The Economics of Tax Behavior: The Absence of a Reflexive Ethical-Economic Agent

Abstract

In this article I do four things. First, I set out and critique mainstream economic theory of tax evasion and its lack of a reflexive ethical-economic agent. Second, I set out more innovative work on “tax morale.” Third, I establish that insofar as work on tax morale draws on behavioral economics it has more continuity with previous work on evasion than one might expect, given that the implications of tax morale for an economic agent seem different. Fourth, drawing on concepts of deliberation, moral economy, and positional objectivity, I provide brief discussion of alternatives that encourage and respect a reflexive ethical-economic agent.

Keywords:

• tax evasion
• tax morale
• moral economy
• positional objectivity

JEL Classification Codes:

• B40
• B52
• K34
• L29
• A11
• A12
• A13

In this article I sketch the origins and illustrate the development of the problematic relation of ethics to economic theory of tax. I first set out and critique Michael Allingham and Agnar Sandmo’s formative paper on the economics of tax evasion (Allingham and Sandmo 1972). Tax evasion is the illegal withholding or non-payment of tax. This is typically contrasted with “aggressive” tax avoidance, which is the deployment of strategies to legally radically reduce tax paid, creating an “effective tax rate” that is significantly less than the headline rate would imply (e.g., Bergin 2021). I then explore more recent work on “tax morale” (Alm 2019; Luttmer and Singhal 2014). Theory of tax morale typically rejects an approach to theory of decision making solely using a standard utility function. It seeks to incorporate “non-pecuniary” factors and it typically adopts a “nudge” approach from behavioral economics to improve payment of tax. I argue that neither treats the economic agent as a realistic reflexive ethical-economic agent and that there is more continuity between the two approaches than it might at first appear. Before concluding, I then suggest that deliberation, Andrew Sayer’s moral economy, and Irene van Staveren’s use of Amartya Sen’s “positional objectivity” offer interesting ways forward insofar as they move beyond the standard economic agent to treat people as grown-ups (Sayer 2015 and 2007; Sayer and Morgan 2022; van Staveren 2007). It should be noted that the purpose of the article is to highlight issues and problems and not to provide a substantive theory of ethics or comprehensive discussion of all points made.

Allingham and Sandmo and the Economics of Tax Evasion

Mainstream economic theory of tax behavior does not treat the economic agent as a reflexive ethical-economic agent. This point can be readily misunderstood and requires elaboration to avoid traducing economists. The claim is not that theorists have no interest in ethics or are unethical. Mainstream economic theorists of tax typically conceptualize their work as contributing to solving problems of evasion and avoidance. However, the way an agent is conceptualized has inherent problems and this begins with the formative work of Allingham and Sandmo on tax evasion (1972).¹

Rational Evaders

As one might expect from a mainstream economics paper written in the 1970s, Allingham and Sandmo’s “Income Tax Evasion: A Theoretical Analysis,” begins by suggesting the first problem is how to formalize tax evasion and that the main question is whether a utility function can be constructed (see Allingham and Sandmo 1972, 323–324, 337–338). This is not a “why” question. It becomes clear as one reads the paper that it is not, in any serious sense, asked whether a utility function is an appropriate way to conceive decision making, it is simply self-evident within a methodological individualist frame of reference that this is the first step in addressing the problem. Concomitantly, the paper adopts the familiar architecture of caveats, the approach is a “theoretical analysis” which awaits empirical confirmation or application, it is a simplification using standard concepts within a standard method. The paper then further bypasses the problem of justification by stating all key variables are exogenously given, and so one need not be concerned with detailed discussion of how the behaviors leading to evasion are produced—one merely has a subsequent need to statistically confirm them.

According to Allingham and Sandmo, the economic agent must decide whether to be a tax evader or more specifically whether to pay tax to the tax authority or to withhold that payment. The decision, however, is a calculation. The economic agent calculates the probability that the authority will detect that a person is not paying tax, if that person were to decide not to pay tax. So, there is some observed possibility of being caught via an informant or audit. The agent then weighs this against the benefit of non-payment versus the relative cost of any penalty. Penalties (fines) constitute income loss. The model thus uses standard economic concepts: a calculative, self-interested agent with all necessary information who then processes that information in a goal-directed fashion in order to achieve an optimal outcome. In this case, maximized utility focused on tax behavior. The economic agent is “rational” as understood in mainstream economics, which means no more than the agent consistently follows rules whose format prefers more to less and most to least in order to achieve a goal.² Consider the implicit nature of the evasion calculation:

1. It is permissible to break the law.

2. Breaking the law is what one would do if not prevented.

3. One is motivated to break the law even if one does not do so.

This is the thought process being attributed to the rational evader, but who is this evader? The model does not explore or explain what makes some disposed to be evaders and some not. By this I mean there is no discussion of socialization, relationality in the form of culture, attitudes, and so on, that any Original Institutionalist might think were important contextualizing matters. The model is unspecified in this sense and so implicitly assumes anyone could be this evader. Moreover, insofar as these issues are omitted, the model seems to imply everyone is a potential evader. It is, from the point of view of the agent, merely the level of loss/gain and risk tolerance that matters and the level of scrutiny and sanction applied by the tax authority.

