International wildlife trade can represent a major threat to biodiversity conservation. Annually, billions
of plants, animals and their products are traded across international borders, with legal trade alone
estimated to be worth over 320 billion USD per annum (TRAFFIC 2009). CITES, the Convention on
International Trade in Endangered Species of Wild Fauna and Flora, regulates and monitors trade in
~35,000 species. CITES has 183 signatory countries (‘Parties’), all of which must provide annual
reports detailing their international trade in CITES-listed species, culminating in more than 18 million
trade records. This wealth of data, reported from 1975 to-date, is maintained in a central, freelyaccessible
database, the ‘CITES Trade Database’ (trade.cites.org), managed by the UN Environment
World Conservation Monitoring Centre (UNEP-WCMC) on behalf of the CITES Secretariat. In recent
years, many scientists have utilized this database to try to understand the wildlife trade, and its
implications for conservation of threatened species, resulting in at least 114 peer-reviewed publications (Supporting Information). However, given the vast and international nature of the dataset,
properly interpreting the data is highly complex, and incorrect interpretation can lead to erroneous
conclusions. This is of conservation relevance, particularly as such studies may form the basis for
management decisions and recommendations. The ‘Guide to Using the CITES Trade Database’
(UNEP-WCMC 2013) provides technical instruction on utilizing the database. However, here we
discuss major challenges of analyzing and interpreting CITES trade data, highlight common areas of
confusion arising in the scientific literature, and provide guidance on how these can be avoided.

General analyses of CITES trade data

Various studies have sought to understand trade dynamics of CITES-listed species (e.g. Carpenter et
al. 2014; Li & Jiang 2014). However, numerous factors affect trade dynamics, including (amongst
others); countries joining CITES at different times (contributing data from different periods),
new/amended CITES listings, taxonomic changes and national and international regulatory
interventions (e.g. quotas, suspensions). Therefore taking the trade data at face value can sometimes
be misleading. Figure 1 (a) illustrates a sharp increase in reptiles imported to the EU since 2006.
However, closer interpretation reveals that this is caused by the inclusion on Appendix III of a
previously unlisted genus, whose trade prior to this time was simply not reported because it was not
CITES-listed.

Whilst some factors affecting trade can be cross-checked utilizing resources including the CITES
website and the Species+ database (www.speciesplus.net), it can often be complex, or there is simply
insufficient information, to reliably identify specific drivers of trends. This emphasises the need for
careful interpretation of the CITES data. Below we highlight four key areas we feel require further
consideration in future studies.

1. Importer vs exporter reported figures

Countries provide data for both their imports and exports, resulting in two data sources for any data
query; 1) that reported by importing countries, and 2) that reported by exporting countries. For several
reasons, these do not always match. Whilst data in annual reports should be based on trade that
actually took place (Annex to CITES Notification 2017/007), sometimes countries report data based on permits issued. As the quantities actually traded may be lower than those permitted, this can result
in lower importer-reported trade levels. On this basis, some studies have analyzed data reported
predominantly by importing countries (e.g. Rhyne et al. 2012). However, countries are not required
under CITES to issue import permits for Appendix II species (although several do as part of stricter
domestic measures), and therefore imports of Appendix II species will not always be reported. Figure
1(b) illustrates how under-reporting of imports can lead to substantially higher trade levels reported by
exporters than by importers. Where it is important to understand quantities that export countries have
authorized, or when the completeness and accuracy of reporting is considered higher by exporters, it
may be more appropriate to use exporter-reported data.

Other reasons for inconsistencies in import and export figures include different use of trade ‘terms’
such as ‘source’, ‘purpose’, ‘unit’ etc. (see Table 1 and UNEP-WCMC 2013), whereby different Parties
apply the ‘terms’ differently. Additionally, export permits issued at the end of the year may not be used
(and not reported) by importing countries until the following year, leading to discrepancies between
years. Finally, data reported by both importers and exporters may be subject to reporting errors, and
therefore neither indicates actual minimum or maximum numbers in trade as indicated in Fialho et al.
2016 and Foster et al. 2014. Where possible, both should be considered, as the existence of
discrepancies is in itself information that may reveal aspects of interest.

2. ‘Comparative’ vs ‘Gross’/‘Net’ trade reports

Trade data can be downloaded as ‘comparative’ or ‘gross/net’ trade reports, the choice of which
requires careful consideration. In general, ‘comparative’ reports provide the most comprehensive
picture of the trade, as they present imports, exports, re-exports, and all trade ‘terms’, as reported by
both importing and exporting countries, allowing side-by-side comparison.

