Our paper contributes to a flourishing literature on the link between taxenforcement activities and taxpayer behaviour. Theory suggests that taxpayers become more com- pliant when audit frequencies rise and fine rates increase (see Allingham and Sandmo (1972) for the seminal work). In recent years, several papers empirically tested for this presumption in both, laboratory and field experiments (see e.g. Mascagni (2018) for a survey). The latter studies commonly investigate income reporting after randomly selected taxpayers were treated with tax authority mailings that convey enforcement information or explicit audit threats. The results of these studies unanimously show that increasing the rates or salience of penalties and audits significantly raises treated individuals’ tax reporting relative to taxpayers in the control group (see e.g., Drago et al. (2001), Kleven et al. (2011b), Fellner et al. (2013) and Dwenger et al. (2016) as well as Hallsworth (2014) and Slemrod (2018) for surveys).
ln (∆ 𝑖𝑖 ) = 𝛼𝛼 0 + 𝜷𝜷 ′ 𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝒔𝒔 𝑰𝑰 + 𝜸𝜸 ′ 𝒀𝒀𝑰𝑰𝒀𝒀𝒀𝒀 + 𝜹𝜹 ′ 𝑪𝑪𝑪𝑪𝒀𝒀𝒀𝒀 𝑰𝑰 + 𝜽𝜽 ′ 𝑺𝑺𝑺𝑺𝑺𝑺𝒀𝒀𝑰𝑰𝑰𝑰 𝑰𝑰
+ 𝝑𝝑′(𝑺𝑺𝑺𝑺𝑺𝑺𝒀𝒀𝑰𝑰𝑰𝑰 𝑰𝑰 ∙ 𝒀𝒀𝑰𝑰𝒀𝒀𝒀𝒀 ′ ) + 𝝁𝝁 ′ 𝑺𝑺𝑰𝑰𝒀𝒀𝑰𝑰𝑰𝑰 + 𝜀𝜀 𝑖𝑖
The bold expressions denote vectors. Error terms are clustered at the state level to correct for spatial correlation of error terms across countries. 𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝒔𝒔 includes a changing set of variables that mirror the taxenforcement incentives. Altogether four specifications are tested. In specification S2.1, 𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝒔𝒔 comprises a single variable: the taxpayer-specific internalized marginal tax revenue, 𝐼𝐼𝐼𝐼𝑇𝑇. The specification thus complies with the optimality condition (8) from the theoretical model. According to the model, we should expect a negative regression coefficient: The higher the incentive to enforce the tax law, the lower the granted tax deductions should be. Of course, to isolate the effect of 𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝑰𝒔𝒔 on granted tax deductions it is important to control for other potential determinants of granted tax deductions. To control for period effects, the vector 𝒀𝒀𝑰𝑰𝒀𝒀𝒀𝒀 includes two period dummies for 2001 and 2004. 𝑪𝑪𝑪𝑪𝒀𝒀𝒀𝒀 comprises the characteristics of the tax unit: the number of tax-relevant children, age, marital status and church membership. 𝑺𝑺𝑺𝑺𝑺𝑺𝒀𝒀𝑰𝑰𝑰𝑰 is a vector of seven dummies. Each dummy indicates whether the tax unit earned income from a particular income source. This is because the German income tax law distinguishes among seven different income sources, and for each, there are particular regulations. A dummy is one if the taxpayer has some positive income from the particular income source; otherwise it is zero. To capture changes in the tax law, we interact
of availability is scraped every one to three days to detect any changes. A change in avail- ability suggests a booking has occurred, which can be verified when a renter writes a review of the host and property after his or her stay. The primary concern with this approach is that we may incorrectly infer that a booking occurs, and thus over-measure the number of nights booked, when a host no longer wants to rent out his or her property for a particular night and blocks that night. This type of measurement error can lead to noisier estimates on the quantity of nights booked, but would only bias our estimates if the enforced tax rate is correlated with the measurement error. This could be true if, for example, the introduction of a taxenforcement agreement causes hosts to reduce their stated availability and those reductions are incorrectly inferred to be bookings. However, given that these Airbnb taxenforcement agreements reduce hosts’ marginal costs, supply responses are likely to be pos- itive rather than negative. A related concern is the possibility that stated availability does not accurately reflect actual availability as discussed in Farronato and Fradkin (2018). In particular, the authors point out that hosts may be better at updating their stated availabil- ity during periods of high demand. If true, this implies that we might over-measure nights booked during such periods. However, in our preferred specification discussed in Section 4, we are able to alleviate this concern by including metro-month-year fixed effects to absorb the effects of idiosyncratic demand shocks.
