Tax with AI …. Anwar Kashif Mumtaz
IN December 2023, Pakistan’s tax revenue reached $17.079 billion, marking a significant increase from the previous figure of $7.622bn reported in September 2023. Tax revenue data for Pakistan is updated on a quarterly basis, with an average of $10.565bn recorded from September 2000 to December 2023, spanning 94 observations.
What is even more significant is that only 2.74 million individuals in Pakistan file personal income tax, representing merely 4.1 per cent of the labour force and 1.3pc of the total population. It is pertinent to mention that 35pc of these individual filers are exempt from paying income tax as their income falls below the taxable threshold, resulting in zero tax liability for the year. Between 2008 and 2021, Pakistan’s federal government collected between 36pc to 39pc of its tax revenue from direct taxes, while the rest came from indirect taxes.
Expanding the tax net, equitable tax, restructuring the tax brackets, etc, have been part of conversations, but that is where such actions have stopped. We are still challenged by the small community of taxpayers, overburdened salaried class and large fields of agriculture and real estate being exempted. We have the super-impressive examples of Malaysia, Singapore, and South Korea — economies that have risen on the basis of solid monetary policies, with far fewer resources.
Tax audits in Pakistan have long been associated with compromise, connivance, and rampant corruption. This pervasive culture not only undermines the integrity of the tax system but also perpetuates tax evasion, depriving the government of much-needed revenue. The discretionary powers wielded by tax officers further exacerbate the problem, raising concerns about their potential for abuse. The complexity of tax laws often hampers the effective enforcement of tax regulations, allowing for loopholes to be exploited.
Unlike human auditors, AI operates without bias or influence.
Pakistan’s population number goes in its favour. Imagine crowd-funding by 240m people. Imagine 50pc of the population paying taxes — Pakistan becomes the golden goose. This argument compels us to look for the quickest, and most effective transformative solution that can restore faith in the tax system and ensure compliance with the law — artificial intelligence (AI), a game changer in the realm of tax audits.
AI is not a mechanical way replacing human error, but a made-to-order algorithm that can cater to and adapt to all local frameworks with optimisation of service, transparency, and above all, accessibility. Smartphone penetration which is currently around 31pc needs to be supported to take it up to 100pc and then to use that platform for reaching out to the banked and unbanked, documented and undocumented, in order to bring them on board.
Unlike human auditors, AI operates without bias or influence. It adheres strictly to the provisions of the tax code, without succumbing to external pressures or considerations. By leveraging advanced algorithms and data analytics, AI can efficiently analyse vast amounts of financial data, identifying patterns and anomalies indicative of potential tax evasion. This analytical prowess not only enhances the effectiveness of audits but also serves as a powerful deterrent against non-compliance.
AI-driven audits prioritise transparency and accountability. Every decision made by the AI system is based on objective criteria, devoid of any subjective judgement or favouritism. This not only minimises the risk of corruption but also ensures fair and equitable treatment for all taxpayers.
Moreover, AI facilitates the simplification of tax laws by streamlining complex regulations into easily understandable guidelines. This not only reduces ambiguity but also empowers taxpayers to fulfil their obligations with confidence and clarity. Critically, AI-driven audits shift the focus from mere revenue generation to long-term deterrence of tax evasion. By identifying irregularities and enforcing compliance, AI helps create a culture of accountability and integrity within the tax system.
Additionally, AI improves data management by automating data ingestion, extraction, and organisation, facilitating a ‘no-touch tax return’ concept, and enhancing audit processes.
This is the only route for us towards equitability of taxation in Pakistan. With the right mechanics, we can investigate the integration of AI into tax audits for the much-needed paradigm shift.
By eliminating the spectre of compromise and connivance, tax revenue can be enhanced in selected fields. AI can expand the tax net, instil trust and confidence in the tax system and foster a conducive environment for economic growth and development. As we embrace the transformative potential of AI, we pave the way for a fairer, more efficient, and more equitable tax regime for all.
The writer is president of the Pakistan Tax Bar Association.
Courtesy Dawn, April 22nd, 2024