The Concept of Development Conventions: Some Suggestions for a Research Agenda

Abstract:

Development conventions are broadly defined as socially shared systems of rules of thought or behavior (institutions) regarding development, with the properties of conformity with conformity and arbitrariness. Conformity with conformity means that the fact that others have adopted a convention, or are expected to adopt it, plays a role in leading someone else to adopt it. Arbitrariness means that a non-inferior alternative exists or is conceivable. This article identifies several social mechanisms through which the conformity of others leads one to conform. It also relates arbitrariness to the possibility of assessing development from various viewpoints, to the inconclusiveness of evidence, and to uncertainty about future consequences.

Keywords::

• conventions
• development
• institutions
• mental models

JEL Classification Codes::

• B52
• E02
• O10

Notes

¹ Erber owes the initial idea of development conventions to Antonio Barros de Castro (1993). Barros de Castro (1993) does not use the term “development convention,” but explicitly discusses what he calls “the convention of guaranteed growth” and “the convention of stability,” while writing about development in Brazil. Barros de Castro (1993, 208n) does not present a concept of convention, treating convention very broadly as synonymous with “shared belief,” and makes only a very brief reference to John Maynard Keynes.

² Alas, Erber died in February 2011. I do not question the compliments paid to Erber by several people who knew him, most notably Fernando Carvalho (2011). His presumably most elaborate writings on development conventions (Erber 2011, 2012) were published posthumously, by friends and colleagues. His untimely death prevented Erber from having enough time to revise these writings, submit them for publication, and then benefit from the feedback of referees who might have urged him to properly study the literature on conventions or, more simply, to eliminate the references to conventions and stick to shared conceptions or models. In addition, Erber’s predominant interest in applied issues regarding the Brazilian economy probably led him to a cursory reading of only very few texts on conventions and to a superficial treatment of the convention concept (for a different view, see Modenesi and Modenesi 2015, 139).

³ All translations from French and Portuguese are mine.

⁴ See Lewis (1969, 69-70). The existence of this alternative is also a necessary characteristic in Robert Sugden’s (1986, 33) account of convention, among others.

⁵ André Modenesi and Rui Modenesi (2015, 137), who borrow Erber’s concept of development convention, quote a passage from an article by Jean-Pierre Dupuy (another conventionalist) that includes this clause referring to an alternative regularity R’. Modenesi and Modenesi (2015) do not notice, however, that (a) in the passage cited, Dupuy (1989, 369) actually described Lewis’s definition of convention, not his own (Dupuy, like Orléan, is critical of Lewis’s concept), and (b) Erber omitts such a clause from his definition.

⁶ The mere mention of coordination problems and uncertainty (Erber 2011, 32) is not enough, much less when accompanied by the statement that these problems are dealt with by means of institutions, “the rules of the game” (as in North [1990], approvingly cited by Erber) — institutions are not the same as conventions.

⁷ There is an ongoing debate on whether and how a game-theoretic approach to institutions, including conventions, should define them with reference to (or even as) shared beliefs and representations (see, for example, Hindriks and Guala [2015] and Aoki [2015], among other authors published in Volume 11, Issue 3 of the Journal of Institutional Economics). Having inadvertently borrowed a game-theoretic concept of behavioral convention, Erber was unaware of the rounds of this debate that occurred before his death.

⁸ Erber (2012, 4) seems to mistakenly assume that Denise Jodelet (1989), whose work is cited by some conventionalists, uses her notion of social representation to define a convention, when in fact she does not refer to a convention in that book chapter.

⁹ The only quotation Erber (2011, 2012) provides is actually the misquotation of Orléan (2004, 12), mentioned above.

¹⁰ A fourth deficiency can be added that would hopefully be less difficult to remedy: a series of inconsistent and unclear statements (see, for example, Erber [2011, 33] on metaphors, rules, and representations). Contradictorily, Erber states that rules are derived from metaphors, before describing metaphors themselves as rules. He also fails to explain how a metaphor can be equated with rules, how a historical metaphor can be a theory, and how behavioral rules can be part of what constitutes a representation. The same is true of Erber’s statement on rules being structured by “a behavior” (“on the behavioral and cognitive planes”) (Erber 2011, 32).

¹¹ Some of these criticisms are explained below.

¹² Modenesi and Modenesi (2015) approvingly refer to my concept of conventions (among others), but do not explore its implications for the concept of development conventions. Instead, their discussion of the concept of development convention closely follows Erber’s statements.

¹³ By contrast, if — as Erber suggests — the concept of development convention is highly dependent on coordination effects to explain conformity with conformity, and at the same it is argued that two development conventions exist within the same country or region, what exactly is the coordination problem that each convention supposedly solves? In particular, regarding development, is the belief that all the other members of a population P conform to a certain rule of behavior enough reason for an agent to also conform? If so, how? Erber’s conceptualization does not specify any mechanisms. Furthermore, as discussed below, it is hard to imagine such a mechanism, especially if two or more development conventions coexist in a given country.

¹⁴ By contrast, if it is argued that the unanimously shared belief in unanimous conformity gives each and every agent enough reason to conform, why would some members of a population P abandon the prevailing convention, as Erber claims (2012, 9)? More generally, with a game-theoretic concept and its strong link to a game equilibrium, how would there be any change in individual adoption and in a convention? Evolutionary game theory (unsatisfactorily) offers a way out via suboptimal deviation, but Erber is unaware of the issue.

¹⁵ The conformity of the rest of population P may not be enough reason for individual conformity, even in the case in which population P consists of all private agents in a country’s economy. It is extremely less likely to be enough reason for individual conformity if population P represents only a fraction of a country’s whole population. The latter condition characterizes the situation considered by Erber, in which two or more development conventions coexist within the same country, each development convention implicitly defined for a specific subset of the country’s population. Consequently, a crucial clause of the concept of behavioral convention that Erber borrows from game theory does not fit the situation that he wants to examine.

Additional information

Notes on contributors

David Dequech

David Dequech is a professor of economics at the University of Campinas, São Paulo (Brazil).