An exploratory analysis of portfolio managers' probabilistic forecasts of stock prices

This study reports the results of an experiment that examines (1) the effects of forecast horizon on the performance of probability forecasters, and (2) the alleged existence of an inverse expertise effect, i.e., an inverse relationship between expertise and probabilistic forecasting performance. Portfolio managers are used as forecasters with substantive expertise. Performance of this ‘expert’ group is compared to the performance of a ‘semi-expert’ group composed of other banking professionals trained in portfolio management. It is found that while both groups attain their best discrimination performances in the four-week forecast horizon, they show their worst calibration and skill performances in the 12-week forecast horizon. Also, while experts perform better in all performance measures for the one-week horizon, semi-experts achieve better calibration for the four-week horizon. It is concluded that these results may signal the existence of an inverse expertise effect that is contingent on the selected forecast horizon.     

Combining expert‐adjusted forecasts

It is well known that a combination of model‐based forecasts can improve upon each of the individual constituent forecasts. Most forecasts available in practice are, however, not purely based on econometric models but entail adjustments, where experts with domain‐specific knowledge modify the original model forecasts. There is much evidence that expert‐adjusted forecasts do not necessarily improve the pure model‐based forecasts. In this paper we show, however, that combined expert‐adjusted model forecasts can improve on combined model forecasts, in the case when the individual expert‐adjusted forecasts are not better than their associated model‐based forecasts. We discuss various implications of this finding.     

Do experts' adjustments on model‐based SKU‐level forecasts improve forecast quality?

Model‐based SKU‐level forecasts are often adjusted by experts. In this paper we propose a statistical methodology to test whether these expert forecasts improve on model forecasts. Application of the methodology to a very large database concerning experts in 35 countries who adjust SKU‐level forecasts for pharmaceutical products in seven distinct categories leads to the general conclusion that expert forecasts are equally good at best, but are more often worse than model‐based forecasts. We explore whether this is due to experts putting too much weight on their contribution, and this indeed turns out to be the case. Copyright © 2009 John Wiley & Sons, Ltd.     

Subjective confidence in forecasts

Forecasts have little value to decision makers unless it is known how much confidence to place in them. Those expressions of confidence have, in turn, little value unless forecasters are able to assess the limits of their own knowledge accurately Previous research has shown very robust patterns in the judgements of individuals who have not received special training in confidence assessment: Knowledge generally increases as confidence increases. However, it increases too swiftly, with a doubling of confidence being associated with perhaps a 50 per cent increase in knowledge. With all but the easiest of tasks, people tend to be overconfident regarding how much they know These results have typically been derived from studies of judgements of general knowledge. The present study found that they also pertained to confidence in forecasts. Indeed, the confidence-knowledge curves observed here were strikingly similar to those observed previously. The only deviation was the discovery that a substantial minority of judges never expressed complete confidence in any of their forecasts. These individuals also proved to be better assessors of the extent of their own knowledge Apparently confidence in forecasts is determined by processes similar to those that determine confidence in general knowledge. Decision makers can use forecasters assessments in a relative sense, in order to predict when they are more and less likely to be correct. However, they should be hesitant to take confidence assessments literally. Someone is more likely to be right when he or she is ‘certain’than when he or she is ‘fairly confident’; but there is no guarantee that the certain forecast will come true.     

Do Experts’ SKU Forecasts Improve after Feedback?

We analyze the behavior of experts who quote forecasts for monthly SKU‐level sales data, where we compare data before and after the moment that experts received different kinds of feedback on their behavior. We have data for 21 experts located in as many countries who make SKU‐level forecasts for a variety of pharmaceutical products for October 2006 to September 2007. We study the behavior of the experts by comparing their forecasts with those from an automated statistical program, and we report the forecast accuracy over these 12 months. In September 2007 these experts were given feedback on their behavior and they received training at the headquarters office, where specific attention was given to the ins and outs of the statistical program. Next, we study the behavior of the experts for the 3 months after the training session, i.e. October 2007 to December 2007. Our main conclusion is that in the second period the experts’ forecasts deviated less from the statistical forecasts and that their accuracy improved substantially. Copyright © 2013 John Wiley & Sons, Ltd.     

