Dear Parul Bajaj,

Missing data can be dealt with a number of ways. There are certain built-in techniques in software that help to interpolate the data. However, as mentioned by Dr. Hubert Escaith, the use of mirror data can also help in calculating various indices.

Besh Wishes,

Actually, your question is not so clear. LIke when we deal with trade there are many countries appeared with zero trade and you can calculate them by adding constant or in gravity just take the values at level. 

1. What type of Data you have (Panel?)

2. May i see your equation?

Dear Parul,

Missing values in trade data for some countries is a real problem. In this case I would refer to international organizations to fill up the gaps such as CIA world fact book

https://www.cia.gov/library/publications/the-world-factbook/fields/2050.html

or any other international organizations that can provide information.

Otherwise if older data for previous years are available you can use them to fill up the gap or create a prediction equation for the following years.

Best regards

If data is missing, then gap can be filled by a prediction equation. if there is no trade then i think adding a constant would work. First you have to explore that data is missing or trade is zero for that particular year or a country.

Thank you for your responses. Basically I am calculating trade indexes like RCA and RTA for few nation's from UNCTAD data but for some nation's the data is not available like Myanmar and Vietnam for product categories at HS 4 level.. How can I calculate the average to make comparisons in this case. 

Thank you

Good morning.

Missing trade data (or strongly asymmetric data where the value reported as exported from A to B is very different from the value reported by B as imported from A) are a big issue, indeed. 

The easiest solution when dealing with least-developed countries exports is to use mirror data from large importing countries. The advantage is to capture the exports from Special Zones, that are often excluded from national export data when the country adopted the special trade reporting system. But it means more data processing work, as you will need to aggregate data from the main (20 or 30) main importers for each H4 category.

Here, you will have to use mirror data (imports) from UN's COMTRADE (now with open access, as long as you do not require bulk downloads). Import data are usually CIF, so they inflate the exported FOB value, but this is only a small issue compared to facing scant national export data. 

Another (more elaborated) approach is to use secondary database developed by experts who looked already into the missing/asymmetric trade data issue. One of them is CEPII' BACI database:  http://www.cepii.fr/CEPII/en/bdd_modele/presentation.asp?id=1

BACI is free to access as long as you have a COMTRADE access code.

If your objective is to compute RCA, there is no need, in my opinion, to look into the issue of true 0 trade vs. missing data and impute missing data. Unless, obviously, you wish to dig deeper and do some gravitation modelling. But on such a weak statistical basis, I'd stay clear from sophisticated modelling..

Hope it helps,

Best, Hubert

Dear Parul Bajaj,

Missing data can be dealt with a number of ways. There are certain built-in techniques in software that help to interpolate the data. However, as mentioned by Dr. Hubert Escaith, the use of mirror data can also help in calculating various indices.

Besh Wishes,

An addition to my answer:

have a look at the following paper on trade data:

https://mpra.ub.uni-muenchen.de/81572/9/MPRA_paper_81572.pdf