Interview Question: An Insufficient Test

Sample Question #177 (econometrics – panel data)

Explain the Breusch-Pagan Lagrange multiplier test for the random effects model. What’s the null hypothesis? How does the test work? (Can you write out the test statistic?) What’s its major defect?

Advertisements
This entry was posted in Sample Qs. Bookmark the permalink.

One Response to Interview Question: An Insufficient Test

  1. Brett says:

    ANSWER
     
    The null hypothesis is that the variance of the random-effect term is 0; in other words, the null specifies a regular OLS model with a constant intercept. The test is based on OLS residuals. Under the null, the LM test statistic is distributed as chi-squared(1). The test’s major defect is if the null is rejected, we still cannot claim the model is random effects because it’s still possible a fixed effects specification is correct. The test is appropriate only when we know fixed effect is inappropriate to begin with, so our only two options were between random effects and OLS.
     
    (Bonus question: how do you test for fixed vs. random effects?)
     

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s