We are interested in the extent of convergence between paradigms in labor economics. Proponents of both neoclassical and institutional paradigms now embrace the common premise that labor markets cannot be adequately understood in terms of the auction market described in the model of perfect competition. The persistence of phenomena such as internal labor markets and involuntary unemployment necessitate some explanation other than the neoclassical competitive model. While institutionalists have long grappled with these issues [Freeman 1988; Kerr 1988], serious attempts by neoclassicals, via efficiency-wage and insider-outsider models [Akerlof and Yellen 1986; Lindback and Snower 1988! have appeared only in the past decade.
This paper focuses narrowly on whether recent work by neoclassical economists marks an acceptance of the institutional paradigm.(1) We presume that existing institutional explanations of labor markets are adequate - indeed both sufficient and persuasive - and that recent empirical work does little to challenge long-held institutional tenets [Freeman 1988]. We argue that recent neoclassical models that depart from neoclassical orthodoxy, while marking an advance over the recently dominant Chicago Sschool, are still largely "ceremonial" efforts deemed necessary for acceptance among practitioners of their paradigm, rather than a search for an instrinsically robust analysis of labor market phenomena.
Distinguishing Neoclassical and Institutional Paradigms
In two papers, Yngve Ramstad [1987; 1993] has developed a clear set of criteria for distinguishing neoclassical and institutional methods in analyzing labor markets. Ramstad identifies two different questions that are raised when trying to distinguish neoclassical and institutional treatments of the labor market: (1) Must one acquire a detailed understanding of labor market institutions in order to explain labor market outcomes? and (2) Is there a neoclassical "core process" revealed in labor market outcomes? [Ramstad 1987, 16-19].(2)
Conventional efforts to distinguish paradigms typically apply only the first question [see Kaufman 1993, chap.1; Kerr 1977]. Ramstad's key insight is that the answer provided by asking this question is still incomplete@ one must also ask the second question, i.e., about the presence or absence of a commitment to neoclassical theory. Thus, it is possible to have the appearance of being an "institutionalist" by answering yes to the first question (i.e., making the effort to acquire a detailed understanding of labor market institutions), while retaining the analytical goals and methods of neoclassical theory.(3) Ramstad calls this approach neoclassical-plus" and cites Dunlop as a leading example.
The crucial distinction between neoclassical and institutional theory is revealed by the answer to the second question. The importance of this question turns on the ontological status of "institutions" in each paradigm. In neoclassical theory, the concrete institutional context of a market is seen as so many "frictions" that inhibit the functioning of the core process and result in "suboptimal" outcomes. This interpretation carries a normative judgment: institutions that are "frictions," such as collective bargaining agreements, create inefficiency and produce social outcomes in Which at least some agents are worse off.
Conversely, the institutional paradigm makes no such normative judgment. Rejecting the existence of a neoclassical "core process" is really part of a broader rejection of methodological individualism in favor of one of a number of "social" explanations of economic behavior [Wilber and Harrison 1979]. Market behavior is understood as an expression of an underlying rule structure embedded in institutions, analysis of labor market behavior thus begins with uncovering the rule structure of the relevant institutions [Ramstad 1987, 16-19].(4) Unlike neoclassical …