For tax authorities, efficiently monitoring and sanctioning VAT evasion is a challenging task. The 2015 VAT Gap report estimates that in 2013, EUR 168 billion in VAT revenues were lost in the European Union due to non-compliance or non-collection. The so called VAT Gap, the difference between the expected VAT revenue and the VAT actually collected, equated to 15.2% of total expected VAT revenue of the 26 member states at that time. The VAT GAP varied between 4% in the Netherlands, Sweden and Finland, and 41% in Romania.
A common strategy for business owners to evade VAT is not to issue the invoice of a business transaction and to pocket the equivalent amount of VAT. In recent years, tax authorities in different countries have devised a new behavioral approach to engage consumers in the fight against tax evasion. Tax lotteries were organized to motivate consumers to request receipts for goods and services they have paid for. To isolate the causal effects of tax lotteries on VAT revenue is not an easy undertaking, but an analysis of the tax lotteries impact suggests a significant positive increase in net tax revenue a few years after their implementation (Naritomi, 2013; Wan, 2010).
In this paper we argue that an evaluation of the welfare effects of tax lottery policies based solely on short-term variation of VAT revenue may overestimate their welfare effects by overlooking important long-term inefficiencies. Our analysis moves from the current undisputed evidence that material incentives may crowd out non-financial motives, such as moral values and civic virtues, while simultaneously determine unpredicted effects on behavior (Bowles, 2008). This has been referred to as the "crowding-out effect" (Frey and Jegen, 2001).
The argument advanced here is that implementing a tax lottery to encourage consumers to ask for sales receipts may be counterproductive due to the crowding-out of the intrinsic motivation to engage in voluntary third-party tax enforcement. An evaluation based only on the short-term effects of tax lotteries disregards possible long-term indirect effects. While the introduction of the tax lottery may determine a temporary raise of the net tax revenue, the crowding-out effect may increase over time by spilling over to future generations. This raises a fundamental question as to whether VAT losses caused by the crowding-out effect are larger than the increase in tax revenue generated by the tax lottery (taking into account the expense of the lottery prize).
Furthermore, tax lotteries may impact on the intergenerational transmission of norms regarding tax morale, defined here as the intrinsic motivation to comply with tax duties or the moral imperative to be honest regarding taxes (Feld and Frey, 2002; Lewis, 1982). This is especially likely to happen if the crowding-out effect is concentrated on people who play a pivotal role in shaping future generations' social norms and values, such as highly educated members of the population. Hence, irrespective of short-term positive impact evaluations, the choice to implement tax lotteries may still create long-term inefficiencies and even a permanent welfare loss.
The argument advanced here is supported by the findings of an empirical study conducted in Portugal, which suggest that the tax lottery has mixed effects on the propensity to engage in third-party tax enforcement. The lottery may motivate some consumers to enforce the emission of invoices, but it also crowds-out the willingness to engage in voluntary third-party tax enforcement in a sizeable fraction of the population. The crowding-out effect is concentrated on the most educated people, who are more likely to have leading roles in transferring tax morale to future generations. Hence, while evidence shows that the Portuguese tax lottery boosted tax revenues and accordingly is considered to be a success by the government, policy makers should be aware of this short-run evaluation, which may compromise the long-run welfare effects of the tax lottery policy.
Although tax lotteries have been used in different countries to fight tax evasion, they have received limited research attention. Moreover, to the best of our knowledge, no study has focused explicitly on the possibility that tax lotteries crowd-out the willingness to engage in voluntary third-party tax enforcement. The purpose of this study is to fill this gap. The article proceeds as follows. It begins with a literature review of the rationale behind the use of tax lotteries and a brief overview of work on tax evasion. The main focus is on recent articles that analyze tax lottery policies as an instrument to enhance tax compliance. The following section elaborates the claim that the crowding-out effect caused by tax lotteries, if concentrated on people with higher level of education, may generate a long-term reduction in VAT collection by hampering the intergenerational transmission of tax morale. Section 4 describes the institutional details of the Portuguese tax lottery and the results of our survey are discussed in section 5. In the last section, a conclusion is drawn and possible avenues for future research are suggested.
LOTTERIES AGAINST TAX EVASION
Neoclassical economic models consider agents' decision to engage in tax evasion as the result of cost and benefit calculations (Allingham and Sandmo, 1972; Becker, 1968; Yitzhaki, 1974). This has been disputed, but if this were the case, to achieve compliance would be prohibitively expensive (Blumenthal and Slemrod, 1992). Past regulation-related research that analyses the difficulties in contrasting frauds and tax evasion includes work by Alleyne and Elson (2013), Blazovich et al. (2013) and Drucker et al. (2012) to name a few. This line of research confirms that tax evasion remains a widespread problem (Feige, 2007), and that the costs of deterrence methods of tax enforcement are high and their effectiveness is low. This explains the growing interest of scholars and policy makers in new approaches that complement traditional enforcement methods (Feld and Frey, 2007; McMahon et al., 2016).
The current study focuses on VAT evasion. A key proposition advanced by prominent scholars is that the presence of third-party tax enforcers supplements the centralized tax enforcement system and thus constitutes a key ingredient in preventing tax evasion (Kleven et al., 2006). However, in the context of VAT evasion, consumers have no material incentives to enforce the emission of invoices. The VAT involved in the transaction will be redistributed to the entire population, but the individual benefits only from a negligible fraction of the tax paid by the business owner. On the other hand, the consumer bears material costs in requesting for a receipt. First, because the consumer forgoes the possibility to engage in a collusive agreement with the business owner which could give a discount equal to a fraction of the VAT evaded. Second, if the social norm is positive towards tax evasion, the consumer by acting as a third-party tax enforcer faces moral costs in requesting the emission of invoices (McGee, 2011).
Therefore, consumers enforcing VAT payment bear an individual cost and produce a positive externality that they do not internalize. This situation is well captured by the incentives scheme typical of a public good dilemma--voluntary contribution in the form of third-party tax enforcement would be optimal and would enhance the social good, but the dominant strategy for an individual payoff maximizer is to free-ride on contribution. Given this, it is plausible to think that consumers engage in a suboptimal level of invoices emission enforcement (Fabbri and Hemels, 2013).
Different strategies have been tried to encourage third-party enforcement of invoices emission. Some European countries, for instance Belgium and Italy, attempted to oblige consumers to request a receipt by imposing sanctions. This strategy was difficult to implement and was ultimately ineffective. Conversely, a tax lottery system based on rewards instead of sanctions may induce consumers to act as third-party tax enforcers. This is the rationale behind the tax lottery incentives scheme. Consumers want to participate in a lottery to win a prize, and accordingly are willing to request business receipts, which is the only means to participate in the lottery. If the increase in VAT revenue generated by the lottery is larger than the prize awarded, net tax revenue increases.
Although tax lottery policies have been common in some parts of the world since many decades, they have been rare in Europe. So far, and only in recent years, Malta, Portugal and Slovakia are the three EU countries that implemented a tax lottery to increase VAT compliance. It should be noted, however, that the interest of European governments for the use of lotteries in order to reinforce desirable behaviours has increased (Fooken et al., 2015). The interest of governments and policy makers in tax lotteries is due to the belief they...