Percentage tax designation institutions. On Sugden’s contractarian account

“Percentage Tax Designation Institutions”, also known as “Percentage Philanthropy Laws”, are fiscal institutions through which taxpayers can freely designate a certain percentage of their income tax to organizations whose main activity is of public interest: churches, third-sector organizations, political parties, etc. A comprehensive explanation of such systems is still lacking. In The Community of Advantage, Robert Sugden provides an original theoretical account of the Italian “8 × 1000” institution as one of those forms of regulation that “would be justified as ways of expanding opportunity for mutually beneficial transactions” and, more particularly, as a liberal and “contractarian approach to the provision of public goods”. This article is an attempt to expand and deepen the understanding not only of the 8 × 1000 but also of the 5 × 1000 and 2 × 1000 institutions, by reflecting on and possibly refining Sugden’s contractarian account, at least with regard to the part that relies on and develops the voluntary exchange tradition (Wicksell, Lindahl and Buchanan). To remain faithful to two normative premises of Sugden’s approach—the opportunity criterion and the correlated freedom of choice—we must introduce some theoretical adjustments to take into due account the way in which taxpayers’ freedoms—not only freedom of choice but also autonomy—are affected by default rules and the related redistribution procedures. In addition, these institutions also go beyond the voluntary exchange tradition—insofar as they go beyond its basic assumptions: the benefit principle of taxation, taxpayers’ self-interest and the very logic of exchange—and, at the same time, they can be read as a new form of voluntary tax justice.


The Italian percentage tax designation institutions
In Italy there are three different institutions through which taxpayers can allocate their taxes to public utility purposes: (1) 8 × 1000: designation to the state or a religious confession; (2) 5 × 1000: designation to the third sector and many other entities pursuing public utility purposes; (3) 2 × 1000: designation to political parties. Taxpayers can express their choice by signing a specific section of the tax return where they can also indicate the specific beneficiary of their tax money. The expressed choice is anonymous, and therefore, the beneficiary cannot know who its "benefactor" is.
It seems important to me, before I go into Sugden's analysis, to dwell on some essential characteristics of these institutions, also because there is no scientific literature in English explaining the functioning of the Italian PTDIs, and it is probable that the lack of knowledge of these mechanisms has largely compromised the possibility of studying and understanding them adequately.
These three institutions fall into the category of PTDIs, since in all three cases taxpayers are given the right to choose whom to give a share of their income tax.

3
However, these institutions do not only differ from each other in relation to the different types of beneficiary organizations. They also differ in their historical, political and economic-social grounds, the different purposes pursued by the legislator in each case, the rules governing the taxpayers' choices or non-choices and the revenue redistribution mechanisms deriving from these choices.

Basic facts
The 8 × 1000, which came into force in 1990, is the share (0.8%) of the income tax revenue that the Italian state redistributes, based on taxpayers' choices or nonchoices, between itself and religious confessions. Currently, 12 religious confessions participate in the 8 × 1000. 4 From the beginning, the most "voted" organization has always been, for obvious reasons, the Catholic Church. Religious confessions are required to spend the sums deriving from the 8 × 1000 for defined purposes: people's worship needs; support of the clergy; charitable, humanitarian, socio-cultural activities in favour of Italian society and the Third World, etc. The state, on the other hand, should use these sums for extraordinary expenses or interventions, for example, in dealing with the problem of world hunger, natural disasters, assistance to refugees, etc.
The 5 × 1000 was introduced in 2006 as an institution implementing the constitutional principle of horizontal subsidiarity (and, implicitly, as a new form of fiscal subsidiarity) in virtue of which taxpayers could play a central role in providing a financial contribution mainly to organizations belonging to the third sector, but also to other types of organization providing public goods and services. It was also meant to give relief to the (never-ending) crisis of the welfare state in the logic of an outsourcing of public services to the third sector. Currently, in the 5 × 1000 scheme there are thousands of beneficiaries clustered in seven sub-sectors or sub-fields: (1) third sector, (2) university and scientific research, (3) health research, (4) social policies pursued by municipalities, (5) amateur sports associations, (6) activities for the protection, promotion and enhancement of the cultural and landscape heritage, and (7) protection and management of environmental areas. The beneficiaries most appreciated by taxpayers are usually health and/or cancer research organizations; renowned international third-sector organizations, such as Doctors Without Borders, Emergency, Save the Children and UNICEF; and many other voluntary organizations operating at national and local level which provide various forms of assistance to the poor, disabled, young and old people, families, etc. 5 4 (1) Catholic Church, (2) Assemblies of God in Italy, (3) Italian Union of Seventh-day Adventist Churches, (4) Union of Methodist and Waldensian Churches, (5) Lutheran Evangelical Church in Italy, (6) Union of Italian Jewish Communities, (7) Baptist Evangelical Christian Union of Italy, (8) Greek Orthodox Archdiocese of Italy, (9) Apostolic Church in Italy, (10) Italian Buddhist Union, (11) Italian Hindu Union, Sanatana Dharma Samgha and (12) Soka Gakkai Italian Buddhist Institute. 5 Data on the redistribution of the 5 × 1000 revenue to beneficiaries are published yearly by the Italian Tax Authority. The redistribution made in 2020 (for the 2018 fiscal year) can be found here: https ://www. agenz iaent rate.gov.it/porta le/elenc o-compl essiv o-benefi ciar i-2018.

3
Percentage tax designation institutions. On Sugden's… To get an idea of the scale and impact of the 8 × 1000 and 5 × 1000 in terms of the number of taxpayers and organizations involved and the amount of resources redistributed, let us consider the following. Every year, almost half of all taxpayers (nearly 18 million) express the choice to designate their taxes. More precisely, recent research showed that this equates to 70% of those who have the duty to fill in their tax return, and 0.6% of those who do not have this obligation 6 (Silvestri et al. 2020: 67). In the last five years alone, on average about 1.2 billion euros per year have been redistributed through the 8 × 1000, and about 500 million euros per year through the 5 × 1000, for a total of 8.5 billion. In the 8 × 1000, the beneficiary organizations number 13, that is, the state and the 12 religious confessions. In the 5 × 1000, the number of beneficiaries has more than doubled since the introduction of the mechanism, reaching over 60,000 entities in recent years (of which about 8,000 are local municipalities), the vast majority of which are large and small organizations belonging to the third sector.
The 2 × 1000 was introduced in 2014 as a form of public funding to political parties with the aim of replacing the electoral reimbursement system. However, it has enjoyed very little success among Italian taxpayers in terms of the overall number of expressed choices in favour of political parties and, accordingly, the overall amount redistributed to them (on average, only 2-3% of taxpayers choose to give their tax money to political parties). 7

