This map and figure use data in the GTA, UNCTAD NTM TRAINS, and WTO Notifications to highlight country and compositional differences. In the case of the WTO notifications, the imputed dataset of Ghodsi et al. (2017) has been used. As a metric we use 6-digit HS line - measure combinations.
Sectoral NTM indexes
As the sectoral composition indicates, the different databases exhibit different profiles of measures across sectors. Francois et al. (2018) combine the information in the different databases using weights generated by a principal component analysis (PCA). By and large, this follows and extends the approach used by Nicoletti, G., Scarpetta, S., Boylaud, O. (2000) for creating OECD PMR indexes.
The schematic view shows how the different indicators in the table below are constructed, which can be summarized as follows:
sep: MAST specific indicators on the basis of within sector variance (not scaled),
--> Highest score (normalized = 1) is country with highest weighted score per MAST chapter within each sector,
Comment: for sector-MAST combinations with only info from one database, original data has been weighted by country contributions to variance only
comb: MAST specific indicators on the basis of across sector variance (not scaled),
--> Highest score (normalized = 1) is sector-country combination with highest weighted score per MAST chapter,
Comment: count data has been divided by the number of HS2016 6-digit lines per HS sector before running PCA
sepall: Overall and PCA factor indicators on the basis of within sector variance (scaled),
--> Highest score (normalized = 1) is country with highest weighted score weighted score for all MAST chapters combined within each sector
coall: Overall and PCA factor indicators on the basis of across sector variance (scaled),
--> Highest score (normalized = 1) is sector-country combination with highest weighted score for all MAST chapters combined,
Comment: count data has been divided by the number of HS2016 6-digit lines per HS sector before running PCA
The complete dataset including sub-indexes and non-normalized scores can be downloaded here.
PRONTO has received funding from the European Union’s Seventh Framework Program for research, technological development and demonstration under grant agreement no 613504.
The schematic view shows how the different indicators in the table below are constructed, which can be summarized as follows: