مقاله انگلیسی رایگان در مورد استفاده از اعداد گرد شده برای شناسایی فرار مالیاتی – الزویر ۲۰۲۴

Abstract

Australian taxpayers display a clear preference for round numbers for end-of-year tax refunds, bunching at positive and salient thresholds such as the tens, hundreds and thousands. Bunching appears to be driven by tax evasion. Data from audited returns shows that bunching is present in returns before audit, but does not persist post-audit. Tax preparers play an important role, being twice as likely to deliver positive round-number refunds as individuals who file their own tax returns. Preparers with greater propensity to bunch deliver larger refunds by lifting deductions and lowering reported income for return items where audits are costly. This highlights how bunching behaviors can help identify tax evasion, including tax preparers who facilitate it and the tax return items which are manipulated.

Introduction

In 2018 a typical Australian taxpayer was 84 per cent more likely to be owed a dollar by the government than to owe a dollar, and 70 per cent more likely to be owed $A1000 than $A999. We document the widespread bunching of tax refunds at positive and salient round-number thresholds such as multiples of $A10, $A100 and $A1000. Drawing on nearly three decades of administrative tax data linking individuals, their tax returns and their preparers, coupled with random audit data, we use bunching behavior to draw out broader insights on the role of tax preparers in the tax system and on tax evasion.

Using data from random audits conducted by the Australian Taxation Office, we observe a systematic relationship between bunching and audit outcomes. Amongst individuals whose balance is adjusted downward post-audit, we observe pre-audit bunching but no post-audit bunching. Amongst those whose balance is not adjusted, or adjusted upwards, neither post- nor pre-audit balances show any statistically significant bunching.

Tax preparers play an important role in this bunching behavior. People who use tax preparers are twice as likely to bunch at round number refunds as those who do not use tax preparers. Individuals who receive a refund at or just above one of these thresholds are more likely to remain with their tax preparer in the following year, though are not charged higher fees. The extent of bunching has grown dramatically over time, likely reflecting these dynamics and the advent of electronic returns.

Conclusion

In this paper we have shown that Australian taxpayers have a clear preference for round-number tax refunds and that their refunds bunch at positive and salient thresholds. Using random audit data, we have shown that this bunching is driven by tax evasion. Finally, we have explored the role of preparers. Taxpayers who use tax preparers are twice as likely to bunch as those who do not. High-bunching tax preparers assist people to bunch by lowering reported income, increasing deductions and, in the process, generating higher refunds. These tax preparers target income and deductions that are difficult for the tax authorities to audit. The main effect of such tax preparers is on behavior at the point of tax filing and does not influence the future behavior of taxpayers. These observations are consistent with tax preparers helping clients to evade taxes rather than helping them to manage their tax affairs in a legitimate way to reduce their taxable income. Our paper has an important policy implication: tax authorities might want to target round-number refunds and “high-bunching” tax preparers in compliance activity, while these patterns hold true.

Bunching behavior has increased dramatically over time. Individuals who self-prepare are bunching more than in the past as are tax preparers. Bunching also increases over time as more people move to “high-bunching” tax preparers. “High-bunching” preparers do not charge higher fees, but they have stronger client retention. This may provide an incentive for the observed behavior of preparers.