Customs data are purchased from commercial vendors, though the ultimate source are the customs authorities of each exporting country.

There are no ‘gold-standard’ datasets with which to compare the volumes reported in customs data.

Here we compare our cocoa customs data for 2018 against comtrade and ICCO statistics, as a qualitative assessment of data quality.

ICCO statistics

ICCO are the International Cocoa Organization, a body with a mandate to work towards a sustainable world cocoa economy - they also publish some cocoa statistics.

ICCO export data were available for the 2017-18 season. The season varies between countries, but in West Africa (the main origin of cocoa), approximately runs from September - September, and is compared with our 2018 customs data (2017 in the case of Brazil).

COMTRADE

Comtrade is the UN’s source of international trade statistics, in theory submitted by customs authorities.

These data are incomplete, however, including no data for the following product/countries:

## # A tibble: 4 x 4
##   COUNTRY_OF_ORIGIN YEAR  TYPE         PRODUCT_TONS
##   <chr>             <chr> <chr>               <dbl>
## 1 COTE D'IVOIRE     2018  COCOA BEANS            NA
## 2 GHANA             2018  COCOA BUTTER           NA
## 3 PERU              2018  COCOA BEANS            NA
## 4 PERU              2018  COCOA POWDER           NA

They also do not always match ICCO statistics.

Compare volumes

Volumes reported per product/country by different sources. Missing data are marked by a *.

Volumes reported per product/country by different sources. Missing data are marked by a *.

The following contexts have more than 20% difference vs both COMTRADE and ICCO statistics:

TYPE COUNTRY_OF_ORIGIN MT_20_PC_OUT
COCOA BUTTER COTE D’IVOIRE TRUE
COCOA PASTE COTE D’IVOIRE TRUE
COCOA PASTE PERU TRUE
COCOA POWDER COTE D’IVOIRE TRUE

These cases make up only 7.21 % (by weight) of trade in our sample.

Value to the left of the line are where our customs data have values lower than the alternative source; vice-versa for data to the right of the line. Points in red have more than 20% difference for both sources, blue have less than 20% difference for at least one source. Excludes the outlier for Cote D'Ivoire cocoa powder > 400% difference.

Value to the left of the line are where our customs data have values lower than the alternative source; vice-versa for data to the right of the line. Points in red have more than 20% difference for both sources, blue have less than 20% difference for at least one source. Excludes the outlier for Cote D’Ivoire cocoa powder > 400% difference.