State of the field: Paper tools

Paper occupies a special place in histories of knowledge. It is the substrate of communication, the stuff of archives, the bearer of marks that make worlds. For the early-modern period in particular we now have a wealth of studies of ‘paper tools’, of the ways in which archives were assembled and put to use, of the making of lists and transcribing of observations, and so on. In other fields, too, attention has turned to the materiality of information. How far is it possible to draw a stable methodology out of the insights of literary and book historians, bibliographers, anthropologists, and those working in media studies? Do these diverse fields in fact refer to the same thing when they talk of paper, its qualities, affordances and limitations? In attempting to answer these questions, the present essay begins in the rich territory of early-modern natural philosophy – but from there opens out to take in recent works in a range of disciplines. Attending to the specific qualities of paper is only possible, I argue, if it is understood that paper can be both transparent and opaque depending on the social world it inhabits and helps to constitute. Paper flickers into and out of view, and it is precisely this quality that constitutes its sociomateriality.     

Scientists as experts: A distinct role?

The role of scientists as experts is crucial to public policymaking. However, the expert role is contested and unsettled in both public and scholarly discourse. In this paper, I provide a systematic account of the role of scientists as experts in policymaking by examining whether there are any normatively relevant differences between this role and the role of scientists as researchers. Two different interpretations can be given of how the two roles relate to each other. The separability view states that there is a normatively relevant difference between the two roles, whereas the inseparability view denies that there is such a difference. Based on a systematic analysis of the central aspects of the role of scientists as experts – that is, its aim, context, mode of output, and standards, I propose a moderate version of the separability view. Whereas the aim of scientific research is typically to produce new knowledge through the use of scientific method for evaluation and dissemination in internal settings, the aim of the expert is to provide policymakers and the public with relevant and applicable knowledge that can premise political reasoning and deliberation.     

Extraneous expert information

When a Bayesian decision maker has to choose among information sources, he should consider the anticipated impact that the information will have on his posterior distribution. In some cases he may determine at the outset that an information source will have no effect on his posterior beliefs, no matter what that source says. Such an information source is called extraneous. In this paper we discuss Bayesian conditions for extraneous information sources, and show a hypothetical example involving experts with overlapping information. Analysis of U.S. weather forecasts demonstrates how this concept can be operationalized to test hypotheses concerning the use of information by forecasters.     

Knowing what would happen: The epistemic strategies in Galileo's thought experiments

While philosophers have subjected Galileo's classic thought experiments to critical analysis, they have tended to largely ignored the historical and intellectual context in which they were deployed, and the specific role they played in Galileo's overall vision of science. In this paper I investigate Galileo's use of thought experiments, by focusing on the epistemic and rhetorical strategies that he employed in attempting to answer the question of how one can know what would happen in an imaginary scenario. Here I argue we can find three different answers to this question in Galileo later dialogues, which reflect the changing meanings of ‘experience’ and ‘knowledge’ (scientia) in the early modern period. Once we recognise that Galileo's thought experiments sometimes drew on the power of memory and the explicit appeal to ‘common experience’, while at other times, they took the form of demonstrative arguments intended to have the status of necessary truths; and on still other occasions, they were extrapolations, or probable guesses, drawn from a carefully planned series of controlled experiments, it becomes evident that no single account of the epistemological relationship between thought experiment, experience and experiment can adequately capture the epistemic variety we find Galileo's use of imaginary scenarios. To this extent, we cannot neatly classify Galileo's use of thought experiments as either ‘medieval’ or ‘early modern’, but we should see them as indicative of the complex epistemological transformations of the early seventeenth century.     

