As Ben Meiselman notes in the introduction to his new study of taxpayer compliance, there’s no marginal cost to making a tax authority’s written communication to a taxpayer more persuasive. In an era of likely increasing austerity in tax enforcement budgets at all levels of government, finding low-costs methods to increase compliance will become all the more important. Field experiments have produced some information about effective messages, and Meiselman’s study adds to this research, but as I discuss after reviewing the paper’s findings, Meiselman’s account also raises questions about Detroit’s capacity to operate an income tax.
In the spring and early summer of 2016, Detroit sent both a postcard and a certified letter to 7,142 suspected “ghosts,” 2014 federal taxpayers with Detroit addresses who had failed to file a Detroit income tax return for that tax year. A control postcard informed taxpayers that they would receive a letter in a few days about filing a tax return, and the subsequent letter informed the taxpayer that Detroit’s records indicated that she was a resident of Detroit and did not file a city tax return in 2014.
The experimental postcards and letters included boxed messages developed by Meiselman, testing a variety of compliance frames: penalty salience, punishment probability, compliance cost, and civic pride. The punishment salience frame included the message “Failure to file a tax return is a misdemeanor punishable by a fine of $500 and 90 days in jail.” The punishment probability frame included a message about the taxpayer’s reported federal income, suggesting Detroit already knew something about the taxpayer’s income for 2014. The compliance frame included a blank tax form and a return envelope, and the civic pride frame reminded taxpayers of the importance of tax collection to Detroit’s comeback. Some correspondence also combined the penalty salience and punishment probability messages.
Consistent with other studies, Meiselman found that moral persuasion (cast here as civic pride) had little impact on taxpayers. The penalty salience message, however, tripled the response rate of the control group and was better than most other messages. (There was no statically significant difference between the response rate to the penalty salience message and the combined penalty salience/punishment probability message.) This penalty salience message was more effective on a variety of metrics; not only did it increase the response rate to the correspondence, but taxpayers receiving this message responded more quickly than other taxpayers, were more likely to admit tax due (rather than claim a refund), and were more likely to actually remit the tax due. Meiselman’s research also reveals which taxpayers might be more successfully targeted with this message, finding that higher income taxpayers, older taxpayers, and taxers who had not previously been non-filers all had higher response generally.
While the paper’s conclusions focus broadly on what tax authorities can learn about messaging, I found equally interesting Meiselman’s account of Detroit’s tax enforcement efforts and his attempt to calculate the welfare effects of additional enforcement efforts. Detroit levies a flat 2.4% tax on resident income, and non-residents working in Detroit owe a 1.2% tax on income earned in the city. (There is a $600 exemption per filer, spouse, and dependent available to both residents and non-residents.)
Unsurprisingly, given Detroit’s fiscal capacity, the city has limited enforcement capabilities. At the height of Detroit’s fiscal crisis, the city had trouble suing to collect tax delinquencies because of limited municipal court staff. Meiselman estimates that in 2014, almost half of those required to file a city return failed to do so. Part of the problems seems to have been technical limitations that prevented Detroit from accepting electronic returns. Detroit residents and non-residents subject to Detroit’s income had to separately mail a paper return to Detroit. One could easily imagine many non-filers either not knowing that they had to take this separate step or simply forgetting about it.
Paper filing is also considerably more costly to process than electronic returns, adding additional compliance costs. Because some of the non-filers were actually owed refunds and the response rate to even the penalty salience letter was around 10%, Meiselman estimated that each penalty salience letter raised about $8, net administrative costs. Meiselman estimates that each letter cost about $4.70 to send, including staff time, with the majority of that cost, $3.30, attributable to the charge for certified mail. There are also additional administrative costs to processing any returns received.
Meiselman suggests that additional enforcement efforts may have a negative effect on social welfare, accounting for both the loss of private spending and the value of taxpayer time in finding and completing Detroit’s tax return. This claim is sensitive to his compliance cost estimates, and depending on how quickly a given taxpayer was able to locate their 2014 federal return, its possible he overestimates the time it takes taxpayers to respond or underestimates the value of additional public revenue. Nevertheless, his calculations suggest that Detroit’s income tax has been very expensive to administer.
Times may be changing. As of 2015, Michigan took over processing Detroit’s tax returns, and for the first time city returns could be filed electronically. This may create a dramatic increase in compliance with the city’s income tax, but if it doesn’t, it would seem the city needs much more than to improve its messaging.