Normalizing Criminal Behavior

It should be immediately obvious here that what is missing is a realistic reflexive ethical-economic agent capable of deciding it is wrong to evade tax and right to pay it. In a representative democracy this decision is not isolated and not purely personal. Paying tax is a subset of following the law, and we follow laws not just because there is an individual pay-off in the given instance but because a society of laws both constrains and enables us all and this has potential emergent properties—a situation where one can trust and be trusted and where one need not rely on close monitoring and continual enforcement for any kind of activity to be possible.³ Moreover, the first presumption of paying tax is that it is a “pro-social” act which signals a commitment to the reproduction of social infrastructure, and, subject to society-specific formulations, is part and parcel of agreed socio-economic political goals, such as automatic stabilizers, welfare, redistribution and so on.⁴ None of this requires an individual to “like” paying tax, but it does require them to acknowledge it can be right to do so and that if one wishes a different tax system, one mobilizes and votes for change. Of course, though the presumption may be “pay,” there can always be instances where dissent creates a reason not to pay—but this is a publicly signaled act and so not “evasion” in the sense of dissimulation. Mass failure to pay the “community charge” or “poll tax” in the UK 1989 to 1993 was not evasion as such.⁵ It was protest in the form of civil disobedience.

By contrast, the Allingham and Sandmo model inadvertently normalizes criminal behavior.⁶ This is subtly different than the original intent, which is no more than to suggest tax evasion is rational behavior. The model implies tax evasion is simply what one would do if not prevented.⁷ To be clear, however, Allingham and Sandmo do also acknowledge that there are “non-pecuniary factors in the taxpayer’s decision” (Allingham and Sandmo 1972, 326) and that in addition to direct income loss from the sanction there can be other factors. They state “these factors may perhaps be summarily characterized as affecting adversely one’s reputation as a citizen of a community” (Allingham and Sandmo 1972, 326). Three points are relevant here. First, the statement does not lead Allingham and Sandmo to expand the nature of their inquiry regarding tax behavior to discuss matters of socialization, relationality and so on, and then to adopt other methods that place a break on the model—our point from before still holds. Instead and second, Allingham and Sandmo fit these “non-pecuniary” factors to their utility function.⁸ As such, the model invites any later empirical application to adopt a quantifiable approach to reputation loss and thus tacitly reduces reputation to indirect measurable effects on income (so the frame of reference is eventually pecuniary and individual). Third, reputation is an exterior effect not an internal state of mind. It presupposes that what matters is public perception, and this is different than whether in fact one has a concept of the right thing to do in an ethical sense. Put another way, reputation is subtly different than doing the right thing, since it is about appearances. It still turns on attitudes to being caught. A concern with reputation does not preclude evasion, it merely suggests reputation management strategies might be advisable.⁹

Insofar as the model agent’s concern is “reputation” then none of the three points lead readily within the model format to a reflexive ethical-economic agent. One might, of course, argue that the model’s calculative economic agent is implicitly consequentialist and thus they are an ethical agent in this sense. One might think this perverse and thus not worthy of comment, but it is at least worth clarifying why the argument is a non-starter, if only to establish that a contrarian radical libertarian could make little use of the model to make such a claim.

A consequentialist claim is untenable for at least three reasons. First, like any other economic agent constructed for the purposes of conformity to the requirements of a utility function, the agent is essentially a set of axiomatic capabilities. No real human has the calculative and information processing capabilities of this economic agent.¹⁰ The lack of specification in the model implies anyone might be a tax evader, but the nature of the model agent implies no one could be an effective consequentialist if this were dependent on the characteristics of the model agent. Second, the model agent does not intentionally adopt a set of ethical rules whose purpose is both the individual good and the social good as a product of the individual good (a standard formulation of well-functioning consequentialism). Allingham and Sandmo are simply adopting a standard economic concept to a specific subject matter and its economic agent is not choosing an ethics, consequentialist or otherwise. Third, one cannot apply a line of “as if” reasoning to circumvent the problem of non-adoption of an explicit ethics, since it would be an odd argument that suggested a general presumption of antisocial behavior was implicitly a social good. The model agent may be rational, as an economist understands that term, but they are not reasonable from a societal point of view.

The Turn to Theory and Practice of Tax Morale

Ken Devos’s book, Factors Influencing Individual Taxpayer Compliance Behavior, includes an extensive and excellent literature review of the development of theory of tax evasion (Devos 2014). As Devos notes, and the more recent work of Leonard Seabrooke and Duncan Wigan confirms (e.g., Seabrooke and Wigan 2023), Allingham and Sandmo’s work has been highly influential if always problematic. According to data on the journal site for Journal of Public Economics, in which “Income Tax Evasion: A Theoretical Analysis,” was published, the paper has been cited (as of September 2023) 2,472 times, including in 360 policy reports and documents. Google’s more expansive data scraping procedures picked up 8,420 citations (though this likely includes some multiples).