The benefits of acquiring interactional expertise: Why (some) philosophers of science should engage scientific communities

Philosophers of science are increasingly arguing for and addressing the need to do work that is socially and scientifically engaged. However, we currently lack well-developed frameworks for thinking about how we should engage other expert communities and what the epistemic benefits are of doing so. In this paper, I draw on Collins and Evans' concept of ‘interactional expertise’ – the ability to speak the language of a discipline in the absence of an ability to practice – to consider the epistemic benefits that can arise when philosophers engage scientific communities. As Collins and Evans argue, becoming an interactional expert requires that one ‘hang out’ with members of the relevant expert community in order to learn crucial tacit knowledge needed to speak the language. Building on this work, I argue that acquiring interactional expertise not only leads to linguistic fluency, but it also confers several ‘socio-epistemic’ benefits such as the opportunity to cultivate trust with scientific communities. These benefits can improve philosophical work and facilitate the broader uptake of philosophers' ideas, enabling philosophers to meet a variety of epistemic goals. As a result, having at least some philosophers of science acquire interactional expertise via engagement will likely enhance the diversity of epistemic capacities for philosophy of science as a whole. For some philosophers of science, moreover, the socio-epistemic benefits identified here may be more important than the ability to speak the language of a discipline, suggesting the need for a broader analysis of interactional expertise, which this paper also advances.     

Collaboration,interdisciplinarity, and the epistemology of contemporary science

Over the last decades, science has grown increasingly collaborative and interdisciplinary and has come to depart in important ways from the classical analyses of the development of science that were developed by historically inclined philosophers of science half a century ago. In this paper, I shall provide a new account of the structure and development of contemporary science based on analyses of, first, cognitive resources and their relations to domains, and second of the distribution of cognitive resources among collaborators and the epistemic dependence that this distribution implies. On this background I shall describe different ideal types of research activities and analyze how they differ. Finally, analyzing values that drive science towards different kinds of research activities, I shall sketch the main mechanisms underlying the perceived tension between disciplines and interdisciplinarity and argue for a redefinition of accountability and quality control for interdisciplinary and collaborative science.     

Artful Physics

This paper gives an account of the establishment and expansion of a Faculty of Science at the Calvinist ‘Free University’ in the Netherlands in the 1930s. It describes the efforts of a group of orthodox Christians to come to terms with the natural sciences in the early twentieth century. The statutes of the university, which had been founded in 1880, prescribed that all research and teaching should be based on Calvinist, biblical principles. This ideal was formulated in opposition to the claim of nineteenth-century scientific naturalists that there was an inherent conflict between science and religion. However, despite their selection on the basis of their strict Calvinist beliefs, the first science professors attributed a certain independence to the domain of science. They agreed with the criticism of the conflict thesis, and tried to defuse the tensions between science and religion, although mainly at the level of philosophy and history, looking for example for harmony between science and religion in the past. Ironically, as a result of this approach, the Calvinist scientists mainly contributed to the acceptance of mainstream science in Dutch Calvinist circles, contrary to developments in other countries (notably the USA) where the conflict between science and orthodox Christianity has reasserted itself.     

The collaboration of Emil Artin and George Whaples: Artin’s mathematical circle extends to America

In his biography of Emil Artin, Richard Brauer describes the years from 1931–1941 as a time when “Artin spoke through his students and through the members of his mathematical circle” rather than through written publications. This paper explores these seemingly quiet years when Artin immigrated to America and disseminated ideas about algebraic number theory during this time in his collaboration with George Whaples, a young American mathematician who had just completed his Ph.D. at the University of Wisconsin. The main result of their work is the use of the product formula for valuations to come up with an axiomatic characterization of both algebraic number fields and algebraic function fields with a finite field of constants. These two families of fields are exactly the fields for which class field theory is known to hold. We situate their mathematical work in the broader context of algebraic number theory and their lives within the broader historical context.     

Evaluating community science

Community science—scientific investigation conducted partly or entirely by non-professional scientists—has many advantages. For example, community science mobilizes large numbers of volunteers who can, at low cost, collect more data than traditional teams of professional scientists. Participation in research can also increase volunteers’ knowledge about and appreciation of science. At the same time, there are worries about the quality of data that community science projects produce. Can the work of non-professionals really deliver trustworthy results? Attempts to answer this question generally compare data collected by volunteers to data collected by professional scientists. When volunteer data is more variable or less accurate than professionally collected data, then the community science project is judged to be inferior to traditional science. I argue that this is not the right standard to use when evaluating community science, because it relies on a false assumption about the aims of science. I show that if we adopt the view that science has diverse aims which are often in tension with one another, then we cannot justify holding community science data to an expert accuracy standard. Instead, we should evaluate the quality of community science data based on its adequacy-for-purpose.     

