Signing at the Top: The Key to Preventing Tax Fraud?

In filling out self-reported documents such as tax forms, we declare the information truthful with our signature, but usually we sign at the end of the form. Researchers Francesca Gino and Lisa Shu discuss whether governments and companies can bolster honesty simply by moving the honesty pledge and signature line to the top of the form, before people encounter the opportunity to cheat.
by Carmen Nobel

Among the most worrisome issues for the Internal Revenue Service is the federal tax gap, which is the difference between what Americans should pay on their taxes and what they actually do pay. According to the IRS's most recent estimates the gap amounts to around $350 billion annually, largely due to understatement of income and overstatement of expenses. Obviously, many taxpayers give in to the temptation to cheat on their taxes—in spite of the requirement for them to sign a statement declaring that the information on the form is "true, correct, and complete."

As with most self-reported documents, the signature line is at the end of the tax form, after the form has been filled out and, consequently, after the potential cheating has occurred. (It's analogous to a witness first testifying in a court trial and then swearing on a Bible, "For the last twenty minutes I have told the truth, the whole truth, and nothing but the truth.")

Now research suggests that the federal government might be able to encourage honest reporting simply by moving the signature line to the top of the form, such that signers declare that they will tell the truth rather than that they have told the truth. The findings, based on several experiments, are published in a new paper, When to Sign on the Dotted Line? Signing First Makes Ethics Salient and Decreases Dishonest Self-Reports, written by Lisa L. Shu, Francesca Gino, and Max H. Bazerman of Harvard Business School, Nina Mazar of the University of Toronto, and Dan Ariely of Duke University.

"A lot of prior work has focused on the consequences of unethical decision-making and the factors that lead people to be unethical," says Lisa Shu, a doctoral candidate in Organizational Behavior at HBS. "This drove us to want to explore the flip side. We know a lot about when and why people cheat based on our lab and field studies; we thought it was now time to examine how to prevent people from cheating."

The key, according to the researchers, lies in increasing ethical salience: inducing people to pay greater attention to their moral standards and examine the integrity of their behavior when it counts the most.

"A signature is a way to highlight the fact that you're about to do something important, and that it's going to be a reflection of the self," says Francesca Gino, an associate professor in the Negotiation, Organizations & Markets Unit at HBS. "Attaching a signature to a pledge of honesty is a way of effectively linking identity to morality."

Testing Honesty

The research team conducted a series of experiments in which participants could benefit financially from dishonest self-reporting. In each case, the researchers compared the effectiveness of signing one's name at the beginning of the form with signing at the end.

In one experiment, they teamed up with an automobile insurance company, manipulating the signature line on the policy review forms that customers receive at the end of each year. The form behooves policyholders to report low mileage because the fewer miles driven, the lower the annual premium—presenting a temptation to lie.

Controlling for the number of cars per policy, the researchers found that the average reported mileage per car was much higher among policyholders who signed at the top of the form than those who signed at the bottom, where the signature line traditionally is located. (The difference averaged 2,747.8 miles per car.)

In another experiment conducted in a university research lab, the researchers tested the possibility of increasing honesty on tax return forms. The 101 participants—mostly students—received instructions to complete an incidental problem-solving task, which included the following information: "For the problem-solving task, you will be paid a higher amount than what we usually pay participants because you will be taxed on your earnings. You will receive more details after the problem-solving task." The instructions went on to say that the participants would receive a tax credit for the cost of their commute to the lab.

After the problem-solving exercise, the participants went into another room and filled out a "research study tax return" form based on the IRS's Form 1040. As with the previous experiment, some participants were asked to sign an honesty statement at the top of the form, while others were asked to sign the statement at the bottom. The experiment also included the added wrinkle of a control group, in which the forms had no place for a signature at all.

By covertly using information from the participants' individual workstations, the researchers could figure out who had understated their earnings or overstated their commuting costs.

Again, the results were significant in indicating the efficacy of a top-of-the-form honesty statement. For the signature-at-the-top group, 13 out of 35 participants cheated on their tax forms (37 percent). For the signature-at-the-bottom group, 26 out of 33 participants cheated—79 percent. This percentage was actually higher than the control group participants who didn't sign at all (64 percent, or 21 out of 33).

"There is a lot of empirical evidence that suggests that we do care, as individual beings, about being good and moral," Gino says. "But there is equally compelling evidence that we often cross ethical boundaries when we have the opportunity to do so. Our research identifies an effective way to curb cheating: We just need to be reminded to care at the right times. This reminder can be a very simple one, and it is something that could be applied in many different contexts."

How Practitioners Can Help

The researchers plan to share their findings with the IRS, with the aim of shrinking the tax gap by moving the honesty pledge to the top of Form 1040. "It's a long shot, but we're hoping they'll be interested," Gino says.

In the meantime, they are looking to test the top-of-the-page signature hypothesis on other self-reported documents such as expense reports, medical information forms, and time sheets.

"My hope is that this practice will be picked up by a lot of businesspeople," Gino says.

Note to executives: Are you interested in finding out how a top-of-the-page signature would increase honesty among your employees as they fill out expense reports and other corporate forms? And would you like to participate in the next phase of the aforementioned research at Harvard Business School? If so, please send a note to professor Francesca Gino at and doctoral candidate Lisa Shu at

About the Author

Carmen Nobel is the senior editor of Harvard Business School Working Knowledge.
    • Jules Muis
    • VP&Controller and DG internal Audit, former World Bank and European Commission
    Prof. Jennifer Berdahl of the Rotman School of Management, Toronto, kindly brought this article to my attention.

