Tax Notes logo

Can IRS Change Taxpayers from Procrastinators to Payors By Drafting Letters that Make Taxpayers Feel Bad?

Posted on Jan. 13, 2015

An earlier version of this post appeared on the Forbes PT site on January 12, 2015

What if the IRS could change taxpayer behavior and collect billions of dollars in unpaid taxes just by using words in correspondence that raise the taxpayer’s psychological costs of not paying what is owed?

Seems pie in the sky? Social science research suggests that people are motivated beyond factors such as risk of detection and severity of punishment. Research suggests that people want to view themselves as good and honest people, but if they can do the wrong thing without too much psychological cost, well, then people will often do the wrong thing. A recent research paper considering a field experiment on hundreds of thousands of UK taxpayers suggests that the content of letters may make a material difference and nudge taxpayers who are otherwise procrastinating to pay the IRS what they owe.

I have previously discussed that Congress and the IRS should more readily use insights from behavioral economics to increase psychological costs among preparers and taxpayers themselves (such as due diligence rules, the framing of questions and prompts on forms and software and more direct oversight over preparers). See for example my post What Kids Peeing in the Pool Can Teach Us About Tax Compliance (suggesting that a targeted duty of inquiry for return preparers can serve to increase both taxpayer and preparer psychological costs and thus reduce errors).

A recent research study called The Behavioralist as Tax Collector: Using Natural Field Experiments to Enhance Tax Compliance published as an National Bureau of Economics Research (NBER) working paper (link to digest only, NBER Working Paper 20007, March 2014) using field experiments on hundreds of thousands of UK taxpayers suggests that by drafting letters to delinquent taxpayers that describe that most other taxpayers in the UK pay taxes on time and by personalizing the letters so that the taxpayers understand that they specifically are a distinct minority in the UK that did not pay on time led to a material difference in collections.

The research is part of an important strand in social science research that analyzes variables other than probability of detection and severity of sanction as ways to improve tax compliance. While other research has previously looked at the impact of letters on tax compliance, this study differs in that it focuses on compliance relating to tax payment, whereas most other studies looked to measure the effect of agency-drafted letters on reporting behavior (rather than paying behavior) following tax agency correspondence (I might add that the results of those earlier studies have been mixed, including a study looking at state tax compliance in Minnesota by examining the impact of letters sent to taxpayers prior to the state’s tax return filing deadline).

Before summarizing the experiments and the results, I will give a brief discussion of the underpayment tax gap and the challenges the IRS faces in finding out which taxpayers genuinely cannot pay due to financial constraints and which taxpayers can pay but are choosing to not pay for reasons apart from liquidity concerns (e.g., procrastination, favoring other creditors).

Underpayment Tax Gap

The underpayment aspect of the tax gap relates to the amount of the tax that is properly reported but not paid. It comprises about 16% of the overall tax gap. The most recent comprehensive Treasury report (looking at tax year 2006) shows that the underpayment tax gap is $46 billion compared to $376 billion in the underreporting aspect of the tax gap, and $28 billion for nonfiling.

In the upcoming filing season, the IRS will soon receive hundreds of thousands of tax returns that will reflect billions of dollars in taxes due without any question relating to the amount of the underlying liability. The IRS has the difficult job of sorting those taxpayers into two categories, those who cannot fully pay what they owe due to genuine financial challenges and those who can pay but choose not to. For both groups, the IRS will send letters explaining what they owe, what their rights are in the collection process and also what are the consequences of not paying. Some taxpayers will respond and pay fully; others will try to respond; and others will simply ignore or possibly not even receive correspondence relating to the past due amounts declared on their tax returns(not a complete list).

What Did the Researchers Do to Goose Payment Among the “Won’t Pays”

In the UK most individuals do not have to file a tax return, as employers collect income tax at source under a “Pay as You Earn” system. In the UK, taxpayers who are self-employed (and certain others), however, have to submit an annual self-assessment tax return. The UK system requires filers to make two payments a year, one at the end of January and the other at the end of July. Taxpayers who fail to make payments receive a reminder letter, correspondence and phone calls. As with the IRS, HMRC can take a variety of enforced collection actions.

The researchers sought to focus on whether messaging following the initial declaration of non-paying can make a difference among the so-called won’t pay taxpayers, rather than the can’t pay group. In its parlance, the research refers to the “can’t pay taxpayers” as those with liquidity issues and the “won’t pay” taxpayers as procrastinators. The researchers are Michael Hallsworth, John List, Robert Metcalfe and Ivo Vlaev. Metcalfe and List are economists with the University of Chicago; Hallsworth and Vlaev are with the Imperial College of London and all worked closely with with Her Majesty’s Revenue and Customs and the Behavioural Insights Team (a social purpose company previously part of the UK government and now partly owned by the UK government and self-described as a “a world-leading social purpose company whose mission is to help organisations in the UK and overseas to apply behavioural insights in support of social purpose goals.”)

