Question

3.为了观察100家公司各自的投资y和利润x，已知y =α+βx+ e，并建议通过OLS估算α和β。

一种。假设样本中的每个公司都有相同的利润。这会产生什么问题？

b. If the distribution of profits over firms were not normal, we would not be able to apply the CLR model. True, false, or uncertain? Explain.

c. If the conditional variance of investment (given profits) were not the same for all firms, we would unable to rely on the CLR model to justify our estimates. True, false, or uncertain? Explain.

Answer 1

Solution:

a. Assuming here the term CLR, referred as Classical Linnear Regression. We wont be able to apply the CLR. So its a TRUE statement.

The basic assumtion of CLR model is the variable is normally distributed. We may sucecced to run the CLR in the package but it will give biased coefficient estimates and not reflect the actual variation in the Model.

b. If the conditional variation is different then we can not rely on CLR model. It is TRUE statement.

Due to variantion in conditional variance we will have a problem of Heterscedasticity. Which will lead inconstient estimates.