Ppml gravity model-panel data

[Gabriel Kwenda]

Hi everyone,

I am working on the following topic for my master's thesis "IMPACT OF COMESA AND SADC ON AFRICA’S AGRI-FOOD INTRA-TRADE: A CASE STUDY OF MALAWI OIL SEEDS EXPORTS".


My data, in summary, is as below:

1. Data from 2001 to 2019 (The limitation is due to the available data on the CEPII: it ends with the year 2019)
2. Countries compose of all SADC and COMESA members (Including those that are no longer but were members in the 2001 - 2019 time span)
3. the model adopted is the PPML gravity model

I face the following obstacles:

1. After using interpolation to deal with the missing values for the Explanatory variables, there are a few countries that remain with the missing values. This is due to the extent of missing values. For example, Sudan has no data on its population and hence no data to interpolate from.
2. The Oil seeds exports had missing values too. Extracted from UN COMTRADE, I did not find any documentation that explained the difference in missing values from zero values when the data was recorded. Due to the sensitivity of the model i did not interpolate for fear of overestimating the trade potential rather I recoded the missing values as zeroes.
3. Lastly, I need guidance on which tests will be necessary to be checked.

Note:

• the command being used is the "ppmlhdfe

Please, your assistance will highly be appreciated.

Best regards
Gabriel

[Joao Santos Silva]

Dear Gabriel Kwenda,

I am afraid I cannot help you with the data, but if you use the PPML estimator, the only test I would consider is the RESET.

Best wishes,

Joao

[Gabriel Kwenda]

Dear Joao Santos Silva,

Thank you for the guidance.

Best regards
Gabriel