Data and Analytics for Tax Agency Transformation

By Shaun Barry, Deborah Pianko, Robert Morison, May 01, 2021

Download the PDF

Discussion Overview

Tax agencies manage the essential flow of revenue to governments. That core mission has expanded even as their resources remain constrained. And they face the challenges of any major financial services organization – serve customers, operate efficiently, prevent fraud, and adjust to changing economic and social conditions. To explore how data and analytics can assist agencies in meeting those challenges and forwarding their missions, IIA spoke with Shaun Barry, Director of Government Fraud & Security Intelligence Solutions at SAS Institute, and Deborah Pianko, Principal, Government Fraud & Security Intelligence Solutions at SAS Institute.

What are the biggest challenges that tax agencies and their leaders face today?

Shaun Barry (SB): For starters, globalization. Transactions by enterprise and individual taxpayers routinely occur all around the world, and that creates a big problem for governments and tax agencies. It’s hard for them to track economic activity globally, while it’s easy for taxpayers to locate transactions in places where there are low or no tax liabilities.

The increasingly global economy is also increasingly a digital economy. For tax agencies, that plays out in two ways. Tax avoiders can move money quickly and disguise its trails. At the same time, there are rising expectations among taxpayers that services be online. They get instant responses from Amazon and Google, and they expect similar speed and service in all their online interactions, including with government agencies. Of course, digital also represents a big opportunity for agencies to modernize.

A third challenge is that people’s propensity to comply with tax laws is declining. In many places, it’s more socially acceptable to cheat than it has ever been.

And finally, the missions of tax agencies continue to expand. It’s about a lot more than collecting taxes. They also help control and even disburse social benefits. In the U.S., the earned income tax credit and economic and trade development incentives are built into the tax code. Now agencies are tasked with dispersing pandemic relief funds to individuals, families, and businesses. These role expansions make it more complex to manage the mission of a tax agency.

Deborah Pianko (DP): At an operational level, one of the biggest challenges is having to do more with less, including performing more of those additional roles. No tax agency seems to have the resources it had 10 or 20 years ago. Over the last decade, the funding and staffing of the IRS have each dropped by about 20 percent. Some 95 percent of federal government revenues come in through the IRS, yet the agency is increasingly hampered. It can audit less than 1 percent of taxpayers each year, and that percentage just gets lower.

The way to do more with less is by digitizing effectively. But even with all the computerization in tax agencies, there are still information silos and insufficient coordination between departments. Returns processing still doesn’t know what customer service is doing, and customer service doesn’t know what compliance is doing. You can digitize locally, but if there’s a lack of shared data as connective tissue, you’re not really going to make operational and productivity improvements. So, data and process fragmentation remain major challenges.

At the same time, we can see how integrated digital platforms, with good data and innovative analytics, are transforming tax agency performance. In a major European country, the taxing authority was charged with issuing COVID-19 payroll protection payments to employers, while monitoring to prevent payments from being fraudulently obtained. It took only three weeks to deploy a new fraud detection program, and over the course of 2020 they had protected the equivalent of hundreds of billions of dollars of payment activity.

Please say more about the roles of data and analytics in addressing these broad challenges.

SB: I think of data and analytics as the great equalizers. If the digital economy has made it easier for taxpayers to move money quickly to low-tax jurisdictions, and made it harder to catch them, and willingness to comply is declining, then tax agencies need ways to level the playing field. Data and analytics are the ways. If you think about it, governments are the most prolific collectors of data in the world – data on economic activity, demographics, consumer behavior. And within government, tax agencies lead the way as prolific data collectors. That enormous amount of data creates opportunity to keep pace in what has become a digital arms race in tax evasion and compliance.

DP: I see just as much opportunity on the taxpayer service side. At the end of the day, a tax agency’s role is to collect what is owed and make that process as painless as possible for the taxpayer. Those are the two things that you’ve got to get right. Digitizing taxpayer service is largely about using data and analytics to enable taxpayer self-service. Enable them to find information, get guidance, and perform transactions. Not everyone is tech-savvy, and some taxpayers still need a lot of handholding. But agencies should give taxpayers more credit for their digital ability when thinking about how to serve them. A basic way to do more with less is to have the customer do more – as long as it’s secure and productive for both parties.

SB: The key to taxpayer service is to understand your taxpayer. Highly digitized enterprises like Amazon do a great job of understanding customers and their behaviors. Tax agencies have the ability to do something equivalent and then customize interactions, including self-service experiences, for taxpayers. In the old way of thinking, tax agencies have always said, “We need to treat everybody the same.” And that has great appeal in our society. But it may not make sense. Instead of treating everybody equally, you want to treat people who have equal circumstances equally. A taxpayer filing late for the first time in 40 years is very different from one who is filing late for the tenth year in a row. Agencies can use data and analytics to treat people of different behaviors and different circumstances differently and more effectively, in essence to mass customize the taxpayer experience.

