For all the time and effort we spend predicting the future in our personal and professional lives, doing so is becoming a significant focus for many tax departments in business and government. In this installment of Tax Technology Corner, we try to get a better handle on the direction and impact of predictive capabilities for tax. Our senior editor, Michael Levin-Epstein, interviews Amit Ringshia, a principal in KPMG’s tax ignition practice and a recognized expert in technology, about predictive analytics for tax.
Levin-Epstein: We’re going to be talking today about predictive analytics. What’s your definition of “predictive analytics”?
Ringshia: Predictive analytics is a spectrum of advanced statistical pattern recognition methods and technologies that attempt to make predictions about future outcomes based on the analysis of a broad variety of data, including current and historical facts. To understand modern predictive analytics, it is helpful to reflect on traditional efforts at modeling outcomes. Traditional analytical models tend to be rules-based and require clean, structured data sets. The resulting solution is often inflexible, requires quite a bit of data prep, and is slow to generate predictions in new areas. Predictive models start with the basic capabilities of analytic models, but add a variety of new capabilities, including models mimicking human judgment and experience as factors in decision-making and the ability to use much larger volumes of somewhat messy data. And with this power comes greater speed, flexibility, and insights. The possibilities for the tax function are just beginning.
Levin-Epstein: That was an excellent description. How are tax departments using predictive analytics today?
Ringshia: Predictive techniques can be widely applied in complex areas within and around tax that generally [have] required human knowledge and judgment but suffer from lack of adequate human time or interest for processing, [for example], evaluating large volumes of transactional data. One such area is financial statement planning and forecasting. By applying key market indicators and likely business decisions to historical business data and outcomes, predictive techniques can provide a near real-time understanding of a company’s likely financial path. This capability is allowing businesses to more accurately forecast revenue and earnings targets. It is also creating a path for several tax enhancements: speed and accuracy of tax provisions, reduction in true-ups needed, and significant enhancements to transfer pricing and compliance activities. This is a rapidly evolving area. We expect to see more integrated predictive models that project the possible tax consequences of business events like new business ventures, tax reform, environmental, social, and governance (ESG), or mergers and acquisitions. There’s a strong appetite for companies to leverage their tax data in a forward-looking and meaningful way. In a recent KPMG study [see sidebar], we gleaned some insights on how companies are leveraging their tax data. Interestingly, many execs are not using tax data in a way that could help them align with ESG priorities or help them make better decisions when it comes to M&A. We’re also beginning to see predictive models aimed at deriving more tax value from existing business operations, such as supply chain analyses that can lower tax duties and enable companies to obtain quicker refunds. Other tax areas such as indirect tax and global mobility and transfer pricing policy can all benefit from predictive analysis and forecasting to improve business outcomes over traditional techniques.
Predictive Tools & Technologies
Levin-Epstein: What are some common predictive technologies and tools?
Ringshia: Predictive analytics are not new. Predictive concepts like machine learning, data mining, predictive modeling, and probabilistic reasoning have been around for quite a few years. However, the underlying technologies that enable predictive analytics today have changed significantly. Unlike the behemoth systems of yesterday that were designed for limited purposes and could only be used by mathematicians and early computer scientists, modern predictive analytics are enabled by vast cloud computing power and accessible through intuitive, user-friendly interfaces designed for the skills of the modern business professional. There is a broad range of tools available for predictive analytics. Python is an example of one of the more prominent programming languages that are being deployed for a variety of general applications. All the major cloud service providers ([like] Microsoft, AWS, Google) have developed large suites of user-friendly tools. Even more traditional tools like Excel and Alteryx that many tax professionals already use include a growing list of predictive analytics capabilities.
Levin-Epstein: How can tax professionals encourage their companies to make better use of predictive analytics?
Ringshia: In general, as tax professionals we’re already very attuned to data, numbers, and statistical thinking. However, companies can increase their predictive analytics capabilities through training and talent acquisition. A good way to begin is through just-in-time training courses. Professionals can enhance the value of their subject matter expertise by augmenting it with a forward-looking data capability. There is a broad variety of quality training readily available online, and many organizations have internal nontax training programs that can be accessed for these purposes as well. The benefit of this approach is to put the capability of the technology in the hands of professionals with the knowledge to ask relevant questions and the skill to assess the value of the answers. This real-time approach provides just enough technology ability at a time when it is needed. Another approach to accessing these capabilities is through new talent acquisition with a more data-focused talent profile. In addition to maintaining traditional tax expertise, companies are beginning to augment their talent pool with data science, basic programming, and business analyst skills. Having convergent skills in tax and data science domains can bring new ways of thinking to tax, including effective ways to apply predictive technologies. In that KPMG survey I referenced earlier, we found significant shifts in talent acquisition profiles. C-suite executives now value coding as a sought-after skill—more so than having an accounting background or knowing Excel. Finally, consider external parties or service providers who focus on these areas in the tax space and can help you accelerate the learning and adoption of these practices within your organizations.
Levin-Epstein: That’s a good way to end it, Amit.
Sidebar: KPMG Survey
Tax Reimagined 2022: Perspectives From the C-Suite
Here are the data findings from the KPMG survey of 300 C-level executives at companies with more than $1 billion in revenue.
Coding ranks as the tax skill in greatest demand
Among these executives, 46% said coding is the most important skill needed in the tax function to ensure it operates at its full potential, an 18-point gain over last year’s survey results. Coding ranked higher than an accounting background (43%), data analysis skills (42%), mastery of Excel (42%), and strategic thinking (27%).
Leveraging the predictive power of data
When asked how their organizations currently use tax data, 51% of respondents said they do not use it to align their companies with environmental, social, and governance (ESG) priorities but are interested in doing so. Similarly, 49% said they do not use tax data to inform overall business strategy but would like to, and 47% do not leverage tax data to help decision-making on M&A deals but would like to.