BEHAVIORAL INSIGHTS FOR TAX COMPLIANCE: A Whole-House Approach to Improving Taxpayer and Tax Official Beliefs, Attitudes, and Behaviors s TA X © 2025 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The bound- aries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissem- ination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@ worldbank.orgmailto:pubrights@worldbank.org. Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be con- sidered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank, therefore, does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. 4 BEHAVIORAL INSIGHTS FOR TAX COMPLIANCE: A Whole-House Approach to Improving Taxpayer and Tax Official Beliefs, Attitudes, and Behaviors. POLICY NOTE - Jun 2025 s CONTENTS Acknowledgements  7 I. Context: the domestic revenue mobilization challenge  8 II. A behaviorally informed approach to tax compliance  10 1. An overview of the behavioral approach to tax compliance  12 2. A conceptual framework for improved compliance  18 III. How Can Behavioral Insights Solve Tax Compliance Challenges?  21 1. Diagnostics for understanding tax compliance challenges  21 2. Solutions iteration for improved compliance: The behavioral taxpert’s toolkit  23 2.1 What evidence exists for the nudge?  28 2.2 What evidence exists for the budge?  33 2.3 What evidence exists for the trudge?  37 III. Looking forward: expanding the use of behavioral insights in revenue administration  40 1. What have we learned from the literature and World Bank experience?  40 2. What direction should behavioral insights take for revenue administration in the future?  41 IV. References  44 V. Annex: the approach and experience at the World Bank  48 1. How the eMBeD team supports improved compliance  48 2. Case studies  51 Case Studies  51 Glossary  84 ACKNOWLEDGEMENTS This policy note was prepared by Jonathan Karver (Economist, eMBeD Unit, Development Impact), Ana Maria Rojas (Extended Term Consultant, eMBeD Unit, Development Impact) and Fatima Karout (Short-Term Consultant, eMBeD Unit, Development Impact). The note is a com- panion and update to “Behavioral Insights for Tax Compliance” (Dalton et al., 2021). The current policy note was prepared to provide a broader description of an approach to tax compliance informed by behavioral science, based on recent experience by the World Bank Mind, Behavior and Development (eMBeD) Unit as well as developments in the literature. This policy note is a product of many useful conversations, insights and invaluable feedback from several colleagues at the World Bank and beyond. The authors are incredibly grateful for inval- uable feedback from Benu Bidani, Renos Vakis, Ana María Muñoz Boudet, Zeina Afif, Chris Hoy, Thiago Scot, Jorge Luis Castañeda, Abigail Dalton, Laura Zoratto and Anne Brockmeyer. It has also benefitted tremendously from conversations with internal and external colleagues including Oyebola Okunogbe, Mahvish Shaukat, Mario Negre, Alan Fuchs, Marco Hernandez, Ceren Ozer, Stephen Davenport, David Palmer, Jonas Fallov, Artur Tamazian, Anahit Basentsyan, Gabriel Tourek, Guillermo Cruces, Kadri Jogiste, Njomza Ilazi, and Ramadan Islami, among many others. 8 I. CONTEXT: THE DOMESTIC REVENUE MOBILIZATION CHALLENGE Domestic revenue mobilization is a critical -and increasingly relevant -element of the develop- ment policy agenda. The increasing importance of domestic revenue mobilization in international development comes from a growing need for sustainable financing coupled with recent global trends, primarily impacting low- and middle-income countries. A declining trend in the tax-to-GDP ratio for low- and middle-income countries, estimated to have fallen by nearly 1 percent between 2010 and 2019, highlights the critical nature of this challenge. In many of these countries, the tax-to-GDP ratio has fallen below 15 percent, considered to be the minimum for a country to finance the basic functions of the state. This challenge puts pressure on sustainable sources of financing to support reaching the Sustainable Development Goals (SDGs) (World Bank, 2023). There is substantial potential for governments around the world to increase tax revenues to meet development goals, including through tax policy (e.g. tax rates and tax expenditures) and tax administration reforms (IMF, 2024). In most contexts, raising revenues requires an efficient, effective tax system (World Bank, 2019). Domestic revenue mobilization (DRM) -the generation of revenue from domestic activities -is highly dependent upon tax revenue, which in turn requires compliance to tax obligations among all segments of society and the efficient collection of taxes. But standard policy measures for improving taxpayer compliance and tax collection – such as reforming legislation and adminis- tration – may not be enough. They may be time-consuming or infeasible due to technology in- frastructure, limited resources, and political constraints. Given these limitations, there is typically not a silver bullet to improve tax compliance and collection. We define compliance as following legally established fiscal obligations of individuals and businesses, including registration, filing, and payment, and define collection as the government’s ability to ensure the collection of tax liabilities to fund government activities. Governments need sustainable funding to fulfill their social contract with citizens, but ensuring an efficient and effective tax system to secure this funding is a challenge everywhere. Domestic revenue mobilization is at the center of the social contract because citizens expect quality public goods and services that improve individual and societal welfare in return for paying taxes (Dalton et al., 2021). For this reason, tax compliance is crucial for the social (namely, fiscal) contract: wide- spread non-compliance of tax obligations reduces a government’s ability to provide and improve societal welfare, weakening the social contract. Moreover, if the resources captured through taxes are not used effectively or if civil servants do not comply with their own obligations to serve the public, the sense of responsibility from citizens will diminish and compliance will drop, creating a vicious cycle (Dom et al., 2022). As such, effective revenue collection requires diverse tools beyond traditional strategies. Compliance challenges take many forms, from individual or businesses registration with the tax authorities, to accurate reporting of their activity, filing, and payment of their obligations. As will be discussed, the compliance journey starts with registration and ends with (accurate and timely) Behavioral Insights for Tax Compliance 9 payment, but there are many intermediate steps where compliance challenges will emerge. Each of these create compliance challenges in the form of informal workers or businesses, non-filing taxpayers, inaccurate declarations, late filers and payment delinquency. These tax compliance challenges translate into revenue collection challenges, ultimately reducing domestic revenue mobilization and limiting the scope for government spending. A thorough understanding of behavior and how individuals interact with policies and programs can provide creative and cost-effective solutions to these challenges. Insights from behavioral science (henceforth behavioral insights) are those that emphasize the importance of context for decision-making and behavior: namely, the social, psychological, and economic factors that influence decisions and actions. They address details in bureaucracies, technologies, and service delivery that are sometimes overlooked. Moreover, they help policymakers avoid decision traps and biases, sparking innovation and low-cost solutions. This policy note provides evidence from the global experience testing solutions applying behavioral insights (either explicitly or implicitly) in the field to improve tax compliance as well as an overview of World Bank experience in this area. These are robust evaluations of tax administration policy and strategy that consider social, psychological, and economic factors that influence taxpayer decision-making. Global experience (well-documented in the literature) highlights the role of behavioral science in many of the innovative solutions that have been tested in low- to high-income countries. These interventions show that a deeper understanding of behavior can help individuals and firms better process information, make decisions, and submit their tax declarations accurately and on time, and help tax administration officials improve the way they do their jobs (as collectors and public servants). Complementary studies carried out by the Bank in collaboration with revenue authorities and other partners in Costa Rica, Guatemala, Poland, Latvia, Kosovo, Indonesia, Albania and Armenia demonstrate how context-specific, behaviorally informed messaging can offer an immediate, low-cost solution to bureaucratic and technological barriers. At the center of behavioral science for improved tax compliance is an evidence-based approach, in particular one harnessing micro tax data to understand tax compliance more thoroughly. A behaviorally-informed approach to tax compliance is a data-driven exercise relying on both insights from fieldwork and administrative tax data. Administrative tax data is indispensable for the design, implementation and evaluation of behaviorally informed solutions, which requires a deep investment in micro tax data. As a result, there are important complementarities and spill- overs from the behavioral approach and an evidence-based approach using micro tax data. For example, low cost, behaviorally-informed interventions (which tend to be popular with revenue administrations) can provide an important entry point for a broader collaboration on evidence based policymaking as well as motivate access to data that can be valuable for studying broad- er policy questions around tax policy, tax administration, and even non-tax policy (e.g., labor market analysis). The overarching contribution of behavioral insights into revenue administration is to provide complimentary and low cost (financial and political) solutions to improve the efficiency and effectiveness of revenue collection. Behavioral insights are not meant to replace traditional 10 tools -such as tax rates, IT systems or risk models -but rather to compliment these by making their design and implementation more efficient and impactful as well as by introducing parallel strategies meant to strengthen them. The approach centers around holistic engagement with taxpayers and the population more generally, considering direct and indirect engagement to build awareness and educate, facilitate, change norms and narratives, and increase the credibility of enforcement actions. The objective of this note is to summarize the role of behavioral science in tax compliance and tax administration, elaborating on a previous note outlining the World Bank’s past work in this area (Dalton et al., 2021). This note will provide an overview of a behaviorally informed approach to tax compliance and highlight how behavioral insights can be useful for diagnosing and confronting tax compliance challenges (section II), provide thoughts on a way forward for integrating behavioral insights into revenue administration, summarizing evidence from the field (section III), and present the approach and experience of the World Bank in this space (Annex). The note will place particular emphasis on the experimental and quasi-experimental literature on behaviorally-informed solutions to improve compliance and collection and group these into three generalized tools: the nudge (modifications to the decision environment), the budge (elevating new narratives around tax compliance), and the trudge (taking an inward-looking approach to improving compliance). We also developed a Practitioner’s Guide where we provide detailed “how-to” on applying behavioral insights for revenue collection.1 II. A BEHAVIORALLY INFORMED APPROACH TO TAX COMPLIANCE The application of behavioral science has grown exponentially over the past decade, with governments adopting it in various ways. Behavioral science has taken hold globally at local and central government levels, appearing in various shapes and sizes with similar objectives in mind. How the behavioral science agenda develops at the country level and across institutions is context-specific and can emerge from a sectoral or a sub-national focus (Afif et al., 2018 and Manning et al., 2020). Despite these differences, applications cover a similar set of theories and tools centered around the notion that a thorough understanding of human behavior can greatly improve the design and implementation of policy. The behavioral insights approach acknowledges that standard development policy often only focuses on financial resources, incentives, laws, and information provision and not enough on mindsets, decision-making frames, and the social environment. Public policy tends to be designed under the assumption that beneficiaries or agents behave fully rationally (i.e., making a fully informed decision based on expected costs and benefits, and the utility this provides 1 https://documents.worldbank.org/en/publication/documents-reports/documentdetail/099342007172539760 Behavioral Insights for Tax Compliance 11 them) and these agents have a full set of information to act rationally. The traditional models of tax compliance based on expected utility theory part from the assumption that taxpayers are rational and act selfishly to optimize their utility (e.g., see Allingman and Sandmo, 1972 and Yitzhaki, 1974) minimizing the amount of taxes they pay given the likelihood of getting caught (audit, fines, or worse). However, this assumption implies that tax authorities have complete information about taxpayers and the resources to enforce compliance when evasion occurs. Traditional policy tools can benefit from a context-specific, theory-driven, and evidence-based approach to understanding a problem, drafting potential solutions, testing these, and scaling up what works best (World Bank, 2015). One of the key assumptions driving the relevance of behavioral science for policy is that agents do not make fully rational decisions, but rather make decisions automatically, socially, and using mental models. Behavioral economics introduced the notion of bounded rationality, wherein de- cisions are made as rationally as possible given cognitive, time and other constraints; satisfying, rather than optimizing, utility. Behavioral insights further elaborate on this notion by highlighting that people think and act automatically in various ways and use mental shortcuts (also known as heuristics) that facilitate decision-making. People also think and act socially. They follow es- tablished norms and trust information coming from trusted social contacts and networks. Lastly, people think and act with mental models (everyone’s idea of how the world works); their unique worldview is shaped by history and where they see themselves within it. Insights from behavioral science can be effective at addressing complex policy challenges, including domestic revenue mobilization (DRM). Behaviorally informed policy emphasizes the importance of context for decision-making and behavior that goes beyond standard policy in- struments such as reforms to legislation or public administration. It evaluates and incorporates individual, social, and structural factors that influence decision-making (World Bank, 2015). In doing so, it addresses features in bureaucracies, technologies, and service delivery that are sometimes overlooked. Moreover, it helps policymakers avoid decision traps and biases, spark innovation, and identify solutions to specific policy problems. In the context of domestic reve- nue mobilization, and tax compliance in particular, these insights can be used to influence tax compliance and informality decisions and actions (Dalton et al., 2021). Integrating the behavioral lens into tax compliance helps explain and improve a society’s fiscal contract, which centers around the mutual trust between taxpayers and governments (Timmons, 2005). The Innovations in Tax Compliance framework (Prichard et al., 2019 and Dom et al., 2022) outlines the three main areas of focus to build strong fiscal contracts, each of which benefits from a behavioral lens: enforcement, facilitation, and trust. Enforcement refers to the tools to identify, coerce, and/or punish non-compliers. Facilitation revolves around removing transaction costs to improve service delivery (i.e., make it easier to comply).2 Trust explains the motivation to comply at various levels and is linked to tax morale (the intrinsic and extrinsic willingness and motivation 2 E-filing and simplified declaration processes (simplified forms and auto-filling of information) are two common examples of this. 12 to pay taxes).3 Behavioral insights can be used to improve each of these revenue administration “functions” by focusing on improving decision-making frames, making information more salient and credible, and improving interactions between taxpayers and tax officials, among others. Evidence is at the center of an informed understanding of human behavior, and policymakers can harness this to help solve complex development challenges. Attitudes, beliefs, and mindsets are central to how policymakers confront these challenges, and understanding their relationship to behavior is key to finding effective and impactful solutions. Evidence not only tells us what works, but also how it works, why it works, and for whom. A greater understanding of challenges and potential solutions through evidence can greatly improve policy design and implementation. Many policies are designed using inaccurate assumptions of beneficiaries or agents subject to a policy rather than using evidence, weakening their impact. Evidence can be used to reject weak assumptions and strengthen policy. The application of behavioral science to improved revenue administration in particular shows the promise of this evidence-based approach for identifying improvements in policy design and implementation. The existing evidence-based4 literature, as well as a wealth of observational and survey-based studies, suggest that a more holistic approach to taxpayer behavior can improve compliance substantially. Integrating behavioral insights into revenue administration is not new, but it has developed substantially over time. Effective behavioral science strategies to improve tax com- pliance include making tax declaration and payment process easier to navigate (facilitation), incorporating tax morale and social incentives to comply and encourage others to as well (trust), and implementing deterrence measures more effectively, such as making the private costs of evasion more salient (enforcement) (Dalton et al., 2021). The rest of this section provides an overview of the behavioral approach applied to tax compli- ance while introducing concepts, tools, and methods that are required to design and implement behaviorally informed activities to improve compliance. This section also provides a selected review of the relevant literature on behavioral science applied to tax compliance. 