, the error terms are called homoskedastic. This is not always the case in economic data, for example the variation in a person's wage will vary with their level of education -- someone who is a high-school dropout will not have much variation in their wage, where people with Ph.D.s may see very different wages.

No Serial CorrelationEdit

The error terms are independently distributed so that their covariance is 0. Cov(u_i, u_j | x_i, x_j) = 0 \forall i \ne j.

Normally Distributed ErrorsEdit

The error terms are normally distributed. u_ ~ N(0,\sigma^2)

Last modified on 3 April 2014, at 04:02