✅ 1. Zero Mean

• The expected value of the error term is zero:

✅ 2. Constant Variance (Homoscedasticity)

• The variance of the error term is the same across all observations:

• If not, it leads to heteroscedasticity, which affects standard errors.

✅ 3. No Autocorrelation (Independence)

• The error terms are uncorrelated across observations:

• If this assumption is violated (especially in time series data), it results in autocorrelation.

✅ 4. Normality of Errors

• The error terms are normally distributed:

• Important especially for inference (e.g., t-tests, confidence intervals).

✅ 5. Errors are Uncorrelated with Independent Variables

• Error terms should not be correlated with the predictors:

• Violation of this leads to endogeneity and biased estimators.

✅ 6. Linearity in Parameters

• The relationship between the dependent variable and the parameters (coefficients) must be linear.

✅ 7. Correct Model Specification

• The model must include all relevant variables and exclude irrelevant ones.

• If the model is misspecified (e.g., omitting a key variable), it leads to specification bias.

✅ Summary Table