The taxpayer's opportunity sets, the rules of choice and the redistribution mechanisms
The way that the choice can be made is regulated by a few basic rules which form what I would call the "taxpayer's opportunity sets". This choice procedure, in turn, can be understood as a decision-making process subdivided into several levels and sub-levels.
First, taxpayers are faced with three binary choices: to give or not to give the 0.8%, 0.5% or 0.2% parts of their income tax. It is important to specify that these three choices do not mutually exclude each other, so that taxpayers can allocate up to 1.5% (i.e. 0.8% + 0.5% + 0.2%) of their income tax.
(1) If taxpayers choose not to give (default setting), then: • the 0.5% (and 0.2%) of their income tax will go to the state coffers by default; • "their" 0.8% or, more precisely, the so-called inoptato, "non-opted funds", calculated as 0.8% of the tax revenue from income tax that taxpayers have not chosen to designate, will nevertheless be redistributed among the state and religious confessions in proportion to the number of choices expressed by other taxpayers.
1 3 (2) If taxpayers choose to give (expressed choice), they have to sign the specific form on the tax return. In this case, they are faced with a sub-set of possible choices: • they are free to choose among the three non-mutually exclusive macro-sectors, i.e. the 8 × 1000, 5 × 1000 and 2 × 1000; • in the case of the 8 × 1000 and 2 × 1000, they can choose only one beneficiary among those listed on the tax return form, respectively: either the state or one religious confession, or one political party; • in the case of the 5 × 1000 they can select between a "generic" and a "specific" choice: • the generic choice can be expressed by signing one and only one of the boxes corresponding to one of the seven sub-sectors (the third/voluntary sector, university and scientific research, health research, etc.); • the specific choice can be expressed in favour of a specific beneficiary, by signing the box corresponding to one of the seven sub-sectors and inserting the beneficiary's "fiscal code" (tax identification number) in the corresponding box.
When taxpayers only express a generic choice by giving their 0.5% in favour of a sub-sector, this tax money ends up in the so-called inoptato, the "non-opted fund" which collects the total amount of resources deriving from the expressed generic choices of taxpayers for a given sub-sector. In this case, the state redistributes the non-opted fund of each sub-sector to each individual beneficiary belonging to that sub-sector in proportion to the number of specific choices expressed by taxpayers in favour of that beneficiary. This means that the most frequently named beneficiaries will take a larger portion of the non-opted fund.
The aforementioned explanation highlights one of the main differences between the 8 × 1000 on one hand and the 5 × 1000 and 2 × 1000 on the other, that is, the different redistribution criteria linked to taxpayers' choices and non-choices, and, above all, the connected redistribution of the "non-opted" funds to (non-opted or non-chosen) beneficiaries. It is important to distinguish between the redistribution of the funds deriving from expressed choices and the redistribution of non-opted funds deriving from non-expressed choices, in order to better understand the relative weight of the respective types of choice.
In the case of the 5 × 1000 and 2 × 1000, the individual taxpayers' choices determine the designation of their 0.5% or 0.2% share of income tax, and the state redistributes the tax money deriving from these choices to the beneficiaries.
In the 8 × 1000 system, the Italian state redistributes 0.8% of the total tax revenue from income tax. The overall redistribution of this share of tax revenue to the state and religious confessions is calculated in proportion to the taxpayers' expressed choices. (This redistribution is the result of a single calculation mechanism, which I will explain later: Sec. 4.4.) This implies that the individual taxpayer's choice does not determine the designation of "his/her" 0.8% share of income tax, but an average share that is the same for all taxpayers. In this way, the sum total given by a taxpayer to a specific beneficiary does not depend on his/her income tax. For this reason, it is not entirely appropriate to say that the 0.8% share is his/hers. Above all, it is wrong to say that this share is calculated as 0.8% of the taxpayer's income tax (that is why I use the possessive pronoun referring to the taxpayer(s) between single quotation marks).
In the case of the 5 × 1000 and 2 × 1000, the taxpayer's non-choice implies that his/her 0.5% or 0.2% will go to the state coffers, while in the case of the 8 × 1000 "his/her" 0.8% will be redistributed to the beneficiaries (state/religious confessions) in proportion to the other taxpayers' expressed choices. As I will explain later on, this is a fundamental difference, as in the case of the 8 × 1000 the taxpayers' freedom not to choose is conditioned by the "default option" of this mechanism.
I will return to the implications of these rules later, that is, after having provided a detailed account of Sugden's explanation of the 8 × 1000, also with the aim of understanding if, and to what extent, his explanation takes these rules and the issues they raise into due consideration.

Sugden's contractarian account
Sugden develops his reflection on the 8 × 1000 in the context of chapter 7 of TCOA, devoted to "Regulation" and giving a more general analysis of regulatory approaches that would justify the supply and financing of public goods. The fact that Sugden deals with the 8 × 1000 in a single paragraph might suggest that this topic is of wholly marginal importance in the context of TCOA. Yet, in my opinion, his treatment of the 8 × 1000 helps us better understand not only his broader and particular contractarian approach and his intellectual debt towards Buchanan-considering that "the book as a whole" is "deeply influenced" by the latter (Sugden 2018: xi) 8 -but also the way he furthers his theory with respect to Buchanan in particular and to the tradition of the voluntary exchange paradigm in general.
Furthermore, although it is true that Sugden's normative approach is essentially contractarian (see also Sugden 1986Sugden , 1990Sugden , 2019, it is also true that it is relatively easier to justify the market in the light of a contractarian approach based on certain axiological principles and assumptions-such as mutual advantage, individual freedom and responsibility, 9 and the opportunity criterion-than to justify the rules by which these values and principles can also work for the regulation, provision and financing of public goods. In this connection, as I find Sugden's contractarianism particularly convincing for the case analysed here, I will not discuss it in more detail. Nevertheless, there is one note that seems important to me. Although some may legitimately ask what kind of contractarianism Sugden's is (see Vanderschraaf 2020), confining it to mere normative justification of the market only captures one of its aspects. 10 It seems to me, instead, that the kind of justification for the regulation and provision of public goods provided by Sugden's contractarianism is importantand even more so in the case of PTDIs-as it shows its ability to deal with the radical diversity of individuals and their inclinations as to which public goods deserve funding 11 (see also Sec. 3.3).
It is perhaps no coincidence that the chapter on "Regulation" begins by recalling and summarizing the guiding ideas of the book incorporated in its very title, inspired by John Stuart Mill's conception of the market as a "community of advantage". This conception encapsulates three core components of a liberal tradition of economic thought: that, in a well-ordered society, cooperation for mutual benefit is a governing principle of social life; that a competitive market is a network of mutually advantageous transactions; and that, in cooperative relationships, it is for each individual to judge what counts as his or her benefit (140).
One might therefore think that Sugden's analysis of the 8 × 1000 can almost be considered a kind of "stress test" of his contractarianism.

Public goods
How can the principles on which a "community of advantage" rests work when it comes to finding forms for the regulation and provision of public goods, given the well-known free riders and hold-out problems? Sugden proposes as a regulatory ideal that public goods should be supplied if and only if the total willingness to pay of beneficiaries exceeds the total cost, and that costs should be apportioned between beneficiaries in such a way that, for each individual separately, willingness to pay exceeds actual payment (165).
It is a regulatory ideal that Sugden traces back to Dupuit (1844Dupuit ( /1952, as the founding father of cost-benefit analysis, and to the tradition of the voluntary exchange of public goods theory, as developed by Knut Wicksell (1896/1958), Erik Lindahl (1919Lindahl ( /1958 and James Buchanan (1968), which tries to overcome the socalled preference revelation problem, that is, the problem of how to reveal an individual's willingness to pay in collective decisions regarding public goods. Furthermore, considering that the assumption of integrated preferences-as a prerequisite for cost-benefit analysis and the potential Pareto improvement criterion intended as methods for making decisions about the provision of public goods-has long been questioned, in this chapter Sugden also tries to bypass the problem of non-integrated preferences "by taking an opportunity-based approach to regulatory economics" (166).