What are some of today’s interesting and innovative use cases where agencies can deploy better data and analytics?

DP: First let’s acknowledge that many agencies are using statistical data analysis in fraud and identity theft detection on one hand and prioritizing audits and collections on the other. But there are so many more opportunities. For example, tax agencies in some European countries are conducting VAT reporting at point-of-sale. In most places, the taxes are totaled, reported, and paid by businesses monthly, and it’s largely an honor system. Instead, sales can be reported by cash registers, individually and instantly, to the tax agency. Along with data about the sale, the VAT could automatically be paid in real time. This process provides honest and simplified reporting for the complying business. And the tax agency can analyze patterns in that streaming data for revenue forecasting as well as anomaly and fraud detection, things they try to do today retrospectively after month end.

SB: Here’s a specific example of customizing taxpayer interactions – predictive call centers. If I call a tax agency, it can often identify me and my basic circumstances from my phone number. Based on my recent activities, including online, it can anticipate my need and automatically route my call to a tax agent best able to help me. Maybe it’s about a late payment or a new tax credit that I claimed for the first time. It’s especially useful to know what a possibly frustrated taxpayer was attempting to do online prior to calling the service center. This kind of responsiveness is becoming common in private-sector enterprises. It requires knowing the taxpayer to begin with, then analyzing the data sources instantly.

DP: I mentioned prioritizing audits. There’s also a lot of untapped value in using sophisticated analytics in the audit process itself. For example, imagine that a large telco is undergoing its regularly scheduled sales tax audit. There are a variety of sales tax lines on a typical customer’s bill, and calculating and accumulating the right amount of sales tax gets complicated. Part of the auditor’s job is to determine whether or not that telco has calculated and assessed the correct sales tax for each of its customers’ invoices. This is made more complex when there are multiple sales tax rates in play, such as state, county and local. The amount of data obtained by the tax agency from the telco in order to facilitate this evaluation can be overwhelming, especially if Excel is still the auditor’s tool of choice. What the agency really needs is the ability to ingest very large volumes of data, automatically check the calculations for accuracy, compare them to the previously-filed tax returns, and detect anomalies and signals of non-compliance or unintentional mistakes. There are similar opportunities to enhance the audit process across tax types.

Here’s a different example in the property tax domain. Tax bills are based on appraised value. There’s technology and analysis involved in the appraisal process, but it’s still partly subjective, based on the appraiser’s sense of properties and neighborhoods. And that can introduce bias. Today we can do machine-learning-based appraisal calculations and compare them to the traditional ones and determine whether assessments are accurate, consistent within and across neighborhoods, and free of undue bias. Recognizing and removing systemic biases is of growing importance in our society. Mathematical processes can provide useful counterparts to human viewpoints.

SB: Another innovative use of advanced pattern recognition is the risk cataloging of tax returns. Banks are required to identify and quantify their risks around solvency, fraud, and other issues. So they catalog and monitor those risks on a regular basis. Some innovative tax agencies are starting to do the same thing with tax returns. For example, Chile’s version of the IRS, the Servicio de Impuestos Internos or SII, analyzes every line item on every return to identify and catalog their respective risks. Analytical models both detect those risks and help mitigate them. The idea is to be structured about compliance risk, but also about common errors and taxpayer behaviors. Enforcing this discipline requires analytics.

The list could go on. OCR technology and analytics to “clean up” hand-written tax returns. Satellite images and drone video analyzed to ensure severance tax compliance by mining and oil companies. Text analytics to save agency staff from having to review volumes of documentation behind specialized tax credits. I’d just summarize by saying that analytics can be used to manage almost all aspects of a tax agency, including overall revenue management and policy simulation around tax codes and rates.

How has the COVID-19 pandemic changed the scene and the pressures on tax agencies?

SB: The pandemic hasn’t introduced new challenges as much as it has amplified and accelerated the challenges that tax agencies already faced. They were under pressure to provide better digital taxpayer service. Now with tax agency offices offering limited services and many employees working remotely, a lot more business has to be conducted either online or by phone. Agencies had to rapidly rethink their processes and accelerate their digital services. The one major new challenge, as we’ve mentioned, is the responsibility to distribute pandemic relief funds and try to minimize associated fraud. The pandemic has brought out the fraudsters, as the U.S. has seen in the small business loan program. Tax agencies around the globe had to shift in a matter of weeks from collecting the money to finding ways to pay it out to the right people in the right amounts.

