Contrived desires, affluence, and welfare: J.K. Galbraith's Pigovian redistribution argument reconsidered

Abstract

I argue that John Kenneth Galbraith's theory of the “dependence effect” in The Affluent Society provides a way to rescue A.C. Pigou's argument for wealth redistribution from a powerful objection. The objection is based on the unprovability of statements making interpersonal comparisons of utility. Galbraith's dependence effect theory allows him to present a version of the Pigovian argument that requires no such statements to be made. I argue that Galbraith's main piece of advocacy in The Affluent Society was for income redistribution, despite the fact that he claimed to be in favour of greater spending in the public sector rather than redistribution as such. I then show how my reading of the dependence effect theory helps to defend it against objections from Hayek and Rothbard. I end by discussing what improvements in economics a proper test of the theory would require and showing how my reading of it helps to reveal the ongoing importance of The Affluent Society to the understanding of political economy.

Keywords:

• welfare
• utility
• Pigou
• Galbraith
• Hayek

Notes

¹ See Harsanyi (2008).

² Gibbard's claim that logical positivism was an influence upon Pareto's English-speaking followers requires further historical substantiation. Logical positivism holds, very roughly, that statements that cannot be tested empirically are either meaningless or they are true or false by definition. Note that Robbins’ argument makes no reference to this thesis. He does, however, betray a commitment to some sort of positivism, meaning the determination to justify all substantive claims by empirical evidence.

³ The claim that this amounts to a general empirical test of preference is in fact deeply problematic. See also Rosenberg (1981, 58--67).

⁴ See Rosenberg (1981, 57--60).

⁵ I use the most recent (1999) edition of this work, which preserves the main argument while speaking in terms tolerably familiar to a contemporary audience.

⁶ See Friedman (1978).

⁷ Pressman (2008, 478).

⁸ For a discussion of how Galbraith's views fit into broader discussion about advertising in economics from this period, see Edwards (Forthcoming).

⁹ George Batten, Roy Sarles Durstine, Alex Faickney Osborn, and Bruce Fairchild Barton were the founders of the still prominent advertising agency, BDDO. For its potential clients, it promises to do “work that changes consumer behavior” (see http://www.bbdo.com).

¹⁰ Galbraith (2008, 494). The reference is to Mirowski (1989). In fact, the stability of preference fields, at least for most kinds of goods, was challenged before The Affluent Society; see Gorman (1953). I respect this work a great deal, but I am not thereby led to insist that it must be continuous with that of Galbraith than I am to insist that all my friends must be friends with each other.

¹¹ Of course, any transfer of wealth from the richer to the poorer person will in real cases most likely inhibit the richer person's ability to satisfy some of her real wants as well as her contrived wants. It is barely possible to suppose that real wants are so much more urgent in the richer person than in the poorer person as to override any net-welfare benefits arising from the elimination of waste in the transfer of income. Scepticism about cardinal utility is one thing, but this is turbocharged quibbling.

¹² In using the term “liberal” here I conform to the American usage. The term actually means something like “social democrat,” but for historical reasons political labels involving the word “social” are best avoided in the United States. See Judt (2010, 4--6).

¹³ To this extent, I believe that Lamdin's assessment of Galbraith's advocacy of public spending in terms of its effect on productivity somewhat misses the point. See Lamdin (2011, 608--9).

¹⁴ See Galbraith (1999, Chapter 22).

¹⁵ See Friedman and Friedman (1982, Chapter 12).

¹⁶ Dutt, who provides a model in which this is the result, explains the basic logic as follows: “If advertising expenditures increase workers consumption, they may experience a rise in income in the short run because of the expansionary effect of higher spending by both workers and capitalists. But with increases in capitalist income and the desire by workers to emulate them, workers become borrowers and eventually net debtors. As their debt increases, workers experience a decline in their net income, and with a redistribution of income (from interest payments), there is a contractionary effect on demand, because capitalist interest earners have a higher saving propensity than workers.” (Dutt 2008, 547) This explanation involves the other side of the dependence effect – envy – and the role of debt – two topics I have not been able to address in this article.

¹⁷ A far more detailed discussion of various criticisms of the dependence thesis, as well as rebuttals of them, can be found in Dutt (2008). My aim here is only to show specifically how my reading of the theory can help to fend off some prominent objections.

¹⁸ Two fairly mainstream economists (from opposite sides of the political spectrum), Milton Friedman and Paul Krugman, have criticised Galbraith, but their criticisms were mostly directed at a different theory than the one discussed here, namely his theory of the “technostructure,” first outlined in The New Industrial State (Friedman 1978, Krugman 1995, 13--5).

¹⁹ Indeed, for my preference to influence my rational choice it must at least be combined with my belief that I in fact have such a preference.