3
Percentage tax designation institutions. On Sugden's… The opportunity criterion expresses the idea that "it is in each individual's interest to have more opportunity rather than less. An individual's opportunities can be thought of as the set of options from which he can choose, options being described in such a way that they are mutually exclusive and jointly exhaustive." (84). When an individual is confronted with two opportunity sets O and O′ and O′ is a superset of O, the opportunity criterion implies that it is in the individual's interest that the set of opportunities is O′ rather than O.
As such, the opportunity criterion is inevitably linked to the normative value of freedom of choice insofar as more opportunity implies more freedom of choice. This is also why the "opportunity criterion", formalized in the paper "The opportunity criterion: consumer sovereignty without the assumption of coherent preferences" (Sugden 2004), and variously referred to in TCOA from the introduction on, as well as in the chapter specifically dedicated to it (Ch. 5), is fundamental to Sugden's normative approach.
The strategy followed by Sugden is a reformulation of the willingness-to-pay principle, but without having to resort to the assumption of integrated preferences. In any case, the problem remains that "when a good is public, willingness to pay and actual payment cannot be linked so directly" (166).
If an opportunity-based approach is to make any headway in dealing with public goods, we need to find some setting in which people can make genuine individual choices about whether or not to take and pay for the benefits of public goods (166).
It is in the light of this problem that Sugden claims to follow the strategy set out in the voluntary exchange theory of Wicksell, Lindahl and Buchanan. Sugden does not claim to have "found a universally applicable and operational criterion for decisions about the provision and financing of public goods" (166) but suggests two possible ways-respectively his treatment of the public goods problem and the 8 × 1000-in which progress might be made. In both cases, Sugden sees the possibility of mutually beneficial transactions between members of society. However, such transactions "would probably need to be intermediated by some agency with the authority to impose compulsory charges" (167), if only to avoid the aforementioned free riders and hold-out problems.
The case of the 8 × 1000 is, however, treated separately from the more general problem of public goods, also because taxpayers may have non-self-interested and non-instrumental preferences for the "public goods" that can be chosen with the 8 × 1000. 12 And the specificity of the 8 × 1000 also lies in the fact that it could enable a solution to "the problem of how to ensure that, when people have non-instrumental 12 In order to defend his version of willingness-to-pay and his contractarianism, Sugden begins the paragraph on the 8 × 1000 by arguing against some criticisms of the use of willingness-to-pay measures of value in cost-benefit analysis when applied to public goods (Hausman 2011), especially where such measures are supposed to lead to commodification and to thinking of public goods in instrumental terms (Anderson 1995;Sandel 2012). This seems to me to be another important aspect of the merits and potential that Sugden sees in the 8 × 1000. However, I cannot address this here. See also Hausman's reply (Hausman 2020), which, however, does not take into account Sugden's treatment of the 8 × 1000. preferences for public goods, mutually agreeable transactions (and only such transactions) are intermediated" by some sort of public authority (171).
Above all, the 8 × 1000 has two advantages from a theoretical and a practical point of view. On the one hand, it "illustrates the logic of a contractarian approach to the provision of public goods". On the other hand, "it could usefully be adapted to many cases in which there is divergence in people's beliefs about which kinds of public goods are valuable and which are not" (171). Furthermore, in a very short note, he adds: "in Sugden (1990) I discuss some of the merits and demerits of mechanisms of this kind". To better understand Sugden's explanation of the 8 × 1000, it is therefore necessary to go back to that 1990 paper, while nevertheless specifying that in that year the 8 × 1000 had only just been introduced and, therefore, he could hardly have known or explained it.

Bettering the voluntary exchange paradigm
In that paper-"Rules for Choosing Among Public Goods: a Contractarian Approach"-Sugden provided an important "contribution to the contractarian enterprise", by introducing a "new criterion for decision making" (Sugden 1990: 63). In my view, this criterion is a rather innovative development of the voluntary exchange tradition. So, it seems to me, when Sugden claims to follow this tradition in the TCOA, he is either being too modest or perhaps he has lost hope that this old paper, which, as he once complained, "few people have read", 13 would ever be redeemed from the "strange" 14 oblivion into which it had fallen. I will now try to summarize its key ideas.
The contractarian theories of the time-mainly those of Rawls (1971) and Gauthier (1986)-limited themselves to finding arguments to justify compulsory taxation to finance public goods, given the free-rider problem. They simply assumed that a public good was to everyone's advantage, but completely neglected "the much more problematic-and much more common-case of a public good that is to some people's advantage but not to others" (Sugden 1990: 64).
In this regard, the only significant theoretical development is represented by the voluntary exchange tradition inaugurated by Wicksell (1896/1958), later developed by Lindahl (1919Lindahl ( /1958, and then further refined by the so-called Virginia School (Buchanan and Tullock 1962;Buchanan 1968Buchanan , 1975Brennan and Buchanan 1980).
In "A New Principle of Just Taxation", Wicksell (1896) tries to give a new institutional-political formulation of the principle of tax justice known as the "benefit principle", which he interprets in terms of citizen-taxpayers' mutual advantage: taxes to be paid for public goods expenditure should be based on a politically revealed willingness to pay for benefits received. The solution envisaged by Wicksell was a constitutional rule in virtue of which proposals for the provision of public goods had to be linked to specific proposals for tax finance as well as to an "approximate unanimity" as a decision-making rule, so that no one was made worse off.
This idea was later developed by Buchanan 15 and Tullock (1962), with further variations and readjustments by Buchanan (1968), in the attempt to resolve the difficulties resulting from the "transaction costs" of the unanimity rule-imperfect information and opportunistic behaviour (free-riding and hold-out threats, the preference revelation problem, veto power, etc.)-and to find a way to implement forms of lessthan-unanimity voting and, at the same time, avoid the tyranny of the majority (i.e. by passing on the costs of public goods to the minority), for example by introducing a pre-determined constitutional rule of proportional taxation on income (Buchanan 1968). 16 Nevertheless, as Sugden noticed, a crucial issue remains to be settled: what rule is to be used to decide which public goods should be financed out of any given total budget? Buchanan and Tullock's proposal is in effect to use the same rule as is used to determine the size of the budget. For Buchanan and Tullock, the size of the public good budget is not a matter for explicit decision, but is simply the consequence of whatever decisions are made about the supply of the various public goods (Sugden 1990: 71).
Hence, Sugden's proposal is that: "the size of the budget and how it is spent should be regarded as separate decisions, for which different kinds of decision rules are appropriate" (Sugden 1990: 71).
As said, Sugden's proposal takes a significant step forward in the tradition of the voluntary exchange paradigm: despite taking Wicksell's principle as the ideal, he implements it in a new way. The voluntary exchange paradigm is thus reframed in a way that can be summarized and subdivided into three rules, by keeping in mind the aforementioned general rule proposed by Sugden, the two sides of the public budget and the respective decisions to be made: 1. tax-raising decisions (about the size of the budget and/or tax rates for public good spending) are separated from tax-spending decisions (about which public goods to supply); 2. tax-raising decisions are made collectively; 3. tax-spending decisions are made individually.
Note that, contrary to what Wicksell wanted, the second rule implies that individual choice about overall tax/spending levels is ruled out. However, it has the advantage of disarming free-rider and hold-out threats. The third rule allows each individual to allocate his/her own tax among the public goods according to his/her choice.
It is worth mentioning that Sugden calls his basic idea "proportional spending". At the bottom of the above-mentioned decision-making process, there is an "analogy with the idea of proportional representation" and the connected voting mechanism (Sugden 1990: 72). It is as if tax revenue were allocated to public goods in proportion to individuals' "votes" rather than as the outcome of a centralized collective decision. Indeed, Sugden draws other important analogies between his rules for choosing among public goods and other procedures such as tax relief for charities or state support for churches and political parties as in the case of Germany, where, for example, the redistribution to political parties is proportional to the number of votes each party received [I will try to say something more on the analogy between Sugden's rules for choosing among public goods and the voting mechanism based on proportional representation later (Sec. 4.3)].

