Ambiguities in Applications of Economic Theory

Mary Morgan (London School of Economics) - There are a number of important umbrella terms in economics that are well defined, and even invite quantification in their framing, yet whose measurement is, in practice, deeply problematic. They have acquired measurements, in the sense that the state and scientific structures have devised numbers for them. But the difficulties in numbering the elements crowded under those umbrella terms mean that their usage in both scientific and public policy domains invites questions about integrity and accountability of those measuring regimes and therefore the actions taken using them. Concepts like the ‘national income,’ ‘poverty’ and ‘development,’ provide good examples of the terms and the problems involved. They are umbrella concepts that are characterised by a considerable combination of elements, so the ‘measuring instruments’ used to anchor those terms into numbers must access a wide range of measurements for the elements associated with the concept. And then — if possible — economists must figure out how to integrate the individual measurements in some way that would capture the overall concept. For ‘national income,’ these multiple measurements are held together in a technical accounting framework that maps onto the conceptual level — the measuring system passes some basic kind of ‘representing’ account. But when it comes to providing measurements for ‘poverty’ or ‘development,’ the many characteristics involved are numbered separately in ‘indicators’ (of poverty, or of development) and these indicator numbers have no clear inter-relational structure that integrates them to provide the kind of single measurement with representing power that might be desirable for the overall concept. However, representing power should not be the only criteria. These numbers are regularly used for public action, and so the accountability, as well as the integrity of the measuring systems needs consideration. Whereas the accounting framework of the national income data provides a regime of internal audit, there is no obvious external accountability beyond the institutions that oversee them. It might seem that those numbers assembled to measure poverty and development would be even more problematic, but each of the separate ‘indicator’ numbers can be used by special interest groups to hold governments to account — so that there may be accountability without audit.

Rebecca Livernois (University of British Columbia) - An externality in economics is broadly defined as an unpriced spillover effect between agents. There is little consensus, however, among economists or philosophers on a more precise characterization of the term (Arrow 1969; Meade 1952; Berta 2017; Lagueux 2010; Papandreou 1994). In this paper, I offer a clarification of the concept of an externality and argue that the bridging conditions are unclear when forging policies. An externality is modelled in a multi-agent optimization framework where an unpriced activity taken by one agent affects another agent. In this model, an externality is a case of untapped gains from exchange. That is, it occurs when an agent’s willingness to pay for a reduction in the unpriced activity diverges from another agent’s willingness to accept a payment for the reduction. Unlike an externality in the model where agents have perfect information, however, externalities in the world are invariably accompanied by imperfect information. As a result, it is unclear whether the willingness to pay in the world that constitutes an externality refers to an agent’s actual willingness to pay for an activity or an agent’s willingness to pay conditional on their having true beliefs about the activity. I argue that both interpretations are problematic. As such, it is unclear how to interpret an externality when applied in specific policies such as a carbon tax.

Julian Reiss (Durham University) - The main claim I am making in this paper is that there are few if any uncontroversial facts in economics. I provide three reasons to back up this claim. First, fact/value entanglement is pernicious in economics. Second, socio-economic facts tend to be sensitive to minute changes in context and specification. Third, there are no accepted standards for testing claims in economics. I speak of fact/value entanglement whenever ostensibly factual claims such as ‘The current inflation rate is 2%’ or ‘Raising the minimum wage will have no negative employment effects’ or ‘Free trade spurs growth’ cannot be established without making value judgments about what is and what is not desirable. I speak of pernicious fact/value entanglement when it is widespread, irreducible, or significant enough to make a real difference to outcomes and unlikely to be resolved by agreement on values. Fact/value in entanglement is pernicious in economics. Socio-economic facts are sensitive to context and specification. That is, whether any of the exemplary claims stated above are true may depend on (a) the time and place of application (this issue has been called the problem of ‘historical specificity’); (b) the horizon (what is true of the short run may be false of the mid or long run); (c) the choice of contrast (what is true relative to a specific contrast may be false relative to another); and (d) the choice of the measure or indicator. (Note that I do not defend anti-realism about economic facts in this paper: I do maintain that factual questions have true or false answers. However, these answers are highly context-sensitive and depend on the exact interpretation of the question, which is not always transparent.) There are no commonly accepted standards for testing claims in economics. Ever since the separation of the discipline from philosophy, there have been methodological controversies, which have often focused on the right balance between inductive and deductive elements in inference. This debate is unresolved. Three conclusions stand out. First, a certain degree of epistemic humility is called for. A disagreement about socio-economic facts does not mean that one party is irrational, ignorant or ill motivated. Second, we should be wary of claims to specific epistemic authority in economics. Third, we should not make socio-economic institutions dependent on the existence of uncontroversial socio-economic facts.