Modeling for fairness: A Rawlsian approach

In this paper we introduce the overlapping design consensus for the construction of models in design and the related value judgments. The overlapping design consensus is inspired by Rawls’ overlapping consensus. The overlapping design consensus is a well-informed, mutual agreement among all stakeholders based on fairness. Fairness is respected if all stakeholders’ interests are given due and equal attention. For reaching such fair agreement, we apply Rawls’ original position and reflective equilibrium to modeling. We argue that by striving for the original position, stakeholders expel invalid arguments, hierarchies, unwarranted beliefs, and bargaining effects from influencing the consensus. The reflective equilibrium requires that stakeholders’ beliefs cohere with the final agreement and its justification. Therefore, the overlapping design consensus is not only an agreement to decisions, as most other stakeholder approaches, it is also an agreement to their justification and that this justification is consistent with each stakeholders’ beliefs. For supporting fairness, we argue that fairness qualifies as a maxim in modeling. We furthermore distinguish values embedded in a model from values that are implied by its context of application. Finally, we conclude that for reaching an overlapping design consensus communication about properties of and values related to a model is required.     

基于USD—IOWA算子的方案优选群决策方法

针对方案优选决策中评估专家经多轮交互后没有达成群体共识的问题，提出了相对稳定偏好的概念。通过对评估专家偏好信息的计算获得专家的效用稳定度，并以此作为USD—IOWA算子的诱导分量完成对单个评估专家相对稳定偏好信息的群集结，得到评估专家群体的相对稳定偏好矩阵。在此基础上采用基于量化优势度的方法对备选方案进行选择。最后给出一个算例说明方法的可行性。     

De-anthropomorphizing energy and energy conservation: The case of Max Planck and Ernst Mach

Discussions on the relation between Mach and Planck usually focus on their famous controversy, a conflict between ‘instrumentalist’ and realist philosophies of science that revolved around the specific issue of the existence of atoms. This article approaches their relation from a different perspective, comparing their analyses of energy and energy conservation. It is argued that this reveals a number of striking similarities and differences. Both Mach and Planck agreed that the law was valid, and they sought to purge energy of its anthropomorphic elements. They did so in radically different ways, however, illustrating the differences between Mach's ‘historical’ and Planck's ‘rationalistic’ accounts of knowledge. Planck's attempt to de-anthropomorphize energy was part of his attempt to demarcate theoretical physics from other disciplines. Mach's attempt to de-anthropomorphize energy is placed in the context of fin-de-siècle Vienna. By doing so, this article also proposes a new interpretation of Mach as a philosopher, historian and sociologist of science.     

Moral trust & scientific collaboration

Modern scientific knowledge is increasingly collaborative. Much analysis in social epistemology models scientists as self-interested agents motivated by external inducements and sanctions. However, less research exists on the epistemic import of scientists’ moral concern for their colleagues. I argue that scientists’ trust in their colleagues’ moral motivations is a key component of the rationality of collaboration. On the prevailing account, trust is a matter of mere reliance on the self-interest of one’s colleagues. That is, scientists merely rely on external compulsion to motivate self-interested colleagues to be trustworthy collaborators. I show that this self-interest account has significant limitations. First, it cannot fully account for trust by relatively powerless scientists. Second, reliance on self-interest can be self-defeating. For each limitation, I show that moral trust can bridge the gap—when members of the scientific community cannot rely on the self-interest of their colleagues, they rationally place trust in the moral motivations of their colleagues. Case studies of mid-twentieth-century industrial laboratories and exploitation of junior scientists show that such moral trust justifies collaboration when mere reliance on the self-interest of colleagues would be irrational. Thus, this paper provides a more complete and realistic account of the rationality of scientific collaboration.