²⁰ Gibbard argues that it is impossible to properly measure welfare in terms of preferences, even if these are restricted to those involving no false beliefs (ideally informed preferences, as he calls them) (Gibbard 1986, 170). Here he dissents from standard economic welfare theory more than Galbraith does. Various objections to the preference satisfaction view of welfare are summarised and explained in Hausman and McPherson (2006, Chapter 8). There various alternatives are presented.

²¹ See Mill (1991), II. Mill was working within a cardinal utility theory and trying to include within it the idea of levels as well as quantities of utility. There is no obvious way in which this idea is directly relevant to Galbraith, who, I have argued, largely respects the methodological commitments that lead to the rejection of cardinal utility theory.

²² The dangers of taking it as anything more than this are given a compelling airing in Emmett (2006). This article is Emmett's attempt to deduce what Frank Knight might have written, had he lived long enough to respond to Becker and Stigler.

²³ This is how, in reply to James Galbraith's objection cited above, it can be possible for stable preferences to exist even for goods that do not yet exist. One's preferences are directly for the products of “household production.” But if there are yet-undiscovered raw materials that can be used for such production, one's preferences for household goods will precisely determine one's preferences for such raw materials, given specific information about their availability, cost, etc.

²⁴ See Hayek (1945).

²⁵ Some discussion of this issue from a sociological point of view can be found in Dupuy (1989).

²⁶ Lamdin surveys a variety of studies attempting to identify the direction of causation between advertising and consumption. Hsu et al. (2002), Schmalensee (1972), and Simon (1970) find that the causation runs from consumption to advertising, which goes against Galbraith's thesis. The pathbreaking (Jung and Seldon 1995), however, partly supports Galbraith's thesis, as does Lamdin's own research.

²⁷ For surveys of discussion of this issue, see, for example, Fullbrook (2002), Mason (1998), and Pollak 1976.

²⁸ Of course, any model capable of representing these kinds of scenarios would be of a non-linear dynamic, possibly chaotic system. It is hard to see how one could use such a model to judge the efficiency of various outcomes the way one can with a stable equilibrium model. It is hard, after all, to judge the efficiency of outcomes when the outcomes refuse to stay still. Becker suggests that this kind of situation can be modelled as a stable equilibrium system, so that one can solve for maximum efficiency in Becker (1974). But note that his model assumes the degree and direction of interdependence among preferences – what Becker calls “social capital” – to be constant and exogenous to the system, rather than being an endogenous variable. This assumption is the mere hope that reality will constrain itself to the mathematical capacities of its modellers.

²⁹ Daniel Kahneman's latest book provides a recent basic survey of some of these (Kahneman 2012). Discussions from other points of view can be found in Earl (1990), Easterlin (2002), and many other sources. Some of this literature is discussed in relation to Galbraith in Dutt (2008, 536).

³⁰ Indeed, my personal belief is that the most compelling cases are advanced within Modern Monetary Theory (MMT), which emphasises the extent to which the constraints on public spending in countries like the United States are now entirely different than they were under the Bretton Woods regime of fixed exchange rates. The Affluent Society was, of course, written under this regime, and much of what it has to say about the need to raise revenue for public spending no longer apply in the same way. See Mosler (2010) and Wray 2012. It is interesting to note that James Galbraith is one of a handful of academic economists who appreciate the insights of MMT.

³¹ On the first argument, see Wilkinson and Pickett (2010). On the second, see Hausman and McPherson (2006, Chapter 8) and Sen (1977).

³² A philosophical case against it can be found in Rosenberg (1992). A set of even more troubling criticisms, less dependent on contentious philosophical ideas, can be found in the first few chapters of Keen (2011).

³³ To take a currently prominent example: when Thomas Piketty states that “no matter how justified inequalities of wealth may be initially, fortunes can grow and perpetuate themselves beyond all reasonable limits and beyond any possible rational justification in terms of social utility” and that “[e]very fortune is partially justified yet potentially excessive” (Piketty 2014, 440--1), he employs standards of reasonability, social utility, and fairness going well beyond Pareto measures of preference maximisation.

³⁴ James Galbraith claims that his father, probably rightly, repudiated the macro/micro division (Galbraith 2008, 497). Even so, it is interesting that the argument discussed in this article can work within the terms of standard microeconomics as conceived by those who uphold the division.

³⁵ James Galbraith resembles his father in this respect. He also believes that government spending should be promoted as a means of addressing the most urgent social need rather than as a means of fiscal stimulus. Indeed, he is sceptical about the benefits of fiscal stimulus as such, believing that low growth in the developed world today has structural causes that fiscal stimulus is powerless to address. However, he argues, this makes public spending on social insurance programmes all the more vital. See Galbraith (2014, 246--51).

³⁶ See Milanovic (2011), Stiglitz (2013), Wilkinson and Pickett (2010), and of course Piketty (2014). This is another sense in which James Galbraith carries on his father's work (see Galbraith 2012).