Taxenforcement is an essential dimension of any tax system and has therefore been the sub- ject of an extensive literature. The enforcement environment consists of several components such as auditing, self- versus third-party reporting, penalties and documentation requirements. Each element is arguably important, but in recent decades many countries have moved specif- ically towards a heavy reliance on stricter documentation requirements and third-party re- porting to assure compliance. Given the significant role of this policy dimension, assessing the behavioural responses to the reporting environment is important both to understand the magnitude of misreporting and to be able to invest efficiently in enforcement initiatives. Moreover, the enforcement environment can have a significant effect on how other traditional policy instruments, such as tax subsidies, affect the behaviour of taxpayers. Understanding the size and composition of such behavioural responses is essential for designing optimal poli- cies but also for evaluating the efficiency of a tax system. If elasticities are not structural parameters determined by preferences, but instead parameters controlled by policy choices of the government, such as the level of enforcement, this affects how we should interpret and utilise such parameters. Further, it affects the external validity of estimates across the litera- ture that are dependent on a specific form of tax policy design.
3 A model of a small open economy
This section develops a model of a small open economy with a representative house- hold and a government. The model embeds production with different technologies in an otherwise standard framework ` a la Eaton and Gersovitz ( 1981 ) where a benev- olent government with limited commitment issues debt on behalf of the household. The predictions of the model derive from the interaction of two frictions. There is imperfect taxenforcement: transfers from/to the household are made through a distortionary instrument that affects the cost of raising revenues in the future. There is limited commitment from the government to reimburse its debt, and de- fault risk limits the government capacity to transfer consumption from the future to the present. Consider a government willing to reduce debt. An increase in tax rates raises the distortions associated with future tax collection. Bondholders revise downward their expectations on future payments that could be sustained by the government. The increase in tax rates decreases debt through a direct effect but affects the borrowing capacity of an economy through this indirect channel.
3. The Theoretical Framework
Here I seek to identify the possible externality in tax administration due to terrorist activity. I develop a simple framework consisting of a federal state comprising two regions ( = 1,2) of equal size in which the total population is normalized to one. Region 1 is subject to the permanent threat of terrorist activity, while the other one is not. I consider two players: the regional tax authorities and the terrorist organization. Adhering to the most common approach in the literature (see, e.g., Shaw et al., 2009; Slemrod & Yitzhaki, 2002, 1987), I design the tax administrations as revenue maximizing agencies that set the taxenforcement rate ∈ (0, 1) in their regions. Here I focus on the potential externality effect of terrorism on taxenforcement policies, and so I restrict my attention to one tax instrument, , while assuming the tax rates in the two regions to be exogenously set. In line with the literature on extortion by mafia-type criminal organizations (see, e.g., Alexeev et al., 2003; 2004), I design the terrorist organization as a competing, revenue-maximizing tax collector that finances its violent activity in region 1 through the extortion of regular payments from its population. Individuals face an income tax on an exogenously fixed and normalized-to-one tax base and decide the share ∈ (0, 1) of income to declare maximizing their utility. To ensure an interior solution, tax evasion is assumed to be costly for the individual. Since the effectiveness of a taxenforcement policy largely depends on the way it is perceived by taxpayers 8 , I assume the enforcement rate to enter into the individual’s objective function through his perceived probability of
Abstract. In many federations, fiscal equalization schemes soften fiscal imbalances across the member states. Such schemes usually imply that a member state internalizes only a small fraction of the additional tax revenue from an expansion of the state-specific tax base, while the remainder of the additional tax revenue is redistributed horizontally or vertically. We address the question as to which extent state-level authorities in such a federation under-exploit their tax bases. By means of a stylized model we show that the state authorities in such a federation have incentives to align the effective tax rates of the state residents to the internalized marginal return from a stricter enforcement of the tax law. We empirically test the model using two approaches. In a state-level approach, we explore whether the state-specific internalized marginal returns matter for the states’ investments in taxenforcement. In a micro-econometric approach, using OLS regressions and natural-experiments, we explore whether internalized marginal returns matter for the effectiveness of the states’ taxenforcement activities, captured by the tax deductions granted to tax units. All our estimates support the results from our theoretical model.
