Investigating under-invoicing with trade data: Red herring or red flag?

“There is nothing more deceptive than an obvious fact”: Sir A. C. Doyle

The Pakistan Business Council wrote to the Minister of Finance a few days ago, documenting discrepancies in Pakistan’s trade data as probable evidence for under-invoicing of imports. The data was used to put numbers on the scale of the problem and the loss to the exchequer from forgone customs duties and taxes.

Under-invoicing is indeed a serious concern for revenue authorities, and for businesses that compete with under-invoiced products. Anecdotal evidence on this abounds. Electronic Data Interchange agreements with trade partners are also uncontroversial and are likely to improve data accuracy. But neither is the trade data discrepancies been described in the letter as evidence of under-invoicing nor can we estimate the size of under-invoicing from them.

PBC have compared exports to Pakistan reported by partner countries (in this case, Singapore, the UK, Germany and China) with imports from these partner countries reported by Pakistan. Theoretically, both values should be identical as they measure the same flow.

Yet, Pakistan’s average discrepancy across these four partners is around 28% of the reported import value. Is this a Pakistan specific issue? If we leave Pakistan out and do the same exercise for Singapore, with the same set of partners (i.e., Germany, China and the UK but not Pakistan), the discrepancy is around 30% of the import value. China’s (for Germany, the UK and Singapore) is around 45% and Germany’s (for Singapore, the UK and China) is around 14%. It isn’t Pakistan specific.

Perhaps there is something anomalous about this set of countries and we can expect much smaller differences between, say, the European Union country pairs that are part of a common market, share a border, use the same currency, and have little incentive or opportunity to under-invoice. But that is not the case.

In 2022, while France reported importing goods worth USD 84.6 billion from Germany, Germany reported exporting goods worth USD 72.9 billion to France. There were no exact matches of product line and value in 2022 in the ITC Trade Map data, even at the 2-digit level. (The more disaggregated a view you take, i.e., going from HS 2-digit level to 4, 6 and 8-digit level, the larger the discrepancies become. When aggregated, the negative and positive discrepancies cancel each other out to some extent and the differences are, to that extent, masked)

The point is that this is a common phenomenon with trade data. So, what exactly is going on?

Firstly, as mentioned in the PBC letter, imports are recorded with cost of insurance and freight (CIF), while exports are “free on board” (FOB). This can cause a difference of 10-20 per cent. Secondly, there is usually greater accuracy in the measurement of import data (which has revenue repercussions) than in the measurement of export data.

Thirdly, there can be differences in the detailed HS code categorization of a specific good between the two trade partners - that also contributes to the problem mentioned above of discrepancies getting worse as you get into more detailed HS categorizations.

Fourth, differences in the commodity classification system used, or the version of the commodity classification system used, can lead to differences when converting between them. Fifth, the partner country is typically the country where the product originated from, not the country that directly exported the product. Sixth, there may be differences in the exchange rates used when converting to USD.

Seven, there is variation in data quality among different countries, even if procedural instructions are identical, and finally (phew!) there is a higher propensity for missing data for codes at a higher level of disaggregation, which some countries may withhold as confidential (but which may still be included at an aggregated level such as at the HS 2-digit level).

So, while under-invoicing may be one of the culprits behind differences in reported values, it has got quite a bit of company! PBC’s research and letter should be taken as a spark for a deeper discussion on controlling under-invoicing and outright smuggling, and on the data and digitization needed for this. These are important priorities, even if we cannot - for now - put specific numbers on the scale of under-invoicing and potential revenue gain using this trade discrepancy data.

Copyright Business Recorder, 2023