"Otto per mille"
In this light, it is understandable why Sugden sees the 8 × 1000 as displaying several strengths and a potential for theoretical and practical development. We can summarize and subdivide his explanation into five claims. We have already seen the first two, but it is worth remembering them as general theoretical premises. With reference to the other three, I will add only a few clarifications where necessary for the purposes of the subsequent analysis.
The 8 × 1000: 1. "illustrates the logic of a contractarian approach to the provision of public goods" (171); 2. "could usefully be adapted to many cases in which there is divergence in people's beliefs about which kinds of public goods are valuable and which are not" (171); 3. presents a very interesting aspect "from a contractarian viewpoint: […]: each taxpayer can choose which fund on the list will receive her 0.8 per cent" (172); 4. "provides a rough and ready response to the free-rider and hold-out problems" (172); 5. "is also a rough and ready way of ensuring that people contribute to the costs of only those public goods that they actually value" (172).
The connection between the first two claims should now be clear, since "contractarian arguments are addressed to people who are trying to negotiate mutually acceptable agreements without necessarily resolving all their moral and political differences" (171). This connection should be even clearer and more significant when we talk about the supply of public goods, especially if we consider goods and services provided by religious confessions as public goods and that disagreements 1 3 Percentage tax designation institutions. On Sugden's… among citizens(-taxpayers) over which religious confessions should be financed are not likely to be settled.
The contractarian approach can be (so to speak) "completed" if we read the first two claims along with the fifth. While the first two seem to refer to the horizontal dimension of the contract among citizens, the fifth claim can be read not only as a variation on (Sugden's view of) the willingness-to-pay principle but also along the vertical dimension between citizen-taxpayers and the state, and as an implicit reference to his critique of welfare economics 17 present throughout TCOA: a contractarian would say that how Seventh-Day Adventists choose to spend their own money is no business of Catholics, and vice versa. In saying this, the contractarian is saying that each taxpayer's 0.8 per cent is that person's own money, and not the property of the Italian state. The Italian government is not subsidizing religious organizations according to its judgements about their relative merits; it is simply intermediating [mutually advantageous and agreeable] transactions in which individuals pay for activities that they value (172).
The third claim might also be read not only as a concrete objection to paternalistic and welfaristic approaches, but also as an aspect of Sugden's opportunity criterion and the implicit value of the freedom to choose.
The rationale and background of the fourth claim, on the other hand, have already been analysed in the previous section.
Note, incidentally, that all the considerations linking together claims 1, 2 and 5-namely (in sum) the problem of the likely disagreements over which public goods are to be provided and financed and the contractarian approach solution-are equally valid for and applicable to the 5 × 1000 and 2 × 1000 institutions.
In this connection, one may wonder whether all public goods and services provided by the beneficiary entities of PTDIs can rightly be termed "public goods", also because the Italian legislator makes use of other expressions-such as "public utility purposes", "humanitarian purposes" (in the case of the 8 × 1000) and "socially relevant activities" (in the case of 5 × 1000)-which, in some cases, seem to be more appropriate. In general, the answer is "yes", to the extent that the goods and services provided by the beneficiary entities generate positive externalities and suffer from the free-riding problem. Nevertheless, some of these entities may be thought of as "clubs", for example churches or amateur sports associations, but for several reasons they may either be reluctant to introduce pricing schemes or prevented from doing so because, legally speaking, they are non-profit. 18 18 In the case of churches, Sugden (1990: 76) already noticed that "for reasons of theology, many churches are reluctant to introduce pricing schemes. This leaves them very vulnerable to the free-rider problem". The inclusion of "amateur sports associations" in the 5 × 1000 institution is due to the Italian legislator's intention to have them legally framed within the voluntary sector. Political parties too can be considered as providers of public goods (Sugden 1990: 76). Again, though education services provided by universities generate positive externalities, technically speaking they are excludable and rivalrous. Nevertheless, it is worth noting that the 5 × 1000 is aimed at financing "scientific research". 17 Note that in the same period in which Sugden was putting together the 1990 paper, he was also already developing a critique of welfare economics: "Maximizing Social Welfare: Is it the Government's Business?" (Sugden 1989).
Apart from these considerations on the nature of public goods, it would seem that what counts most in Sugden's reflection is to point to this type of institution as a reference model, precisely because (it is worth repeating) they "could usefully be adapted to many cases in which there is divergence in people's beliefs about which kinds of public goods are valuable and which are not". I therefore believe that claim 2 plays a crucial role in Sugden's treatment. It is perhaps no coincidence that he concludes the paragraph dedicated to the 8 × 1000 with a final objection to Sandel's (2012) critique of "market reasoning" and the alleged superiority of public discussion (or "public reasoning", in Amartya Sen's terms 19 ) even on issues on which it is not at all easy to reach a universal consensus: in the spirit of Sandel's critique of "market reasoning", one might object that the tendency of the otto per mille procedure is to empty public life of moral argument and to drain public discourse of moral and civic energy. That is probably true. Political argument would be more intense and energetic if a collective decision had to be made about which (if any) religion was, by virtue of the truth of its doctrines, worthy of state support. But in the light of centuries of experience of wars of religion, one might conclude that energy was a mixed blessing (173).
We must now consider some important but little-known differences between the 8 × 1000 and 5 × 1000, which need to be analysed in order to "test" the validity of Sugden's account. In the following sections, I will focus mainly on the 8 × 1000 and 5 × 1000, since the 2 × 1000 is identical to the 5 × 1000, at least as regards some of the characteristics I intend to analyse, namely taxpayers' freedom and opportunity to choose and the redistribution mechanisms linked to expressed and non-expressed choices.

Considering some objections and improving Sugden's account
I would like to point out at the outset that it makes no sense to criticize Sugden for his possible lack of knowledge of some technical details or rules of these mechanisms-the fact, for example, that he does not consider the 5 × 1000 (or 2 × 1000) at all-because, as mentioned, there is no literature on the subject in English, and it is likely that not even all Italian taxpayers are aware of such technicalities. However, I would like to analyse these aspects and put forward possible objections and counterobjections, in the spirit of clarifying and, if possible, improving Sugden's explanation and theoretical perspective.
For this purpose, it is necessary to distinguish several aspects: taxpayers' freedom; the redistribution criteria for expressed and non-expressed choices and the connected criteria for the redistribution of opted and non-opted funds; and the implications of these criteria on the same taxpayers' freedom. These are closely related aspects, but I analyse them separately to better understand all the implications. I will intertwine these considerations with the analysis of other principles and assumptions of Sugden's contractarian explanation: the opportunity criterion (more opportunity is better than less) and the related normative value of freedom of choice, the analogy with voting and the voluntary exchange tradition.