1. AN OVERVIEW OF THE BEHAVIORAL APPROACH TO TAX COMPLIANCE Taxpayers, both individuals and businesses, are subject to psychological and behavioral biases that policymakers should consider when aiming to improve tax compliance. As discussed above, taxpayer rationality is based on many assumptions that are not supported by empirical evidence (see Pudney et al., 2000; Dhami and Al-Nowaihi, 2007). As it turns out, taxpayers demonstrate a 3 Trust can be decomposed into elements of fairness (are tax systems fairly and competently designed and administered?), equity (does everyone pay their fair share?), reciprocity (do taxes translate to public goods and services?) and accountability (are tax authorities and related agencies accountable to taxpayers?). 4 Here, we define evidence-based literature as that which incorporates impact evaluation methods; namely experimental and quasi-experimental methods to identify the causal impact of interventions on behaviors (the gold standard being ran- domized controlled trials). See https://dimewiki.worldbank.org/Randomized_Control_Trials Behavioral Insights for Tax Compliance 13 spectrum of decision-making tendencies, influenced by various cognitive biases that can lead to inattentive decision-making. As such, the decision to comply with tax obligations is not a discrete one based on a simple estimation of costs and benefits. There are both structural and behavioral barriers to compliance, some of which fall outside of the rational taxpayer model of compliance, that appear at different stages. By structural barriers, we mean those pertaining to financial, legal, or informational constraints to compliance. By behavioral barriers, we mean those pertaining to social, psychological, cognitive and contextual constraints to compliance. Many of the former materialize as biases that influence the way taxpayers make decisions and take actions. While these biases are generally common across all taxpayers, an important distinction can be made between individuals and businesses since each will face unique challenges to comply with their obligations. Individual taxpayers -such as employees, self-employed or individual entre- preneurs- and business taxpayers are subject to different processes and tax regimes as well as a different set of incentives to comply. Individuals generally file and pay only once per year if it all, since in many countries employees are not required to submit their own declarations unless they have additional non-wage income. Businesses, on the other hand, file regularly, on a monthly or quarterly basis, depending on the type of business and the taxes to which it is subject. Both types of taxpayers can outsource their obligations to an accountant or tax preparer, though the involvement of an individual will differ from the involvement of a business representative, which can be the owner, an in-house accountant, an external accountant or some other employee within the firm. These factors will influence how compliance decisions will be made at various stages. The simplest way to show the taxpayer decision-making process is to illustrate the journey that they must navigate to fully comply with their obligations. Figure 1 below presents a stylized representation of the individual taxpayer journey showing that tax compliance involves several complex decisions and actions surrounded by both structural and behavioral barriers. Taxpayers, especially new ones, must understand the process (step 1), but this could be complicated by a lack of accessible and clear information about obligations (both a structural and behavioral bottle- neck). Once they have full information, they then need to make an active decision to declare and thus, eventually pay their taxes. However, passing this stage could be influenced by low salience of reciprocity (“I pay my taxes and get something valuable in return”), social norms around tax compliance (“why should I pay my taxes if no one else does?”), or the credibility of enforcement actions from the revenue authorities (another structural and behavioral bottleneck). Assuming a taxpayer wishes to comply, they then need to act on it by declaring their obligations. This may entail interacting with declaration platforms that could be confusing, complex, or frustrating, potentially creating information or bandwidth overload. In many cases, reaching this point is not the end; declaring is one action, and paying is another. Payments can be complicated by choice overload, procrastination, or simply liquidity (financial) constraints. Box 1 summarizes many of these biases that influence taxpayer decision-making. 14 Figure 1. The individual taxpayer journey from the behavioral science perspective TAX COMPLIANCE PERCEPTION OF CONS EQUENCES Social, psychological, and WILL I GET PUNISHED economic factors influencing IF I DON’T PAY?? taxpayer decision-making EIVED SOCIAL NORMS PERC DO OTHERS PAY THEIR TAXES? K OF KNOWLEDGE LAC SALIENCE OF RECIPROCITY LOW AM I REQUIRED TO DECLARE MY TAXES? WHAT DOES THE GOVERNMENT OR FRAMING DO WITH MY MONEY? PO WHAT EXACTLY DO I NEED TO DO? Understanding Deciding Tax Processes to Pay Taxes Source: Dalton et al. (2021) Behavioral Insights for Tax Compliance 15 LIQUIDITY CONSTR AINT BANDWIDTH OVERLO S AD CAN I AFFORD TO PAY HOW DO I PAY? MY TAX BURDEN TODAY? PROCESS COMPLE XIT PROCRASTINA TIO N Y HOW MUCH DO I OWE? CAN I DELAY MY PAYMENT? INFORMATION OVERLO CHOICE OV AD ERLO AD WHERE DO I PAY? WHICH IS THE BEST PAYING OPTION? Declaring Paying Taxes Taxes 16 A similar journey can be outlined for businesses, with additional steps and bottlenecks that merit attention. Business taxpayers (those in accounting regimes) will need to engage in bookkeeping to keep track of relevant information about transactions with businesses and consumers, as well as deductible costs associated with their operation. At this stage, businesses may face high levels of cognitive burden, especially new ones or those newly “graduated” into accounting regimes. This is unlikely to be the case for small businesses in presumptive or simplified tax regimes, who are typically taxed on sales/turnover/gross sales, and thus are not required to keep detailed accounts of their business activity. This stage of the taxpayer journey typically involves a decision around working with an external accountant, hiring an accountant for the firm, or having someone within the firm engage in this task. The journeys outlined above are of course not independent of tax officials, who will appear at key interaction points as well as have their own journey to follow. In each stage of the taxpayer journey, there are potential interactions with tax officials (in this case, taxpayer facing tax officials such as those working in service centers, call centers, or field offices) that can facilitate or com- plicate progression to subsequent steps. For example, bad interactions with tax officials in the first two stages can erode trust and influence declaration and payment decisions. Alternatively, having no access to a “live” interaction with a tax official might discourage taxpayers from gath- ering information about their obligations. Beyond these interactions, we can also imagine a tax officer journey that is also subject to decisions, actions and bottlenecks. Different types of taxpayer-facing tax officers (field staff, call center staff, service center staff, among others) will need to make decisions about who to visit (which taxpayers to prioritize), how to engage with these, and how to respond to taxpayer requests. These decisions and actions will be influenced by various biases such as confirmation bias and implicit bias that encourage decision-making that might ignore reality or prejudice certain groups (Lourenço et al., 2022). Behavioral Insights for Tax Compliance 17 Box 1. Common behavioral and psychological barriers to compliance • LACK OF KNOWLEDGE: when taxpayers do not know why, when, where, or how taxes should be paid. • POOR FRAMING: when tax information is presented in a confusing or overwhelm- ing way, which can lead taxpayers to misunderstand or ignore their obligations, thus reducing compliance. • PERCEIVED SOCIAL NORMS: when taxpayers believe others are not tax compliant and might default on their debts as well. • LOW SALIENCE OF RECIPROCITY: when taxpayers are not perceiving or under- standing the direct benefits they receive in return for their tax contributions, which can diminish their motivation to comply. • BANDWIDTH OVERLOAD: when taxpayers face too many decisions or informa- tion at once, leading to decision fatigue and reduced ability to focus on completing tax-related tasks effectively. • PROCESS COMPLEXITY - HASSLE FACTORS: when taxpayers find that the process of paying taxes has unclear steps, leading them to abandon the process all together. • INFORMATION OVERLOAD: when taxpayers solicit information about tax collec- tion, the different types of taxes and due dates might make it harder for them to make an actionable plan to pay taxes. • INTENTION-ACTION GAP DUE TO PROCRASTINATION: when taxpayers might intend to pay taxes but leave it for later. • CHOICE OVERLOAD: when taxpayers are presented with too many payment op- tions or decision points, which can lead to paralysis in decision-making and poten- tially delay or prevent tax compliance. Additional barriers that may emerge along the journey map: • LACK OF SALIENCE: taxpayers might not be reminded of tax payment deadlines in their day to day, nor when the deadline is near. • MENTAL MODELS: taxpayers might perceive that the chance of getting caught for not paying taxes is low. • PRESENT BIAS: taxpayers might prioritize using funds for immediate needs instead of paying taxes that could benefit them in the long term. • MISTRUST: towards authorities and low satisfaction with public service delivery due to perceptions of institutional corruption. 18 2. A CONCEPTUAL FRAMEWORK FOR IMPROVED COMPLIANCE The journey described above highlights challenges at various stages of taxpayer compliance, but also hints at entry points for policy action, which can be summarized through a stylized conceptual framework around improved compliance. The Innovations in Tax Compliance (ITC) Enforcement, Facilitation and Trust (EFT) framework (Prichard et al., 2019, Dom et al., 2022) described below provides a useful tool to understand the elements of tax compliance that can serve as binding constraints or enablers to improved compliance. Traditional revenue authorities’ strategies focus on enforcement and facilitation, so the framework places additional weight on the factors that explain trust. Figure 2 below highlights the EFT elements, while disaggregating trust into sub-components that are each relevant for improved compliance (fairness, equity, rec- iprocity, accountability). The EFT framework treats trust as the fundamental factor behind tax morale (alongside norms and values) and posits that improvements in trust can boost tax morale through increased political support for tax reform. Figure 2. The Enforcement, Facilitation, and Trust Framework ENFORCEMENT POLITICAL EXPANDED SUPPORT FACILITATION TAX FOR REFORM COMPLIANCE TRUST Fairness Equity Reciprocity Accountability Source: Adapted from Prichard et al. (2019) and Dom et al. (2022) Behavioral Insights for Tax Compliance 19 Combining the behavioral approach and the EFT framework, we can develop a stylized theory of change aimed at enhancing tax compliance to reinforce the fiscal and social contract. This theory emphasizes behavior changes in both taxpayers and tax officials, requiring mutual guarantees: taxpayers will increase compliance with a heightened sense of responsibility, civic pride, and trust in government, while tax officials will become more transparent and efficient in response to societal demands and revised perceptions of taxpayers. Informality complicates this challenge, involving not just the taxpayer-authority relationship but also consumers who prefer informal markets, creating a barrier to formalizing business activities. Addressing informality, which encompasses both tax evasion by known taxpayers (intensive margin) and activities by unregistered entities (extensive margin), requires a comprehensive approach that considers taxpayers and consumers as agents of change. The deeper exploration of why taxpayers decide and act the way they do (diagnostics) and the testing of solutions to compliance challenges (interventions) we discuss throughout the remainder of this policy note address both tax compliance and informality as two sides of the same coin. Interventions to improve tax compliance over the past decade have both explicitly and implicitly leveraged behavioral insights in their design, implementation, and evaluation. A useful review of the evidence is Antinyan & Asatryan (2024), who conducted a meta-analysis of 71 randomized controlled trials applying “nudging” approaches to tax compliance, finding that effects from these experiments are small in magnitude but meaningful, with the largest impacts found from deter- rence-based language meant to influence perceived audit probabilities and the private costs of noncompliance (fines). Another example of this is Alm et al. (2023), who explicitly focus on the role that behavioral insights have played in improving tax compliance. While these reviews point to the limitations of behavioral insights (mostly in terms of the nudging approach), many studies show how key elements from behavioral science can be incorporated into tax compliance/evasion models with sustainable impacts, with evidence from low- to high-income countries (Hallsworth, 2015; Hallsworth et al., 2017; Rees-Jones, 2018; Cranor et al., 2020). The extensive literature points to meaningful and typically cost-effective impacts of targeted communication from tax authorities, highlighting how the effectiveness of different approaches (e.g., language applied, channels used) is highly sensitive to the context (a recent policy insight note from J-PAL5 highlights the benefits of reminder messages to improve tax compliance). Evidence from multiple World Bank projects in collaboration with revenue authorities and other partners since 2014 highlights the value added of strategic notifications or nudges while emphasizing how behaviorally informed framings can have a significant impact on various measures of compliance among individuals and businesses. The case study section in the Annex provides country case studies summarizing the impact of these interventions. 