Tax Authorities Embrace Big Data – But is it Working?

Just like corporations and individual taxpayers, tax authorities around the world increasingly rely on digital technologies to navigate today’s complex and rapidly evolving tax landscapes. That’s especially true when it comes to tackling noncompliance and fraud, two of the biggest drivers of tax gaps globally. The stakes are high. A European Commission report estimates the E.U.’s 2018 VAT collection shortfall at €140 billion. In the U.S., estimates prepared by the IRS put the gross tax gap for tax years 2008–2010 at $458 billion. 

The response that authorities have typically adopted is to collect more data, faster, in order to enable massive, automated cross-checking of information. It’s basically a big data strategy, one that may pull in details of purchase and sales transactions, financial records and inventory, as well as information on the movement of goods. This approach leverages a variety of methods, including mandating real-time B2B e-invoicing; integrating data from other sources, such as financial institutions; and mandating the use of accounting systems that send information to the authority.

There’s some evidence that efforts like these can yield benefits in terms of increasing revenue and reducing the size of the informal economy. For example, Russia achieved a 38% increase in retail sector VAT collection between 2016 and 2017 when electronic cash registers were introduced, and Mexico has achieved considerable success in bringing micro-businesses into the formal economy through e-invoicing initiatives. But elsewhere, the results have been less encouraging. India has had to backtrack on its efforts to tie out input and output VAT, and VAT gap remains stubbornly high in the E.U.

Plus, in cases where a data-driven approach has worked well, it’s unclear whether those results are due to signaling effects or increased communication with taxpayer communities (as opposed to the initiatives themselves). What’s more, many authorities are starting to find themselves awash with data, which brings its own costs related to storage and cybersecurity. 

So what can tax authorities do to improve their digital efforts and pull more value from large, disparate datasets? I’ll look at two powerful strategies – machine learning and blockchain technology – in another post.