Intensive and Extensive Margins (Hummels and Klenow, 2005)

[Juliana Sales]

I would like to discuss some questions about the commands of Intensive and Extensive Margin (Hummels and Klenow, 2005)in Stata.

In my database I have the value of exports (tradevalue) of 5000 varieties of products (commodities) for 15 years (year), the five BRICS countries (exporter) to 59 importing countries (importer). So I wonder if the correct way to calculate the extensive and intensive margins would be:

hummels tradevalue, j (exporter) m (importer) i (commodities)

or

hummels tradevalue, j (exporter) m (importer) i (commodities) by (year) outfile

The estimated values of the margins should be between 0 and 1?


Best regards,

Juliana de Sales Silva

[Clyde Schechter]

Please read the FAQ for advice on how to post questions in a way that maximizes the number of people who will respond and the number of people who can learn from them. Specifically, you are asked to provide complete references. Hummels and Klenow, 2005 may be folklore in your shop, but most of us on this interdisciplinary, international forum have no idea what you are referring to here and you give us insufficient information to find out. Also, -hummels- is not part of official Stata. It is fine to post questions about user-written commands, but you are asked to identify the command as being user-written and also indicate where it comes from.

Thank you.

[Anabia Maham]

I guess, second one is the right command because we need to mention the time as well according to the given coding at Stata.

[Anabia Maham]

You can confirm it by typing "Search hummel" in the command window of Stata.

[Soeren Oestervig]

Dear Juliana.
I have recently faced the same problem so even though this answer is a bit delayed, others might benefit from it.
Yes, according to Hummels and Klenow the margins are indices between 0 and 1.
Sadly, the 'hummels' code is not correctly computing the margins. I suggest reading the H&K paper carefully to discover why, but the short answer is: it assumes that the margins are the ratio between value of the exports from country i to country j compared to the value of world exports to j minus the exports from country i. For computing the margins correctly, the world exports must include all countries. Also country i. I realize that the full explanation of the margins are more complicated, but as i don't know your data i cannot provide a single sequence of code for computing them)
To compute the margins, i suggest using the egen var1=group(var2 var3) to match your criteria to relevant margins, and then using total( ) for summing the relevant flows based on your defined groups.
best
Søren

[Nivedita Mullick]

Hi!
I am in search of Stata commands that can help in calculating the extensive and intensive margins as doing it manually without any software is a very tedious task. Could someone please help out.

Thanks in advance!

[Mihai Paraschiv]

In case additional details on the HK methodology are needed, the replication files for HK (2005) are available at https://www.openicpsr.org/openicpsr/.../aerfinalcodes