5 https://www.povertyactionlab.org/policy-insight/improving-tax-compliance-through-reminder-messages-taxpayers 20 Behavioral Insights for Tax Compliance 21 III. HOW CAN BEHAVIORAL INSIGHTS SOLVE TAX COMPLIANCE CHALLENGES? Solutions harnessing insights from behavioral science require a deep understanding of the context, challenges, and barriers or enablers to compliance; typically, this means that one- size-fits-all policies do not fit-in to an approach informed by behavioral science. The value of a behavioral insights approach is that it starts with analyzing the context, both in terms of the target population (e.g., considering barriers and solutions for specific segments of beneficiaries) and the setting (macroeconomic, geographic, cultural, etc.). The benefit of this approach is that solutions are tailor made to each specific circumstance. The cost is that this requires a deeper understanding of the challenge (who it impacts most, what sustains it, and what possible levers can confront it). For this reason, behavioral insights are both about diagnostics and solutions iteration. Solutions must be considered for the specific target population, which can be some variation of an individual taxpayer, a business taxpayer, a tax official, or even a consumer (as a citizen rather than a taxpayer per se). While there are broad sets of solutions that can work for all types of taxpayers, in most cases these need to reflect the peculiarities of the taxpayer segment or target audience. As discussed in section II, individual and business taxpayers face unique journeys and thus a unique set of decisions, actions, and barriers to overcome compliance challenges. As such, diagnostic activities must explore the different barriers individual and business taxpayers or tax officials face. Solutions need to reflect the unique taxpayer environment in various ways. On the one hand, how information is presented and framed (and what messages are prioritized) will depend upon the type of taxpayer. For example, an individual taxpayer is likely to be much more receptive to messages about reciprocity than a business, since the individual might benefit from public goods more or differently than a business. On the other hand, the individual exposed to the solution will likely dictate the type of response. For example, an intervention targeting businesses that primarily outsource their tax obligations will need to be designed with this lack of connection between the decisionmaker and the entity (and its owner) in mind. 1. DIAGNOSTICS FOR UNDERSTANDING TAX COMPLIANCE CHALLENGES A behaviorally informed approach to tax compliance starts with problem definition; that is, a concrete statement of the compliance challenge we would like to address. Often, the default starts and ends with a policy outcome rather than a policy challenge (e.g., working forward from “revenue collection in country X is low”). Behavioral insights emphasize the need to make the context the fundamental piece to understand a policy challenge. Increased granularity in the problem definition will reveal concrete bottlenecks and impactful solutions. Consistent with the journey mapping approach outlined above, a good rule of thumb is to formulate the problem 22 statement along the lines of what action or behavior is desired from a taxpayer or citizen; for example, by filling in the following blanks: “I would like [agent X] to engage in [activity Y].” Once a problem has been adequately framed, the next step is to engage in diagnostic activities to understand the scope of the challenge and the factors that sustain it. Diagnostic activities are a key component of a behaviorally informed approach to improved revenue administration. In the tax compliance space, these can range from literature reviews and analysis of secondary data to key informant interviews, focus group discussions, in-depth interviews, non-participant observations and quantitative surveys. Such fieldwork diagnostics help understand and map the beliefs and attitudes of a specific segment of taxpayers or citizens and how these might influ- ence decisions and actions. Taxpayers, for example, can be probed about their expectations of the behavior of others, the difficulty or ease of carrying out various compliance-related actions, beliefs about the fiscal contract, levels of trust in relevant authorities, experience with revenue authority engagement, and preferred communication channels. These can be explored with survey data to understand the magnitude of beliefs and attitudes to get a sense of the magnitude of certain types of behaviors, and qualitatively to understand what is behind beliefs and attitudes.6 Administrative data diagnostics are another fundamental component of a behavioral diagnostic: these center around an exploration or thorough analysis of taxpayer filing and payment data to identify observable predictors of problematic behavior among existing taxpayers. Finally, systems assessments allow for the identification of internal barriers or frictions to improved revenue outcomes (such as those related to the filing environment). Diagnostic activities will help identify the binding constraints or enablers of the revenue collection challenge, which could be structural, behavioral or both. It is difficult to understand what to solve without understanding the binding constraints to the desired behavior. Fieldwork, administrative data diagnostics and systems assessments shed light on the binding constraints to the desired behavior, which will lead to entry points for innovative solutions. While behavioral science can identify highly impactful solutions to societal challenges, it is also a highly successful approach to managing complex challenges by addressing individual bottlenecks to a desired behavior. As such, it also works towards chipping away at the policy challenge in a connected way. A detailed review of behavioral diagnostic activities can be found in the Practitioners’ Guidebook, section 3. 6 Refer to the partitioners guide to see examples of quantitative and qualitative instruments. https://documents.worldbank. org/en/publication/documents-reports/documentdetail/099342007172539760 Behavioral Insights for Tax Compliance 23 2. SOLUTIONS ITERATION FOR IMPROVED COMPLIANCE: THE BEHAVIORAL TAXPERT’S TOOLKIT In this note, we classify solutions into three sets of tools that encompass various behavioral science elements with the ultimate objective of improving tax compliance, increasing revenues, and boosting the efficiency of tax authorities: (i) the nudge (“to touch or push gently”7) revolves around strategic notifications to taxpayers or changes in the filing environment to alter a spe- cific behavior and/or reduce sludge (frictions in the choice architecture); (ii) the budge (“to move, shift, or give way”8) revolves around awareness raising and holistic engagement more generally to alter the dialogue and existing narratives; and (iii) the trudge (“to walk or march steadily and usually laboriously; a long tiring walk”9) revolves around interventions directed at civil servants working on DRM that change mindsets, reduce biases and change their behaviors. These tools complement traditional administration and policy levers available to revenue administrations and ministries of finance. In many cases, these solutions are traditional but include insights informed by behavioral science. These tools are defined to be progressive in nature; that is, solutions un- der each are meant to be progressively more profound, focusing on short-term (e.g., nudges) to longer-term (e.g., budges and trudges) approaches. In particular, the nudge has typically been used as a one-off, short-term solution, whereas the budge and trudge can be complex and require more time to materialize. As such, these tools should not be viewed as independent options but rather three progressively increasing options in terms of impact and sustainability. The nudge is an effective tool used across time and space to improve tax administrations’ com- munications and engagement with taxpayers directly and indirectly (e.g., in the user experience and decision making process), leading to more honest and timely tax declarations and payments. Nudging is best described as an intervention to alter the choice architecture (i.e., how information or options are presented to us) to promote certain behaviors over others without significantly changing economic incentives (Thaler and Sunstein, 2009). In this context, communicating or presenting information to taxpayers about their obligations to induce voluntary compliance rath- er than enforcing compliance through audits, fines, or (the threat of) business closures framed in specific ways, has shown positive impacts. Most existing literature has focused on strategic, targeted communication to taxpayers to induce a change in declaration behavior, including ex- amples from the World Bank (summarized in the Case Study section). Nudging goes beyond reminders and notices, particularly in how tax authorities engage with taxpayers indirectly; for example, through facilitated compliance. A smaller, but important ap- proach revolves around the declaration process itself. How declaration processes are presented to taxpayers can serve to enable or dissuade taxpayers from carrying out their obligations fully. This approach typically focuses on simplifying filing processes and facilitating compliance through design elements, particularly through e-filing and simplified or prepopulated10 tax forms, and 7 https://www.merriam-webster.com/dictionary/nudge 8 https://www.merriam-webster.com/dictionary/budge 9 https://www.merriam-webster.com/dictionary/trudge 10 Fonseca & Grimshaw 2017. 24 addressing common manifestations of sludge (discussed in more detail below). The budge assumes that beliefs, norms, and behaviors cannot be changed through targeted communications alone and require a more holistic approach to changing the existing narrative. While impactful, simple nudges are less likely to generate systematic changes in repeated be- havior as they aim to facilitate a behavior in the short term (Brandon et al., 2017). As such, they must be part of a wider strategy targeting the fiscal contract to improve tax compliance or more broadly the social contract to increase institutional trust across government, potentially having an indirect impact on tax compliance. Trust-building interventions identify and test engagement strategies that alter the dialogue around the fiscal contract, raising awareness about the benefits of compliance (and costs of non-compliance) in a way that challenges existing beliefs. A recent set of interventions in Kosovo led by the Tax Administration of Kosovo (TAK) -with support from the World Bank eMBeD team -tested this approach.11 Between 2020 and 2021, TAK implemented a series of targeted tax morale interventions, from an online competition for middle school students about the importance of taxes in society and the costs of informality, to multimedia content by local influencers and cartoonists to confront established attitudes about tax compliance and informality.12 Such interventions can raise awareness on the benefits of compliance, influence tax morale, and influence ongoing dialogue at the societal level. The trudge acknowledges that changing the behavior of taxpayers is not enough, but also requires a shift in the behavior of public servants working in the tax space. Trust building is a two-way street where values, attitudes, beliefs, and norms of tax authorities also need to be addressed. In fact, civil servants are also subject to similar factors influencing citizen’s behavior (Lourenço et al., 2022), as discussed above in the taxpayer journeys. Both agents need to hold their end of the fiscal contract, which means that the decisions and actions of policy professionals can be equally important as the behavior of taxpayers. Tax collectors, effectively the frontline workers of tax administrations, are frequently the first and only face-to-face interactions of taxpayers with the tax authority. As such, how they engage with taxpayers can significantly influence trust, tax morale, and compliance (OECD, 2019). This process can be enhanced through capacity building for tax administration staff, including by confronting their assumptions about taxpayer behavior and their own biases.13 The underlying objective of these exercises is to change attitudes and improve behavior that supports improved efficiency in tax administration and ultimately an improved col- lection of domestic revenue. As such, the trudge is centered around institutional strengthening. The three set of tools can and should be applied throughout the taxpayer and tax officer jour- neys, and can build off each other for more meaningful and sustainable impact. Nudge type interventions are valuable across the entire taxpayer journey and can also be relevant for tax 11 Implemented in partnership with the World Bank and Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). 12 These interventions are outlined in an activity report available under P169627. The activities were a follow up to those covered under the report in Hernandez, et al. 2019. A similar intervention was implemented in Armenia, showing the nexus between taxes and government investments. https:/ /blogs.worldbank.org/governance/nudging-under-pressure-behavior- al-insights-tax-compliance-during-global-pandemic 13 The eMBeD team has supported multiple tax administration teams over the past three years in Kosovo, Tanzania, Georgia, Kenya, Mauritius and Pakistan, with additional capacity building under development. Behavioral Insights for Tax Compliance 25 officers before or while they engage with taxpayers. Budge type interventions can be introduced both at initial taxpayer stages (when a taxpayer is considering registering, for example) as well as in the final stages (when the decision to pay, and how much, is made). Trudge type interventions can be embedded within all stages of the taxpayer journeys (to improve interactions) as well as within the tax officer journey. Moreover, these tools are not mutually exclusive: in many cases, nudges and budges can be combined, for example, within the same intervention (e.g., the World Bank’s experience with calendars in Indonesia, described in the Case Study section). An illustra- tion of the three tools is provided in Figure 3 below. 26 Figure 3. Putting the nudge, budge and trudge into perspective NUDGE BUDGE • What is the tool? • What is the tool? Strategic notifications and alternations Broad societal engagement (awareness in the taxpayer choice architecture. raising, education). • Who are we targeting? • Who are we targeting? Taxpayers (registered and unregistered). Taxpayers and society as a whole. • What are we targeting? • What are we targeting? Filing and payment behaviors. Beliefs and attitudes (first order); registration, filing and payment (second order). Example: Example: Letter reminders to taxpayers highlighting Social media campaign to confront the benefits of compliance. misperceptions around the tax system. Behavioral Insights for Tax Compliance 27 TRUDGE • What is the tool? Targeted actions to improve institutional functions. • Who are we targeting? Tax officials (tax/customs officers at all levels of revenue administration). • What are we targeting? Officer beliefs and attitudes, decisions and actions having a direct impact on taxpayers. Example: Debiasing tax officers to improve fairness of inspections 28 2.1 What evidence exists for the nudge? The nudge approach is the most common and most documented solution harnessing behavioral insights for tax compliance, partly due to its ease of implementation and the quick evidence it can generate. Many of the systematic and meta-analytic studies have focused on the impact of nudging on tax compliance (e.g., Alm et al., 2023 and Antinyan & Asatryan, 2024). However, most of these are centered around strategic notifications (e.g., reminder and notification letters). We define nudges as any alterations in how information is presented directly to taxpayers that can fall under one of three groups: (i) strategic notifications, (ii) alterations in the compliance choice architecture and (iii) sludge reductions. We discuss evidence from each below. Strategic notifications can serve many purposes within the nudge a. Strategic framework, which depend on when, how, and with what type of notifications language these are presented. Strategic notifications are meant to either remind taxpayers of their obligations, or flag suspicious (e.g., underreporting or income/sales) or non-compliant (e.g., failure to submit a tax declaration) behavior. These can be delivered in a variety of ways -postal letter, e-mail, SMS, phone call, or in-per- son -and the decision of the communication channel will depend upon taxpayer preferences and institutional capacity. Many rev- enue administrations are moving to fully digital communication, such that letters are delivered digitally by e-mail or e-filing plat - form messages (OECD, 2023). The tested communication chan- nels have highlighted the context-specific nature of these notifica- tions, particularly when thinking about the notification credibility. For example, in Colombia, in-person visits were more effective at motivating delinquent taxpayers than emails and letters (Ortega and Scartascini, 2015). In North Macedonia, deterrence messages increased compliance among non-compliant property taxpayers (Benković et al., 2023). Similarly in Guatemala, behavioral inter- ventions with deterrent and social norms messages significantly increased tax compliance (Kettle et al., 2016). Lastly, in Rwanda, visual nudges, particularly those focusing on deterrence, improved compliance attitudes and behavior (Santoro and Mascagni, 2022). The content and framing of these notifications are where behavioral insights are most evident in past studies; messages can take a softer (motivational) or stronger (enforcement) tone. As outlined in Dalton et al. (2021), strategic notifications leverage different insights, and these can be tested against one another to identify the key enablers of compliance behavior among specific segments of taxpayers. For example, notifications can: Behavioral Insights for Tax Compliance 29 • Help taxpayers avoid bandwidth overload by simplifying the messages taxpayers receive: Reducing compexity makes it easier for citizens to comply with what might be an otherwise overwhelming pro- cess. This is true for both communication strategies and the tax declaration process itself. For example, using easy-to-understand images alongside messaging with information about the tax process. This strategy was tested in Rwanda providing better information to taxpayers (Mascagni et al., 2017). Similarly, a central London local authority tested the impact of message simplification on tax payment rates, finding that simplification increased the number of people paying by four percentage points (John and Blume, 2018). • Help taxpayers identify the clear steps needed to comply with tax declarations: The Behavioural Insights Team deployed simplified letters to increase tax compliance among doctors and dentists in the United Kingdom. The letters, which delineated the actions re- quired of an individual, saw a 15 to 30 percent higher response rate than other types of messaging (Behavioural Insights Team, 2012). Similarly, in a an RCT in Papua New Guinea, an intervention involved sending SMS messages to taxpayers, which included reminders of the due dates for tax declarations, as well as information about the public benefits of paying taxes (Hoy et al., 2024). Specifically, one treatment provided firms with a reminder about what they were required to do to comply with tax legislation. This reminder SMS message was sent 10 days before Salaries and Wages Tax (SWT) declarations were due and again 3 days before the due date. Both nudges doubled the number of on-time SWT declarations made by zero filers, who are taxpayers that had filed a declaration in the previous 15 months and claimed they had no SWT to pay because they have no employees. • Harness social norms and introduce social comparisons