Taxpayers' freedom: autonomy and choice
To understand in what sense, in the case of PTDIs, we can speak of taxpayers' "freedom", let us address three possible objections: taxpayers' freedom is rather limited because: (1) PTDIs remain within a mandatory state framework; (2) the amount of taxes over which the taxpayers have a decision-making power is very small; (3) their freedom of choice is merely binary: to give or not to give. I will use these objections not only to further explain the normative value of PTDIs, but also to address some aspects of Sugden's contractarianism that might need some clarification or improvement, in particular through the addition of some notion of freedom, in the sense of autonomy.
Sugden's opportunity criterion and the implicit freedom of choice has appeared to some scholars to neglect freedom in the sense of "autonomy", "agency", "selfconstitution" or "self-creation". According to them, this freedom as autonomy should complement freedom as choice. This has raised an interesting debate, on which I cannot dwell more herein (Schubert 2015(Schubert , 2020Sugden 2015Sugden , 2019Dold and Rizzo 2020).
For our purposes, namely, to understand the taxpayers' freedom implied by TPDIs, the following considerations might suffice. Sugden seemed to be willing to accept a "minimal" notion of autonomy as "a sense of volition. It is a person's subjective perception of herself as the cause of her own actions-her perception that she has the power to act in ways that in fact she chooses not to do" (Sugden 2019: 27). Some of these scholars (Dold and Rizzo 2020: 3-4) have further noticed how it is implicit in Sugden's adherence to Buchanan's notion of "liberty as individual sovereignty" that "the requirements of "sovereignty" necessarily go beyond the presence of opportunities. Sovereignty requires a sense that the choices are "owned" by the individuals, that they were not merely puppets of outside influences". To me these latter considerations suffice to put freedom of choice in relation to this sense of autonomy, insofar as we accept the following minimal understanding of their relationship: choice and autonomy "mutually reinforce one another: we value autonomy in part because of the freedom to choose that it validates, and we value choice in part because it contributes to our autonomy" (Dan-Cohen 1992: 221). 20 In the light of these considerations, we can now try to understand the taxpayers' freedom implied by TPDIs by addressing the aforementioned three objections.
The objection that PTDIs remain within a mandatory state framework misses the difference between the situations ex-ante and ex-post the introduction of these mechanisms. We have seen that the 8 × 1000 is conceived by Sugden as one of those forms of regulation that "would be justified as ways of expanding opportunity for mutually beneficial transactions" (ix, italics added) and that this claim can equally be applied to PTDIs in general. In the ex-post situation taxpayers' "freedom" is undoubtedly "expanded", note, however, not only in terms of their opportunity sets and freedom of choice, but also in terms of their autonomy. It is one thing to pay taxes, stop, and another thing to pay them while having the freedom to decide autonomously and choose whom to give (a share of) them to. Here "autonomously" can be taken in the minimal sense that it is the taxpayers-and not or no longer the state-who have the right to decide how to spend their tax money, or that the state does not or no longer chooses in their place. I think this is consistent with Sugden's acceptance of the notion of autonomy as the "person's subjective perception of herself as the cause of her own actions".
Note that these considerations concerning the "expansion" of taxpayers' freedom apply equally to the second objection, as the taxpayers' decision-making power, however small, is greater than the ex-ante situation. It is not my intention to deny the positive and "quantitative" correlation between taxpayers' freedom and the share of income tax over which they can exercise their choice (I will say something more about this in Sec. 4.3). But the problem of freedom is also and above all of a qualitative nature.
With regard to the third objection, to reduce taxpayers' freedom to a merely binary choice amounts to neglecting the relevance of both the varieties of choice among and within the three existing institutions, as well as taxpayers' freedom as autonomy.
As for taxpayers' autonomy, it exists both from a legal-institutional perspective and from the taxpayers' subjective perception. First of all, by introducing these institutions in the name of the constitutional principle of horizontal subsidiarity, the legislator intended to restore some fiscal sovereignty to taxpayers (Antonini 2012: 12-14). Moreover, note that it is the law that recognizes taxpayers this type of freedom, giving them this sovereignty, "power" or "authority" to choose, which, properly speaking, is a "right" to choose. I think, therefore, that in this case the link between freedom as autonomy and freedom as choice is evident, to the point that we could speak of autonomy-in-choosing or choosing-in-autonomy.
With regard to the taxpayers' perception, a recent survey has shown that the vast majority of taxpayers perceive the choice of the 5 × 1000 designation as a form of decision-making autonomy regarding the allocation of their taxes and that they appreciate the 5 × 1000 mainly because it is they (and not the state) who decide "who part of my taxes goes to". Such decision-making autonomy has been assumed as a proxy of the taxpayers' fiscal sovereignty : 56ff and 105ff).
All these reasons also explain why these types of institution can be framed conceptually as forms of quasi-voluntary taxation, where the expression "quasi-voluntary" implies that, as noted above, while the tax levy remains mandatory, the choice of how to spend a part of one's taxes is free and left to the taxpayer's will. 21 Furthermore, the freedom of choice among and within TPDIs depends on the scope of the taxpayer's opportunity set and on the level or sub-level at which the choice is made. The binary choice-to give or not to give-only refers to the first level of choice, and, to be more precise, taxpayers have not one, but three binary choices at this level, or, properly speaking, we should say that taxpayers have three different opportunity sets-the 8 × 1000, 5 × 1000 and 2 × 1000. Beyond Sugden's possible lack of knowledge of the 5 × 1000 and 2 × 1000, we may say that his account is still fitting if we consider (as we should) the three PTDIs as a whole and, above all, that the taxpayers' choice of one designation does not exclude the others. Also, within TPDIs, consider for example the 5 × 1000: it gives taxpayers a very wide choice among thousands of beneficiaries. 22 Therefore, overall, the taxpayers' opportunity sets as well as their freedom of choice, both at the first level and above all at the sub-levels of choice, are indeed extremely ample.
In virtue of all these considerations, contractarians have more than one good reason to propose the introduction of PTDIs to citizens and legislators: because they expand not only their opportunity and freedom of choice but also their autonomy.
However, the problem changes if we consider the different redistribution criteria for the expressed choices and, above all, the non-expressed choices in the case of the 8 × 1000 and 5 × 1000.

Expressed choices
We have seen that in the case of the 8 × 1000, the state redistributes 0.8% of the total tax revenue in proportion to the taxpayers' expressed choices.
Though Sugden seems to be aware of the existence of this redistribution rule, he seems to believe that the taxpayer allocates "her 0.8 per cent [of income tax]" (172). But we have also seen that the 8 × 1000 redistribution rule implies that the 21 This is also the reason why it is important to distinguish between "quasi-voluntary" and "voluntary" taxation, where the latter is usually understood as voluntary contribution or gift-giving (see Sloterdijk (2010), contra: Honneth (2009) and Zizek (2012)). On the relevance of Sloterdijk's proposal see Silvestri (2015) and Gribnau (2017). All things considered, it is still possible to conceive the giving of the 0.5% as a gift-giving (Sect. 4.5). 22 The very wide variety of choices in the case of the 5 × 1000 may raise a further problem, known as "choice overload". I do not think it is a severe problem either in general or with regard to the 5 × 1000. First, because I agree with Sugden that in many real-world scenarios, for example in retail markets, there are mechanisms "which favor the emergence and persistence of conventions that reduce the complexity of consumers' choice problems" (TCOA: 146). In the case of the 5 × 1000, there is sporadic, but rather reasonable evidence that taxpayers tend to repeat the choice made in the past following a logic of trust and loyalty towards the beneficiaries. Such evidence emerged from several focus groups held with representatives of tax consultants (who have a crucial role in reminding taxpayers to exercise their choice) carried out in view of preparing for some research on the 5 × 1000 ).