2 This makes shaming the fourth most used instrument of tax debt enforcement ( OECD 2017 ). Figure A.1 provides an overview
of its worldwide use for tax collection. The popularity of shaming is not limited to tax compliance. There are numerous examples from different areas of policy, including the naming-and-shaming of water wasters (South Africa), sex offenders (United States), and speeding drivers (Australia) (see, for instance, Pawson 2002 ). Related are transparency policies that reveal the behavior of every entity involved. Transparency policies include public health initiatives where each restaurant has to display its most recent rating of the government hygiene inspection. Fung et al. ( 2007 ) discuss the advantages and disadvantages of transparency policies.
In this paper we investigate how accounting for the ability of individuals to avoid tax, as well as to evade tax, alters the conclusions for optimal enforcement of models in which only tax evasion is possible. In our model individuals can engage in tax evasion by under-reporting their income, but can also, at a cost, participate in a tax avoidance scheme that permits them to further lower reported income. Additional to the financial cost of avoidance, both forms of non-compliance are assumed, when detected, to impose psychic harm in the form of a social stigma cost. The nature of the avoidance scheme is not unambiguously prohibited by law, but is unacceptable to the tax authority. Accordingly, if the tax authority learns of the scheme, it will move to outlaw it ex-post. If a taxpayer is audited the tax authority observes whether they are using a tax avoidance scheme and also the extent of any tax evasion. The taxpayer is fined on the evaded tax, but the tax authority has no grounds to impose a fine on the avoided tax (it can only take measures to outlaw the scheme and then recover the tax owed on the avoided income). In this context we characterise the audit function first for a linear penalty function, and later for a general penalty function. The tax authority can condition its audit function only on the amount of income declared; it does not observe the amount of non-compliance or how it is split between evasion and avoidance. We therefore look for a taxpayer such that, if this taxpayer (weakly) prefers to report truthfully rather than hide an amount of income, then all other taxpayers will also wish to report truthfully.
net and by almost ten-fold in terms of the volume of transactions subject to it. I use this variation to identify the causal impact of VAT on firm compliance, along both the intensive and extensive margins.
For this purpose, I focus on manufacturers—the firms who enter the tax net at the earliest—and see how their outcomes respond as the tax gets extended to their buyers and suppliers. To the extent that VAT is self-enforcing, the compliance of existing firms will improve as their trading partners become subject to the tax (intensive-margin re- sponse). The increasing exposure to VAT will also push informal firms into the formal sector as their returns from operating informally squeeze (extensive-margin response). One useful feature of the Pakistani setting is that I can compare these enforcement spillovers with the effects of another policy experiment through which taxenforcement in the country was tightened using more traditional measures. Pakistan launched a countrywide tax survey from May 2000 during which inspectors from the tax admin- istration and other law enforcement agencies visited firms to assess their compliance with the tax laws. The survey, which continued for more than twenty-four months, arose out of political compulsions of the country at the time and was not connected with the planned trajectory of VAT in the country. It, however, was a large enforce- ment shock in the sense that within a short span of time the majority of firms in the country—both registered and unregistered—were visited by tax inspectors. I estimate the effects of the survey on firm compliance and compare them to VAT spillovers. The comparison helps me put the VAT spillovers into perspective to see how significant they are relative to the direct enforcement.