around timely and honest compliance: People are social creatures and highlighting the behavior of others can have a strong influence. When a field experiment in the United Kingdom invoked social norms, such as referencing the average number of people who pay on time, pay- ment rates for overdue taxes increased. The payment of declared tax liabilities was boosted by up to 5.1 percentage points after 23 days of delivery, compared to a control group (Hallsworth et al., 2017). Similarly, the decision not to comply may be further influenced by feelings of unfairness. As individuals will be less inclined to pay one’s taxes if they believe others are not paying their fair share, strategies can focus on demonstrating that others are complying (Frey & Torgler, 2007; Bazart & Bonein, 2014). Moreover, if a firm is 30 to adapt their behavior, it may be necessary for them to understand how similar firms act. In Bangladesh, when an intervention informed firms that details of their compliance would be shared with others, VAT payments increased. This is true for firms in clusters where at least 15 percent were complying at baseline. A similar exploration found that an SMS message threatening to publish non-compliant firm names online increased filings significantly (Chetty et al., 2014). Similarly, participants received social norm messages focusing on descriptive and injunctive norms, with approval-framed and disapproval-framed messages, resulting in a 2 percent increase in reported taxes. This suggests that emphasizing the moral approval or disapproval of certain behaviors can be an effective strategy in promoting compliance. • Emphasize the inherent value of paying taxes: Helping taxpayers see how their actions can positively impact their own community may make a difference. One way to do this is by specifying what public goods one’s taxes are used for. An experiment showed that public service messages in Rwanda that emphasized how taxes help provide education, healthcare, and safety led to persistent tax compliance (Mascagni et al., 2017). Similarly, in Democratic Republic of the Congo, citizens in neighborhoods where tax payments were enforced were more likely to attend townhall meetings and demand public goods and accountability. As the enforcement mechanisms signaled state capacity, taxpayers felt more confident that there would be returns to their compliance (Weigel, 2018). This was also tested in Papua New Guinea (Hoy et al., 2024) where flyers were sent along with a standard letter from the tax administration to taxpayers. The flyers included information about how tax revenue helps provide access to free education for over 2 million children and free healthcare for all citizens. The flyers were designed to draw upon a public-benefit motivation for why taxpayers should comply with tax regulations. This intervention primarily increased the number of SWT and VAT declarations made by zero filers, who are taxpayers that previously reported they had no employees in the case of SWT and those who reported that their revenue stream was below the exemption threshold or were of an exempt class in the case of VAT. • Make suspicions and consequences more salient to increase the cred- ibility of enforcement: Strategic notifications taking a “hard-tone” approach are the most common framing in the literature, likely reflecting more traditional approaches by revenue authorities to elevate the consequences of non-compliance and salience of Behavioral Insights for Tax Compliance 31 enforcement reach. Signaling an increase in enforcement mecha- nisms can make it feel risky and costly for an individual to engage in non-compliance. For example, U.S. taxpayers were more likely to accurately self-report their taxes when they perceived a higher likelihood of being audited. In Denmark, those that were to be au- dited with certainty were significantly more likely to comply (almost double) compared to those with a 50 percent audit probability. In Uruguay, providing firms with details on past audit statistics and penalties increased compliance, and in Chile, notifying firms that they are being monitored and may be audited ultimately increased VAT payments (Slemrod et al., 2001; Kleven et al., 2011; Bergolo et al., 2017; Pomeranz, 2015). Other insights less tested • Elevate inclusion and participatory decision-making on spending priorities: Evidence from lab and lab in the field studies suggests that compliance may increase when taxpayers have a say in how tax revenue is spent, including through the provision of a feedback channel (Torgler, 2004). • Appeal to a taxpayer’s moral compass: People prefer to think of them- selves as honest and moral. This may be why individuals are less comfortable when their unethical behavior is described as a choice rather than simply a failure to take action (Ritov & Baron, 1990). Framing non-compliance as intentional has increased compliance among taxpayers, likely by reminding taxpayers of their desire to be a good citizen. • Anchoring tax liability amounts to address cognitive burden: Calculating tax liabilities might be complicated for businesses, espe- cially smaller or newer businesses without a dedicated accountant. New evidence from Kenya (Hoy et al., 2024) shows that providing an anchor for tax liability amounts tailored to specific business clusters (sector and location) can serve to improve compliance. An RCT identified an increase in the amount of tax paid by 12 percent relative to a control group. Careful attention should nonetheless be paid to the equity implications of such exercises, since these can result in relatively higher taxes being paid by smaller taxpayers (as the study in Kenya found). 32 Alterations in the taxpayer choice architecture and sludge assess- b. Alterations in the ments, while considered very important in the literature, are as- compliance choice sociated with few robust studies showing their causal impact. architecture and While strategic notifications are a form of targeted communica- sludge reductions tion to taxpayers to motivate compliant behaviors, alterations in the choice architecture are meant to influence taxpayers’ deci- sions at key decision-making junctures (e.g., on filing systems). As such, they nudge taxpayers when these are making decisions and taking actions and can serve to reduce sludge. These can be decisions made in face-to-face interactions or through digital or traditional platforms. Such interventions are typically directed at individual or small business taxpayers who would be more sen- sitive to changes in taxpayer choice architecture (compared, for example, to larger businesses who likely work with accountants). In many cases, these interventions focus on improving the use of technology to improve taxpayer services and thus compliance (a useful summary is presented in Okunogbe and Tourek, 2023). Some common examples of this are: • Digitizing filing (e-filing): E-filing can reduce transaction costs as- sociated with filing (paper filing can be time-consuming, require accessing or printing a large number of pages, and requires deliv- ering or presenting this paperwork to revenue authorities) as well as simplify procedures by reducing errors in entering information. In Tajikistan, the introduction of e-filing among businesses led to a reduction in the time spent on taxes by 40 percent and doubled the amount of taxes paid among firms previously more likely to evade (though has the opposite effect on firms less likely to evade in the past) (Okunogbe & Pouliquen, 2022). In a separate study in Uganda, e-filing was estimated to increase the tax base in the presumptive tax system as well as collection (when combined with complimentary taxpayer education initiatives) (Jouste et al., 2021). • Simplifying tax forms: As discussed previously, tax forms are generally complicated to fill out, at times even for tax accountants, and can be burdensome and frustrating to complete, especially for small and medium-sized businesses (OECD, 2015 and Benzarti, 2020). Simplifying these by streamlining fields or introducing formulas in digital forms to reduce errors can improve compliance by improving accuracy and motivation. Presenting information on these (whether they are tax declarations or tax bills) can serve the same purpose: for example, an experiment conducted in Belgium highlighted the positive impact of improving the presentation of tax bills by income tax taxpayers (De Neve et al., 2021). Behavioral Insights for Tax Compliance 33 • Pre-populating tax declarations: The prefilling or pre-populating of tax forms can address default bias, the tendency to stick with choices made by others (in this case, the revenue authority), and can curb evasion by signaling the monitoring capacity of a revenue administration (Hesami et al., 2024). For example, pre-populating values using third-party data (employer declarations, information from financial institutions, etc.) can serve to reduce the likelihood of claiming deductions when these should not be taken (Fochmann et al., 2021). Experimental evidence from lab studies highlights a positive impact conditional on the pre-populated amounts being accurate (see, for example, Fonseca & Grimshaw, 2017; van Dijk et al., 2020; and Fochmann et al., 2021 and Doxey et al., 2021, Schwebke et al., 2024 and Schwebke, 2024 in the US context). 2.2 What evidence exists for the budge? The budge, while being a highly used solution to improve compliance, has been experimentally less tested in the literature, partly given the difficulties in evaluating larger-scale campaigns. The budge is concerned with communications and awareness raising in a way that changes the narrative around compliance and informality across segments of society. We classify budges into two groups: (i) awareness-raising interventions and, (ii) transaction-based interventions. The former broadly targets multiple stakeholders (taxpayers and citizens more generally) or specific groups while the latter engages with actors in economic activity (e.g. consumers, and vendors). Transactions-based interventions are typically considered traditional tools to improve compliance such as enforcement. Yet, behavioral insights can boost their impact, for example, by using tax lotteries as opportunities to confront the narrative around compliance. We discuss evidence from both below. Awareness raising from the behavioral science perspective treats a. Awareness-raising communications more holistically as engagement, allowing for interventions two-way communication between stakeholders. Such activities are generally concerned with addressing beliefs and attitudes that lead to problematic compliance behaviors in the hope that a change in these will lead to a change in behavior. Traditionally, these have revolved around behaviorally informed communication campaigns (sometimes referred to as SBCC, or social and behavior change communication). Beyond the traditional (S)BCC approach, innova- tive budge interventions have revolved around taxpayer services and education. Some specific examples of the budge are: • Behaviorally informed communication campaigns: BCC type inter- ventions can serve to raise awareness through different channels, 34 highlighting how taxes finance critical public goods and services and their overall role in society. Such campaigns can communicate details of how taxes finance highly valued public services and prime house- holds and businesses for their priorities in public service delivery to increase a sense of inclusion and citizen engagement. The commu- nications can also highlight which taxpayers face the highest burden to signal the tax system’s progressivity. Revenue administrations globally implement diverse communication and awareness initiatives to enhance taxpayer compliance. African nations commonly organize National Taxpayer Days and annual campaigns, which include vari- ous components like internal recognition of revenue authority staff, moral appeals, and efforts to raise awareness about the importance of taxation. Mozambique’s annual campaign, “Together We Make Mozambique”, launched in 2010, aimed to familiarize citizens with tax matters, particularly targeting the informal sector, while Senegal hosts a range of activities targeting the informal economy during campaign days, including conferences, information booths, media coverage, and involvement of high-ranking officials to underscore the significance of tax compliance (Mascagni and Santoro, 2018). The World Bank’s own work in Indonesia (described in the Annex) highlights how communications campaigns can incorporate specific behavioral insights to elicit compliance through a nudge+budge approach (distributing calendars). • Educational interventions: Educational interventions can serve two purposes: to fill knowledge gaps of existing or future taxpayers in terms of their obligations or elevate the role of taxes and the importance of tax compliance (also directed at current and future taxpayers), building on the SBCC type interventions described above. Taxpayer education initiatives globally are well document- ed, covering everything from school competitions directed at future taxpayers to edutainment style interventions for current and potential taxpayers (see OECD 2021) and more recently har- nessing different social media platforms to promote positive tax culture (Araki & Claus, 2014), including the World Bank’s recent experience in Kosovo. Taxpayer education is both about helping taxpayers navigate complex tax systems to address the cognitive effort needed to comply (Mascagni et al., 2024) as well as building a strong tax culture (OECD, 2021). A recent quasi-experimental study from Rwanda found that educational training of recently registered taxpayers (educating on the basic elements of the tax system) had a positive impact of various measures of compliance (Mascagni et al., 2024). One innovative area that has recently made headlines is Behavioral Insights for Tax Compliance 35 the “Mr. Taxes”14 initiative in Colombia, where the current director of the revenue authority (DIAN), Luis Carlos Reyes, uses social me- dia platforms (especially TikTok) to answer questions from citizens, including frequently humorous or non-sensical questions that are provided with quasi-serious responses. Social media engagement around these social media posts has been substantial. While not evaluated, such initiatives can prove impactful, particularly in build- ing trust and making tax obligations more relatable. • Social recognition and public disclosure: Social recognition of com- pliance or public disclosure of non-compliance can change the narrative around tax compliance and informality by challenging and potentially replacing existing dialogue within society or at a minimum among taxpayers. In Pakistan, a quasi-experimental study found that public disclosure and social recognition of compliance improved tax payments meaningfully at the federal level (between 9 and 17 log points, respectively) (Slemrod et al., 2022). A quasi-experimental study in Slovenia found that shaming tax evaders led to a reduction in tax debt by 8.5 percent (Dwenger and Treber, 2018). However, caution is necessary, as illustrated by a study conducted in Uganda, which revealed backfire effects (Manwaring and Regan, 2023). They found that efforts to shame tax evaders reinforced non-compliant behavior in certain communities, leading to a decrease in tax com- pliance among those who felt unfairly targeted. Social recognition has focused on rewarding taxpayers with the largest amount of taxes paid, though this typically means rewarding very large businesses and multi-national corporations representing a small proportion of taxpayers. Ex-ante, the expectation would be that social recogni- tion can be more effective if the compliant behavior of small and medium-sized businesses were elevated (since this would induce more peer comparison across taxpayers). Transaction-based interventions to improve tax compliance can b. Transaction-based be thought of as boosted traditional tax interventions that ad- interventions dress both structural and behavioral bottlenecks. Transactions- based interventions introduce stimuli into everyday transactions between stakeholders that involve some element of taxation (e.g., business-to-business transactions or business-to-consumer trans- actions). These interventions aim to change the attitudes and be- 14 https://www.eltiempo.com/politica/gobierno/la-historia-de-mr-taxes-el-director-de-la-dian-que-es-fenomeno-en-re- des-3325154 36 haviors of the various agents in the direction of positive norms and habits about tax compliance. • Consumer tax lotteries: Consumer tax lotteries incentivize consum- ers to demand that transactions with businesses get fiscalized (i.e., the transaction is properly recorded and any sales taxes due are charged and properly accounted for). These incentives indirectly place consumers in the role of auditors given that it is in their best interest to ensure fiscal receipts are provided with all purchases (and thus a paper trail created). Lotteries can also change the nar- rative around tax compliance and informality by making consumers aware that each transaction has a fiscal component and challenging existing norms that tolerate tax evasion by businesses. As such, tax lotteries can shift consumer preferences towards formal businesses and formalized transactions, reflecting an increase in tax morale. The only documented example of a robust lottery evaluation was from a consumer lottery conducted in Sao Paolo, Brazil, where the lottery was associated with a 9.3 percent increase in revenue collection net of prizes (Naritomi 2019). Tax lotteries can be even more impactful if incentives are provided to both consumers and businesses since the latter must also be convinced that fiscalizing transactions is in their best interest. Transaction-based interventions represent an area for further ex- perimentation within the budge approach, including considering sticker campaigns, buying formal initiatives, and taxpayer seals of approval. While consumer tax lotteries have been implemented extensively (albeit few rigorously), other opportunities exist to harness business-to-consumer transactions to shift the narrative around tax compliance for consumers and businesses. Point-of- sale interventions are common in the public health space -for ex- ample, product placement and messaging to encourage healthier decision-making by consumers in purchasing food products -but similar interventions have not been tested in the tax compliance space. One example is sticker campaigns on cash register machines reminding consumers and store clerks of the importance of fis- calized sales. Another example is “buy formal” campaigns encour- aging consumers to shop at registered and compliant businesses. Similar to social recognition interventions, this approach leverages “seals of approval” for businesses to place outside of their store- front to communicate to consumers that they are recognized for their compliance. Behavioral Insights for Tax Compliance 37 A final area of interventions that could be included under the budge is the “one-stop-shop” approach to taxpayer services, which aims to facilitate compliance through education and service de- livery. One-stop-shops can further be used to educate, inform, and raise awareness about tax compliance, playing an important role in changing the narrative around compliance and informali- ty. One example of this approach can be found in Jamaica (Small and Brown, 2020), where taxpayer service provision in the form of dedicated support to large taxpayers (through Large Taxpayer Offices) through a variety of services (tax advisory services, pro- cessing of tax compliance certificates, stamping of documents, facilitating tax seminars and workshops, filing and payment re- minders, filing and payment facilitation, and registration and rec- onciliation of tax accounts) had mixed effects on tax compliance. The provision of taxpayer services had a positive impact on filing and payment compliance for the general consumption tax (GCT) in Jamaica, where enforcement is stronger. However, this effect was not observed for the corporate income tax (CIT), which has weaker enforcement, suggesting that the strength of enforcement in the tax regime may interact with service delivery to influence compliance. In Indonesia, the creation of medium taxpayer offices (MTOs) found meaningful impacts on compliance (more than dou- bling revenue over a nine-year period), partly explained by their role in facilitating compliance (Basri et al., 2021). 