3
individual taxpayer's choice does not determine the designation of "his/her" 0.8% share of income tax, and, therefore, it is not entirely appropriate to say that the 0.8% share is "his/hers". Nevertheless, if we consider the proportional voting mechanism to be a valid analogy, Sugden's approach still stands up: from the taxpayer's point of view, expressly choosing the beneficiary of one's "own" 0.8% is like expressing a vote in favour of that beneficiary. In this way, the more a beneficiary is "voted", the more resources it will receive.
The 5 × 1000, on the other hand, calculates 0.5% of the actual amount of the taxpayer's income tax, and therefore, the 0.5% share varies according to the taxpayer's income. Thus, higher-income taxpayers, who therefore pay higher taxes, will be able to transfer more resources to their beneficiaries than low-income taxpayers. Consequently, the respective beneficiaries will also receive more or less according to the income of the taxpayers who have chosen them.

On the analogy with voting
With regard to these two different mechanisms for redistributing the expressed choices, one would think that a gap, if not a conflict, opens up between the two guiding principles of Sugden's approach: on the one hand, the principle of proportional spending based on the analogy with the proportional voting mechanism (but neither the principle nor the analogy is taken up by Sugden in TCOA); on the other hand, the tradition of voluntary exchange (taken up and explicitly referred to in TCOA), which instead moves from the principle of tax justice known as the benefit principle. However, in the 1990 paper he had already dealt with a possible objection, which now, albeit with hindsight, I think can be reread and reformulated to address this type of problem. Some readers will have noticed that my proposal ties taxation and representation together in a way reminiscent of eighteenth-century thinking. To say that each individual should retain control over the destination of his own tax payments is to say that decisions about public spending on public goods should be taken only by taxpayers. And the more tax a person pays, the more influence he has on these decisions. So, it might be said, my proposal amounts to disenfranchising the poor. (Sugden 1990: 77).
Although this objection is somewhat weak from Sugden's viewpoint (ibid.), he is willing to introduce a "variant of the principle of proportional spending".
Instead of treating each person as having an entitlement to the taxes he pays, we may treat everyone as having an entitlement to an equal share of the total public-good budget. As before, the size of this budget is a matter for collective decision, say by majority vote, and then each citizen chooses how to allocate his share of the budget between different public goods. The difference is that every citizen has the same amount of money to allocate, irrespective of the amount he has paid in taxes. (This is effectively how the West German system 1 3 Percentage tax designation institutions. On Sugden's… of subsidizing political parties works: a party receives the same amount per vote, irrespective of how much tax its voters pay.) (Sugden 1990: 78).
In the light of the aforementioned redistribution criteria for expressed choices, as well as Sugden's proposal to deal with the possible objection of the disenfranchisement of the poor, one would think that the general approach he proposes holds up more strongly on the benefit principle and, therefore, works better to explain the 5 × 1000. The variant (or exception) proposed for the principle of proportional spending, that is, "treat[ing] everyone as having an entitlement to an equal share of the total public good budget", on the other hand, is more suitable to explain the 8 × 1000.
In truth, the issues are more complex and, therefore, several clarifications must be made. First of all, and more generally, it is the analogy between the principle of proportional spending and the proportional voting system which in the case of the 8 × 1000 and 5 × 1000 institutions has some limitations (and perhaps it is no coincidence that Sugden does not take up this analogy when analysing the 8 × 1000 in TCOA).
To explain the point, let us recall the last two (of the three) rules for choosing among public goods proposed by Sugden: (2) tax-raising decisions are made collectively; (3) tax-spending decisions are made individually. In the case of the 8 × 1000 and 5 × 1000, all citizens, having the right to vote, have a form of indirect control (through their representatives in parliament) over the tax-raising decision. However, not all citizens but only citizen-taxpayers-more precisely those who have incomes above a certain threshold and therefore have the duty to fill in the tax return-have the "right" to make tax-spending decisions, that is, to choose how to allocate a share of (their) income tax. For the same reason, the analogy with the German system of subsidizing political parties should also be used with caution, since in this case there is an immediate and direct link between the exercise of the right to vote and the redistribution of resources to the parties according to the "decisions" (the votes) of all voting citizens.
The fact that only the citizen-taxpayers can choose is also one of the reasons why we need to be cautious in qualifying these mechanisms as a form of fiscal democracy (Fazekas 2000;Bullain 2004;Antonini 2005Antonini , 2007. As I have proposed elsewhere, it seems to me more accurate to qualify these mechanisms, at least from an institutional point of view, as forms of "tax allocation without representation", for at least two reasons: (a) because it is an institution that implements a tax allocation based on an individual and decentralized decision-making process, that is, without any public discussion on the matter (and even less "democratic participation") and without the mediation of a central authority; (b) because it is the first and only mechanism (to my knowledge) that goes, at least "partially", beyond the modern principle of "no taxation without representation"; "partially", that is, the representatives in parliament are only "bypassed" by taxpayers with regard to tax-spending decisions (Silvestri et al. 2020: 58-59).
Finally, in support of Sugden's concern about the objection that these institutions can disenfranchise the poor, it seems to me that this objection is misplaced for (at least) two reasons. First, all citizens and, therefore, also the poor, have at least some form of control over tax-raising decisions through their representatives in parliament, in particular over how much of the public budget must be spent and for which types of public goods (8 × 1000, 5 × 1000 and 2 × 1000), as established by the law that implemented these fiscal institutions. However, it should be noted that this particular form of control can only be exercised at the time of the parliamentary discussion to enact the laws or in moments of subsequent amendments to these laws (or any future repeal of the same). It is, in any case, an important form of control, all the more so if the tax revenue resources on which even the poor can make this kind of general decision are not "theirs" but other people's, even though one day they could be theirs, for example if they emerge from their condition of poverty and become taxpayers.
Secondly, and reasoning in comparative terms between the decision-making "power" of the poor and that of those with taxable incomes (among whom, it should be noted, the "rich" are only a very small minority), the latter do not have such great decision-making power over tax spending. Moreover, this power cannot be exercised on the entire tax revenue, since the taxpayers only decide on a very small percentage of their income tax.
The latter claim may give credit to the already encountered objection (Sec. 4.1) that taxpayers' freedom is rather limited, precisely because their decision-making power over their income tax is not that great. As said, I cannot deny that there is a positive correlation between taxpayers' freedom and the share of income tax over which they can exercise their decision-making power, even though I have also added that considerations concerning freedom cannot be reduced solely to a quantitative discourse. This opens up a possible trade-off or conflict of values between taxpayers' freedom and other fairness or equality considerations regarding the poor. In other words, the larger the share of income taxes on which taxpayers can decide, the more the objection of disenfranchising the poor and any attempts to mitigate it can make sense. However, this is a political, ethical, legal and contingent problem, dependent both on the law that establishes this share or percentage, and on the will of the people (including the poor) who legitimize this law. By voting, the people can always try to modify the existing law in order to increase or even, if they like, decrease this share.

