Dear All, on the same wave some of you helped me yesterday on a similar topic, I need you help again as I'm relatively new to the use of stata.
I need to calculate an index called REVEALED TECHNOLOGICAL ADVANTAGE, somehow similar to RCA (Reveald Comparative Adv). It ranges from 0 to 1+ (it can be above 1, but it usually stays < 5)
Following OECD's definition, it is calculated as a country's share of patents in a particular technology field divided by the country's share in all patent fields.
This is a data example:
since I have data in long format (panel data) and a tremendous number of technological classes I created 1digit technological class as in A,B,C etc.
To calculate the RTA I have tried the following:
where nr_pat = NUMBER OF PATENTS
But I'm not sure that the sum of patents in the first line is correct, can someone help in coding this index?
A paper on US patents use this formula, but it looks quite complex to code for me, any idea is very much appreciated.
Link to paper for those interested: https://www.sciencedirect.com/scienc...791?via%3Dihub
Thanks a lot for any help or idea you may give me!
Dalila
I need to calculate an index called REVEALED TECHNOLOGICAL ADVANTAGE, somehow similar to RCA (Reveald Comparative Adv). It ranges from 0 to 1+ (it can be above 1, but it usually stays < 5)
Following OECD's definition, it is calculated as a country's share of patents in a particular technology field divided by the country's share in all patent fields.
This is a data example:
Code:
* Example generated by -dataex-. For more info, type help dataex clear input int time str37 country str9 ipc str1 tech_class_1dig float nr_pat 1990 "France" "C03B" "C" 12.1833 1990 "France" "G09D" "G" 0 1990 "France" "H03H" "H" 5.0833 1990 "France" "B29L" "B" 2.084 1990 "France" "F24C" "F" 1.9786 1990 "France" "E03C" "E" 0 1990 "France" "C12Q" "C" 6.3845 1990 "France" "C14C" "C" 0 1990 "France" "F02F" "F" 1.6643 1990 "France" "H03F" "H" 8.1095 1990 "France" "C22F" "C" .8333 1990 "France" "B60R" "B" 3.8295 1990 "France" "F25J" "F" 3.425 1990 "France" "B29C" "B" 19.0836 1990 "France" "C08C" "C" 0 1990 "France" "D01H" "D" 2.1429 1990 "France" "B31B" "B" .6 1990 "France" "A44C" "A" 0 1990 "France" "F17D" "F" .9929 1990 "France" "A63J" "A" .4 1990 "France" "F28C" "F" 0 1990 "France" "B05D" "B" 1.7872 1990 "France" "G01L" "G" 4.2833 1990 "France" "F02D" "F" 1.5869 1990 "France" "G01C" "G" 2.3095 end
since I have data in long format (panel data) and a tremendous number of technological classes I created 1digit technological class as in A,B,C etc.
To calculate the RTA I have tried the following:
Code:
bys country time:g tot = sum(nr_pat) bys country tech_class_1dig: gen tp = sum(nr_pat) bys country time : g num = tp/tot summ num bys country : g den = total_patents/tot summ den bys tech_class_1dig: g rta = num/den
But I'm not sure that the sum of patents in the first line is correct, can someone help in coding this index?
A paper on US patents use this formula, but it looks quite complex to code for me, any idea is very much appreciated.
Link to paper for those interested: https://www.sciencedirect.com/scienc...791?via%3Dihub
Thanks a lot for any help or idea you may give me!
Dalila
