ADVERTISEMENT

Rajasthan Shows States How To Use AI And Analytics To Plug Revenue Leaks

Rajasthan is using data analytics and artificial intelligence to help track tax revenue.

<div class="paragraphs"><p>(Photo: Unsplash)</p></div>
(Photo: Unsplash)

Indian states are looking at ways to plug revenue leaks as they stare at curtailed borrowing limits and the possibility of no extension in goods and services tax compensation. One way is to rely on data analytics and artificial intelligence to track taxes.

Tracking fake claims, evasion of interest payments and royalty charges could improve compliance and plug in the leakages of state tax systems. Something Rajasthan has tried with some success.

“In 2017, we piloted the advance analytics software in the State Directorate for Revenue Intelligence," Anand Swaroop, director-general of Rajasthan Revenue Intelligence, told BQ Prime over the phone. "Rajasthan was the first state to do it. Now in the last year, the state of Telangana has also come up with something similar.”

Emailed queries to Telangana chief secretary and excise and commercial tax commissioners elicited no response.

The central government has been relying on data analytics for income tax and goods and services tax compliance. But such tracking and predictive profiling is not common for states.

Efforts to plug revenue becomes critical for states when there is increased scrutiny of their off-budget debt. The central government is likely to ask states to include such borrowing through its agencies as part of state fiscal deficit, Bloomberg reported. And GST compensation, created at the time the nationwide indirect tax regime was rolled out to make up for lost revenue, is ending this year.

Rajasthan's advanced data analytics team has identified over Rs 3,700 crore of missed state revenue for FY22 from its five main departments—commercial tax, mining, stamp duty and registration, transport, and excise duty, according to the data shared by the state's Directorate for Revenue Intelligence.

"The advanced analytics system mainly works in two ways—one is effective information management by integrating data from multiple points and identifying evasions, non-compliances," said Vipul Kaushik, associate partner, analytics, government and public sector, EY, told BQ Prime. "And the other is using rule based and predictive analysis and risk profiling,"

EY serves as the software partner for Rajasthan and has a six-member team associated with each revenue department including the state's policy intelligence team, Center of Excellence for Revenue Research and Analysis.

"The system helps us track if the quantity of minerals transported checks out with the e-way bill issued and the capacity of the vehicle as registered with Vahan, the centralised vehicle registration with the Ministry of Road Transport and Highways," Swaroop said. "This way, we eliminate misreporting of quantities transported and ensure due overloading penalties are paid."

According to Kaushik, what makes Rajasthan different from economic intelligence teams in other states is the integrated approach of applying data analytics in each stream of revenue spanning across multiple state departments.

Opinion
What Some States Plan To Do After GST Compensation Cess Period Ends

“Rajasthan is one of the largest states of India with a potential of high economic growth," Arvind Mayaram, vice chairman of Rajasthan Economic Transformation Advisory Council and former Finance Secretary, said in an emailed response to BQ Prime. He is also associated with the policy intelligence unit CoERRA that uses the data analytics to relook at the policy framework.

CoERRA conducts risk analysis of taxpayers, sector-wise revenue receipts, potential leaks and pilferages and ways for plugging evasions.

Huge amount of data is generated on a daily basis through various applications, but they are in departmental silos, according to Mayaram. There is limited mechanism to monitor economic activities across the departments he said, explaining that such tracking is important as policies, especially tax rates, duties and various fees are decided based on previous rates.

In its revised estimates for 2021-22, Rajasthan estimated its own tax revenue at Rs 82,803 crore. While the missing amount identified through the data analytics mechanism contributes 4.5%, economists agree that the initiative that will bear fruit in time.

“It’s definitely a welcome step," Manish Gupta, associate professor at National Institute of Public Finance and Policy, said. "And while the quantum is not much, it’s important that the department undertakes such exercises so that you can identify areas of loopholes and issues, and future evasions and misreporting can be avoided.”

Still, all identified gaps do not translate into incremental revenue immediately.

Long-Term Benefit

“Not all the amount from the cases identified are realised, many are further appealed by the taxpayers, and some remain pending,” said Swaroop.

For FY22, Rajasthan has assessed cases worth Rs 695 crore but realised 72 crore, with majority of such cases stuck either in a litigation or in the process of being assessed. Interest payments on late filing, false input tax credits claims beyond due date, underreported royalties from mining are some of the common cases the state's Revenue Intelligence Department is pursuing.

“The more states start reporting this data along with the realisation time, the amount spent on recovery and issues in the systems also come into perspective," Gupta said. “In the long term, states should look at innovative ways to improve the efficiency of revenue collection and expenditure efficiency.”

According to Saurabh Bandyopadhyay, researcher at the National Council for Applied Economic Research, models that predict those at higher risk of not paying and tracking people to effect positive compliance measure will help in improving revenue collection and lowering tax avoidance in the longer run.

Another positive trend, he told BQ Prime, is how state revenue departments are moving towards more interaction with taxpayers through such actions than before.

Rajasthan’s early success has made other states take notice. Officials at the Finance Ministry of Jharkhand are looking at AI-backed system to track revenue. The state depends heavily on mining for its tax income.

Swaroop hopes artificial intelligence can be used to keep track revenue beyond state jurisdictions. Inter-state data sharing and integration could better track GST cases in cases the point of consumption is different from the origin of goods and services, he said.