Allingham and Sandmo’s work encourages a cost-benefit analysis approach to evasion prevention policy—a deterrent approach that orients on adjustments to penalties and audits as well as inducements for informants. This has flourished in tandem with work on setting marginal tax rates, which is by no means to suggest work on marginal tax rates focuses exclusively on evasion, it has multiple contexts and numerous theoretical debates. Perhaps the highest profile and most controversial over the decades since the 1970s has been the adequacy of the Laffer theorem (popularized through the “Laffer curve”), though one might also make a case for theory of tax competition (e.g., on both, see Jaakola, Ylönen, and Saari 2023; Bergin 2021; Morgan 2021; Saez and Zucman 2019; Berman and Milanes-Reyes 2013; Rixen 2008).

As any empirical survey of countries will indicate, tax evasion is a persistent problem, as is avoidance. For example, in the UK, HMRC estimates the “tax gap,” of which evasion is one component, at 4.8% of “theoretical tax liabilities,” and the most recent figure for this in 2023 is a gap of £35.8 billion for 2021–2022 (Richard Murphy uses a different methodology and argues the HMRC approach is a massive underestimation).¹¹ So, while cost-benefit analysis type economic theory of tax evasion has become a mainstay of theory and policy, tax authorities have had reasons to remain open to alternatives that offer something different that might help them close this gap.

As many prominent mainstream commentators, such as Dani Rodrik in the US and Diane Coyle in the UK, have noted, economics has experienced an “empirical turn” during the twenty-first century (Coyle 2021; Rodrik 2015). The mainstream has diversified to incorporate information-theoretic economics, game theory, and behavioral economics and claims to be more evidence-based. In particular, behavioral economics and the work of Cass Sunstein, Richard Thaler, Daniel Kahneman and Amos Tversky on bounded rationality and prospect theory, and subsequent work that has elaborated the themes they highlight, has provided a critique of more “neoclassical” mainstream economics, of which Allingham and Sandmo’s work, and its descendants are examples. Unsurprisingly then, behavioral economists and fellow travelers in business schools with an interest in psychology, have turned their attention to the problem of tax evasion and policymakers have been receptive to this. This alternative field is termed “tax morale.”

Tax Morale

By 2014 tax morale was sufficiently developed as an alternative to warrant an essay in the American Economic Association’s Journal of Economic Perspectives. The journal was established in 1987, its remit is to publish surveys and syntheses of notable areas of new research, new fields and “state-of-the-art” thinking. Essays are typically solicited by the editors. Erzo Luttmer and Monica Singhal simply title their essay in the journal “Tax Morale” (Luttmer and Singhal 2014). According to Erzo Luttmer and Monica Singhal, while Allingham and Sandmo has been the “benchmark” for over forty years, there has been a growing “disconnect” between this formal theory and more diverse strategies of compliance adopted by tax authorities. The Allingham and Sandmo model focuses on penalties and is pecuniary. It is essentially about costs as threats designed to compel compliance and deter evasion. However, tax morale is concerned with strategies designed to encourage voluntary compliance. According to Luttmer and Singhal, this has implications for the kind of economics brought to bear on the problem. They define and illustrate tax morale as:

broadly as an umbrella term capturing nonpecuniary motivations for tax compliance as well as factors that fall outside the standard, expected utility framework. For example, individuals may have some intrinsic motivation to pay taxes or feel guilt or shame for failure to comply. They may comply due to reciprocal motivations: the willingness to pay taxes in exchange for benefits that the state provides to them or to others even though their pecuniary pay-off would be higher if they didn’t pay taxes. Individuals may be influenced by peer behavior and the possibility of social recognition or sanctions from peers. Cultural or social norms can affect the strength of these intrinsic motivations, reciprocal motivations, or the sensitivity to peers. (Luttmer and Singhal 2014, 150)

As with many recognitions in mainstream economics, acknowledgement follows years of changes to practice in the subject area. In addition to the UN, the OECD is the main international source of work on tax standards and policy. For example, the OECD is a source of model tax treaties, standards and mechanisms for transfer pricing and was engaged by the G-8 (now G-7) to undertake the widely publicized Base Erosion and Profit Shifting project (BEPS), whose principal aim was to tackle corporate tax avoidance. As the Tax Justice Network and other critics note, the OECD has limited influence, countries and regional organizations may choose to adopt its recommendations and it is by no means a seat of radical thinking, since it relies heavily on input from the world’s largest and most influential tax, accountancy and consultant firms and derives much of its specialist personnel from them in a revolving door system of employment and secondment.¹²

In any case, and the previous comment notwithstanding, as early as 2001 the OECD “Principles of Good Tax Administration” began to draw attention to the promotion of voluntary tax compliance as “good practice.” In 2012 the OECD produced a working paper (Daude, Gutiérrez, and Melguizo 2012) and a year later a paper surveying policy, “Tax and Development—What Drives Tax Morale?” (OECD 2013).¹³ This was followed by the Luttmer and Singhal paper for Journal of Economic Perspectives, and other similar journals have followed suit, such as Journal of Economic Surveys (Alm 2019). The Luttmer and Singhal paper has now been cited more than 350 times. Titles of some of the papers citing it provide a flavor of how the field has developed since:

• “Shaming for Tax Enforcement,” (Dwenger and Treber 2022).

• “Tax Morale and Fairness in Conflict, an Experiment,” (Engel, Mittone, and Morreale 2020).

• “You’ve Got Mail: A Randomized Field Experiment on Tax Evasion,” (Bott, Cappelen, Sørensen, and Tungodden 2020).