Non-expressed choices
Let us now return to the taxpayer's freedom-opportunity of choice in relation to the problem of the redistribution of non-opted funds deriving from non-expressed choices. The question mainly concerns the different functioning of the default rules (or default setting) in the event of non-choice by the taxpayer.
In the case of the 5 × 1000, without prejudice to the redistribution criteria envisaged for the expressed choice of a generic sector, which, however, is still an expressed choice, if I do not choose (in favour of a generic sector or a specific beneficiary), my 0.5% remains a tax in all respects and therefore flows into the state coffers.
In the case of the 8 × 1000, however, if I do not choose, "my" 0.8% share will be redistributed among the beneficiaries in proportion to the number of choices expressed in their favour by those who do instead choose. This means that the redistribution is only determined by the taxpayers who express their choice, who therefore also decide for those who have not expressed any choice.
Let us take an example. Imagine a simplified world in which: (a) there are 100 taxpayers and only the state and the Catholic Church among the beneficiaries; (b) the budget corresponding to 0.8% of the tax revenue is 100 euros; (c) 10% of taxpayers express a choice in favour of the state and 40% in favour of the Catholic Church; (d) the remaining 50% of taxpayers do not express a choice.
The redistribution criterion for the non-opted funds means that 50% of that budget will be redistributed in proportion to the taxpayers' choices. The data are shown in the table below. I am not sure that Sugden knows these implications of the 8 × 1000 redistribution mechanism, since, if he did, he would also have had to consider the implications for the taxpayer's freedom of non-choice and draw some important conclusions.
If we take into account the effects of non-choice and the consequent mechanism to redistribute the non-opted fund, it should be said that the 8 × 1000 is a mechanism that risks not adequately representing the intentions of those taxpayers who, for various reasons, do not choose. The 8 × 1000 is compulsory, which means that the redistribution will take place in any case and regardless of the intentions of those who do not choose, and will be carried out according to the intentions and decisions of those who made the choice. Therefore, those taxpayers who do not choose end up giving a part of "their" 8 × 1000 to the Catholic Church and, to a lesser extent, to the state and other religious confessions.
Unsurprisingly, the redistribution mechanism for non-expressed choices and non-opted funds has been at the centre of controversy and criticism. For example, the Corte dei Conti (Court of Auditors, in particular the "Central Section' that controls the state administrations' accounts management) has highlighted some critical issues in the functioning of the 8 × 1000. Among these, the court has often focused on the problem of non-expressed choices and the lack of publicity of the quota allocation mechanism: the mechanism neutralizes the non-choice. In this way, everyone is involved, regardless of their will, in the financing of confessions […] This has a further consequence: the effective weight of a single choice is inversely proportional to the total number of those who make the choice. The redistribution of the non-expressed choices benefits, above all, the major beneficiaries.
[…] The system, therefore, is not entirely respectful of the principles of proportionality, voluntariness, and equality. The state administrations provide little information on this peculiar manner of redistribution. In fact, taxpayers-even with average diligence-can reasonably be led to believe that funds are only redistributed and assigned when an explicit choice is made. (Corte dei Conti, Deliberazione no. 16/2014/G: 75).
So, to make a slightly brutal simplification, one would therefore say that the 8 × 1000 default rule works like a nudge, such as in the case of nudging organ donation through an opt-out system, where the consent of the "donor" is merely presumed. This is something Sugden would certainly not like to know, given his criticism of Sunstein and Thaler's (2008) libertarian paternalism (see, above all, TCOA, Ch. 4). It is also true that, by recognizing the risk of manipulating or exploiting the status quo bias intrinsic in these kinds of default rules, many libertarian paternalists have recently been invoking the need for an expressed choice or consensus. Even though in the case of the 8 × 1000 an opt-out solution is not possible, since the choice of the 8 × 1000 is mandatory for everyone, one could nevertheless argue that an expressed choice is always better than a presumed choice in order to better redistribute the 8 × 1000 revenue according to the taxpayers' intentions. Note that the solution put forward by the Corte dei Conti-an information campaign aimed at increasing taxpayers' awareness-might resolve the problem only partially. This is because, as we have seen, roughly 30% of taxpayers do not express their choice, as they do not have the duty to fill in their tax return, and in order to express the choice they have to follow a more complicated procedure that could cost them more time and effort. For example, they would have to go to a tax consultant, or send the form with their expressed choice to the Italian tax authority through the post office.
A possible counter-objection to these issues is that one can apply Sugden's analogy between the voting mechanism and rules for choosing among public goods (Sugden 1990) to the 8 × 1000 redistribution of non-expressed choices: those who do not go to vote will simply submit to the choices of those who vote. To put it another way, this can be understood through Buchanan's (1962) distinction between the constitutional and post-constitutional levels of public choice. But this might be considered a bit of a stretch since the "rules of the game" established by PTDIs are not usually introduced or enacted at constitutional level but through ordinary laws.
Last but not least, Sugden may rightly claim that if taxpayers do not express a choice between the Italian state and a list of humanitarian/religious causes, there does not seem to be any plausible reason why the default assumption should be that they intended to send their tax money to the Italian state. 23 But if I am not mistaken, he would not accept that others choose in my place, especially on matters that affect my religious beliefs. I think this is implicit in Sugden's claim that "a contractarian would say that how Seventh-Day Adventists choose to spend their own money is no business of Catholics, and vice versa", and at the same time is no business of the government (172). In this respect, I still believe that allocating the tax money deriving from taxpayers' non-choices to the state coffers to finance the wide variety of public goods and services, like in the 5 × 1000, is a more "laic" or more "neutral" default rule, at least compared to the 8 × 1000.

