The Normative Problem of Merit Goods in Perspective

Abstract

In his Theory of Public Finance (1959), Musgrave invented the concept of merit wants to describe public wants that are satisfied by goods provided by the government in violation of the principle of consumer sovereignty. Starting from Musgrave’s mature discussion (1987), I construct two categories to classify the explanations of merit goods. The first strand of thought attempts to justify merit goods within the New welfare economics, by modifying its assumptions to accommodate irrationality, uncertainty, lack of information, and psychic externalities. The second category encompasses more radical departures from consumer sovereignty, drawn from philosophical critiques of economics. In the third part of the paper, I argue that the two strands might be represented by a non-individualistic social welfare function. I also show how this solution echoes Musgrave’s early views on public expenditures before he coined the concept of merit wants. From an historical perspective, the survival of the concept highlights the persistence of a social point of view in welfare economics.

Keywords:

• merit wants
• merit goods
• Richard A. Musgrave
• consumer sovereignty

JEL classifications:

• H40
• B29
• B40
• A13

Acknowledgments

Earlier versions of this paper were presented at the Summer School in History of Science and Economics in Montréal (July 2015), at the GATE seminar in Saint-Etienne (October 2015), and at the AOH seminar in Paris (November 2015). I am grateful to Antoinette Baujard, Annie Cot, Jérôme Lallement, Joseph Mazor, and Eric Brandstedt for their feedback. The paper has also benefited from the constructive criticism of the editor of this special issue, Stefan Mann, as well as from two anonymous referees. I am solely responsible for the remaining errors and disagreements. My visit to the LSE was made possible by a mobility grant from the Swiss National Science Foundation.

Notes

¹ Due to space limitations, I select only important contributions to the debate on the normative status of merit goods. For surveys focusing on different aspects, see Andel (1984); Walsh (1987); Head (1991); Ver Eecke (2007); Clément, Moureau, and Vidal (2009); Sturn (2015).

² Namely, Condition 2: “Positive association of social and individual values” and Condition 4: “non-imposition” of the SWF.

³ As early as 1934, Hutt assumed that the validity of the competitive equilibrium rested on the assumption that consumers were rational. In the discussion on merit goods, the sovereign individual is conceived in welfarist and non-comparable utility terms, to use the expressions of Sen (1979).

⁴ Note that the first dimension can be interpreted as positive, but the second one is resolutely normative. Desreumaux (2013) convincingly argues that consumer sovereignty is the central normative principle of (paretian) neoclassical economics. McLure (1990, 179) gives a definition of consumer sovereignty very close to the one I use.

⁵ Galbraith’s (1958) essay had just been published. Musgrave did not refer to it, but the rise of the mass consumer society and its side effects must have been an obvious phenomenon to keen observers by the end of the decade.

⁶ In fact, as new individualistic models of redistribution appeared (Hochman & Rodgers, 1969; Pauly, 1970), Musgrave (1970) stressed that a sizable part of the redistribution of wealth and income could not be justified in such Pareto-optimal ways.

⁷ Charles M. Tiebout was a student of Musgrave’s at the University of Michigan in the 1950s (Musgrave, 1999, 158).

⁸ Their analysis suffers from serious problems that have been pointed out by Ballentine (1972) and Braulke (1972/2007).

⁹ For a similar position, see Burrows (1977/2007, 29).

¹⁰ Similar ideas are also discussed by Head (1988).

¹¹ This lead is also followed in the surveys of Clément et al. (2009), Sturn (2015), and Kirchgässner (2015).

¹² For the formal definitions, see Munro (2009, 6 ff). Following this definition, Munro provides in the seventh chapter of his book a review of the optimal taxation models of merit goods, especially the recent ones such as Racionero (2000) who follows in the footsteps of Sandmo (1983) by focusing on information provision.

¹³ In his review of the Theory of Public Finance, Wiseman (1960, 266) noted that Musgrave failed to directly address the problems of political philosophy he raised.