• “Requiem for a Nudge: Framing Effects in Nudging Honesty,” (Dimant, van Kleef, and Shalvi 2020).

• “Nudging for Tax Compliance: A Meta-Analysis.” (Antinyan and Asatryan 2020)

A pioneer and perhaps the best-known academic working on the subject has been Benno Torgler (e.g., 2022; 2016; 2007; 2003).¹⁴ It should, however, be clear that insofar as tax morale has been adopted by economists, it has been mainly within behavioral economics and has involved use of field, natural, and laboratory “experiments,” randomized control trials and so on.

A common approach is to try to shape the payment decision through timed and targeted messaging in order to trigger feelings likely to make one positively disposed to pay and/or induce negative feelings if one fails to pay. So, communication may be made during the month income tax filings are due. It may orient on zip codes/postcodes, income groups, and occupation categories, such as the self-employed, or enterprise types, such as small leisure, hospitality, and tourism businesses. It can involve different strategies and media: letters, texts, and advertising that highlight the numbers of people who pay, illustration of specific local instances of what tax “pays for,” statistics of the increased burden on others because of failure to pay, survey results of attitudes to those who do not pay, etc. These are all ways of framing the payment decision. The intent is to “nudge” the payer by shifting the decision process from an abstract frame of reference (focused on how much it costs the payer) to a meaningful and perhaps personal sense of its purpose. For a behavioral economist the use of control groups or “before and after” comparisons provide an opportunity for creation and use of datasets, and for model formation, hypothesis testing, and so on, according to the standard construction of controlled trials and experiments.

If one returns to the quote from Luttmer and Singhal it should be clear that tax morale speaks to the point we made in the previous section—the Allingham and Sandmo model agent pays no attention to socialization, relationality in the form of culture, attitudes and so on that one might think were important contextualizing matters—while tax morale does seem to pay attention to them. As such, tax morale seems to inherently repudiate the Allingham and Sandmo model agent. However, as anyone familiar with the development of behavioral economics and critiques of that development is likely aware, matters have not quite turned out this way and this is true also of tax morale.

Tax authorities, of course, are mainly interested in what works and tend to be agnostic about the adequacy of economic theory. For a tax authority, the Allingham and Sandmo approach provided a convenient focus on the efficacy of penalties (the stick), while tax morale offers more of a persuasion approach (the carrot), and so the two can readily seem to be policy complements. From this point of view economists may, tacitly at least, be relatively unconcerned that one approach inherently problematizes the agent of the other. Moreover, the mindset that seems to have followed the “empirical turn” also encourages this attitude. As both Dani Rodrik and Diane Coyle have argued, the “empirical turn” has led to a proliferation of models and the general inference they draw from this is that one selects the appropriate model from the available suite.

In any case, behavioral economics has not, in reality, involved a decisive break with the model agent one finds in Allingham and Sandmo. As James Galbraith recently observed, behavioral economics has mainly innovated by “relaxing restrictive assumptions” rather than building an alternative from the ground up (Galbraith 2021). This is by no means a new observation, numerous critics interested in philosophy of economics, political economy, history of economic thought and in the sociology of knowledge have pointed out that “new” approaches in mainstream economics tend to take existent core commitments as points of departure—and thus the new accommodates to the old, restricting change despite the more radical implications of critique (e.g., Davis and Morgan 2018; Milonakis and Fine 2009; Fine and Milonakis 2009). For example, Luttmer and Singhal suggest no more than that “the Allingham and Sandmo (1972) model does not capture the full range of factors relevant for compliance” (Luttmer and Singhal 2014, 151). This brings us to tax morale and the realistic reflexive ethical-economic agent.

Tax Morale: Measurement, Treatment, and the Neoliberal Subject

There are three linked and specific points one might make here in terms of work on tax morale and in combination they suggest that the approach still lacks a realistic reflexive ethical-economic agent and that this is ultimately self-limiting or counterproductive.

First, the overwhelming focus of tax morale work is empirical. This in itself is not a criticism, since it would be absurd to suggest theory that looks to test outcomes and has a strong sense that evidence is required was a bad thing. It would be an inadequate response to the “empirical turn” to repudiate evidence in general (even if one were not reaffirming formalism). Evidence, however, does not speak for itself—this would be, as the Quine-Duhem thesis holds, an empiricist error—assumptions and theory matter to the adequacy of evidence. While there is not the space here to detail the many critiques of behavioral economics and its methods, one can provide some sense of the main issues, and these begin from Galbraith’s point in the previous section—there is a tendency to take the existent model agent as the benchmark and then relax restrictive assumptions. As Blair Fix, following Jason Collins, points out, behavioral economics tends to:¹⁵

1. Begin with the model rational utility-maximizing agent and claim this is false.

2. Create a test of the falsity of the model agent and confirm that it is false.

3. Keep the model agent as a benchmark and label the behavior isolated in the test a “bias.”

4. Repeat the process for a new context of behavior.

The result is a growing list of biases or deviant behavior, but this is an inherently strange process precisely because a model agent who is known to be false—someone we have never been and can never be—remains the point of departure and “what we seemingly do” becomes sub-optimal and thus in the economic sense “irrational”—a quite different connotation than that which underlies the language used and that is the claim that what we do follows some context specific reasoning.

