Catherine Greene

This article argues that it is far from trivial to convert social science concepts into accurate categories on which algorithms work best. The literature raises this concern in a general way; for example, Deeks notes that legal concepts, such as proportionality, cannot be easily converted into code noting that ‘The meaning and application of these concepts is hotly debated, even among lawyers who share common vocabularies and experiences’ (Deeks in Va Law Rev 104, pp. 1529–1593, 2018). The example discussed here is recidivism prediction, where the factors that are of interest are difficult to capture adequately through questionnaires because survey responses do not necessarily indicate whether the behaviour that is of interest is present. There is room for improvement in how questions are phrased, in the selection of variables, and by encouraging practitioners to consider whether a particular variable is the sort of thing that can be measured by questionnaires at all.

This article argues that it is far from trivial to convert social science concepts into accurate categories on which algorithms work best. The literature raises this concern in a general way; for example, Deeks notes that legal concepts, such as proportionality, cannot be easily converted into code noting that ‘The meaning and application of these concepts is hotly debated, even among lawyers who share common vocabularies and experiences’ (Deeks in Va Law Rev 104, pp. 1529–1593, 2018). The example discussed here is recidivism prediction, where the factors that are of interest are difficult to capture adequately through questionnaires because survey responses do not necessarily indicate whether the behaviour that is of interest is present. There is room for improvement in how questions are phrased, in the selection of variables, and by encouraging practitioners to consider whether a particular variable is the sort of thing that can be measured by questionnaires at all.

de Bruin et al. The Stanford encyclopedia of philosophy, Stanford University, Stanford, 2018) write that it is a philosophically interesting question “whether there is such a thing as an “intrinsic” value of financial assets” noting that the calculation of any intrinsic price will depend, in part, on subjective elements. McCauley suggest that there are at least five different notions of the ‘true value’ of an asset in finance theory, and argues, consistent with de Bruin et al. that in many cases the calculation of such values “makes impossible information demands on our knowledge of future dividends and returns”. This paper explores some of the subjective elements involved in calculating the intrinsic value of an asset, and their implications for the Law of One Price. The Law of One Price states that investors should not pay different prices for the same investment. Consequently, the existence, and persistence, of arbitrage opportunities is often attributed to investor irrationality. However, if subjectivity is involved in the calculation of intrinsic value then two investors can, rationally, disagree about the value of a security. This implies that arbitrage opportunities may not always be instances of irrationality, or genuine mispricing, but reflect different investor’s perceptions of the security. In economics, value is taken to be subjective, and Postigo describes how goods have value because of the role they play in the satisfaction of our needs. By applying this to finance, this paper will argue that a security can play a number of roles in a portfolio, depending on an investors’ strategy. Given these different roles, different investors will be working with different data in their assessment of intrinsic value. In order to establish that an arbitrage opportunity exists it is therefore necessary to understand what information is relevant to pricing the investment in different locations, or markets. It may, in fact, be the case that the same security can trade at different prices because different investors do not see it as the ‘same thing’ at all. This analysis can also be used to throw doubt on the emphasis that fundamental analysis receives in discussions of rationality.

Counterfactual analysis is an interesting feature of thought experiments, because it requires the imagination of alternative states of the world (see also publications by Fearon, Lebow and Stein, Reiss, and Tetlock and Belkin, who suggest the same). In historical analysis, the use of imagination is often the focus of criticisms of such counterfactual analysis. In this article, I consider three strategies for constraining imagination: making limited counterfactual changes, limiting counterfactual changes to the decisions of important figures, and using evidence to restrict the scope for imagination. Given the focus of this special issue, I will relate this discussion to Lewis’s and Woodward’s analyses of counterfactuals in the philosophy of science. I show that counterfactual analysis in historical cases has some resemblance to Lewis’s and Woodward’s analyses, but that what Lewis calls “transition periods” cannot be left entirely vague, as Lewis suggests, nor can counterfactual changes be seen simply as interventions, as Woodward suggests. I propose that efforts to limit imagination in historical counterfactuals are ultimately problematic, but that imagination can nevertheless play a useful role in counterfactual analysis.

The observation that concepts used by social scientists are often problematic is not new; they have been described as Ballung concepts, cluster concepts, essentially contested, and reflexive; however, the need to work with these concepts remains. This article addresses the problem of variable choice in the social sciences by exploring and extending Woodward’s recommendations. This article demonstrates why Woodward’s criteria are difficult to apply in the social sciences and proposes an alternative, but complementary, framework for assessing variables.

The concept of ‘information’ is central to our understanding of financial markets, both in theory and in practice. Analysing information is not only a critical part of the activities of many financial practitioners, but also plays a central role in the Efficient Market Hypothesis (EMH). The central claim of this paper is that different data can count as information in fi-nancial markets and that particular investors do not consider all of the available data. This suggests that firstly, saying the price of a stock should be $X once a particular piece of information has been incorporated in the price is only justifiable if we know that this information really is relevant for other investors. Sec-ondly, the EMH is often tested by looking at market behaviour after the release of information; again, this is only justified if we know that other market participants share the view that this is information. The purpose of this chapter is to suggest that there are good reasons for thinking that we do not know this. Finally, this chapter also suggests that bubbles in financial markets are best understood in terms of salience of information, rather than irrationality.