‘Gross’ and ‘net’ reports provide more simplified data summaries, excluding information on source,
purpose and country of origin, but can overestimate trade. Gross reports combine both exports and
re-exports of traded specimens, thereby often double counting individuals. Re-exports can be
identified by their country of origin, which will differ from the country of export. In the database, direct
trade can be isolated by selecting data with no origin country listed indicating that the exporting
country is its origin. Rivalan et al. (2007) use gross reports to assess the effect of trade bans, but as this includes re-exports and details are not provided regarding the source of specimens (e.g. wild or
captive-bred), the analysis may not provide an accurate picture of the trade or of shifts in reported
sources that may occur following trade bans. Instead, for research questions aiming to consider actual
numbers in trade, ‘net’ reports account for double counting by reporting the difference between gross
exports and gross imports. However, where discrepancies exist between importers and exporters,
gross and net reports both take the larger value, and therefore net reports can also inflate trade
figures, albeit by a lesser amount.

Regardless of the report type, a misconception is that each row in the data represents an individual
shipment/trade transaction (e.g. as in D’Cruze & Macdonald 2015), which is not the case. All data for
the year, concerning the same taxa, exporter, importer and trade terms are aggregated into one row
(see UNEP-WCMC 2013).

3. Terms and units

CITES regulates trade in whole animals and plants as well as their parts and derivatives. The different
commodities in trade are defined by standardised ‘terms’ (Table 1), which can in turn be reported in
different ‘units’ (e.g. number of specimens, kilograms), and many of these terms and units cannot be
meaningfully summed in one analysis. For example, Jiang et al. (2013) summarise China’s trade in
Ptyas mucosus as ‘number of pieces/specimens’, but the figures presented combine ‘terms’ such as
live animals with parts such as leather products. Additionally, Mieres & Fitzgerald (2006) provide
figures of Tupinambis trade that result from the addition of number of skins with kg and cm of skins.
Units should not be combined unless they can be directly converted (e.g. grams to kilograms) or
conversion rates are available (e.g. kilograms to cubic meters based on known density) (e.g. ArroyoQuiroz
et al. 2007).

4. Purpose codes

Confusion has arisen in the literature over the use of purpose codes, which can be particularly
challenging. For example, trade for ‘personal’ purposes is often used for non-commercial movement
of pets (e.g. holidays, emigration), but this excludes large numbers traded for the pet trade, which are
traded as ‘commercial’. Indeed, Harrington (2015) demonstrates that ten times the number of carnivores and primates are traded as ‘commercial’ than ‘personal’, and Robinson et al. (2015) report
that 99.2% of Appendix II live reptiles are traded as ‘commercial’. Consequently, in seeking to
understand global trade in exotic pets, the data in Bush et al. (2014) are only representative of a tiny
fraction of the trade because only transactions traded as ‘personal’ were used. In addition, CITES
‘purpose’ categories are used for multiple types of trade and can overlap. So, whilst ‘commercial’
encompasses specimens traded for the commercial pet trade, specimens supplying research facilities
or zoos, for example, may also be traded as ‘commercial’ by some countries, even though alternative
codes can also be used (e.g. ‘scientific’ and ‘zoo’, respectively).

The use of ‘purpose’ ‘hunting trophy’, often used in combination with a variety of ‘terms’ (e.g. body,
skin, skull), can also result in confusion because a specific ‘trophy’ ‘term’ also exists, which is
sometimes used in combination with ‘purposes’ different from ‘hunting trophy’. Consequently, an
analysis that only uses the ‘term’ ‘trophy’ but discounts other ‘terms’ with ‘purpose’ ‘hunting trophy’
(e.g. Di Minin et al. 2016) may underestimate trade. In addition, different parts of an animal (e.g. skin,
skull, feet and horns) are sometimes reported as separate trophy items, which if incorrectly interpreted
can lead to overestimates of the number of animals.

Conclusion

The CITES Trade Database provides a powerful tool for understanding the wildlife trade for listed
species, but it must be carefully analyzed and interpreted. Additionally, as with all data sources, there
are limitations relating to the use of CITES trade data, and therefore we recommend users refer to the
database guidelines and wider literature for further discussion (e.g Thomas et al. 2006; Herrel & van
der Meijden, 2014; Lopes et al. 2017). Notwithstanding, the database provides an unparalleled tool for
monitoring trade in wildlife and wildlife products across borders, and with more than one million
records added annually in recent years, it represents an invaluable resource with enormous potential
for understanding global wildlife trade.

Read full report: robinson & sinovas