2.3 What evidence exists for the trudge? The trudge places a unique emphasis on improving compliance by targeting taxpayer-facing tax officers accountable for low compliance. Civil servants working on revenue administration, generally referred to as tax officers, can become barriers to compliance. For example, binding constraints to compliance originating from tax officers relate to lack of capacity, lack of financial incentives, or the existence of various biases (e.g., beliefs about certain groups), among others (Lourenço et al., 2022). Typically, trudge interventions directed at tax officers are designed to improve beliefs, attitudes, and behaviors related to taxpayer registration and tax collection, either directly through incentives or indirectly through debiasing/sensitization or training on behavioral science. In other words, the trudge are nudge and budge type interventions directed at tax of- ficers rather than taxpayers and citizens. Among the three tools, the trudge has the least number of examples, partly given the complexity of these interventions (difficult to ensure institutional support) and partly given the difficulty in introducing randomization. Below are some examples of trudge interventions (while most literature in this area is not explicitly interested in behavioral insights, such interventions can be adapted in various ways with insights from behavioral science): 38 • Debiasing and training of tax officers: Debiasing or sensitization re- volves around confronting beliefs and attitudes held by tax officers that could erode trust or create opportunities for collusion and corruption, both potentially having a negative impact on compliance (Devas et al., 2001). Training tax officers can also focus on focus on sensitizing these towards a facilitation (service-oriented approach) and trust-building approach through the building of socio-emotional skills and de-biasing. A recent experimental study conducted across Latin America showed that problem solving and decision-making can be significantly improved with a short training on behavioral science (Rojas & Scartascini, 2024). A field experiment conducted in Ghana to assess the impact of ethics and integrity training on police officers, which aimed to re-activate intrinsic motivations and create a new shared identity, was found to significantly improve officers’ values, beliefs, and attitudes, as well as reduce their pro- pensity to behave unethically (Harris et al., 2024). This suggests that such training can be effective in promoting ethical behavior among tax officers. • Assignment of functions: Problematic tax officers’ beliefs, attitudes, and behaviors can be addressed by pairing tax officers in taxpay- er-facing activities or by offering choice in the assignment of posts. In the Democratic Republic of the Congo (DRC), an experiment on the assignment of tax officers based on their ability found that pair- ing high (low) performing tax officers with each other and assigning these to high (low) propensity to pay properties yields the largest returns in terms of compliance. The authors estimate that optimal assignment can lead to a 26 percent increase in revenue relative to random assignment of teams and properties (Bergeron et al., 2022). In Pakistan, a study conducted in Punjab found that offering performance incentives in the form of post selection (rewarding the highest performing tax officers with their preferred posting) had an impact on property tax revenue of 40 and 31 percent in the first and second years after the intervention, respectively (Khan et al., 2019). While both examples are primarily concerned with structural solutions (ability-based assignments), such interventions show promise as strategic pairing can reduce the expectations and experience of bribery or other unfair practices towards taxpayers, non-transparent enforcement actions, and a general low trust in revenue authorities. • Social and financial incentives: Like the example above on the as- signment of functions, tax officers can be incentivized to perform better with the introduction of social and financial incentives that Behavioral Insights for Tax Compliance 39 reward their behavior in different ways. In a separate study con- ducted in Punjab, Pakistan, financial rewards to high performing property tax officers led to a 46 percent higher growth in revenue collection compared to control areas (Khan et al., 2016). While dif- ferent treatment groups were included to test the marginal impact of a scheme that tied performance pay to taxpayer satisfaction, the incentive based solely on revenue collected performed the best (with no meaningful change on taxpayer satisfaction). While most evidence in the tax space focuses on financial incentives, there is room for experimentation on social incentives as well. A World Bank study in Nigeria found that recognizing exemplary behavior of civil servants in health facilities improved the performance of their peers15. Such an approach could also be applied to taxpayer facing officers in revenue administrations. 15 See https://documents1.worldbank.org/curated/en/440211517949747866/pdf/123249-eMBeD-Nigeria-HC-Brief.pdf 40 III. LOOKING FORWARD: EXPANDING THE USE OF BEHAVIORAL INSIGHTS IN REVENUE ADMINISTRATION 1. WHAT HAVE WE LEARNED FROM THE LITERATURE AND WORLD BANK EXPERIENCE? The evidence from the past decade across regions and countries shows that the role of behavioral insights for improved revenue administration is context-specific. Contextual factors influencing the impact of behavioral insights fir tax compliance signal sociocultural differences across coun- tries and contexts on the one hand, and implementation realities on the other, reflecting both institutional structures and the underlying social contract. The published academic literature highlights the clear value added of behavioral insights for improved compliance and discusses the sustainability of the nudge approach and the role of enforcement vs. non-enforcement-based messages and engagement (Antinyan & Asatryan, 2024 & Hallsworth 2014). In particular, Atinyan & Asatryan (2024), in their recent meta-analysis, identify average impacts of 2.7 percentage points from simple reminders on the probability of compliance and additional impacts of 1.4 and 3.2 percentage points from tax morale and deterrence messaging (respectively). The experience from World Bank technical assistance projects has shown meaningful impacts from behaviorally informed solutions on on-time filing and payment as well as revenue collection more generally (Dalton et al., 2021; Case Studies in the annex). The sensitivity of impacts to contextual realities highlights how pivotal these factors are for the design, implementation, and evaluation of solutions. Despite the positive evidence, many studies signal potential pitfalls that mediate the impact of behaviorally informed interventions, and these should be considered carefully in their design and implementation. One clear example of this is that messaging might be incredibly effective in one context but ineffective (potentially even backfiring) in others. Harnessing national pride, reciprocity in the tax system, or potential enforcement actions might fall short if national sentiment is heterogenous, the reach of public expenditures is limited, or institutional or policy constraints limit enforcement actions that revenue administrations can take. Even worse, empty threats could worsen compliance in the long run given its impact on the credibility of enforcement. These interventions also show that practice makes perfect; that is, the impact of interventions will depend upon the quality of data and the ability to experiment and learn from it. As was outlined in the case of Kosovo (Hernandez et al., 2019), the quality of administrative data (in particular, contact information) and the use and preference for certain communication channels highlight the importance of diagnostics. It will also depend on the target audience. For example, for businesses, the person receiving the message might not fully represent the target businesses’ best interests or might represent multiple businesses, which means that the framing and impact evaluation design need to be considered carefully (e.g., by clustering treatment at a level higher than the business, such as the accountant). Behavioral Insights for Tax Compliance 41 Despite well-documented challenges, revenue authorities are well-positioned to harness behav- ioral insights to help taxpayers understand the tax process and the importance of their tax-re- lated choices. Revenue administrations have the legal authority to assess, levy, and collect taxes. They can use behavioral insights to provide taxpayers with information and education about the tax declaration process. For example, context-specific messaging may motivate taxpayers about the process in a timely, accessible way. The tax authority can iterate on the content of these messages and mechanisms of message delivery, from letters to simple technologies. Doing so can ultimately close the intention-action gap for individuals and firms, reminding them to submit their declarations accurately and on time. Alternatively, behavioral insights might support revenue administrations in their use of traditional tools, such as compliance risk management and audit functions. In many contexts, audits and enforcement actions are viewed as unfair. Behavioral insights can be used to better inform and increase transparency (and the sense of fairness) around the actions being taken and why taxpayers are being pursued based on their behavior. Revenue authorities can improve quality tax systems by reducing sludge and harnessing be- havioral informed choice architecture. Behavioral insights can motivate tax administrators to update their systems, for example, by addressing intentional and unintentional sludge. This will help them better communicate with non-compliers. A commitment to reducing implementation challenges, such as data quality issues, can become self-reinforcing should the tax authority recognize that effective solutions require strong systems. These strong but adaptive systems should be the norm. Lastly, emphasizing tax revenue collection should be a priority for governments, but addressing compliance, which is low in many places, should not be devalued. In low and middle-income countries, many people do not just submit their tax declarations late; they may not submit them at all. As compliance involves action by the taxpayer, it is something that needs to be motivated. Efforts to increase government revenues should focus on how to reach non-compliers, those at risk of non-compliance, and previous delinquents. While the short terms gains from a compli- ance-oriented approach might be muted relative to more structural actions, these will serve to improve revenue collection in the aggregate over the long-term. 2. WHAT DIRECTION SHOULD BEHAVIORAL INSIGHTS TAKE FOR REVENUE ADMINISTRATION IN THE FUTURE? While behavioral science cannot single-handedly solve critical policy challenges in domestic revenue mobilization, it can and should be used to increase our understanding of taxpayer behavior and to identify impactful and sustainable solutions. Behavioral insights incorporate social, psychological, cognitive, and contextual factors into our understanding of various policy challenges and elevate the role of evidence-based decision-making in policy. Behavioral science provides a useful framework to approach policy challenges from problem definition to solutions iteration, paying close attention to the decisions, actions, and interactions across various stake- holders. Revenue administrations can and should continue to harness behavioral insights when confronting complex compliance challenges. 42 Behavioral insights for tax compliance should continue improving and testing nudging strategies for taxpayers but should also generate more evidence on how to harness budge and trudge interventions to increase revenue collection. Nudges are often a proof of concept to apply be- havioral insights within revenue administrations; they are low-cost, generate quick evidence, and can be implemented with minimal institutional disruption. However, they typically only address the tip of the iceberg, which should motivate greater use of the budge and the trudge. Behavioral insights have been used extensively beyond the nudging approach in many policy spheres outside of taxation (e.g., health, education, and employment, among others). These approaches should be considered carefully for tax compliance (and revenue administration more broadly). Alm et al. (2023) elevate the potential benefits of a boosting approach; one that aims to equip individ- uals with the necessary tools to make decisions that benefit them and society as a whole. The boosting approach overlaps with our budge approach, changing how people make decisions in the first place. Alm and coauthors also provide useful examples that could overcome some of the limitations of traditional nudging approaches: for example, a “public goods clock” providing more concrete information about reciprocity than is typically included in strategic notifications (e.g., generic messages about taxes being used to finance important public goods). One important area of further exploration is introducing non-traditional messengers to improve compliance and shift the narrative around compliance in a meaningful way (one promising example is Mr. Taxes in Colombia, described above). If trust in traditional actors is low, budges might be more impactful and sustainable if the messengers themselves are part of civil society (something that was piloted by the World Bank in Kosovo). Behavioral insights should aim to create a stronger link with broader evidence-based policymaking within revenue administrations. A clear benefit of the behavioral approach to tax compliance is that it introduces principles of evidence-based policy design in government in a meaningful and generally painless manner. Deeper exploration of administrative data contributes to improved data quality; in particular, greater use of data helps revenue administrations identify gaps in collection and quality of data, as well as the usefulness of non-tax data to improve compliance monitoring. This improved quality in the analysis of administrative data can serve to benefit government policy analysis around taxes specifically and other priority areas more generally. The recently established DaTax initiative at the World Bank highlights the benefits of this approach and the pioneering role tax administrations play in the transition towards evidence-based policymaking. Behavioral insights for improved revenue administration will need to emphasize a shift from pilot approaches to scaled-up solutions. Scaling what works is needed for behavioral insights to have a lasting impact on revenue mobilization efforts. This will require more investment in under- standing how tested solutions can translate into persistent changes in revenue administration. Many of the proposed approaches can be understood as movements in the direction of scaling (institutionalizing certain aspects of behavioral insights for tax administration, such as recurring diagnostics, simplifications in the choice architecture and reminders to taxpayers prior to filing