I say “seemingly do” because there is also a question mark here over the methods applied. The methods applied—experiment, randomized control trials, etc.—tend to isolate some particular behavior or effect in order to measure whether it exists and what its strength is. This, of course, artificially separates one aspect of a whole, and yet humans are social relational and social reality is not simply an aggregate of isolated behaviors—an additive situation. The isolation then, requires the researcher to look through rather than look at the complex causal mechanisms that stand behind any outcomes. Moreover, models are quantitative, seeking quantities for variables of interest, and model testing often depends on standard analytical statistics. So, given the problem of separation and isolation, the whole leads to spurious precision, as well as a mismatch between a heterogenous, irregular, and processual social reality and the basic demands of standard techniques. For applied economists these become concerns with “internal” and “external validity,” but what the concerns amount to is a high degree of uncertainty regarding the link between given numbers and real world relations, despite that the whole enterprise tends to turn on and highlight the strength of particular behaviors.

To be clear, the point here is not to criticize an interest in evidence, but to question the adequacy of the theory that stands behind evidence and the nature of the evidence produced—a matter that economic philosophers such as Tony Lawson, and critics of behavioral economics, such as Lars Syll, have had a great deal to say about (e.g. Syll and Morgan 2019; Lawson and Morgan 2021a). In any case, as James Galbraith also points out, once one realizes the strength of any behavior is highly contingent, the pay-off in behavioral economics is often banal, amounting to “real people do not appear to fit well into the portrait of autonomous, selfish, commodity-obsessed pleasure seekers that is “economic man”… These things even the dimmest observer of real-existing capitalism already knew” (Galbraith 2021).

One might, of course, argue that tax morale is a special case. There is a very obvious focus, a clear purpose and scope for a measurable effect “in the field.” Tax yields either rise or they do not, surveys of attitudes either indicate people are more disposed to pay tax or they are not. However, as the previous indicates, one must be careful here not to conflate a practical goal and empirical claim with the empirical adequacy of the theorization of the agent, insofar as the agent derives from behavioral economics. This brings us to a second point. Work on tax morale tends to take place within the standard mindset of “thinking like an economist.” It draws on and plays to the skillset, toolkit, and expectations. Few mainstream economists these days refer to themselves as positivists and few defend the positive-normative divide and yet the legacy of both is deeply ingrained (Morgan 2019 and 2018). Mainstream economists see themselves as data scientists. They set up tests of models, propositions and hypotheses and then make “policy recommendations”: a suite of models, a range of possibilities. Mainstream economists have occupied the niche of social engineering legitimator, they represent themselves as technicians with a technical toolkit. This has consequences.

The main aim of tax morale is to improve voluntary payment of tax (thereby reducing evasion), the main method is to home in on some behavioral bias and exploit it—peer effects, shame, and so on—a frame and a nudge. But how one treats a subject is also how one implicitly theorizes that subject. In marketing circles behavioral economists refer to themselves using the somewhat repellent term “choice architects,” they are quite clear that the service they offer is manipulation and arguably this is what tax morale offers. Reducing evasion may well be a “good cause,” but it is also worth pointing out that tax morale presupposes for its efficacy a culturally sensitive socially relational agent who has been socialized to care about what others think, has a peer group, and a sense of right and wrong, and yet the process of intervention treats them as a collection of feelings and behaviors that can be isolated and triggered. This is an inadvertent outcome of the skillset and toolkit of the economist, and the methods used. This may not seem like a major problem until you start to think about what tax morale is doing and what it is not, and this brings us to a third point.

Anyone familiar with the main trends exhibited by work on tax morale will likely agree that the overwhelming focus is on the methods and measurement. But this means that mainstream economists are giving little attention to the ethics of an agent. They are, in the main, neither addressing themselves to the ethical-economic agent or providing resources for ethical debate regarding the status and purpose of taxation. The main academic focus is the experiment and its result—and one finds this in unexpected places. One would, for example, anticipate that Journal of Business Ethics would host a lively debate on the nature and role of taxation, but even here the main focus is experiment and reporting of results (e.g., Mickiewicz, Rebmann, and Sauka 2019) and this is also the case for the few papers on the subject to be published in Journal of Economic Issues (e.g., María-Dolores, Alarcon, and Gare 2010; Martinez-Vazquez and Torgler 2009).¹⁶ Of course, one might point out that the pursuit of this work does not rule out work that addresses or engages the ethical-economic agent. But the relative use of resources and omission are important.

