hypotheses testing: inequlity between coefficients

[chiara piccardo]

Hello everyone,
I have a question on hypotheses testing in stata

I run the following Pooled OLS regression (Stata 14.2):

regress Y   L.D1 L.D2 L.D3 L.D4 L.X1 L.X2 i.X4 X5 X6 X7 i.year,noconstant  vce(robust)

where D1, D2, D3 and D4 are dummy variables identifying exclusive combinations of firms' innovative activities; Y is a measure for firm's performance and Xs are control variables.

I need to test the following hypotheses:

1) H0: b4>b1 (where b4 and b1 are the estimated coefficients for L.D4 and L.D1, respectively)

2) H0: b4-b2>b3-b1 (where b4, b2, b3 and b1 are the estimated coefficients for L.D4, L.D2, L.D3 ad L.D1, respectively)

How can I test these hypotheses using STATA?
I thank you all in advance for your help
Chiara

[Maarten Buis]

If you wanted a two sided test, then both could simply be done with the test command. The second hypothesis is just a linear combination of parameters, so test can handle that without problem.

However, you are looking for a one sided test (and strictly speaking the null hypothesis should include an equal sign, so "larger than or equal to" rather than "larger than"). You can find a description of how to turn the two-sided output from test into a one sided test here: https://www.stata.com/support/faqs/s...-coefficients/