The Deterrent Effect of Whistleblowing on Tax Collections

Abstract

Many countries use tax-related whistleblowing programs, but the evidence on these programs suggests information provided by whistleblowers yields modest tax collections. However, when every citizen could become a whistleblower, deterrence from tax evasion can by itself increase tax collections. We find that tax collections significantly increased after the introduction of the whistleblowing mechanism in Israel in February 2013, although this mechanism directly yielded little or no tax collections. In support of the hypothesis that deterrence led to the increase in tax collections, we find that collections increased in industries with high tax-evasion risk, but not in industries with low tax-evasion risk. Furthermore, the increase in tax collections occurred in corporations, where the timing and magnitude of tax payments are more discretionary, but not from employees, for whom employers directly deduct taxes. Eventually, following reports that the whistleblowing mechanism is ineffective, deterrence diminished and tax collections decreased, suggesting the deterrence effect was temporary.

Keywords:

• whistleblowing
• tax evasion
• deterrence

Acknowledgements

We thank Dan Amiram, Eli Gilbai, Martin Jacobs (Associate Editor), J-P Kallunki, Shirly Levy, Herve Stolowy, Ilay Piterman, two anonymous referees, and seminar participants at University of Oulu (Finland) and Tel Aviv University for comments and suggestions. We also thank the Henry Crown Institute of Business Research in Israel for financial support.

Notes

1 According to the IRS website (https://www.irs.gov), any citizen can fill out a form reporting on tax noncompliance and potentially receive a reward of up to 30% of the tax collected.

2 In the UK, any person can report tax evasions to the tax authorities (HM Revenue and Customs, HMRC). HMRC aims at reducing noncompliance by rewarding whistleblowers (Topham, 2015).

3 Whistleblowing is often considered socially unacceptable (Pacella, 2015), as is also evident from moral justifications whistleblowers voice when disclosing the wrongdoing of others (Stolowy, Gendron, Moll, & Paugam, 2018).

4 See the report of the Israeli State Comptroller on the Tax Authority of Israel, October 2014 (in Hebrew).

5 The TAI, as in other western countries, requires individuals and corporations to file annual income tax returns and pay the taxes due. Many individuals who only receive salaries up to a certain limit are exempt from filing, because the employer deducts their income tax at the source.

6 Although tax evasion and tax fraud are illegal, firms engage in legal tax planning to reduce tax liabilities (e.g., Dyreng, Hanlon, & Maydew, 2008; Dyreng, Maydew, Hanlon, & Thornock, 2017; Amiram, Bauer, & Frank, 2017; and Dyreng, Jacob, Jiang, & Müller, 2017). See Shackelford and Shevlin (2001) and Hanlon and Heitzman (2010) for reviews of the tax literature.

7 According to the article, no whistleblowers received payment during 2013; that is, none of the information tips yielded tax revenues. The source of the information in the article was the TAI.

8 For example, when we include in our models imports of consumption goods and wage per employee, the coefficients of these two economic indicators are not significant and results are similar. When we include lagged tax collections as an independent variable, results also do not change.

9 Changing the tax rate could have two effects, as suggested by an inverted-U-shape Laffer curve. An increase in tax rates will result in an increase in tax revenue collected by the government, but only up to a certain point; an increase in tax rates beyond that point will result in lower tax revenues, because people will either work less or engage in tax evasion (see Romer & Romer, 2010). Brander and Politser (2014) examine the association between tax rates and tax revenues in the Israeli economy and find a positive association.

10 For robustness, we also present analysis for the level of tax collections, using calendar-month dummies as an alternative method of controlling for seasonality.

11 We determine the number of lags for the augmented Dickey-Fuller test using the Akaike Information Criterion. Brender and Navon (2010), for example, also use two lags in an ADF test of the time series of Israeli tax collections.

12 An autocorrelation function analysis suggests tax collections are correlated with collections in the same month of the previous year (lag 12) but not with tax collections two years back (lag 24). Therefore, an autoregressive model with tax collections in the same month of the previous year, or seasonal differencing as we use above, is appropriate. Yet for this test, we wish to emphasize the change in tax collections relative to the pre–Justice Hotline period, and not necessarily relative to the same month in the previous year, and therefore we use monthly indicators to control for seasonality.