The treatment of the subject follows a standard neoliberal practice—the creation of a technocratic sphere of expertise, a conceptual space that citizenry are not often invited into, despite that they are its subject matter (on neoliberalism see Jessop and Morgan 2022). Moreover, the treatment may presuppose a world of socializations where there is an ethical-economic agent, but it does little to encourage reflexivity, since the work is not typically intended as a component of public discourse. It thus serves as an “enabler” of the passive neoliberal subject (albeit in a different fashion than one finds in myth-making explored by Wrenn 2014). In the end this is self-limiting and counterproductive since it can have little impact on the meta-socialization or dominant discourses of the time and their tendencies. As with expertise in general, tax morale is typically approached as a technocratic and narrow “problem-solving” theory (to induce voluntary payment). It thus tends to lack a critical dynamic able to encourage debate about why taxes are paid. One might describe this as the kind of subtle operative activity of power and expertise that mainstream economics has always dealt with poorly despite that economics itself is a significant site of power (see e.g., Ylönen, Jaakkola, and Saari 2021; D’Ippoliti 2021; Javdani and Chang 2023). As such, the point speaks to the defensive discourse that has surrounded taxation in the neoliberal period—we have, as Richard Murphy puts it, lost the “Joy of Tax” insofar as neoliberalism encourages us to think in individual terms and to focus on what it costs us rather than what its role is in society (Murphy 2015). This encourages a resentment towards taxation and politicians on all sides have responded by becoming apologetic about the need for taxation. This works with ideas of individualism while doing little to empower the individual to think constructively about the basic issues which they might be invited to decide upon. The approach is, therefore, not neutral in its effects, it inadvertently conforms to the world of socializations that are in fact observably dissolving social bonds.¹⁷ In this sense attitudes to taxation, and thus behavior in terms of payment, are symptomatic of broader crises in representative democracy and serve to illustrate the important role trust plays as a social glue in society (Lawson 2019; Lawson and Morgan 2021b; Colledge, Morgan, and Tench 2014).

Deliberation, Moral Economy, and Positional Objectivity

There is a great deal more that might be said here but the point I want to emphasize is that the absence of an ethical-economic agent is a significant one for mainstream economics in theory and practice. We may be, as Original Institutionalists aver, creatures of habit and routine, but as they also argue, we are self-conscious beings and thus inherently reflexive. There are numerous resources to draw on in exploring reflexivity and its relation to habit and routine, notably the work of Margaret Archer and Pierre Bourdieu (see sources in Archer and Morgan 2020) and the societies we live in can do more or less to invite us to be reflexive and the social sciences, including economics, can work towards or hinder this—by endorsing or questioning technocratic tendencies and raising issues of “social purpose” (e.g., Baker and Morgan 2021). There is, moreover, a long tradition of theory and practice that seeks to treat its subjects as grown-ups, rather than biddable test subjects. Deliberative democracy, for example, seeks to create conditions where relevant persons can come together in dignity and mutual respect to talk through issues, often in pursuit of solutions to problems: citizen assemblies, summits, panels, juries, and workshops, etc. (e.g., Gastil and Levine 2005). Deliberation is, of course, only as inclusive, representative, participatory, fair, and effective as design and commitment allow (see Elster 1998), and while it is no panacea, nonetheless the basic premise is to treat participants as fully integrated centers of ultimate concern, and this is quite different than the journey mainstream economics of tax has been on since the 1970s.¹⁸

Deliberative fora provide time and space to discuss deeper more contextual matters. A tax morale field experiment might test the degree to which targeted information can exploit guilt or shame to induce “voluntary” payment of tax, but this is quite different than a presentation of evidence and discussion of its significance according to constitution of socio-economic practice and trends. For example, there are different interpretations and claims made regarding the proportionate contribution to tax revenues of the very wealthy. According to the Institute for Fiscal Studies, the top 0.1% of income earners in the UK pay over 10% of all income tax (Johnson 2022). The top 0.1% is around 50,000 people. Other figures suggest the top 300,000 pay more than the next 20 million people in the income distribution combined. From one point of view, these few are vital wealth creators who we are dependent on—a perspective that suggests “do not kill the golden goose.” From another point of view, this dependency is a signal of failure, and the salient question is, why has the UK become so dependent on income tax receipts from the few? Clearly discussion of this issue is intimately bound up with issues of growing inequality, growing levels of working poverty, and so on. Moreover, interrogation of this matter cannot but lead to questions such as, why is it that the options open to the working poor turn on evasion and breaking the law, while those of the rich turn on tax avoidance and bending the law? Here, taxation and justice may well currently part company, since to the majority it likely feels unjust that the well-off need not break the law, it has merely come to reflect their concerns and interests—a quite different take on any kind of dependency that may exist, a quite different frame of reference for discussing “normalizing criminal behavior” than that which economic theory of evasion invokes, and a quite different framework for evidence on proportionality, since the question invites further investigation regarding what is paid compared to what could be paid.¹⁹

In any case, it is not my purpose here to discuss claims and counter-claims in detail, but rather to suggest there is great scope for informed discussion regarding the nature and role of taxation in society and the kind of agent this both presupposes and encourages. Clearly, this scope is intimately related to matters of trust, authority, and legitimacy and these are all matters of current concern, and one might add, something like MMT’s radical change in perspective regarding the function of taxation cannot be pursued along pathways popularized by mainstream economics. Such a huge change in understanding requires public debate and initiatives to develop public understanding—and there is surely great need for this in a period of “climate emergency” if we are to transition from questions along the lines of “how do we pay for this?” and focus on what really matters (it is easy to forget just how odd it is to be fixated on whether we can afford to do things that we ultimately have no real choice over—we either make the changes compatible with survival of civilization or we don’t).