¹⁴ The political dimension is also conceptualized by Pulsipher (1971); Folkers (1974/2007); Mackscheidt (1974/2007); Brennan and Lomasky (1983) and Rüffer (2007).

¹⁵ The quotation is from a letter from Musgrave to W. Ver Eecke, October 12, 2003. Richard A. Musgrave Papers, Box 6, “Correspondence,” Princeton University Library.

¹⁶ The concept of merit good was explicitly discussed by many participants and Musgrave’s (1987) and Head’s (1988) retrospective surveys of the concept were rewritten after the conference, according to the organizers, Brennan and Walsh (1990).

¹⁷ Cooter and Gordley (1995) also develop their merit good model on Aristotelian arguments about excellence.

¹⁸ That being said, following Frankfurt (1971), the idea of higher order goods has already been used by economists who modeled it using ranking of preferences, as was mentioned in the previous section.

¹⁹ This definition would align merit goods with the concept of intermediate goods (see Colm, 1965, p. 215). Musgrave (1969b) already used the concept of intermediate social goods to describe these cases, something that Ver Eecke does not mention.

²⁰ One way to interpret Ver Eecke’s claim that economic actors desire a free market is through his reading of Hegel’s Philosophy of Rights. For Hegel, the market, or civil society, is an ethical institution that emerged with the growth of commercial society in the modern age. The identity of the economic agent is concomitant with the emergence of a free economic sphere. So the market can be understood as promoting individual freedom in the modern world. It realizes the individual’s subjectivity (Ver Eecke, 2008; Chapter 3). Yet, it is not self-sufficient in the sense that the state, as a higher ethical realm, must solve some of the contradictions generated by the market (Hegel 1821/1991).

²¹ Backhouse and Nishizawa (2010b) reveal the plurality of views on welfare in Britain in the 1920s. Yet, by the 1930s, the positivist aspirations of economists had considerably reduced the range of acceptable positions, in particular among the younger generation of LSE economists.

²² Yet, as Habermas (Habermas, 1990, 50 ff) observes, moral theory does not need to be intuitionist or emotivist. A similar point is made by Nussbaum (2000, p. 127): She explains Harsanyi’s hesitation to engage with a normative theory of justice on the grounds that he was convinced that the two moral theories available were G. E. Moore’s theory of mental states and hedonism.

²³ Mazzola (1890) and De Viti de Marco (1934) are representatives of the Italian tradition. Ritschl (1931) is a twentieth-century representative of the older German holistic view.

²⁴ The “Old” welfare character of Musgrave’s dissertation follows from his public household point of view and his acceptance of interpersonal comparisons as a “workable assumption.” Following Pigou (1932), he argues: “The capacity to enjoy benefits is after all but part of the general nature of ‘man.’ It being the generally accepted procedure to define certain general characteristics of men, there is no reason why no typical degree of intensity for the enjoyment of benefits could be assumed” (Musgrave, 1937, p. 274).

²⁵ The SWF thus becomes W = W(U¹, … ,U^H; X₁, …, X_n) where X_j = ∑_h X_j^h.

²⁶ Respectively, defined as $\frac{\mathit{\partial }W}{\mathit{\partial }{X}_j}<0$

²⁷ In terms of the general formulation, where merit goods are indexed to individuals, it would call for individualistic prices.

²⁸ A similar model has been developed by Roskamp (1975) apparently without knowledge of the Pazner (1972) and Folkers (1974/2007) papers. He uses a more cumbersome typology of private, merit, and public goods, and defines the SWF as a function of the individual utilities and of a set of society’s preference functions for the merit goods (rather than just the goods directly entering as argument). Besley (1988), Walsh (1987), and Cooter and Gordley (1995) are among those who also used a similar non-individualistic function for conceptualizing merit goods.

²⁹ The behavior function can be expressed as $B(x,y)=B^0(T(x,y),V(x,y))$.

³⁰ Provided that the individual utilities are represented by behavior functions.

³¹ John Rawls, “Philosophy 273 Primary goods and the concept of material welfare.” October 17, 1984. Papers of John Rawls. Harvard University Archives. Box 37, Folder 14.