Apart from providing an interesting field of study, the concepts and methods used in combating crimi- nal tax in the United States can be used to improve the work of German tax investigation and other related agencies, especially with respect to the orga- nizational and the strategic dimension. The organiza- tional structure of the CID seems superior to the German tax investigation, in that the CID is estab- lished as a separate part of the IRS without neglect- ing the important flow of information between the civil and criminal part of the IRS. In addition, a high level of strategic planning makes it possible to use the scarce resources of the criminal taxenforcement agencies as efficiently as possible.
3) Extract information: Several law enforcement agencies, such as security forces, the secret service, and the military, base at least part of their workings on information- gathering practices in order to achieve their institutional goals. These practices, however, may involve the selective or nonenforcement of the law as a mechanism to extract vital information from informants and selected allies. What probably best epitomizes this phenomenon is referred to in the literature as “dual state” (Fraenkel 1969), “parallel state” (Briscoe 2008; Cucchiarelli and Giannuli 1997; Paxton 2004; Sain 2016), or “deep state” (Mérieau 2016; Söyler 2013; Tunander 2009). The parallel state alludes to a group of state agents holding positions in different branches of the state apparatus, usually the military, police forces, and judiciary, over which civilians have limited or no control. Because of the nature of these state agencies, the parallel state is particularly concerned with security issues and closely bound to the military- industrial complex (Söyler 2013, 311). These state agents are not accountable for their actions and, moreover, their participation in deep-rooted patronage networks grants them ongoing impunity. The covert nature of their work creates high levels of autonomy and provides room for informal rules to grow. Over time, the autonomy, unaccountability, and the informal organization of actions create an asymmetry be- tween the parallel and formal states: a tutelage function over democratic policies. These parallel formations within the state have close relationships with outlawed
A point of departure for our analysis is the acknowledgment that enforcing the law en- tails discretion or discretionary power. We assert that discretion over the laws’ reach is constitutive of state organizations and that state servants’ discretionary power “involves the power to exert choice among a range of alternatives between which the law does not discriminate” (Zacka 2017, 33). By bringing discretion to the fore, we acknowledge two important features of law enforcement measures. First, enforcing the law is a bureau- cratic and political choice that involves a flexibility to interpret and execute actions. As we argue later, such discretion accounts for different forms of law enforcement: full or ef- fective enforcement, selective enforcement, forbearance, or nonenforcement. Departing from the common definition of law enforcement as a system of state efforts to achieve compliance, we further define actions that tamper with enforcement, especially those that lead to nonenforcement, not in terms of inaction (“turning a blind eye”) but as a se- ries of practices that make use of discretionary power. Such practices block, alter, modify, falsify, distort, or eliminate actions prescribed by law, often relying on gaps or loopholes. Studying these practices becomes possible when one no longer sees law enforcement as either complete or failed, but instead understands it as a fluid spectrum of options. Second, the exercise of discretionary power over law enforcement mechanisms also means an exercise of sovereign power (Sarat and Clarke 2008, 390). Drawing on re- search that focuses on professions such as those of prison guards (Liebling, Price, and Shefer 2011, chap. 6), prosecutors (Sarat and Clarke 2008; Joseph 1975; Gelsthorpe and Padfield 2003), and other street-level bureaucrats (Lipsky 1980; Gutiérrez and Costan- tino 2020; Johannessen 2019; Forteza and Noboa 2019), we maintain that decisions about the different ways of enforcing the law “participate in, and exemplify, the logic of sovereignty and its complex relationship with legality” (Sarat and Clarke 2008, 390). In that sense, choices about enforcing the law are decisions aimed at establishing a specific social order, whether in the context of organizations such as prisons, bureaucracies, police stations, or courts, or when they come into contact with “society.” This insight has long been established in organizational sociology (Meyer and Rowan 1977; March and Olsen 1983), but it often slips out of focus. Taken together, these two features of law enforcement decisions – embedded in discretionary power and aimed at establishing social order – shift the conversational register from a problem of weak institutions and corruption to the domain of political decisions and the need of actors to achieve specific objectives or solve pressing problems.