If one were to look for resources, however, and by this I mean for discussion of the role of taxation and ethics rather than for MMT, then a good place to start would be with Andrew Sayer’s work on moral economy (Sayer 2015 and 2007). For Sayer, by moral economy:

what I mean is an approach to economic life which not only seeks to explain its structure and how it works but examines and evaluates the normative justifications for its practices, including their constituent norms (for example, regarding property relations), and their consequences. To do this it acknowledges that all economies are moral economies in the sense that they depend on the acceptance of certain norms regarding what people are expected and allowed to do in economic matters, and that moral beliefs both influence and are influenced by economic practices. (Sayer and Morgan 2022, 6–7)

It is this approach—testing out claims and counterclaims— that Sayer adopts in Why We Can’t Afford the Rich. He emphasizes:

this is not about the politics of envy … but the politics of injustice … The argument is directed not at particular individuals but at the sources of their wealth and power and the ways in which they are justified … focusing just on individuals actually lets the causes off the hook—the rules, institutions and situations that they are allowed to take advantage of … [the purpose, rather, is to] question the legitimacy of their wealth. (Sayer 2015, 2, 9)

Sayer’s work implicitly treats people as grown-ups who can be addressed and persuaded. One finds something similar in Mary Wrenn’s “mental models” (Wrenn 2018) and in Irene van Staveren’s well known essay drawing on Sen’s “positional objectivity” (van Staveren 2007). Ethics are a type of norm and are intrinsic to the human as an active agent producing the social world in which they live. Any adequate social science must come to terms with the values that do and could justify or disallow activities. Objectivity is not impossible or eschewed, but its basis is not data that speaks for itself, rather it is open-minded evidential argument for and within context: particular discourses, the value of which to society and for the theorist can be explicitly stated and then judged according to criteria. The importance of this for rethinking the role of economics and ethics is neatly summarized by van Staveren:

Fact and value are mutually related, both in the real world and in economic analysis [… and] I reject the dichotomy between fact and value, between positive and normative economics and between descriptive and prescriptive analysis […] there is a fundamental difference between distinguishing facts and values (which is necessary for conceptual reasons as well as for doing justice to the different meanings of these concepts) and placing them in a dichotomous relation … Contrary to such dichotomies, values are part of economic analysis, alongside and frequently intertwined with facts, rather than being separated from facts in the form of policy recommendations following a supposedly value-free analysis. (van Staveren 2007, 21–22)

In terms of our argument there is, to reiterate, a need for more discussion of the very nature of taxation and this calls for engagement with reflexive ethical-economic agents.²⁰

Conclusion

In this short article I have critiqued both economic theory of tax evasion and more recent work on tax morale drawn from behavioral economics. I have raised more questions than I have answered, since part of the point has been to make the case for a treatment of economic agents as competent ethical-economic agents, and part of what it means to be that kind of person is a capacity for responsible deliberation on given subjects. Reflexivity is something economists can provide resources for and also respect and to do so would give economists positive value in society—as educators for processes of judgement with an interest in earned agreement—and this would be quite different than is mainly currently the case. This change, as the expert tax activist Nick Shaxson argues, drawing on Veblen’s critique of prevailing economists and his comments on “business sagacity” as “sabotage,” is not anodyne and nor is it impartial, it is economics as political economy (Shaxson 2018).

Disclosure Statement

No potential conflict of interest was reported by the author.

Additional information

Notes on contributors

Jamie Morgan

Jamie Morgan is in the School of Economics, Analytics, and International Business at Leeds Beckett University Business School.

Notes

1 Note a similar case could have been made using the contemporaneous work by J. A. Mirrlees (1971), which might similarly claim to be formative. One might also take a step backwards and begin from Gary Becker’s more general economic theory of crime (Becker 1968).

2 In the context, of course, of information processing capabilities and information availability.

3 Note, “emergence” is a philosophical concept. In simplest form it refers to the idea that the organization of a set of parts results in an entity whose constitution and powers do not reduce to those of any of its parts, that those powers cannot be necessarily predicted or anticipated from those of its parts, and that the emergent entity or state of affairs constrains, enables, or influences what parts now do or can do. The claim is that society and subsets of society are emergent—society and its subsets do not reduce merely to the existence of individuals or of individuals that occupy positions within society. The general term, however, does not in itself tell us how aspects of society arise historically or what constitutes those aspects in particular—and there is a great deal of discussion regarding such matters in social theory. For discussion of the concept see, for example, Dave Elder-Vass and Jamie Morgan (2022, 218–222). See also later references to theory of trust.

4 The point is neutral on whether, in fact, the role of tax in society is commonly misunderstood as MMT advocates assert, since what matters here is public perception for the purposes of reproduction of society. See Richard Murphy (2019).

5 Note, the community charge was a new universal flat rate tax on every adult (hence poll tax—a “head tax”) in a local authority area, introduced by central government, first in Scotland and then in England and Wales in 1990. It was controversial for many reasons, not least because it sought to replace (and thus transfer payment from) the previous system of funding local authorities from the ratable value of housing and was thought to be punitive for the poor. Dogged persistence with the tax despite its unpopularity helped to bring down the Thatcher government. It was replaced by the Council Tax in 1993.

6 As one referee noted, there is an additional line of reasoning one can pursue here. If one drops the implicit assumption in standard economic theory that people will not take criminal action in the pursuit of pure rationality, then incentive structures in capitalism often encourage heinous behavior. For example, George Akerlof and Paul Romer (1993).

