Proper Format of Data for Gravity Model

[Samira Abunemeh]

I am a novice with Stata and I am trying to organize my data for a gravity model of trade. I got separate export and import trade data from the OECD with a column for country and then each column for each month ( Jan 1970, Feb 1970, March 1970..etc). I am wondering if this format is proper to do a regression of the gravity model. I also adding distance and terror data to these datasets. I am wanting to see how terrorism affects imports and exports. Here is an excerpt example of how an export dataset looks in Excel:

country a	Country b	Jan-1970	Feb-1970	Mar-1970	Apr-1970	May-1970	Jun-1970	Jul-1970	Aug-1970	Sep-1970	Oct-1970	Nov-1970	Dec-1970	Jan-1971	Feb-1971	Mar-1971	Apr-1971
afghanistan	Australia	..	..	1000	..	..	9000	..	..	..	..	..	..	1000	..	..	..
afghanistan	Austria	35000	10000	17000	14000	7000	23000	26000	31000	13000	82000	50000	48000	22000	30000	14000	19000
afghanistan	Belgium	..	..	..	..	..	..	..	..	..	..	..	..	..	..	..	..
afghanistan	Canada	12000	17000	2000	5000	55000	11000	28000	19000	105000	18000	46000	77000	693000	2000	38000	5000
afghanistan	Chile	..	..	..	..	..	..	..	..	..	..	..	..	..	..	..	..
afghanistan	Czech Republic	..	..	..	..	..	..	..	..	..	..	..	..	..	..	..	..

I'm sorry, but I don't know how to show it in the Stata Command.

If this is not the correct format, can you please explain to me how I can put this data into the correct format? Please explain as plainly as possible because I am new to Stata.

[Charlie Joyez]

Samira,
Here your data is displayed in "wide" format, which prevents you to run the proper analysis you want in Stata (notably because it prevents you to declare panel data and to run a regression on all country-year observations simultaneously).
Consider reshaping your data into long format, with the reshape long command (type help reshape, and see the many examples of reshaping issues in this forum, via the Search tool)

           long
        +------------+                  wide
        | i  j  stub |                 +----------------+
        |------------|                 | i  stub1 stub2 |
        | 1  1   4.1 |     reshape     |----------------|
        | 1  2   4.5 |   <--------->   | 1    4.1   4.5 |
        | 2  1   3.3 |                 | 2    3.3   3.0 |
        | 2  2   3.0 |                 +----------------+
        +------------+

An alternative to reshape, is to re-extract the OECD data, and going to the extracting option to specify that you want a long format (although the "long" term isn't used in OECD websites).

Then, to perform a real gravity model for trade, you'll have to add at least information on the country size (GDP) and the bilateral distance (you could find them on the CEPII website, in the dataset called gravity), and the terror data you want to investigate.
You'll also have to declare your data to be panel data

xtset country year

Then you could run your gravity model.

I hope this helps,
Charlie

Ps: As asked per FAQ, please use dataex to report your data (available from SSC); Also, you might think that gravity model is a trivial notion everybody knows, but all Stata users are not international economists. So in order to increase the probabilities of answers, don't use such specific terms in your topic name.