Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Problem with GMM command

From   Tyrone Hopwood <>
To   "" <>
Subject   st: Problem with GMM command
Date   Sun, 15 Dec 2013 22:15:55 -0800 (PST)

Dear Statalisters 

I am using Stata/SE 13 for Windows, revision 17 June 2013.

I want to run the GMM estimator using -GMM- and then Hansen's J statistic using the command -estat overid-. I am having trouble producing the correct Stata command to run the GMM test. Please can you assist me by producing the correct GMM command for my needs below.

My sample data representing daily returns of the DOW and a Tech portfolio is as follows:

Date       std dev*      DOW     Tech portfolio 
02-Jun-09     -2      -0.01263   -0.01834411 
07-Aug-11     -2      -0.01205   -0.01801961 
01-Jan-06     -1      -0.00557   -0.01028249 
23-Apr-09     -1      -0.00559   -0.01334217 
12-Dec-04      1      0.012393   0.056200529 
07-Nov-07      1      0.00902    0.024050865 
15-Aug-05      2      0.02709    0.059483887 
21-Sep-10      2      0.041326   0.055315953 

(*each record in the data table above is captured when the DOW and the TECH portfolio register equivalent std deviations away from their respective means.) 

To give some background, the null hypothesis of my test is:

H0 : p+ - p- = 0

(p.s. In the above null hyp the “+” and “-“ after each rho (p) indicates correlation above a certain std dev threshold. So as an example, in a market rally the DOW and the Tech Portfolio register +2 std deviation moves above their respective means (p+2), we document the returns at that move. The DOW then gets hit by bad news and drops significantly, moving -2 std deviations away from its mean (p-2), as well as the Tech Portfolio, we then also document the respective returns. We then use the null hypothesis (p+2 - p-2 = 0) to determine correlation symmetry .i.e. at +2 and -2 std deviations are the correlations equal? This helps us conclude whether the Tech portfolio is more correlated with the DOW for a +2 or a -2 std deviation move.

I apologize in advance for my detail, but I thought it may help you if I contextualize it a bit.


*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index