deadlines, to name a few). But the more innovative approaches are typically tested at a smaller scale and/or only once, leaving room for improvement to find solutions that fundamentally change the way revenue administrations function. Behavioral Insights for Tax Compliance 43 The solutions outlined above should be carefully designed and prioritized to avoid declining impact or cost-effectiveness as they expand. This pitfall is easy to imagine; for example, with reminders to taxpayers before due dates, especially if these serve to shift intentions to comply temporarily through targeted language around enforcement or tax morale. These can be avoided with a replication of pilots, careful selection of target groups and behaviors, and a clear under- standing of what can be sustained over time at the institutional level (List, 2022). As behavioral insights are harnessed for innovative solutions among revenue administrations, practitioners should reflect on these challenges early in the cycle and confront them accordingly. Lastly, for behavioral insights to be sustainable in revenue administration, institutional changes are required to incorporate these tools, and evidence-based policymaking systematically. As summarized in Afif et al. (2018) and Manning et al. (2020), governments across the world have developed unique approaches to institutionalizing behavioral insights into various government functions, with some notable examples in the tax space (e.g., the Australian Tax Office). There is no one-size-fits-all approach, though revenue administrations should seriously consider creating initiatives to incorporate behavioral science into various functions and move towards a sustain- able strategy for bottom-up or top-down institutionalization. Stand-alone experimentation is useful to answer specific questions, but for these insights to permeate institutional functions and last through leadership and staff changes, in-house capacity should be developed and political will must exist to make behavioral science a fundamental element of revenue administration. These efforts should be part of a broader push to establish dedicated research units within rev- enue administration, who can use administrative tax data and fieldwork data more effectively to influence administration and policy actions. Such units might need to precede behavioral insights units, since research units can help establish a dedicated team of researchers who can work effectively with micro tax data, which is fundamental to approaching tax compliance with a behavioral insights lens. 44 IV. REFERENCES Afif, Z., Islan, W. W., Calvo-Gonzalez, O., & Dalton, A. (2018). Brandon, A. and Ferraro, Paul and List, John A. and Metcalfe, Behavioral science around the world: Profiles of 10 countries. Robert and Price, Michael and Rundhammer, Florian (2017). Do World Bank the Effects of Nudges Persist? Theory and Evidence from 38 Natural Field Experiments (March 2017). NBER Working Paper No. w23277. Allingham, M.G., Sandmo A. (1972). Income tax evasion: A theoretical analysis. Journal of Public Economics 1, 323-38. Brockmeyer, Anne, Spencer Smith, Marco Hernandez, and Stewart Kettle (2019). Casting a Wider Tax Net: Experimental Alm, James, Lilith Burgstaller, Arrita Domi, Amanda März, and Evidence from Costa Rica. American Economic Journal: Matthias Kasper (2023). Nudges, Boosts, and Sludge: Using New Economic Policy 11 (3): 55–87. Behavioral Approaches to Improve Tax Compliance. Economies 11, no. 9: 223. Chetty, Raj, Mushfiq Mobarak, and Monica Singhal (2014). Increasing tax compliance through social recognition. Antinyan, Armenak and Zareh Asatryan (2024). Nudging for Tax International Growth Centre Policy Brief 14/0658 August 2014. Compliance: A Meta-Analysis. Economic Journal. Cranor, T., Goldin, J., Homonoff, T. and Moore, L. (2020). Araki, Satoru, and Iris Claus (2014). A Comparative Analysis of Communicating Tax Penalties to Delinquent Taxpayers: Evidence Tax Administration in Asia and the Pacific. Asian Development from a Field Experiment. National Tax Journal 73(2), 331-36. Bank. Dalton, Abigail Goodnow; Manning, L.; Jamison, J.; Sen, I.; Basri, M. C., Dharmapala, D., & Shapiro, N. R. (2021). Tax Karver, J.; Castaneda Nunez, J.; Guedes, L.; Mujica Estevez, S. Administration versus Tax Rates: Evidence from Corporate (2021). Behavioral Insights for Tax Compliance. eMBeD brief. Taxation in Indonesia. American Economic Review vol. 111, no. Washington, D.C. : World Bank Group. 12, December 2021 (pp. 3827–71). De Neve, Jan-Emmanuel, Clement Imbert, Johannes Spinnewijn, Bazart, Cécile, and Aurélie Bonein. (2014) Reciprocal Teodora Tsankova, and Maarten Luts (2021). How to improve relationships in tax compliance decisions. Journal of Economic tax compliance? Evidence from population-wide experiments in Psychology 40 (2014): 83-102. Belgium. Journal of Political Economy 129, no. 5 (2021): 1425- 1463. Behavioural Insights Team (BIT) (2012). Applying behavioural insights to reduce fraud, error and debt. UK Cabinet Office. Devas, Nick, Simon Delay, and Michael Hubbard (2001). Revenue Authorities: Are They the Right Vehicle for Improved Benković, S., V. Dmitrović, and M. Milosavljević (2023). Tax Administration? Public Administration and Development 21, Impact of Behaviour Nudges on Tax Compliance: Age as a no. 3 (2001): 211–22. Factor in Municipalities in the Republic of North Macedonia. Contemporary Financial Management, 2023. Dhami, Sanjit and Ali al-Nowaihi (2007). Why do people pay taxes? Prospect theory versus expected utility theory. Journal of Benzarti, Youssef (2020). How Taxing Is Tax Filing? Using Economic Behavior and Organization 64, 171-92. Revealed Preferences to Estimate Compliance Costs. American Economic Journal: Economic Policy 12 (4): 38–57. Dom, Roel; Custers, Anna; Davenport, Stephen R.; Prichard, Wilson. (2022). Innovations in Tax Compliance: Building Trust, Bergeron, Augustin, Pedro Bessone, John Kabeya Kabeya, Navigating Politics, and Tailoring Reform. World Bank Gabriel Z. Tourek and Jonathan L. Weigel (2022). Optimal Assignment of Bureaucrats: Evidence from Randomly Assigned Doxey, M., Lawson, J., Stinson, S. (2021). The effects of prefilled Tax Collectors in the DRC. NBER Working Paper 30413. tax returns on taxpayer compliance. The Journal of the American September 2022. Taxation Association 43 (2), 63–85. Bergolo, M, Ceni, R, Cruces, G, Giaccobasso, M and Perez- Dwenger, Nadja, and Lukas Treber (2018). Shaming for Tax Truglia, R (2017). Tax Audits as Scarecrows: Evidence from a Enforcement: Evidence from a New Policy. Hohenheim Large Field Experiment. NBER Working Paper No. 23631. Discussion Papers in Business, Economics and Social Sciences, No. 21-2018. Behavioral Insights for Tax Compliance 45 Fochmann, Martin, Nadja Müller, Michael Overesch (2021). Less International Monetary Fund (IMF) (2024). Stepping up cheating? The effects of prefilled forms on compliance behavior. Domestic Revenue Mobilization: A New Joint Initiative from the Journal of Economic Psychology, Volume 83, 2021, 102365. IMF and World Bank. Washington, DC: IMF and World Bank. Fonseca, M. and Grimshaw, S. (2017). Do behavioral nudges John, Peter, and Toby Blume (2018). How Best to Nudge in pre-populated tax forms affect compliance? Experimental Taxpayers? The Impact of Message Simplification and evidence with real taxpayers. Journal of Public Policy and Descriptive Social Norms on Payment Rates in a Central London Marketing. Local Authority. Journal of Behavioral Public Administration 1, no. 1 (February 26, 2018). Frey, Bruno S., and Benno Torgler (2007). Tax morale and conditional cooperation. Journal of comparative economics 35, Jouste, M., Nalukwago, M.I., Waiswa, R. (2021). Do tax no. 1 (2007): 136-159. administrative interventions targeted at small businesses improve tax compliance and revenue collection? Evidence from Ugandan administrative tax data. WIDER Working Paper Hallsworth, M. (2015). The use of field experiments to increase 2021/17. Helsinki: UNU-WIDER. tax compliance. Oxford Review of Economic Policy, 30(4), Pages 658–679. Karver, Jonathan, Shijaku, Hilda & Ungerer, Christoph (2022). Nudging in the Time of the Coronavirus : Evidence from an Hallsworth, Michael, John A. List, Robert D. Metcalfe, and Ivo Experimental Tax Trial in Albania at the Onset of a Global Vlaev (2017). The Behavioralist as Tax Collector: Using Natural Pandemic. Policy Research Working Paper Series 9961, The Field Experiments to Enhance Tax Compliance. Journal of Public World Bank. Economics 148 (April 2017): 14–31. Kettle, Stewart, Marco Hernandez, Simon Ruda, and Michael Harris, Donna, Oana Borcan, Danila Serra, Henry Telli, Bruno Sanders (2016). Behavioural Interventions in Tax Compliance: Schettini, and Stefan Dercon (2024). Proud to Belong: The Evidence from Guatemala. World Bank, Washington, DC. Impact of Ethics Training on Police Officers in Ghana. SSRN Scholarly Paper. Rochester, NY, May 15, 2024. Khan, Adnan Q., Asim I. Khwaja, and Benjamin A. Olken (2016). Tax Farming Redux: Experimental Evidence on Performance Hernandez, M. Jamison, J., Korczyc, E., Mazar, N. & Sormani, R. Pay for Tax Collectors. Quarterly Journal of Economics 131 (1): (2017). Applying behavioral insights to improve tax collection: 219–71. experimental evidence from Poland. Washington, D.C.: World Bank Group. Khan, Adnan Q., Asim Ijaz Khwaja, and Benjamin A. Olken (2019). Making Moves Matter: Experimental Evidence on Hernandez, M.; Karver, J.; Negre, M.; Perng, J. (2019). Promoting Incentivizing Bureaucrats through Performance-Based Postings. Tax Compliance in Kosovo with Behavioral Insights. Washington, American Economic Review 109 (1): 237–70. D.C. : World Bank Group. Kleven, H.J., Knudsen, M.B., Kreiner, C.T., Pedersen, S., Saez, E. Hesami, Siamand, Hatice Jenkins, and Glenn P. Jenkins (2024). (2011). Unwilling or unable to cheat? Evidence from a tax audit Digital Transformation of Tax Administration and Compliance: experiment in Denmark. Econometrica 79, 651–692. A Systematic Literature Review on E-Invoicing and Prefilled Returns. Development Discussion Papers 2023-14, JDI Executive Programs. Lawson, D.F., Stevenson, K.T., Peterson, M.N. et al. (2019). Children can foster climate change concern among their parents. Nature Climate Change 9, 458–462. Hoy, Christopher, Luke McKenzie, and Mathias Sinning (2024). Improving Tax Compliance without Increasing Revenue: Evidence from Population-Wide Randomized Controlled Trials in Papua List, John (2022). The Voltage Effect: How To Make Good Ideas New Guinea. Economic Development and Cultural Change 72, Great and Great Ideas Scale. New York: Crown Currency. no. 2 (January 2024): 691–723. Lourenço, Joana S., Renos Vakis, and Laura Zoratto (2022). Hoy, Christopher; Scot, Thiago; Oguso, Alex; Custers, Anna; Zalo, The behavioral professional: Improving decision-making and Daniel; Doino, Ruggero; Karver, Jonathan; Orgeira Pillai, Nicolas performance in the public sector. World Bank, Washington DC. (2024). Trade-offs in the Design of Simplified Tax Regimes: Evidence from Sub-Saharan Africa. Policy Research Working Manning, L., Dalton, A. G., Afif, Z., Vakos, R., & Naru, F. (2020). Paper; 10909. Behavioral science around the world volume II: profiles of 17 international organizations (English). eMBeD report. Washington, DC: World Bank Group. 46 Manwaring, P., and T. Regan (2023). Public Disclosure and Pudney S.E., Pyle D, Saruc T. (2000). Income tax evasion: An Tax Compliance: Evidence from Uganda. Centre for Economic experimental approach. In: MacDonald A, Pyle D., Illicit activity: Performance No. 1937 July 2023. The economics of crime and tax fraud. Ashgate: Aldershot. Mascagni, Giulia, Christopher Nell, and Nara Monkam (2017). Prichard, W.; Custers, A., Dom, R., Davenport, S., Roscitt, M. One Size Does Not Fit All: A Field Experiment on the Drivers (2019). Innovations in Tax Compliance : Conceptual Framework. of Tax Compliance and Delivery Methods in Rwanda. ICTD Policy Research working paper 9032. Washington, D.C. : World Working Paper 58. Bank Group. Mascagni, G. (2018). From The Lab To The Field: A Review Of Rees-Jones, A. (2018). Quantifying Loss-Averse Tax Tax Experiments. Journal of Economic Surveys, vol. 32(2), pages Manipulation. The Review of Economic Studies 85(2), 1251– 273-301, April. 1278. Mascagni, Giulia, Fabrizio Santoro, and Denis Mukama (2024). Ritov, Ilana, and Jonathan Baron (1990). Reluctance to vaccinate: Teach to Comply? Evidence from a Taxpayer Education Program Omission bias and ambiguity. Journal of Behavioral Decision in Rwanda. International Tax and Public Finance, January 28, Making 3, no. 4 (1990): 263-277. 2024. Rojas Méndez, Ana María, and Carlos Scartascini (2024). Michie, S., van Stralen, M.M., West, R. (2011). The behaviour Debiasing Policymakers: The Role of Behavioral Economics change wheel: a new method for characterising and designing Training. Inter-American Development Bank, April 5, 2024. behaviour change interventions. Implementation Science 6, 42. Santoro, Fabrizio, and Giulia Mascagni (2022). Visual Nudges: Naritomi, Joana (2019). Consumers as Tax Auditors. American How Deterrence and Equity Shape Tax Attitudes and Behaviour Economic Review 109, no. 9 (September 1, 2019): 3031–72. in Rwanda. Journal of Behavioral and Experimental Economics 107 (August 2022): 102118. OECD (2015). Taxation of SMEs in OECD and G20 Countries, OECD Tax Policy Studies, No. 23, OECD Publishing, Paris. Schwebke, Jason M. (2024). Does choice matter? The effect of filing method autonomy on taxpayer aggressiveness in a pre- filled tax return system. Journal of Accounting and Public Policy OECD (2021). Building Tax Culture, Compliance and Citizenship: 43 (2024) 107171. A Global Source Book on Taxpayer Education, Second Edition. OECD Publishing Schwebke, Jason M., William D. Brink, Victoria J. Hansen, and Charles F. Kelliher (2024). Pre-Populated Tax Returns: Individual OECD (2023). Tax Administration 2023: Comparative Taxpayer Adoption and the Effect on Compliance. Behavioral Information on OECD and Other Advanced and Emerging Research in Accounting 36, no. 1 (March 1, 2024): 85–103. Economies. Tax Administration. OECD. Slemrod, Joel, Marsha Blumenthal, Charles Christian (2001). Okunogbe, Oyebola, and Gabriel Tourek (2024). How Can Taxpayer response to an increased probability of audit: evidence Lower-Income Countries Collect More Taxes? The Role of from a controlled experiment in Minnesota. Journal of Public Technology, Tax Agents, and Politics. Journal of Economic Economics, Volume 79, Issue 3, 2001, Pages 455-483, ISSN Perspectives 38, no. 1 (February 1, 2024): 81–106. 0047-2727. Okunogbe, Oyebola, and Victor Pouliquen (2022). Technology, Slemrod, J., Rehman, O.U. and Waseem, M. (2022). How do Taxation, and Corruption: Evidence from the Introduction of taxpayers respond to public disclosure and social recognition Electronic Tax Filing. American Economic Journal: Economic programs? Evidence from Pakistan. Review of Economics and Policy 14, no. 1 (February 1, 2022): 341–72. Statistics, 104(1), pp.116-132. Ortega, Daniel, and Carlos Scartascini (2015). Don’t Blame Small, O., & Brown, L. (2019). Taxpayer Service Provision and the Messenger: A Field Experiment on Delivery Methods for Tax Compliance: Evidence for Large Taxpayers in Jamaica. Public Increasing Tax Compliance. IDB Publications, November 5, 2015. Finance Review, 48(2), 250-277. Pomeranz, Dina (2015). No taxation without information: Deterrence and self-enforcement in the value added tax. American Economic Review 105, no. 8 (2015): 2539-2569. Behavioral Insights for Tax Compliance 47 Sunstein, Cass (2022). Sludge Audits. Behavioural Public Policy, Weigel, Jonathan L. (2018). Taxation, corruption, and 6(4), 654-673. engagement with the formal state: Experimental evidence from the DR Congo. PhD dissertation., Harvard University, 2018. Thaler, Richard H., and Cass R. Sunstein (2009). Nudge: Improving decisions about health, wealth, and happiness. World Bank (2015). World Development Report 2015: Mind, Penguin. Society, and Behavior. Washington, DC: World Bank, 2015. Timmons, Jeffrey F. (2005). The Fiscal Contract: States, Taxes World Bank (2022). Behavioral Sciences Approach to and Public Services. World Politics, Vol. 57, pp. 530-567, July Empowering Women in Forest Landscape. World Bank: 2005. Washington, DC. Torgler, B. (2004). Cross‐culture comparison of tax morale and World Bank (2023). World Bank Support for Domestic Revenue tax compliance: evidence from Costa Rica and Switzerland. Mobilization. Independent Evaluation Group. Washington, DC: International Journal of Comparative Sociology 45(1–2): 17–43. World Bank. United Nations (2021). United Nations Behavioral Science Yitzhaki S. (1974). A note on income tax evasion: A theoretical Report. United Nations Innovation Framework analysis. Journal of Public Economics 3, 201-202. van Dijk, Wilco, Sjoerd Goslinga, Bart W. Terwel, Eric van Dijk (2020). How choice architecture can promote and undermine tax compliance: Testing the effects of prepopulated tax returns and accuracy confirmation. Journal of Behavioral and Experimental Economics, Volume 87, 2020, 101574. 