- Relationship between dependent variable and independent variables is linear.
- There is no linear relationship between independent variables and they are also not random. If there is linear relationship between independent variables, it is called Multicollinearity leading to high R^2 and significant F Stat. This results in inflated Std Err and low T Stat ( opposite of heteroskedasticity effect).
- Expected value of Error term is 0 and Error term is normally distributed.
- Variance of Error term is the same for all observations. If this assumption is not true, it is called Heteroskedasticity.This can be tested by Breuch Pagan test. If there is Heteroskedasticity . , then Fstat is unreliable and SError is understated and T stat is overstated.
- Error term is uncorrelated across observations. If it is correlated, then it is called Serial Correlation. This is tested using Durban Watson (DW) test. Here F stat and T stat is too high. DW = 2(1-r)