DP: These are difficult times operationally for tax agencies, and difficult times for taxpayers interacting with agencies, many of them needing to communicate directly with the tax office for the first time in their lives, when the agencies and their staff are so stretched. Agencies’ reputations may be suffering, and unfairly so. They are expected to do it all – the regular work of tax collection and taxpayer service and fraud prevention and now relief funds disbursal. The U.S. government is very focused on fraud directly related to the pandemic relief payments. But I’d like to see more open discussion, for example, on abuse of tax deductions related to the CARES Acts. COVID-19 may be providing cover for tax evasion and non-compliance, and that complicates the work of tax agencies.

We talked about some examples from around the world. How do tax agency opportunities and challenges vary internationally?

DP: For the most part, tax agency strengths and weaknesses are much the same, as are demands on them from the public and their governments. The conversations I’m having with agency leaders in the U.S. are very similar to the ones I’m having with their counterparts in Asia-Pacific and Europe and elsewhere. That degree of similarity opens avenues for collaboration. For example, there is a consortium of the tax authorities of Australia, Canada, the Netherlands, the U.K., and the U.S., referred to as the “J5.” They work together on major international tax fraud cases. However, especially because most of them have similar tax agency technology platforms, they could be sharing other information, analytics methods, and experience. Their challenges and interests are so similar.

There are two big differences. One is very different mindsets and laws around data privacy. Europeans must comply with GDPR, the General Data Protection Regulation. The U.S. is not as strict at protecting personal data as yet. The other is big differences in the scale of tax agencies themselves. We mentioned lack of resources, but that’s not the case everywhere. One European country of about 17 million people has 30,000 working for the tax authority. California, in contrast, has almost 40 million residents and 5,400 agency staff. A lower-staffed agency, perhaps of necessity, tends to have the greater appetite for innovation and automation. They know they have to do more with less.

SB: I agree that there is little variety in the fundamental mission of tax agencies. There are only so many ways to collect revenue and serve taxpayers. But there’s great variety not just in data privacy, but in data access. In Southeast Asia, specifically Singapore and Malaysia, the governments know a lot about you, and they can get almost any kind of financial information that you can imagine. In Europe and the U.S., payroll-related data is generally captured, but a lot of other tax data is self-reported by the taxpayer.

There’s also wide variety in the cultural norms around compliance. Scandinavian countries tend to have a very high level of voluntary compliance among taxpayers. But if you go to certain countries in the Middle East or Southeastern Europe, Greece for example, tax evasion is a virtual sport. That makes for some very wide differences in tax administration, but in all cases, it really helps to know your taxpayers and their behaviors.

What expertise and experience are needed to implement and capitalize on a robust data analytics platform?

SB: I would suggest four core capabilities that you need. The first is to be able to ingest and make sense of the data that’s coming at you even when it feels like a data tsunami in a tax agency. That includes handling a variety of data sources, recognizing and adjusting for the quality of the data, and putting it in an accessible and understandable format. So we’re talking excellence in data management.

Second is the ability to ask the right questions of that data. That’s a somewhat creative process of sensing what the data might have to say and framing innovative and useful questions around business needs. Then there’s the process of selecting from among the dozens or even hundreds of different types of algorithms or analytics that might be used to answer the questions you have. You need many different types of arrows in your quiver of analytical methods.

The third capability is really around action. It’s great to come up with a new analytical model. You massage the data and you ask the right questions. When you send the results to staff members, what are they going to do with it? Can they act upon it? They may need a combination of education, practice, and coaching to use analytics in their workflows and decisions.

Finally, you want to be able to learn how to do better. So you’ve gotten the data, asked and answered good questions, and taken action. Well, did the action work? If not, or not completely, how do you refine the analytical model and business process? You need the information, process, and discipline to close the feedback loop on each of the functions in a tax agency.

DP: Behind the capabilities Shaun just described lie quite a lot of technological skill and experience, both to develop a platform and to put it to work. The tax agencies I work with would like to have an ample supply of statisticians, data scientists, and other technical experts. But nobody has as many as they need, state and local agencies often have small cadres, and people with these skills are hard to retain.

So agencies are asking both what capabilities they need themselves and how to supplement their capabilities through collaboration. At SAS, we try to be ready to help where needed – data and analytics technology, business processes, up-front strategy, effective implementation. As mentioned with the J5, there’s also opportunity to collaborate with and learn from other tax agencies, or other government agencies. And there’s potential for private-sector partnerships with enterprises that are really good at data and analytics for processing and paying claims, or making collections or detecting fraud, or even helping taxpayers collect and prepare their tax data.

