My research focuses on experimental studies of non-compliance behaviour. Using both laboratory based and on-line experiments that require subjects to complete tax returns, we can examine the effects of changing various features in the tax system on compliance.
The questions central to my research include what effect does customer service by the tax authority have on compliance and what effects might there be from the pre-population of tax tax returns with values reported to the tax authority.
On-line experiments allow us to more easily engage with ordinary taxpayers over traditional lab based methods which are typically targeted at students.
Qualifications
BA MA DPhil(Oxon) MSc PhD (Exeter)
Key publications | Publications by category | Publications by year
Publications by category
Journal articles
Chen JC, Fonseca MA, Grimshaw SB (2021). When a nudge is (not) enough: Experiments on social information and incentives.
European Economic Review,
134, 103711-103711.
Full text.
DOI.
Fonseca MA, Grimshaw SB (2017). Do Behavioral Nudges in Prepopulated Tax Forms Affect Compliance? Experimental Evidence with Real Taxpayers.
Journal of Public Policy & Marketing,
36(2), 213-226.
Abstract:
Do Behavioral Nudges in Prepopulated Tax Forms Affect Compliance? Experimental Evidence with Real Taxpayers
Defaults, in the form of prepopulated fields within the tax form, have been identified as potential mechanisms that tax authorities can use to reduce noncompliance. They achieve this by simplifying the process of filing taxes, thus reducing the scope for errors. However, defaults may increase the scope for evasion if set incorrectly. The authors report experimental data on the effect of correct and incorrect defaults. They find that prepopulating tax returns is a worthwhile policy only if it is done with highly reliable information. Setting default levels that underestimate taxpayers’ true tax liability leads to significant drops in compliance and tax revenue. The authors also study whether nudges that contain messages with descriptive norms about compliance can mitigate the adverse effect of prepopulated returns with incorrect values. Nudges that react to inputs from the taxpayer effectively raise compliance, whereas static nudges do not. This result demonstrates the limits to the applicability of nudges in a public policy sphere as well as possible adverse effects resulting from poor implementation.
Abstract.
Full text.
DOI.
Publications by year
2021
Chen JC, Fonseca MA, Grimshaw SB (2021). When a nudge is (not) enough: Experiments on social information and incentives.
European Economic Review,
134, 103711-103711.
Full text.
DOI.
2017
Fonseca MA, Grimshaw SB (2017). Do Behavioral Nudges in Prepopulated Tax Forms Affect Compliance? Experimental Evidence with Real Taxpayers.
Journal of Public Policy & Marketing,
36(2), 213-226.
Abstract:
Do Behavioral Nudges in Prepopulated Tax Forms Affect Compliance? Experimental Evidence with Real Taxpayers
Defaults, in the form of prepopulated fields within the tax form, have been identified as potential mechanisms that tax authorities can use to reduce noncompliance. They achieve this by simplifying the process of filing taxes, thus reducing the scope for errors. However, defaults may increase the scope for evasion if set incorrectly. The authors report experimental data on the effect of correct and incorrect defaults. They find that prepopulating tax returns is a worthwhile policy only if it is done with highly reliable information. Setting default levels that underestimate taxpayers’ true tax liability leads to significant drops in compliance and tax revenue. The authors also study whether nudges that contain messages with descriptive norms about compliance can mitigate the adverse effect of prepopulated returns with incorrect values. Nudges that react to inputs from the taxpayer effectively raise compliance, whereas static nudges do not. This result demonstrates the limits to the applicability of nudges in a public policy sphere as well as possible adverse effects resulting from poor implementation.
Abstract.
Full text.
DOI.