48 V. ANNEX: THE APPROACH AND EXPERIENCE AT THE WORLD BANK 1. HOW THE EMBED TEAM SUPPORTS IMPROVED COMPLIANCE A behaviorally-informed approach to improved tax compliance revolves around a deeper under- standing of tax compliance challenges -taking into account both structural and behavioral barriers and enablers -and an application of this understanding to the context-specific design of solu- tions. As discussed above, behavioral insights are useful for solutions as well as a diagnosis of the problem to be addressed, starting with the identification of the specific tax compliance challenges (problems). Figure A1 below summarizes the framework used by the World Bank’s eMBeD team to approach tax compliance from a behavioral science perspective (a more detailed description of the figure is provided in the Practitioner’s guide). The design and experiment of solutions is based on a diagnostic exercise which can be thorough (e.g., original, in-depth qualitative and/or quantitative fieldwork, among others), light touch (key informant interviews, literature review, etc.), or somewhere in between. The diagnostic exercise identifying bottlenecks and enablers to the desired behavior of the taxpayer (or citizen) and designs interventions to address these. An experimental evaluation (RCT) where possible is set-up to support a rigorous impact evaluation of the intervention, findings are assessed, and lessons learned for further interventions. Figure A1. eMBeD’s approach to integrating behavioral science to address to policy challenges 01 03 IDENTIFY DESIGN PROBLEMS & EXPERIMENT SOLUTIONS 05 LEARN & ADAPT 02 04 DIAGNOSE ASSESS BOTTLENECKS FINDINGS Behavioral Insights for Tax Compliance 49 On a practical level, the eMBeD team’s approach to improved tax compliance is meant to be embedded within revenue administration functions, such that all stages of the approach outlined above are done in conjunction with government teams. Prioritization of tax compliance and re- lated challenges is completed in collaboration with the revenue authority and other government counterparts (as needed), and diagnostics, solutions design, implementation and evaluation are carried out with technical teams across the relevant institutions. Where possible, solutions iter- ation is built into existing or planned strategies and functions of the revenue administration. In many cases, initial experimentation revolves around developing a low-cost, politically feasible proof of concept to generate demand for more innovative future work (e.g., what has occurred with the team’s engagement in Kosovo and Armenia). In many instances, this collaborative process is facilitated through dedicated capacity building of technical teams across the revenue administration to build capacity and opportunity of staff to engage in a behaviorally informed project. The eMBeD team has developed -as part of its eMBeD Academy -a four to five day in-person capacity building workshop that covers the basics of a behavioral informed project (that is; behavioral science, problem identification and behavioral diagnostics, and solutions iteration, including an introduction to impact evaluation and experimental methods, among other topics) and includes group activities that aim for a set of proposed entry points for behavioral science within the revenue administration that are identified and explored at the technical level. In many client engagements, these brainstorming sessions lead to projects with the technical teams that participated in the workshops. Currently, the team is working towards supporting the institutionalization of behavioral science across partner revenue administrations (e.g., in those where behavioral insights units exists and there is demand for further expansion). CASE STUDIES The following case studies summarize tax experiments conducted by revenue authorities in conjunction with the World Bank. Case studies 1-6 are minor adaptations from Dalton et al. (2021). 52 Behavioral Insights for Tax Compliance 53 When the World Bank and the Costa Rican government collaborated to improve tax filing rates among firms, a simple email reminder made a big difference. The government has contact information for registered firms, so they were able to CASE STUDY 1: reach the taxpayers directly and credibly threaten COSTA RICA consequences for those who did not comply. 54 The Experiment: This experiment looked at the effect of enforcement messaging on income tax filing and related compliance issues among 50,000 non-filers in Costa Rica. Non-filers included firms that were tax-registered but had not filed their income tax declaration for 2014. Two-thirds of the target firms were randomly selected to receive an enforcement email while the remaining received nothing. The emails contained three features: • Making punishment salient, mentioning potential sanctions a taxpayer could be subject to, including shop closure, audit, or online publication of names. • Behaviorally-informed messaging, such as personalization, simplification and highlight- ing social norms. • Use of third-party information, emphasizing that reports from clients, suppliers, and finan- cial service providers could be used to validate a non-complier’s sales and spending. The Results: The intervention improved income tax compliance, as expected. The income tax filing rate for 2014 increased by 20 percentage points within five weeks. Filing rates increased by two addi- tional percentage points for those who received a message mentioning examples of third-party information. This also led to improved accuracy in reporting from unincorporated businesses, which are notoriously evasion-prone. Plus, firms that received an email became 2.5 percentage points more likely to file a third-party report about a client or supplier firm. These reports provide the tax authority with additional information that may help with future enforcement efforts. Behavioral Insights for Tax Compliance 55 WHY IT MATTERS Globally, many firms are formally registered but still fail to file their tax decla- rations. This leads to lost government revenue and may encourage those who already comply to stop doing so. Even if a communications intervention ap- pears successful, there is a risk that a temporary uptick in filings could lead to non-compliance later on. But, in Costa Rica, taxpayers in the treatment group remained significantly more likely to file income tax declarations in subsequent years. This was persis- tent even when measured 2.5 years later. This is likely because of the mention of enforcement tools in the behaviorally-informed email. Facing a shop closure, audit, or phone call about discrepancies on tax returns is viewed as highly cred- ible in Costa Rica. This confirms that people may change their behavior when what they were doing before (noncompliance, in this case) puts them at risk of an undesirable consequence. 56 Behavioral Insights for Tax Compliance 57 Between 2011 and 2015, tax revenue made up only 12 percent of Guatemala’s GDP. That’s less than half the average of other Latin American countries (World Bank, 2012). The World Bank, the Behavioural Insights Team, and Guatemala’s Superintendence of the Tax Administration (SAT) responded to this challenge with a behaviorally informed communications campaign. This nationwide intervention CASE STUDY 2: increased tax filings and improved income tax collection, GUATEMALA with positive results that continue to this day. 58 The Experiment: A randomized control trial targeted 43,387 individuals and firms who failed to pay their income taxes in 2013. Delinquent taxpayers were randomly assigned to receive no letter or a letter with one of five messages: • Reminder: A simple reminder to declare taxes, with no information on how to do so. • Behavioral Design: Information on where to declare, the possibility of paying in install- ments, and the deterrence message: “If you do not declare, you may be audited and face the procedure established by law.” • Behavioral Design + Social Norm: “According to our records, 64.5 percent of Guatemalans declared their income tax for the year 2013 on time. You are part of the minority of Guatemalans who are yet to declare for this tax.” • ”Behavioral Design + Deliberate Choice: “Previously, we have considered your failure to declare an oversight. However, if you don’t declare now, we will consider it an active choice, and you may, therefore, be audited and could face the procedure established by law.” • Behavioral Design + National Pride: “You are a Guatemalan citizen, and Guatemala needs you. Be a good citizen and submit the 2013 annual return of income tax. Are you going to support your country?” The “Behavioral Design” served as the basis for all other letters. The variations in framing helped confirm which messages were most effective in the Guatemalan context. The Results: While all letters increased tax declarations, two were most successful. The deterrence message, which framed non-declaration as intentional, and the social norms message, referring to declaring one’s taxes as the status quo, increased both payments and average amounts paid. For example, the deliberate choice letter helped raise the average amount paid per taxpayer by $17.95, or 269 percent. In addition to changing taxpayer behavior, we saw changes in how tax officials approached communications and policy design. Once our collaboration ended, officials continued integrating behavioral lessons into their communication with taxpayers, demonstrating sustainable shifts in tax officials’ attitudes and behaviors as well. Behavioral Insights for Tax Compliance 59 WHY IT MATTERS Taxes were historically a small part of Guatemala’s GDP, but behavioral insights were able to fuel lasting change and build capacity at the policymaker level. These findings are important, firstly, because this strategy is particularly cost-ef- fective. It was estimated that the social norms message would have generated revenues of approximately US$760,000 if sent to the entire sample. This is 36 times the cost of sending the letters. Secondly, the intervention’s success persists. 12 months later, individuals and firms were still complying with the tax declaration process without needing an additional reminder. Behavioral techniques have also become ingrained in the operations of SAT and the wider fiscal authority. To this day, SAT continues to repeat and revise the experiment while the fiscal authority explores how to ap- ply these learnings to other areas of taxation and business. SAT has continued to work with World Bank teams towards strengthening these efforts. Finally, this study sets an example for other regions with chronically low tax compliance. Initial compliance of just 64.5 percent is just marginally a norm. But, likely due to the perceived culture of tax evasion already existing in Guatemala, the social norm message was still enough to increase the moral cost of noncompliance. 60 Behavioral Insights for Tax Compliance 61 Polish authorities and the World Bank sent behaviorally informed letters to late taxpayers in order to combat extremely low tax compliance rates. The letters’ success emphasizes that investing in enhanced communication processes can be more strategic than adapting legislation. The letters tested ten different reminder messages CASE STUDY 3: and two methods of delivery. They successfully increased POLAND income tax compliance and, in turn, government revenues. 62 The Experiment: This randomized control trial reached 149,925 individual taxpayers across Poland. Taxpayers from both urban and rural areas were randomly assigned to receive either a default reminder letter (known as a Dunning letter) or one of nine behaviorally informed letters. The letters were sent to individuals who had declared their taxes for the 2015 fiscal year but had failed to pay what they owed by the deadline. The behaviorally informed letters included different introductory paragraphs and varied in tone: • Soft-Tone Messages: These messages highlight social incentives. This included reinforc- ing social norms by informing taxpayers about the high percentage of residents that had paid their taxes on time. Positive framing was used to explain how taxes finance schools, roads, and safety, and negative framing outlined how those services cannot be provided without taxes. • Hard-Tone Messages: These deterrent messages highlight sanctions for non-compliance or frame non-payment as an intentional, deliberate choice. Some letters used a blend of these techniques. The method of delivery also varied. Some taxpayers received the letter via regular mail, where the post office delivers the letter to the mailbox of the taxpayer. Others received the original Dunning letter via registered mail, a highly formalized process where the taxpayer confirms receipt with a signature, and the tax office is notified that the letter was successfully delivered. Not all combinations of delivery methods and messages were tested, which would have unnecessarily complicated the experiment Behavioral Insights for Tax Compliance 63 The Results: Behaviorally informed letters have a large impact on inducing tax payments, increasing payment amounts, and reducing tax liabilities in Poland. The most successful letter saw a 20.8 percent increase in the number of compliant taxpayers. Taxpayers in Poland were more likely to pay the income tax due, pay higher amounts of taxes, and reduce their tax liabilities after receiving a harder tone message. Letters sent by regular mail were just as effective as those sent via registered mail. Thus, the tax administration could save substantially by changing its delivery method to the lower-cost option. Finally, demographics make a difference. Women had a higher repayment rate overall and responded best to the hard-toned message. Letters with public goods messages were not as effective among people 50 to 64-years-old but increased compliance among younger people and rural taxpayers. An explanation could be that youth and rural taxpayers see themselves as more likely to benefit from the public goods listed in the letter. WHY IT MATTERS Subsets of the population responded to different types of behavioral messaging, reconfirming how attitudes and behaviors are highly con- text-dependent. While further analysis is needed to better understand the specific reasons for differences in compliance, calibrating reminder messages to different audiences can effectively increase both tax compli- ance and government revenues. Plus, where Poland differs from previous experimental cases is that hard-toned messages worked best. This type of message had been least effective in other studies, including Guatemala and the United Kingdom. In all cases, however, the behavioral messag- ing strategy led to similar cost savings for governments. This exploration demonstrates that policymakers might want to review the delivery meth- ods used to communicate with citizens. This may increase efficiency and cost- effectiveness, especially in the era of digital communication. 64 Behavioral Insights for Tax Compliance 65 It is estimated that Latvia’s shadow economy makes up close to a quarter of the country’s GDP, well above the average of 14 percent among other OECD countries (Hazans, 2011; World Bank, 2017). Part of this shadow economy includes individuals who are partially or fully self-employed and had been delaying or failing to submit their tax returns in the past few years. Our field experiment targeting taxpayers presumed to be part of the shadow economy shows that behaviorally informed messages – this time sent by email – can CASE STUDY 4: improve compliance with the tax declaration submission LATVIA process even among a group that is challenging to reach. 66 The Experiment: This experiment involved sending preemptive, behaviorally-informed email messages to taxpayers presumed to be part of the shadow economy. These individuals are formally part of the tax sys- tem, although they do not receive a regular salaried income. Those targeted, too, had previously delayed or failed to declare tax obligations in one or more of the previous three years. In advance of the 2017 filing deadline, 4,324 taxpayers identified by the State Revenue Service (SRS) were randomly assigned to receive one of three treatment emails or no message, if part of a control group. Consequently, 1,081 individuals received each message or were randomly placed in the control group: • Simple Reminder: The first included three short, easy-to-read sentences reminding indi- viduals of the tax timeline. It provided a link to the online submission system and contact information (i.e., aphone number) in case of questions. Signing off with the name of the Chief Tax Inspector made the message more personal, as the names of the recipients could not be included for technical reasons. • Omission/Commission Approach: The second message stated that previously missed deadlines were considered unintentional and inadvertent (i.e., an honest omission). However, future failures would be considered deliberate (i.e. commission). • Social Norms: The third message highlighted the descriptive norm that an increasing number of taxpayers filed by the deadline each year. While social norms messaging typically includes specific statistics, language on the increasing trend was included as descriptive data was unavailable. Behavioral Insights for Tax Compliance 67 The Results: Tax declaration submission rates were higher than the control group for every treatment group. Individuals in the treatment groups were also more likely to submit their Annual Income Declaration sooner, and before the deadline. On average, those who received the social norm message sub- mitted almost three days sooner than those in the control group. In Latvia, as in Poland, the most successful message included a harder tone that made salient the role of one’s deliberate, active choice to not comply. The harder toned message improved subsequent compliance more than a social norms message by 9.4 percent. Still, both a simple reminder and a social norm message also increased timely submission. The impact becomes stronger when controlling for other drivers of compliance, such as demo- graphics, past income, and tax payment behavior. Females who had higher revenue in the 2015 tax year were more likely to submit by the deadline. Those who delayed submission only in 2015, compared to in years prior, were also more likely to submit by the deadline. . WHY IT MATTERS This intervention again demonstrates that context matters when integrating behavioral science into messaging for taxpayer compliance. But perhaps more important is the unique nature of the target audience. Tax authorities are ac- tively seeking opportunities to reduce the shadow economy. Thus, this study has positive implications on the type of programming and messaging that may be replicated in other regions. 