Big data increasingly enables prediction of the behaviour and characteristics of individuals. This is ethically concerning on privacy grounds. However, this article discusses other reasons for concern. These predictions usually rely on generalisations about what certain sorts of people tend to do. Generalisations of this sort are often under scrutiny in legal cases, where, for example, lawyers argue that people with prior convictions are more likely to be guilty of the crime they are currently on trial for. This article applies criteria for distinguishing acceptable from unacceptable generalisations in legal cases to a number of big data examples. It argues that these criteria are helpful, and highlights three issues that should be taken into account when deciding whether predictions about individuals are ethical.

The social sciences are often thought to be inferior to the natural sciences because they do not have laws. Bohman writes that “the social sciences have never achieved much in the way of predictive general laws—the hallmark of naturalistic knowledge—and so have often been denied the honorific status of ‘sciences’” (1994, pg. vii). Philosophers have suggested a number of reasons for the dearth of laws in the social sciences, including the frequent use of ceteris paribus conditions in the social sciences, reflexivity, and the use of ‘odd’ concepts. This thesis argues that the scarcity of laws in the social sciences is primarily due to the concepts that social scientists often work with. These concepts are described as Nomadic and are characterised by disagreement about what can reasonably be included within the scope of a concept. The second half of the thesis explores the implications of this analysis. It argues firstly, that counterfactual analysis is problematic when using Nomadic concepts. Secondly, it argues that taking an intentional perspective on behaviour often involves the use of Nomadic concepts so, if social scientists do hope to formulate laws, then they are more likely to succeed if they focus on behaviour that is not intentional.

de Bruin et al. (The Stanford Encyclopedia of Philosophy, Stanford University, Stanford, 2018) write that it is a philosophically interesting question “whether there is such a thing as an 'intrinsic' value of financial assets” noting that the calculation of any intrinsic price will depend, in part, on subjective elements. McCauley suggest that there are at least five different notions of the ‘true value’ of an asset in finance theory, and argues, consistent with de Bruin et al. that in many cases the calculation of such values “makes impossible information demands on our knowledge of future dividends and returns”. This paper explores some of the subjective elements involved in calculating the intrinsic value of an asset, and their implications for the Law of One Price. The Law of One Price states that investors should not pay different prices for the same investment. Consequently, the existence, and persistence, of arbitrage opportunities is often attributed to investor irrationality. However, if subjectivity is involved in the calculation of intrinsic value then two investors can, rationally, disagree about the value of a security. This implies that arbitrage opportunities may not always be instances of irrationality, or genuine mispricing, but reflect different investors perceptions of the security. In economics, value is taken to be subjective, and Postigo describes how goods have value because of the role they play in the satisfaction of our needs. By applying this to finance, this paper will argue that a security can play a number of roles in a portfolio, depending on an investor's strategy. Given these different roles, different investors will be working with different data in their assessment of intrinsic value. In order to establish that an arbitrage opportunity exists it is therefore necessary to understand what information is relevant to pricing the investment in different locations, or markets. It may, in fact, be the case that the same security can trade at different prices because different investors do not see it as the ‘same thing’ at all. This analysis can also be used to throw doubt on the emphasis that fundamental analysis receives in discussions of rationality.

The financial crisis has led to calls for increased regulation of the financial sector. In many respects this is uncontroversial because increased regulation should promote the behaviours we want to see, while limiting the behaviours we do not. This article takes issue with the idea that regulation, and guidelines, promote ethical behaviour in the way that we want them to. Firstly, judgement is often required to implement guidelines and regulations, which allows room for unethical behaviour. Secondly, we want financial professionals to behave ethically even when there are significant incentives not to; so, verifying compliance with regulations and guidelines when such incentives are absent isn’t a good gauge of the ethical status of financial professionals. The example of the valuation of the assets (Net Asset Value) of hedge funds illustrates these issues well. There are many rules governing how NAVs should be calculated, and while we can verify that a fund is abiding by the rules, the ability to manipulate the NAV remains. Furthermore, the incentive to do so is often only present during a crisis (usually to understate losses). Therefore, verifying compliance during ‘normal’ times is of little help in judging whether a manager will behave ethically when we most need him to. I first argue that the gap between compliance with regulations and ethical behaviour is best filled by the adoption of a virtue ethics approach, as discussed by Spalding & Oddo (2011), and Graafland & van de Ven (2011). I then discuss the implications of adopting such an approach for the financial industry’s codes of conduct, such as the CFA Institute’s ‘Code of Ethics and Standards of Professional Conduct’. Adopting a virtue ethics perspective would require a significant change in the emphasis of such codes.