    I think your base idea to change the syntax of individual assurance statements/sign offs on matters of communal/public interest so as to include the past, current and prospective state makes a hell of a lot of sense, also in macro prudential controls, hence beyond micro/personal assurances. A system limited to signing off on the past only, leaves the community with the considerable risks - or audit costs- signatories often willing to run de limited residual discovery risks that goes with retrospective-only sign offs.

    Interestingly, the concept of actualizing the sign of, and give it a prospective/normative twist, already plays in macro prudential controls.

    1) I have been advocating for al long time that regulators of financial markets, from central bankers, financial market oversight bodies, up to and including the Financial Stability Board should annually sign off, duly qualified as the case may be, ' that they know of no systemic risks that might impair the orderly functioning of the financial markets', their prime remit as we have so painfully discovered. It is a very effective results based accountability tool and, paradoxically, as a bonus, strengthens regulators independence from undue political pressure. It requires regulators to go beyond their cherry-picked truth in their reporting, but go for the uncomfortable whole truth and nothing but the truth in their reporting, including their own 'don't knows', on essentials.
    2) Equally, multi-lateral organizations such as the IMF, World Bank or the European Union, should demand recipient countries periodic assurance statements to the effect that " adequate controls are in place to assure that the monies received will be used for purposes intended". Please note the prospective character of this statement.

    It has been my experience that any proposal for such instrument is immediately torpedoed by the powers that are, and for very good but unacceptable reasons: They don't know, don't want to know, and don't want the accountability.

    But there is one ray of hope: The European Parliament has asked such statement (1) to be submitted by EU recipient countries in 2005* and 2006**, to little effect and great resistance by EU member states.
    One may wonder why, and need not go too deep.
    So far only 4 of the 27 EU countries have volunteered to work to towards such assurance statement, the UK, Netherlands, Sweden en Denmark.

    As for 1) national and international regulators have shown no or little appetite to hold themselves to account- despite major financial reform - through the issuance of such results based prospective statement. Some progress may be observed in the risk paragraph of their annual report in the last couple of years.

    In short, your thesis is an important one to pursue beyond individual sign offs, and has tremendous macro prudential control potential.

    Keep up the good work.


    Jules Muis

    *The EU budget Commission, in 2005 recommended national assurance statements/declarations on EU fund spending by member states:

    - one declaration should be issued before the money flow begins and the Member State should declare that it has set up control and supervisory systems as required by the EU regulations and that these systems are expected to be efficient;

    - the second declaration should be issued after the Member State has received and used the money. It should be a yearly declaration in which the Member State declares on a political level that the money has been used for the intended purposes and that the supervisory and control systems worked well.

    ** 76. "Calls on the Member States fully to support the Commission's efforts to develop a secure and comprehensive internal control framework for the EU budget, which could finally lead to the creation of a long-overdue mechanism for accountability, particularly on the question of shared management;
    77. Stresses the importance of pursuing the idea, as proposed in its resolution of 12 April 2005 , that a declaration of assurance as regards the legality and regularity of the underlying transactions be provided from each Member State's highest political and managing authority;

    There is a host of documentation to be found on the website of the European Parliament/ COCOBU ( EP Commission Budget Control)
    • Mohammad Khan
    • Associate Manager, Mega & Forbes Group
    Frankly speaking i didn't see a much of a change by Signing forms at the bottom or at the top to undertake the info submitted is wholly correct. It's the system that make it write or wrong if the Tax net in the country is unbiased and wide spread it would not entice people to fraudulently represent themselves in whatever case may be that of Tax , Automobile insurance premium etc.

    Its the system if its strong would never allows ,tempted anyone to indulge in wrong doing.

    Rather main focus would be solely on improving check and balances in the respective system and dig out lope holes to avoid recurring of these malpractices.
    • Kapil Kumar Sopory
    • Company Secretary, SMEC(India) Private Limited
    A person who certifies truthful reporting in the end will do the same at the beginning notwithstanding the reality being otherwise. You can't dissuade the unethical to behave better in this way. Even in courts, witnesses take oath and still tell lies.
    There has to be a mechanism to scrutinize the tax returns, etc. and levy exemplary punishments in relation to the nature/quantum of the tax evasion. Many a time action is taken with lot of avoidable delay. And, if someone goes to court, it takes years ( eg., in India ) to get the final judgements. By then, situation might have drastically changed to the disadvantage of the exchequer.
    • Atul Guglani
    • Director, Mantex Technologies
    The paper is perhaps linked more on " Testing Honesty" then up or down signing .

    Technically, I would agree with Mr Khan that it makes no difference on where you sign as long as the system itself is simple, trusted by the user and has value to the user.

    Honesty becomes very flexible, where a person is dealing
    with Institutions which are impersonal and have no direct relevance to his accountability . Any small gains, which can be pinched out of misreporting is then considered the invisible incentive and shrewdness of that person. Same person who would misreport for small gains may give away large sums to charities and still justify his honesty.

    Remember, the corporate world or otherwise is not a world of saints and neither is it a world of robbers. In between the two extremes fall most people and hence make them sign where ever you want. The values of the person will finally deliver the level of honesty. Unfortunately, there are still no clear measures of Values of a person.