Journal of Behavioral and Experimental Economics 70 (2017) 10–22 Contents lists available at ScienceDirect Journal of Behavioral and Experimental Economics journal homepage: www.elsevier.com/locate/jbee The effects of mass media campaigns on individual attitudes towards tax compliance; quasi-experimental evidence from survey data in Pakistan Musharraf R. Cyan a , Antonios M. Koumpias b,* , Jorge Martinez-Vazquez a a International Center for Public Policy, Department of Economics, Andrew Young School of Policy Studies, Georgia State University, 14 Marietta Street, Atlanta, GA 30303, USA b Department of Social Sciences, University of Michigan-Dearborn, 4901 Evergreen Road, Dearborn, MI 48128, USA a r t i c l e i n f o Article history: Received 11 July 2016 Revised 22 July 2017 Accepted 28 July 2017 Available online 29 July 2017 JEL classification: H26 K42 Keywords: Tax morale Tax compliance Mass media campaigns Information nudging Moral suasion Pakistan a b s t r a c t Pakistan has consistently performed low on taxation, with revenue collection hovering around 10% of GDP. Tax reforms have been attempted but without significant gains in revenue collection. Due to the recalcitrance of tax revenues, the tax authority attempted to enlist voluntary compliance, as one of the avenues for its efforts to increase collections. This paper examines the effectiveness of mass media cam- paigns in the TV and national newspapers used by the Federal Board of Revenue (FBR) of Pakistan to increase awareness, tax filing, and, ultimately, tax morale. We use survey data that were collected in 2014 immediately after these mass media campaigns by the FBR. Using coarsened exact matching, we construct treatment and control groups that are nearly identical in terms of pretreatment balance of demographic and behavioral predictors of high tax morale. We find improved perceptions towards tax compliance in Pakistan for respondents exposed to the TV and newspaper advertisements. The choice of the advertisement’s delivery device is important since the latter is more effective. Our findings provide empirical evidence that well-timed mass media campaigns can enhance voluntary tax compliance. © 2017 Elsevier Inc. All rights reserved. 1. Introduction Can media campaigns enhance individual attitudes towards vol- untary tax compliance (hereinafter tax morale) in Pakistan? The country’s stagnant tax revenue to GDP ratio despite two decades of tax policy reform could be explained by factors typicallly associ- ated with low tax morale (Martinez-Vazquez and Cyan, 2015; Bird et al., 2008). Some simple statistics clearly reflect the gravity of the problem. From 2000 to 2014, registered personal income taxpayers in Pakistan increased from 0.75 to 3.6 million but only 0.98 mil- lion were active filers out of 56.5 million employed individuals of whom 5.7 million earned income above the exemption threshold. 1 The purpose of this paper is to examine the effectiveness of mass media campaigns via newspaper and TV ads in enhancing * Corresponding author. E-mail addresses: cyan@gsu.edu (M.R. Cyan), koumpias@umich.edu (A.M. Koumpias), jorgemartinez@gsu.edu (J. Martinez-Vazquez). 1 IMF (2016); p 5. tax morale in Pakistan. We use a novel, individual-level survey of randomly selected Pakistani income tax filers that was conducted after the media campaigns eliciting attitudes towards tax compli- ance. Because only a fraction of the survey respondents was ex- posed to the media campaigns, the survey gave rise to a natural experiment. While survey participants were randomly selected into the sample, they self-selected into or out of treatment. This was due to their differential propensities to read a newspaper ad and watch a TV ad. This generated non-random variation in treatment assignment in a single tax filing season given a random sample of income tax filers. We use this source of cross-sectional variation to identify the effect of ads aired on TV, printed in the newspapers, or both on tax morale by comparing respondents exposed to the media campaign to those not exposed to it. We employ a quasi-experimental design that combines random survey sampling of the individuals and post-treatment matching techniques. First, our identification strategy exploits randomly se- lected survey respondents that offer a representative sample of in- come tax filers across geographical regions of Pakistan. Then, it http://dx.doi.org/10.1016/j.socec.2017.07.004 2214-8043/© 2017 Elsevier Inc. All rights reserved.