imply a voluntary deprivation of personal liberties aimed at the common good, and yield no tangible individual benefits other than public health (Fehr and Gächter, 2002). To study the political cost of enforcement, we exploit the fact that the survey was fortuitously fielded throughout the same week in which the Italian government dramatically hiked lockdown fines. Given that over the previous two weeks, authorities had caught more than 100,000 people outside for no good reason, or lying on their forms, in the evening of March 24 the Italian Prime Minister announced an increase in the fines for violating lockdown rules up to 3,000€ (3,400$). We first show that trust in government and confidence in the truth- fulness of officially-provided information about the outbreak significantly increase with the amount of the fines that respondents would like to see enforced for risky behaviors. Despite that, the likelihood of considering the government’s reaction to the outbreak as insufficient increases with the fines’ amount. Comparing the opinions of respondents who filled in the questionnaire right before the hike in sanctions (until 6.30 p.m. of March 24) and those who did right after, we find that altruistic punishment relates to support for the government in a seemingly counterintuitive way. Right before the hike in sanctions, the degree of approval for the crisis’ management is significantly decreasing with the amount of the suggested pun- ishment. The higher the suggested fines, the higher the probability of considering the policy response to the outbreak as insufficient. The chosen amount also has a positive and weakly significant correlation with confidence in the officially-provided information and a positive though not statistically significant association with trust in government. The day after the Prime Minister announced the hike in sanctions, instead, we observe that higher amounts of the preferred fine become significantly associated with a higher probability of considering the government’s reaction as too extreme, lower trust in government, and lower confidence in the truthfulness of the official information. The flip in the sign of the measures of support for the government is striking and suggests that the new sanctions could have altered how altruistic punishers perceive the management of the crisis.
In this paper, I analyze incomplete enforcement in a political economy model. I use a contest framework to explain changes in lobbying behavior when special interest groups anticipate the incomplete enforceability of an emission tax. After the tax is proposed, two lobby groups – representing the interests of producers and environmentalists, respectively – seek to influence the legislator in order to prevent or support the approval and implementation of the regulation. I develop a gen- eral framework to demonstrate that the efforts of the interest groups to sway the legislator are motivated not only by the stringency of the proposed policy – as de- termined by the level of the tax – but, importantly, also by the policy’s anticipated enforceability. Using common functional specifications, I then show that incomplete enforcement may not only reduce the industry’s opposition to regulation compared to a situation with full enforcement, but it may – despite the possibility of misre- porting of emissions – also reduce expected environmental damage.
The first of the two survey instruments was sent to enforcement bodies in 29 European countries covering approximately 88% of the population (33 countries). We received useful responses from 17 enforcement bodies, which equal a response rate of 59% on a country level and 52% on an enforcement-body level. The different response rates are because four countries have more than one body that enforces compliance with financial information. In these instances, one enforcement body typically focuses on financial institutions, while the other focuses on all other entities, or one enforces auditors while the other enforces issuers of financial information. Respondents have been selected based on a two-step process. First, the relevant enforcement body is identified. Some of the formal responsible enforcement bodies have chosen to delegate actual enforcement to other governmental agencies. This means that they are not involved in the practical aspects of enforcement. Consequently, the respondents have been selected from bodies that perform the actual financial reporting enforcement to ensure that the respondent is knowledgeable about how enforcement is performed. Second, the respondent must be a senior enforcement official because this, ceteris paribus, ensures that the respondent has a broad and deep knowledge of both the formal and the actual enforcement process. A list of potential respondents was developed by the researchers based on publicly available information. The list was discussed with the insiders, who added both new candidates and valuable comments to the existing candidates. The final list of respondents was then completed by the researchers.