Beyond the voluntary exchange paradigm. A new type of tax justice
Finally, and importantly, I want to show how and why reading the PTDIs with the voluntary exchange paradigm could be misleading, if not wrong, for at least two interconnected reasons. First, because the main basic assumptions of this paradigm-the benefit principle of taxation, taxpayers' self-interest and the very logic of exchange-are not suitable for explaining taxpayers' behavior as regulated by PTDIs. Second, because PTDIs can be read as a new form of voluntary tax justice and redistribution from those who are better off to those who are worse off.
Let us first remember the underlying assumptions of the benefit principle of taxation. As a normative principle, it is based on a notion of commutative justice and reciprocity-namely, equivalence between what individuals pay, through their taxes, and the benefits they receive from the public goods and services provided by the public authorities. Taxes are therefore considered to be the "price" paid in exchange for a good or service. Again, the benefit principle of taxation is often associated with a rule of proportional taxation. This is also the reason why the benefit principle is often criticized by advocates of egalitarian perspectives, who generally prefer progressive taxation. 24 We have seen how Sugden himself (1990) had tried to introduce a variant of the principle of proportional spending in order to address this kind of criticism. We have also seen that, in TCOA, he rightly noticed that taxpayers may have non-self-interested and non-instrumental preferences for the public goods that can be chosen with the 8 × 1000 (I will show how this behavioural assumption can easily be extended to all types of TPDIs). Nevertheless, this behavioural assumption should also have led him to downsize his reliance on the voluntary exchange paradigm when he deals with the 8 × 1000 in TCOA. In other words, I believe that the only part of this paradigm that remains valid, if it has to be used to provide an account of TPDIs, is the important amendment introduced by Sugden (1990) himself, namely the three rules for choosing among public goods according to a contractarian approach. However, if this paradigm is emptied of its basic assumptions, one might reasonably wonder whether we have not gone far beyond the voluntary exchange tradition.
To get to the point of my thesis, let us first set aside a possible and useful objection: that one could be led to believe that the 8 × 1000 is more egalitarian than the 5 × 1000 and/or that the latter is closer to the benefit principle. This is because in the 8 × 1000, the shares allocated by taxpayers are independent of their income, and their choices all have the same weight (at least prima facie), whereas in the 5 × 1000, high-income taxpayers must pay a higher tax, and therefore, their 0.5% share will be higher than that of low-income taxpayers. Therefore, according to the benefit principle, the former should benefit more (from the public goods they choose to finance) than the latter.
But this objection and its reasoning are fallacious for various reasons, and above all because in PTDIs in general, and in taxpayers' choices in particular, it is precisely the logic of exchange as well as the return benefit that are substantially lacking.
First of all, and in general, it was in the "spirit of the law", or in the legislator's "intentions", that these fiscal institutions were to be implemented for social purposes and to help the most disadvantaged. This is intrinsic both in the 8 × 1000 and 5 × 1000, namely in the mission of the churches, the humanitarian and social purposes pursued by the state, the mission of the vast majority of third-sector organizations, and, above all, the very principle of subsidiarity in the name of which the 5 × 1000 was implemented. 25 Secondly, and more in particular, those who choose to allocate their share of income tax are very unlikely to choose because they want to have something in exchange (for this share of taxes); in other words, they are very unlikely to expect a benefit in return. And this is not only because taxpayers' choices are anonymous and, therefore, they cannot even expect a "thank you" from the beneficiary. A recent survey has shown that the vast majority of taxpayers perceive the choice to give their 0.5% as driven by an "altruistic" motive and that there is more than one good reason to think that this choice can be formalized as a distinct type of gift-giving, namely as a "gift-without-sacrifice" (Silvestri et al. 2020: 52ff). Roughly speaking, giving 0.5% costs the taxpayers nothing, but giving or not giving makes all the difference from both the giver's and, above all, from the receiver's point of view: if a beneficiary receives a taxpayer's 0.5%, from the beneficiary's point of view it is a gift in any case. If the taxpayer is "altruistic" and therefore aims to "do good to the other", namely the beneficiary and those assisted by the beneficiary, then he/she has to give. 26 Thirdly, the 5 × 1000 is also egalitarian if we think in terms of redistribution from those who are better off to those who are worse off. Since a progressive income taxation system already exists in Italy, as in many other Western countries, accordingly the high-income taxpayers' 0.5% is progressively greater than that of the low-income taxpayers. Therefore, the former will give progressively more to their beneficiaries. In turn, these organizations will provide more and better assistance services to the people they benefit, including the poor.
For all these reasons, both the 8 × 1000 and 5 × 1000 go beyond the voluntary exchange paradigm and implement a new form of tax justice on an individual voluntary basis.

Conclusions
We have seen how Sugden's contractarian account of the 8 × 1000 in TCOA is articulated around five claims. The 8 × 1000 "illustrates the logic of a contractarian approach to the provision of public goods"; "could usefully be adapted to many cases in which there is divergence in people's beliefs about which kinds of public goods are valuable and which are not"; presents a very interesting aspect "from a contractarian viewpoint: […]: each taxpayer can choose which fund on the list will receive her 0.8 per cent"; "provides a rough and ready response to the free-rider and hold-out problems"; and "is also a rough and ready way of ensuring that people contribute to the costs of only those public goods that they actually value".
I have tried to broaden the analysis perspective to also encompass the 5 × 1000 and 2 × 1000, the sister institutions of the 8 × 1000, while evaluating various aspects and assumptions of Sugden's contractarian account-including the opportunity criterion and freedom of choice, and his development of the voluntary exchange tradition-in order to integrate his explanation with various clarifications and understand where it shows some limitations requiring corrections and improvements.
In particular, I have shown how Sugden's contractarianism and the implicit normative value of freedom to choose can easily be integrated with a minimal notion of freedom as autonomy that duly takes into account taxpayers' sovereignty, and how such an integration would reinforce his contractarianism insofar as it aims to provide normative justification for the introduction of PTDIs.
I have also analysed how the implications of the 5 × 1000 and 8 × 1000 default rules may affect these types of freedom. In the case of the 5 × 1000, if I choose not to choose, my choice and my 0.5% will not be affected by other taxpayers' choices. The 8 × 1000 default rule limits taxpayers' freedom not to choose, at least in the sense that the mechanism pushes me to choose if I do not want 0.8% of "my" tax money to go to religious confessions whose values I do not share. The 8 × 1000 seems closer to Sugden's (1990) analogy with the proportional voting system, at least in the sense that those who do not choose to allocate "their" 0.8% (or do not go to vote) will have to accept or "submit to" the decisions made by others. But this is not entirely compatible (at least for those who do not choose) with the principle of freedom of choice implicit in Sugden's contractarian approach, or with a minimal sense of freedom as autonomy. Is it, perhaps, also for this reason that in TCOA he did not resume the analogy with the proportional voting system proposed in his 1990 paper?
Finally, I have shown how the Italian PTDIs go beyond the voluntary exchange paradigm and implement a new form of voluntary tax justice.
In any case, the contractarian account, as important as it is, is only one among many: much work remains to be done to understand PTDIs as well as taxpayers' behaviour from a theoretical and empirical viewpoint, as well as from several disciplinary perspectives. Various issues are worth exploring, such as similarities and differences between PTDIs and other forms of "tax choice" implying the "taxpayer's sovereignty" (Buchanan 1963;Wagner 2013;Le Grand 2003 or participation in decisions as to how tax is spent (Alm et al. 1993); experiments aimed at testing forms of voluntary taxation (Li et al. 2011); historical, economic, anthropological, legal and philosophical reconstructions of forms of voluntary taxation (Einaudi 1940(Einaudi /2014(Einaudi , 2017Silvestri 2015Silvestri , 2017Gribnau 2017) or "voluntary cities" (Beito et al. 2002). That is not to mention similarities and differences between taxpayers' behaviours in TPDIs and the huge literature on voluntary contributions to public goods, 27 and, above all, on charitable giving, where the behavioural hypothesis of a "gift-without-sacrifice" in the case of the 5 × 1000 ) seems to challenge the utilitarian assumption of the literature on charitable giving-ranging from Becker (1974) to Andreoni's "warm glow" (1989Andreoni's "warm glow" ( , 1990)-at least in the sense that the "absence of sacrifice" reopens the possibility of understanding "altruism" through a new lens.
Again, much work remains to be done to understand how PTDIs challenge our dichotomic way of thinking, being a sort of new "third way" between state vs. market, welfare state vs. third sector (and civil society), institutionalized solidarity vs. spontaneous solidarity, obligation and freedom (Silvestri 2019;Kesting, Negru and Silvestri 2020;Silvestri and Kesting 2021).
And perhaps one day we will discover that these forms of quasi-voluntary taxation really are perceived by taxpayers as bellissime.