Formalization is Dead - Long Live Formalization!
Since the time of Chamberlin and Vernon Smith, use of experiments and other non- mathematical techniques to test theories against data has blossomed. In part, this change has been triggered by advancements in statistics, increases in computing power, and availability of higher-quality data; in part, it is also explained by growing dissatisfaction with the mathematics that had imposed itself as primary mode of economic investigation with Samuelson and Debreu.
At the end of the 1990s, even the media echoed a widespread feeling that economics was devolving into a technical discipline for mathematical virtuosi, incomprehensible to the public and unable to answer the crucial questions facing contemporary societies (Cassidy 1996). Thus in recent years, growing emphasis has been placed on the need for procedures allowing validation of theories against data, rather than (or in addition to) abstract, deductive mathematics. Econometrics, experiments,randomized trials, computer simulation and even brain scans have progressively gained ground; empirical research has earned a central place in the discipline, and today’s economists often pride themselves on their superior capacity to extract information from data. New (though still uncommon) interactions with physics have taken place, most notably in the field of complexity studies and with the help of computer simulation rather than formal theorems. Although a detailed account of these developments would be beyond the scope of this entry, it is worth mentioning their repercussions on the issues that are of import here.
Substantively, the new tendency changes the interpretation of many inherited theoretical constructs. For example, behavioural economics challenges the individual optimization model as a description of reality, but preserves it as a normative benchmark: policies, then, have to induce real-world individuals to make choices that are as close as possible to the optimum (Thaler and Sunstein 2009). In a sense, the new approach reverts to the original, primarily normative interpretation of optimization put forward by Gossen.
Methodologically, critics may contend that today’s non- mathematical methods include some elements of induction, thereby lowering the standards of rigour that the deductive mathematics of Samuelson and Debreu had spread. Yet some of the main attainments of formalization have been preserved in that the new methods are all directly or indirectly based on some form of mathematical or game-theoretic reasoning, and the pursuit of rigour continues, affecting all phases of scientific reasoning from the initial formulation of hypotheses to their final empirical test. To some extent, the new approaches synthesize aspects of both quantification and formalization.
As a matter of fact, today’s methods are as sophisticated and demanding as mathematics in terms of the technical and methodological skills needed to apply them. This is a reason that partly explains their rising status, but also raises concerns that they might turn once again into forms of virtuosity unrelated to the discipline’s capacity to work on relevant social issues - somehow renewing with older worries about formalization becoming an end in itself.
While further developments are likely to occur in the near future, the three-century long history of formalization will continue to make its presence felt. It has contributed to shaping economics and its debates in the past, raising methodological and substantive issues that recurrently reappear even under changed historical and intellectual circumstances. It has left its indelible mark on the discipline, and all evidence indicates that its influence is here to stay.
Paola Tubaro