conclusive position. And the EU’s record of sanctions is short: financial sanctions arrived late to the panoply of enforcement instruments of the EU and have not been applied, for instance, under the Stability and Growth Pact regime. And, of course, the EU has not activated the sanctioning stage of article 7. Conscious of these limitations, some EU governments pressed for EU funds-related sanctions and EU institutions have moved towards introducing rule of law linked conditionality in the use of EU funds via the new Regulation on a general regime of conditionality for the protection of the Union budget. Certainly, this move affects, specifically, funds and, in this way, conditionality can hardly be conceived as a mechanism for safeguarding rule of law at large. However, given that EU funds contribute to the consolidation of an authoritarian equilibrium in illiberal states such as Hungary, the move is welcome (in particular, when seen in relation to the lack of efficacy of previous measures). European funds are essential for Hungary and Poland and also instrumental in
I n view of the increased mobility of capital it is of- ten argued that capital income tax rates tend to be reduced to zero by national governments acting independently of one another. Due to a “race to the bottom” taxes on capital income might no longer con- tribute sufﬁ ciently to the ﬁ nancing of public expendi- ture and it might become difﬁ cult or even impos sible for governments to perform their usual tasks. An under supply of public goods and/or an erosion of the welfare state are feared to be the outcome of tax competition. At the very least, the tax burden might be shifted away from highly mobile capital towards im- mobile factors such as labour; this would raise labour costs and impede the reduction of unem ployment es- pecially in Western Europe. The harmonisation of tax rates is thought to be the remedy.
The German speed limit enforcement campaigns that we analyze differ from other traffic law enforcement campaigns in particular by combining the scheduled one-day mas- sive SLMO with an extensive media campaign that not only documents the exact locations of the temporary speed traps but also explicitly informs the public about the dangers of speeding. The underlying idea is that these information nudges should increase drivers’ awareness about the dangers of speeding and thus alter their driving behavior more sus- tainably. This setting allows us to study in a very narrow time frame whether any effects on road safety can still be found after the one-day SLMO, which could then be traced back to effective information nudges. While this campaign started out in Germany, it became a pan-European effort in 2015. Ireland has run this campaign as the ‘national slow down day’ since 2015. In the German public, the usefulness of the speed limit enforcement campaigns to sustainably affect road safety is highly disputed. Not only drivers but also politicians and interest groups of the police have doubted any positive effects for road safety. As a consequence, several federal sates in Germany have recently opted out of the campaigns, also because of the high planning effort and excessive use of police resources. 1
To that end, we first experiment with excluding states that have passed a statewide Trust Act. Trust Acts are adopted with the purpose of increasing community trust and cooperation with the police following the implementation of tougher immigration enforcement measures, such as the 287(g) agreements that promoted information sharing between local, state, and federal government agencies (Skogan and Frydl 2004, Fagan and Meares 2008, Fagan and Tyler 2008, Tyler 2010). We exclude states with a state-wide Trust Act to more accurately capture the impact of intensified immigration enforcement, which should be lax or close to null in those areas. As a result, we would expect the estimated impact of intensified immigration enforcement to be, if anything, greater in magnitude. Table 6 displays the results from this exercise. The mean increase in immigration enforcement over the period under examination (i.e. 𝜇𝜇 𝐼𝐼𝐼𝐼 = 0.564) raises the overall share of Hispanic children entering the foster care system by approximately 17.62 percent. 30 The 18 percent increase