The researchers looked to communicate with taxpayers who had not paid by July 31, 2011 and who had not responded to an initial payment reminder. It selected a randomized sample of 101,471 people located in England, Wales and Northern Ireland.

HMRC sent a control message to one group that stated just the basic statement that the tax payment was overdue with information on how to pay or communicate if the taxpayers believed he did not owe. NBER digest author Matt Nevisky summarizes the five different messages that the agency sent to the nonpaying taxpayers, with the first three norm-based and last two rooted a public goods appeal:

  1. “Nine out of ten people pay their tax on time,” (invoking a basic norm);
  2. “Nine out of ten people in the U.K. pay their tax on time,” (invoking a country norm);
  3. “Nine out of ten people in the U.K. pay their tax on time. You are currently in the very small minority of people who have not paid us yet,” (invoking a minority norm);
  4. “Paying tax means we all gain from vital public services like the National Health Service (NHS), roads, and schools,” (emphasizing the gain from paying taxes);
  5. “Not paying tax means we all lose out on vital public services like the NHS, roads, and schools,” (emphasizing the loss from not paying taxes).

Interestingly, while all of the letters relative to the control standard tax payment reminder letters had an impact on payment, letter number 3, which addressed a country norm and definitively tagged the recipient as being in the minority, had the biggest impact.

Matt Nevisky in the digest of NBER research summarizes the findings:

Both norm-based and public-good messages increased the likelihood of individuals paying their taxes due, with large differences observed within the various norm-based messages. The basic-norm statement raised the tax payment rate by 1.3 percent, and the country-norm statement by 2.1 percent, within 23 days of communication. The minority-norm statement had an even larger effect, raising the number of taxpayers making payments by 5.1 percent. Both public-good messages, which relied on gain and loss framing, raised tax payments by 1.6 percent.

I include a brief description of the results from the report itself (copyright restriction severely limit what I can reproduce from the report itself):

Overall, the results show that short messages can address non-payment through increasing moral costs. This is an important finding, since most previous studies have concluded that the framing on messages does not matter – or, at least, that only sanction-based messages have an effect…. Viewed through the lens of our model, it seems that framing messages to increase moral cost can reduce procrastination and therefore increase tax payment. In total, we estimate that more than £3 million was collected in the 23-day sample period due to the messages in the first field experiment…If the minority norm approach had been taken for the whole sample, £11.3 million in additional tax revenues would have been gathered by this point.

(emphasis added, and footnotes and references omitted).

In the second experiment, which occurred a year after the first, the researchers sent letters to over 119,000 UK taxpayers. More from Nevisky, who describes how the second experiment

explored the effect of descriptive norms (describing what others do) relative to injunctive norms (describing what others think should be done). In the latter case, the researchers used messages that either stated that paying taxes was the right thing to do, or that most people thought that paying taxes was the right thing to do. Descriptive norms had a significantly larger effect on payments than injunctive norms. The authors estimate that as a result of their experimental intervention, HMRC collected £9 million more over a 23-day period than they would have otherwise.

The second experiment was more nuanced, with fourteen groups studied and smaller sample sizes due to the greater number of groups than the first study. The upshot of the second study is that people tend to respond more positively to letters that are based on norms that tie to what people actually do rather than what behaviors people approve (so that letters that describe how most others pay their taxes are more effective than letters that say most people think that paying their taxes is the right thing to do).


The essential argument in the paper is that tax administrators should incorporate moral costs into tax administration by presenting messages to nonpaying taxpayers that relate to social norms. That proposition at a minimum warrants further study.

To be sure the study is to my knowledge still a working paper, subject to further revision and peer review. It also looks at UK taxpayers, who may respond differently than American taxpayers. Yet the authors carefully build on a growing body of social science research that looks to measure the impact of messages on compliance with laws generally and with tax laws specifically. That research holds hope to tax administrators who may be searching for cost-efficient ways of boosting compliance and addresses an area not often studied, that is taxpayers facing an agreed upon assessed liability.

The study presents interesting possibilities for the IRS. Given that the cost of correspondence is relatively small, and the benefits from influencing behavior even on a relatively small percentage of taxpayers can be great, the IRS should examine this carefully and consider its application on this side of the Atlantic.

Copy RID