Revenue from VAT is collected gradually through the chain of production and distribution. This structure, or network, of interconnected businesses presents opportunities for misreporting and fraud. This project aims to develop methodologies that could be used to inform tax administrations of transactions which merit investigation.
This project received funding from the Turing-HSBC-ONS Economic Data Science Awards 2018.
Explaining the science
Under VAT, business purchasers offset the VAT paid on their purchases against their liability on the sales. This results in that no revenue is collected from the taxation of intermediate goods, but the revenue is collected gradually through the chain of production and distribution. But this structure of interconnected businesses (and the rules associated with application of VAT) is not problem free.
Problems include traders failing to register for the tax, under-reporting of sales, product misclassifying, and carousel ‘fraud’ (an importer purchases a tax-free good from a trader who then sells the good to another business with VAT added but does not remit the VAT to the tax authority and disappears as a ‘missing trader’). There are significant tax revenues lost due to these transactions therefore developing methods which identify fraudulent behaviour is a top priority for tax authorities.
The objective of this project is to develop methodologies which identify abnormal data events in networks when economic agents transact. The methodologies will be tested on a VAT network of economic transactions. The benefit from this project is that it will inform tax administrations of transactions that merit investigation. The project will use data provided from the Bulgarian National Tax Authority.
There are clear advantages to taxpayers, and the economy, of a tax system that is well designed and efficiently administered. The output of this work will advise tax authorities on how they can enhance their policies and administrative practices in order to combat fraudulent transactions.
Project is due to start on 1 February 2019.
Researchers and collaborators