# Interpreting the results of dynamic panel models using GMM estimation

I am working on my master dissertation and I am struggling in interpreting my results properly. I am looking for a good source (book, paper, or website) that provides explanation on how to interpret the results (coefficients, standard errors, etc.) of dynamic panel models.

I am using Stata to fit a dynamic panel model using GMM estimation methods. My problem is twofold:

First, I don't understand why the following was done, and how the findings can be interpreted: Multiplying the coefficient with the standard deviation of the variable in the sample to see the impact of the variable. What would the result of the above operation tell us?

Second, one of the papers I read using this method (Acemoglu et al. 2008) stated, "Although the pooled OLS estimate of the independent variable is quantitatively small, the two standard error bands of the fixed effects estimators almost exclude it." What does it mean? How do I get this?

Reference:

Acemoglu, D., Simon Johnson, James A. Robinson, and Pierre Yared American Economic Review, 98(3), June 2008: pp. 808-42.

• What is a sophisticated regression? Are you working in a particular statistical software like R, SAS, or SPSS? – Chris C Aug 15 '15 at 20:49
• I am using STATA. I have a dynamic Panel Model and I using GMM estimation methods. For example, Actually I have two problems. First I dont understand why and how to interpret this: you multiply your coefficent with the standard deviation of the variable in the sample.. what does this result tell you ?? The second: One paper stated this: although the pooled OLS estimate of the independent variable is quantitatively small, the two standard error bands of the fixed effects estimators almost exclude it. What does it mean ? How do I get this ? I just know the basic interpretation things x increase – Jana Aug 15 '15 at 20:50
• Perhaps you could edit that information into the question so that people here can help you more efficiently. If you do, could you please quote and cite the passage from the particular paper that you read on your second question. Also, to answer your initial question, UCLA has a staggering amount of information available for free on STATA, I would take some time going through their tutorials, particularly the topics series. – Chris C Aug 15 '15 at 20:59
• The paper I am talking about is Acemoglu et al. (2008) Income and Democracy. I am writing my dissertation about this topic and use system-GMM estimation methods. I read on page 818 what I have written in question 2 and do not understand what it implies. Thanks a lot, Chris – Jana Aug 15 '15 at 21:32
• @Jana, you should give us at least the full citation of the paper, so we don't have to hunt it down. Even better if you have the link to a copy where institutional access is not required. Also, it is Stata, not STATA. – Marquis de Carabas Aug 15 '15 at 21:53