68 Behavioral Insights for Tax Compliance 69 Unlike other countries that collect significant non- tax revenues (for example, from natural resources), the Kosovar government relies on taxes for more than 85 percent of its revenues. Thus, it is critical to ensure tax collection is not expensive for the public administration, and all comply with the process. The Tax Administration of Kosovo (TAK) worked with the World Bank and the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) to address this challenge using behavioral insights. Three experiments in 2018 and one in 2019 successfully increased the number and CASE STUDY 5: timeliness of tax declarations while helping identify KOSOVO ways to improve the tax administration’s capacity. 70 The Experiment: Enhancing tax compliance in Kosovo involved sending behaviorally informed letters, e-mails, short messaging service messages (SMS) and phone calles to taxpayers. The messages were designed to induce timely and honest declarations and payments at low cost or no cost. These interventions targeted taxpayers subject to the personal income tax (PIT) and value-added tax (VAT): • Letters: Trial 1 targeted a letter at taxpayers who, by four days after the deadline, had not submitted their PIT declarations for the 2017 fiscal year. • E-mails: Trial 2 targeted all firms required to submit June VAT declarations with e-mail– based reminders. The first reminder was sent well in advance of the monthly due date. The second was sent on the declaration’s due date. • SMS Reminders: Trial 3 targeted all firms required to submit August VAT declarations with SMS-based reminders, sent a few days in advance of the due date. • Phone calls: Trial 4 targeted all firms with missing VAT declarations in the months of September and October 2019 with phone calls from the TAK call center (calls were made in early December 2019 as part of regular compliance improvement activities). A key component of the experimental design centers on the taxpayer as a client, informing the decision to promote positively framed messages, such as the benefits of paying taxes, rather than the negatively framed messages that worked in other contexts. Behavioral Insights for Tax Compliance 71 The Results: With regards to the first three trials (letters, e-mails and SMS), these induced more people to submit their declarations on time or to submit at all within the first 4 to 6 weeks. For example, submissions by PIT taxpayers increased by 2-4 percentage points within one month of the transmission of the letters. But, despite statistically significant results, process errors were rampant. Only 48 percent of letters made it to their intended recipients, and less than one in four e-mails was opened. The messages themselves were not always interpreted as intended as well. These implementation challenges highlighted the need to establish and update tax collection functions and data systems while carefully evaluating word choice in communications. With regards to the fourth trial (phone calls), these induced businesses to submit overdue decla- rations at a rate over 30 percentage points higher than a control group receiving no phone calls. November declaration behavior (in the absence of any additional reminders) was also improved significantly: on-time declaration for November was double the rate for the control group. WHY IT MATTERS Kosovo is one of the poorest and youngest countries in Europe in terms of gross domestic product (GDP) per capita, demographics, and statehood. Between 2011 and 2017, total government revenue amounted to about 14 percent of GDP, below the average of 19 percent among European and Central Asian countries. As a small country that depends greatly on revenue capture, this experiment suggests that investing in process upgrades will help ensure people get the information they need in a timely and motivating way. By participating in the design and implementation of the trials, TAK staff were able to understand the value of behavioral science tools, motivating them to engage in improving the systems that will improve their interactions with taxpayers. 72 Behavioral Insights for Tax Compliance 73 Indonesia is the largest economy in Southeast Asia. Here, the 59 million Small and Medium Enterprises (SMEs) make up 89% of all employment, contribute to 14% of exports and provide 58% of domestic investment. After a recent tax reform in 2018, SME tax compliance has reached 15% nationally, providing the potential to expand further the tax revenue. The Directorate General of Taxes (DGT), the country’s tax authority, worked with the World Bank to look at whether CASE STUDY 6: behaviorally designed information to taxpayers INDONESIA could impact compliance rates and payment amounts. 74 The Experiment: Following the introduction of a new Government Regulation in 2018 (no. 23/2018), SMEs with a yearly turnover of up to Rp 4.8 billion (approximately USD 350,000) are required to declare and pay 0.5% of their monthly gross revenue. To motivate these behaviors among firms, the project provided an official letter accompanied by a behaviorally-informed calendar to 18,000 SMEs from across 40 tax districts in Java, where 61% of SMEs in the country are based. Participating firms were randomly assigned to a control group or one of three experimental groups, each with different calendars. The groups were balanced, with 6,000 firms in each one. The calendars included the following messages: • Salient Information: Providing essential information about the tax regime in a salient way using clear and simple language, including an explanation on how to estimate the monthly payment following a rule of thumb, a reminder of the due date for each month, and a step- by-step guide on how to pay. SMEs were addressed as “BIJAK”, an acronym for a compliant business which also means “wise” in Indonesian. • Salient Information + Public Goods: Emphasizing public investments financed with tax rev- enue through seven messages, such as “34% of tax revenue contributes to an equal income distribution”; “taxes help 19.7 million poor students go to school”; and “taxes provide Rp 410 trillion to build roads, bridges, and infrastructure.” • Salient Information + Deterrence: With eight messages stressing that timely payment helps businesses avoid fines and reinforcing the negative consequences of late or non-pay- ment, such as “the DGT has access to financial information of firms”; “after the due date, you are subject to a 2% interest payment”; “tax arrears can lead to actions like freezing of bank accounts, restriction to travel abroad, confiscation of assets, or arrest.” Letters and calendars were mailed to participants between September and November 2018. Behavioral Insights for Tax Compliance 75 The Results: Fifteen months after delivering the calendars, participating SMEs improved their payment rates by 4%, increasing their compliance rate to 25%. Their monthly paid amount was also improved by 7% to Rp 115,000 (USD 10), compared to the control group. The results also showed that the calendar with deterrence messages was the most effective in attracting SMEs to pay their tax obligations (i.e.,increasing tax compliance), while the calendar providing salient information increased to the largest extent the amount paid by compliant SMEs. At the end of the analysis period, the experiment yielded a total revenue 37 times larger than the net cost of implementation. It also increased the revenue collected for participating offices by 3%. Calendar distribution was highly successful, with a delivery rate of 86% of the intended recipients. Implementation costs were estimated at a modest average of USD 3 per participant. WHY IT MATTERS: The experiment in Indonesia added to the value of testing different content to motivate different behavioral change and information resources to improve tax compliance. However, and most importantly, the experience is unique regard- ing the context and the tax regime considered. First, the experiment’s target population was SMEs, a key element in the economy of any developing coun- try, including Indonesia. Furthermore, the behavior promoted for these tax- payers was that of paying and reporting consistently every month, not merely once during a fiscal year. Finally, by working closely with a network of 40 local tax offices, the tax authority is now eager and prepared to scale up this impact- ful intervention and to conduct its own evaluations that inform its programs and policies. 76 Behavioral Insights for Tax Compliance 77 A global crisis can challenge behaviorally-informed policy in many ways. The experience of the tax authorities in Albania at the onset of the COVID-19 pandemic offers a unique opportunity to evaluate the impact of taxpayer engagement during a crisis. To combat tax evasion in reported wages, the General Directorate of Taxation (GDT) collaborated with the World Bank by sending out letters to both employers and employees suspected of under-declaring salaries. Results suggest that in crisis situations, positively framed language highlighting the benefits from tax compliance might be most effective at generating CASE STUDY 7: immediate (positive) reactions from taxpayers while ALBANIA having meaningful impacts on revenues. 78 The Experiment: A randomized controlled trial targeting 5,423 firms and over 8,000 employees, was designed and implemented in early 2020 to evaluate an existing strategy combating the suspected under-dec- laration of worker salaries in monthly payrolls (personal income tax withholding). Suspected employers and their (select) employees were randomly assigned to a control group (no letter being sent) or one of two treatment groups (receiving behaviorally-informed letters with a strong or soft tone). Letters were drafted based on the assumption that, in some contexts, compliance may be motivated by either the taxpayer’s perception of enforcement or their perceived level of civic responsibility and the benefits of compliance. Accordingly, employers and employees in the treatment groups received a letter employing one out of two framing strategies: • Soft-tone letter: Appealing to taxpayers’ moral obligation and the social importance of contributing through taxes. Strong-tone letter: Emphasizing the penalties associated with under-declaring and the enforcement capabilities of the GDT. While not considered in the design of the trial, letters were delivered during the first week of lockdown in March 2020, at a time when workers were sent home and business activity was drastically scaled down. Behavioral Insights for Tax Compliance 79 The Results: Firm wage declarations for all taxpayers in the trial were evaluated over a 12-month period be- ginning (retroactively) in January of 2020. The study found large, statistically significant increases on subsequent wage declarations for employers receiving the soft-toned notification letter (the impacts of letters directed to employ- ees were not evaluated since these were unlikely to have been received given the lockdown). Firm owners receiving the soft-tone letter reported worker salaries of up to 10 percent higher than those receiving no letter (the control group). This gap in reporting gradually decreases over time but reflects the immediate impact of the letters at a time when firm revenues decreased dramatically. On the other hand, no statistically significant differences were found for employers receiving strong-toned letters, highlighting that a poorly framed notification can induce the same reaction as no letter at all. WHY IT MATTERS: Thoughtful selection of letter recipients for taxpayer communication campaigns and strategic framing are of high importance in inducing tax compliance. The present findings also illustrate the context-sensitivity of behavioral change interventions with soft-toned letters achieving a likely stronger than usual response from payroll declarations. In such unprecedented times, strong-toned enforcement language could have been perceived by recipients as insensitive and non-credible and thereby negatively affecting tax morale. Continued testing is needed to understand the sensitivity of these results to the pandemic. 80 Behavioral Insights for Tax Compliance 81 A thoughtful selection of letter recipients for taxpayer communication campaigns and strategic framing are of high importance in inducing tax compliance. The present findings also illustrate the context-sensitivity of behavioral change interventions with soft-toned letters achieving a likely stronger than usual response from payroll declarations. In such unprecedented times, strong-toned enforcement language could have been perceived by recipients as insensitive and non-credible and thereby negatively affecting tax morale. CASE STUDY 8: Continued testing is needed to understand the sensitivity ARMENIA of these results to the global pandemic. 82 The Experiment: To increase tax compliance among businesses in Armenia, independent randomized controlled trials were conducted for three tax systems - Value Added Tax, Turnover Tax, and Corporate Income Tax -where taxpayers were flagged as requiring corrections to their tax returns (ranging from minor declaration errors to more meaningful inconsistencies in declaration amounts). Targeted taxpayers (a total of 7,857 across the three systems) were randomly assigned to receive one of three notification letters highlighting the immediate need for a correction: • Enforcement letter: Behaviorally-informed letter highlighting the enforcement capacity of the SRC and explicit expectations concerning the obligations of taxpayers. • Public goods letter: Behaviorally-informed letter using positively-toned language and em- phasizing the importance of taxes in financing public goods and the civic responsibility of the recipient. • Public goods letter: Business as usual letter sent in place of a pure control group (i.e., receiv- ing no letter) to understand if (and how) the existing approach could be improved. Behavioral Insights for Tax Compliance 83 The Results: Taking all targeted taxpayers into account, there was no evidence that the behaviorally-informed letters increased businesses’ likelihood of submitting the requested correction to their tax returns compared to the existing letters (in fact, only one of three taxpayers made a correction within one month of receiving the letter). However, there was some evidence that the public goods- framed letter worked best for some segments of taxpayers, highlighting higher sensitivity to the alternative framing and language relative to the standard notification. The positively-framed letter was disproportionately effective vis-à-vis the standard letter for younger taxpayers when compared to older ones in the Corporate Income Tax trial. Taxpayers subject to the Value Added Tax were also more likely to submit a correction when their businesses were situated in the pri- mary and secondary sectors relative to the tertiary sector. Despite no overall significant differ- ences in correction rates between the three letter types, public goods messaging was generally more effective in eliciting swifter reactions times. Across trials, among recipients who made a correction, those receiving the public goods letter were 50 percent more likely to do so within five days of the letter being received (relative to the standard letter). WHY IT MATTERS: Like Albania, the interventions were carried out at a time when the effects of the COVID-19 pandemic were still present and during a military conflict with neighboring Azerbaijan in Nagorno Karabakh. Taxpayer attention was likely captured by more immediate concerns, such as health and security-related issues. The results in Armenia highlight the challenges of inducing behavior change during crisis but show the potential of audience-specific messaging. Understanding the specific needs and motivating factors of different sub-pop- ulations and addressing these using targeted framing and language may lead to substantial increases in efficacy of the design and implementation of policies and reforms. 84 GLOSSARY Diagnostics: The process of identifying behavioral issues or inefficiencies within a system by collecting and analyzing data on behaviors, attitudes, and decision-making processes. This helps understand the underlying psychological factors influencing actions. Interventions: Specific actions or strategies implemented to address identified problems or chal- lenges. Interventions aim to improve outcomes by altering behaviors, processes, or environments within a system or organization. Social Norms: Shared standards of behavior that are considered acceptable within a group or society. These norms influence individuals’ actions and decisions, including compliance with laws and regulations. Heuristics: Mental shortcuts or rules of thumb that simplify decision-making processes, allowing individuals to make judgments and solve problems quickly and efficiently, though sometimes at the expense of accuracy. Choice Architecture: The design and organization of the environment in which people make decisions. This involves structuring choices in a way that influences the outcomes of decisions. Choice architecture leverages principles of behavioral science to present options in a manner that promotes better decision-making. Key elements include the arrangement of options, the use of defaults, the framing of information, and the provision of feedback. Nudge: “any aspect of the choice architecture that alters people’s behavior in a predictable way without forbidding any options or significantly changing their economic incentives.” Sludge: Excessive or unnecessary friction that makes it difficult for people to achieve their goals or comply with regulations. Sludge can include complicated forms, long wait times, or confusing instructions that hinder efficient decision-making and action. Behavioral Insights for Tax Compliance 85 86