Forum: help


RE: Autocorrelation between panel data in SFA [ Reply ] By: Jennifer Beckensteiner on 2019-11-03 19:13 | [forum:47120]![]() |
Thank you Arne for your answer, I have a related question I would like to use a SFA for panel data with inefficiency effects (the EEF model from BC95). While I am using the function sfa() from your frontier package, I am not sure I understand how the function is handling the panel structure. When I am entering the data as a data frame or as a panel data frame (created with pdata.frame), the outputs of the model are exactly the same. The only difference is the calculation of mean efficiency of each year with the panel data frame specification (see code attached for models specification). Therefore, I would like to be sure that the EEF model is taking into account potential correlation between my IDs and YEARs, and is not using my data as cross-sectional observations. I have a feeling that ECF can deal with panel data and, then we can check for a time effect and for adequate distribution of u (truncated or half normal) but EEF is using all observations... I am not sure I understand why it is not possible to have an EEF on panel data with a truncated normal distribution for u, and time effect. Thank you in advance for taking the time to clarify this matter. Best, Jennifer NB: the code I am using is attached |
RE: Autocorrelation between panel data in SFA [ Reply ] By: Arne Henningsen on 2019-11-03 14:10 | [forum:47118] |
Checking for autocorrelation / serial correlation in SFA panel data models has not been implemented yet in the "frontier" package. Checking for autocorrelation / serial correlation in simple linear panel data models is already non-trivial (see, e.g., Croissant and Millo: Panel Data Econometrics with R, Wiley) and probably much more complicated for SFA models. I suggest that you test for autocorrelation / serial correlation in linear panel data models using the "plm" package and then assume that the results would likely be similar for SFA models... |
Autocorrelation between panel data in SFA [ Reply ] By: Jennifer Beckensteiner on 2019-10-29 17:06 | [forum:47092] |
Dear Dr. Arne, I wonder how SFA handles autocorrelation in panel data? Can we check for autocorrelation and endogeneity using the frontier::sfa() function? Thanks for the help. |