7 One might also note that if one applied similar modelling approaches to tax avoidance one normalizes subversion and exploitation of law—and this tendency ought to be immediately familiar to readers from the kind of responses to criticism of the kind of tax strategies and defenses corporations deploy (we pay all legally required taxes—a statement that omits that the corporation lobbies to shape their own environment and chooses to adopt aggressive avoidance strategies). See, e.g., Ylönen and Teivainen (2018); Morgan (2017 and 2016); Morgan and Sun (2017); Zucman (2015).

8 They do state “This is a very simple theory, and it may perhaps be criticized for giving too little attention to nonpecuniary factors in the taxpayer’s decision on whether or not to evade taxes” (Allingham and Sandmo 1972).

9 Again, this ought also to be familiar in the case of tax avoidance behavior. Note also that lobbying to capture the law is subtly different than dissent, public disobedience, and mobilization—tax avoiders tend to shun debate regarding whether the tax they pay or want to pay is just or fair. The whole is shifted into a context of legality and “economic efficiencies,” since from the point of view of reputation management responding that it is entirely just that a corporation pays little or no tax is unlikely to have the desired effect. One might say then that tax avoidance involves normalizing “anti-social” behavior, since avoidance is not criminal (and even if sanctioned for strategies that fall foul of the law, this is likely to be non-criminal areas of law).

10 To be clear, the standard economic agent has unlimited calculative capacity and all possible information—they are unbounded; so, this is different than whether in fact a real person fitted to this utility function could deal with the limited context of any given situation where payment of tax was required. Still, a weaker version of the claim would still be that no realistic economic agent would make decisions according to the strictures of the utility function even if they could. Agents are not just bounded they are diversely constituted, so even the weaker version rejects the utility function approach.

11 For HMRC visit GOV.UK 2012 (www.gov.uk/government/statistics/measuring-tax-gaps/measuring-tax-gaps-2021-edition-tax-gap-estimates-for-2019-to-2020). See also Murphy (2014). This was one-off sponsored research, but the general principles are often referred to on his blog site (Murphy n.d.): www.taxresearch.org.uk/Blog/.

12 For critique over the decades making this point see the work of Prem Sikka, Sol Picciotto, Nick Shaxson, John Christensen, Richard Murphy, Ronen Palan, Jeremy Sharman, and many others. It should be noted though that various expert activists have had some limited influence at the OECD, including Christensen and Murphy.

13 More recently the OECD conducted a series of regional surveys published as regional “roundtables” on tax morale between 2020 and 2021: www.oecd.org/tax/tax-global/tax-morale-roundtables.htm. This followed the report, “Tax Morale: What Drives People and Business to Pay Tax?” September 2019. www.oecd.org/ctp/tax-morale-f3d8ea10-en.htm.

14 According to Google Scholar, as of September 2023, Torgler had been cited over 28,000 times. His monograph Tax Compliance and Tax Morale (Torgler 2007) had been cited 1242 times; the text incorporates several of his early influential papers. In general, citations of these and subsequent papers vary from around 300 to over 1,000 in addition to the figure for the monograph. Note, Torgler traces the history of “behavioral taxation” to the 1950s (Torgler 2022). This is somewhat more complicated than the simpler tale of recent accommodation by the mainstream to behavioral economics and the rise of tax morale.

15 Visit https://rwer.wordpress.com/2021/07/23/weekend-read-is-human-probability-intuition-actually-biased/(Fix 2021) and https://evonomics.com/please-not-another-bias-the-problem-with-behavioral-economics/(Collins 2016).

16 For example, in Journal of Business Ethics, Mickiewicz, Rebmann, and Sauka state “Paying taxes is a social obligation towards the state. Citizens pay taxes to support the state in order to receive benefits: the state, however, is expected to act in a trustworthy and fair manner.” The obligation is expressed as “conditional cooperation” where “paying taxes in full” is a “strategic choice” influenced by the “authority” and “legitimacy” of the intermediary (the state) which spends into the community (Mickiewicz, Rebmann, and Sauka 2019, 79). They then formulate five hypotheses to test using a dataset of 279 phone interviews (with some use of Likert scores) conducted in 2010 with owner/managers, and apply an OLS regression to each (Mickiewicz, Rebmann, and Sauka 2019, 80, 81, 82). Hypothesis 1 provides a flavor of the whole: Business owner/managers who express confidence in the government have higher tax morale. All hypotheses are confirmed by the test at varying strength and degrees of confidence and the whole is an interesting, if limited, exercise in stating the obvious.

17 On the subject of performativity and the “non-benign” consequences of use of economic methods drawing on Bourdieu, see, also, Powell (2021).

18 I should stress deliberative concerns are not incompatible with raising tax morale, but are different than much of tax morale method and foci insofar as these follow trends in behavioral economics—the problem is one of emphasis, use of resources and omission.

19 And there is work published in Journal of Economic Issues on tax and inequality, progressive tax reform and a wealth tax—notably predating Piketty’s work by some years (e.g., Dugger 1990).

20 For a case study example of tax practices and ethics—focused on private equity finance—see Morgan (2022).