Specialized skills may be in short supply, but it’s a mistake to think that only those with very deep data science programming skills can add value. Not that many years ago, analytics needed specialized programmers. Now with graphical interfaces and analytics, business analysts of all kinds can do amazing things on their own.

Why is this a good time to modernize data and analytics platforms and transform tax agency processes?

DP: My motto is “Carpe Diem,” seize the day. The pandemic has been an unprecedented disruption. Individuals, families, and organizations have made so many rapid adjustments. If when the doors finally fully reopen, tax agencies and other organizations revert to old ways of doing things, it will be a great loss, a missed opportunity.

Many of us seem to be working harder and maybe more frantically much of the time. But at the same time, the pandemic has slowed us down. People are taking the occasion to reflect on and talk about how to make the future better. In organizations, there’s a willingness to work together to advance the cause. People really want to step forward, not back. We sense this because we feel increased intimacy with our clients. Working relationships are less formal but more connected.

Maybe it’s not “seize the day,” but “seize the silence,” the pause. I think this is an ideal time to commit to innovation and change.

SB: Absolutely. Another way to look at it is that COVID gives you cover. The cover to continue to experiment and innovate and adjust, to take advantage of the disruption. And there’s so much to do, so many opportunities to improve dramatically. Tax agencies are dealing with more data than ever, and you can’t make sense of it and act upon it without more and better analytics. Now is a great time to continue to modernize your platforms and to transform the ways that a tax agency works.

What are your closing recommendations to tax agency leaders who want to pursue these opportunities?

DP: I again say “Carpe Diem.” Take this moment of different to do something different. Your agency has data and analytics capabilities today. Understand what you can already do. Chances are you can do more and pivot more quickly than you realize. There are lots of ways to infuse better data and analytics into existing processes and get better results. You can be on a better path forward without making it a whole new adventure.

SB: Act now, no question, but don’t be shy. Find your biggest challenges or pain points today and tackle them first. You may find it easy to make significant and measurable improvement. That’s far more valuable than putting minor problems to rest. And as a leader, give cover to your employees. Let them have license to be innovative. Frankly, give them license to fail along the way – as long as they are learning from that failure how to get better over time. The license to be creative and take action will generate transformational ideas and opportunities.

About the authors

Author photo

Shaun Barry, Director of Government Fraud & Security Intelligence Solutions, SAS Institute

Shaun Barry is a renowned expert in fraud and integrity, with a specific focus on government. Shaun has worked for and with federal, state, and local governments around the world for over twenty-five years to foster innovative and efficient business processes through technology. He specializes in tax and revenue, healthcare, social benefits, and motor vehicle functions. Shaun holds a Bachelor of Arts degree in American Studies from the University of Notre Dame and a Master of Public Policy degree from Duke University.

Author photo

Deborah Pianko, Principal, Government Fraud & Security Intelligence Solutions, SAS Institute

Deborah Pianko has 25 years of experience building technology solutions for tax and revenue agencies. She is a subject matter expert in tax administration including collections, audit, return and payment processing, customer service, revenue protection and fraud detection. She has led tax agencies through many tax filing seasons and adoptions of new technology. As a system engineer, Deborah helped to build the systems and data models used by many tax agencies including DC, Tennessee, Arizona, Maryland, Ohio, Nevada, Puerto Rico, Detroit, Australia, and others. Deborah holds a dual Bachelor of Science degree in Economics and Journalism and is a competitive rower in her spare time.

Author photo

Robert Morison, Lead Faculty Member, IIA

Robert Morison serves as Lead Faculty with IIA. He is an accomplished researcher, writer, speaker, and management consultant, and an authority on what happens at the intersections of business, technology, and people management. He has been leading breakthrough research for more than 30 years, collaborating with eminent academics, thought leaders, and management innovators. He has written on topics ranging from business innovation, reengineering, and analytics to workforce management, demographics, and retirement.

Bob is coauthor of three books: What Retirees Want: A Holistic View of Life’s Third Age (Wiley, 2020), Analytics at Work: Smarter Decisions, Better Results (Harvard Business Press, 2010), and Workforce Crisis: How to Beat the Coming Shortage of Skills and Talent (Harvard Business Press, 2006). His 2004 Harvard Business Review article, “It’s Time to Retire Retirement,” coauthored with Ken Dychtwald and Tamara Erickson, received a McKinsey Award. He holds an AB from Dartmouth College and an MA from Boston University.