Apple released the Macintosh Portable in September 1989, one of the first mainstream laptop computers to hit the market. The device featured 1MB of memory, a 10-inch black and white screen, and weighed 16 pounds. Compare that to today’s MacBook that you can purchase with 16GB of memory, a 12-inch retina display and a weight of two pounds — clearly 1989 was a long time ago, especially when it comes to technology. That year was also three after the federal government last made major changes to the tax code.

With the passing of the Tax Cuts and Jobs Act in December 2017, tax auditors now have a gargantuan challenge in front of them. The new law requires them to catch up on 30-plus years of changes in a little more than 30 days as the reform takes immediate effect. At the same time, fraudsters are actively looking for gaps to exploit the system for personal gain.

Luckily for auditors, though, they have a 30-year improvement in technology to help them along the way.

Areas of concern

Analytics have been used effectively to help tax auditors detect fraud. The new tax bill includes a number of key areas that could serve as potential areas for deception. While new areas will certainly emerge — fraudsters are always hard at work — there are a few key types of fraud that tax auditors should immediately watch.

Deductions and exemption. With the elimination of state and local tax deductibility, auditors should keep a closer eye on the itemized deductions of higher earners who may seek to make up the difference. A type of unsupervised machine learning, called clustering, can automatically form groups of “like” taxpayers and apply different business rules and predictive models for each group, spotting larger tax evasion trends.

Alternative Minimum Tax. Changes to the Alternative Minimum Tax (AMT) — including increases in exemption amounts and raising the phaseout thresholds for these exemptions — which affect many taxpayers in the upper class and upper-middle class, may or may not be enough to balance out the loss of several itemized deductions. However, it is likely these taxpayers will be less incented to fudge their numbers as their tax bill goes down. Taxpayers who illegally evade taxes often do so to the degree they feel they are being taxed unfairly by the system and under-report income or overstate deductions.

Offshoring of income. The new tax structure imposes a 15.5 percent one-off tax on offshore cash, coupled with an 8 percent levy on less liquid assets. It is safe to say there is significant overlap between individuals likely to offshore their assets and those who would be subject to the estate tax upon their death. Is it possible that the elimination of the estate tax will provide increased incentive towards transparency? One of the ways this can be observed is through the use of social link analysis technology to visualize and track how these shell companies, bank accounts, and other “creative” investments such as real estate are linked back to the taxpayer in question. Analytics technology has been used in this regard for many years to detect and combat Value Added Tax carousel fraud in several European countries.

Consolidation of tax brackets. The new law consolidates the tax bracket levels from seven to three, meaning that the income range in each bracket expands. Individuals will potentially attempt to manipulate the bracket in which they fall, pursuing fraudulent practices to either drive up — or scale down — their taxable income.

Harnessing analytics

These areas are merely starting points for tax auditors. The reality is that the new tax code will bring with it a host of scams and schemes of bad actors trying to beat the system. The good news is that tax auditors have more tools at their disposal than ever before. From an analytics perspective, this means that auditors and tax agencies themselves can and should consider the following:

  • Accurate detection. A combination of advanced analytics, artificial intelligence and supervised/unsupervised machine learning with the traditional detection methods auditors are accustomed to can help uncover suspicious events. By integrating capabilities such as graph analytics, social network analysis, anomaly detection and text analytics, auditors can have the most comprehensive view of their tax filing systems.
  • Understand risk. This integrated view across all systems will enable a more enhanced risk profile through outlier detection, while revealing hidden relationships and suspicious associations among tax filers, accounts and other entities across the filing spectrum.
  • Adapt to new schemes. Artificial intelligence techniques, including adaptive machine learning methods and unsupervised intelligent agents, will help auditors identify emerging threats and potential new fraud schemes, automatically suggesting new rules and scenarios in real-time, and further reducing risk.

Analytics today can provide in-depth analysis of large amounts of data, helping identify larger trends that humans might miss. The world of tax auditors has changed quickly. In order to keep up with this changing paradigm auditors need to alter how they approach tax fraud. With a strong analytics program and an emphasis on data-driven analysis, these front-line tax professionals can prevent fraud, thwart criminals, and ensure tax money supports the greater good.

Deborah Pianko is a government fraud solutions architect at SAS Security Intelligence Practice.