Skip to main content
Publication records
ESMT Working Paper
ESMT Working Paper No. 22-03
David Ronayne, Roberto Veneziani, William R. Zame (2022)
Subject(s)
Economics, politics and business environment; Management sciences, decision sciences and quantitative methods
Keyword(s)
subjective probability, choice under uncertainty, online experiments
JEL Code(s)
D01, D81, D84, C09
Anscombe & Aumann (1963) offer a definition of subjective probability in terms of comparisons with objective probabilities. That definition - which has provided the basis for much of the succeeding work on subjective probability - presumes that the subjective probability of an event is independent of the prize consequences of that event, a property we term Prize Independence. We design experiments to test Prize Independence and find that a large fraction of our subjects violate it; thus, they do not have subjective probabilities. These findings raise questions about the empirical relevance of much of the literature on subjective probability.
Pages
49
ISSN (Print)
1866–3494
Subject(s)
Information technology and systems; Management sciences, decision sciences and quantitative methods; Technology, R&D management
Keyword(s)
machine accuracy, decision making, human-in-the-loop, algorithm aversion, dynamic learning
Artificial intelligence systems are increasingly demonstrating their capacity to make better predictions than human experts. Yet, recent studies suggest that professionals sometimes doubt the quality of these systems and overrule machine-based prescriptions. This paper explores the extent to which a decision maker (DM) supervising a machine to make high-stake decisions can properly assess whether the machine produces better recommendations. To that end, we study a set-up, in which a machine performs repeated decision tasks (e.g., whether to perform a biopsy) under the DM’s supervision. Because stakes are high, the DM primarily focuses on making the best choice for the task at hand. Nonetheless, as the DM observes the correctness of the machine’s prescriptions across tasks, she updates her belief about the machine. However, the DM observes the machine’s correctness only if she ultimately decides to act on the task. Further, the DM sometimes overrides the machine depending on her belief, which affects learning. In this set-up, we characterize the evolution of the DM’s belief and overruling decisions over time. We identify situations under which the DM hesitates forever whether the machine is better, i.e., she never fully ignores but regularly overrules it. Moreover, the DM sometimes wrongly believes with positive probability that the machine is better. We fully characterize the conditions under which these learning failures occur and explore how mistrusting the machine affects them. Our results highlight some fundamental limitations in determining whether machines make better decisions than experts and provide a novel explanation for human-machine complementarity.
Pages
42
ISSN (Print)
1866–3494
ESMT Working Paper
ESMT Working Paper No. 22-01
Mirko Kremer, Francis de Véricourt (2022)
Subject(s)
Management sciences, decision sciences and quantitative methods
Keyword(s)
congestion, diagnostic accuracy, experiments, partially observable markov decision process, path-dependent decision making, undertesting, task completion bias
To study the effect of congestion on the fundamental trade-off between diagnostic accuracy and speed, we empirically test the predictions of a formal sequential testing model in a setting where the gathering of additional information can improve diagnostic accuracy, but may also take time and increase congestion as a result. The efficient management of such systems requires a careful balance of congestion-sensitive stopping rules. These include diagnoses made based on very little or no diagnostic information, and the stopping of diagnostic processes while waiting for information. We test these rules under controlled laboratory conditions, and link the observed biases to system dynamics and performance. Our data shows that decision makers (DMs) stop diagnostic processes too quickly at low congestion levels where information acquisition is relatively cheap. But they fail to stop quickly enough when increasing congestion requires the DM to diagnose without testing, or diagnose while waiting for test results. Essentially, DMs are insufficiently sensitive to congestion. As a result of these behavioral patterns, DMs manage the system with both lower-than-optimal diagnostic accuracy and higher-than-optimal congestion cost, underperforming on both sides of the accuracy/speed trade-off.
Pages
40
ISSN (Print)
1866–3494
Working Paper
SSRN Working Paper
Louis-Daniel Pape, Christian Helmers, Alessandro Iaria, Stefan Wagner, Julian Runge (2021)
Subject(s)
Strategy and general management
Keyword(s)
Price discrimination, personalized pricing, mobile apps, online games, freemium
JEL Code(s)
D40, L11
We use a unique dataset from a mobile puzzle game to investigate the welfare consequences of price discrimination. We rely on experimental variation to characterize player behavior and estimate a model of demand for game content. Our counterfactual simulations show that optimal uniform pricing would increase profit by +340% with respect to the game developer’s observed pricing. This is almost the same as the increase in profit associated with first-degree price discrimination (+347%). All pricing strategies considered—including optimal uniform pricing—would induce a transfer of surplus from players to game developer without, however, generating sizeable dead-weight losses.
Pages
93
ESMT Working Paper
ESMT Working Paper No. 21-02
Hans W. Friederiszick, Alexis Walckiers (2021)
Subject(s)
Economics, politics and business environment; Strategy and general management
Keyword(s)
economic dependency, bargaining theory, vertical restraints, law & economics, competition law
JEL Code(s)
D43, D86, L42, K210
An increasing number of countries have introduced some form of prohibition of abuses of economic dependence or broadened the scope of their existing legislation. Yet, very little has been written on the economics of economic dependence, that is on economic reasoning, tools or metrics that can be relied upon to identify whether a company is economically dependent on another company. The present paper aims to fill this gap, and argues that bargaining theory and the economics of relative market power can be helpful to characterise economic dependence. We summarise a number of takeaways from this literature, and describe empirical strategies that can be relied upon to try and quantify economic dependence in specific cases.
ISSN (Print)
1866–3494
ESMT Working Paper
ESMT Working Paper No. 19-02 (R3)
Işık Biçer, Florian Lücker, Tamer Boyaci (2021)
Subject(s)
Management sciences, decision sciences and quantitative methods; Product and operations management
Keyword(s)
Product proliferation, lead-time reduction, process redesign, delayed differentiation
Product proliferation occurs in supply chains when manufacturers respond to diverse market needs by trying to produce a range of products from a limited variety of raw materials. In such a setting, manufacturers can establish market responsiveness and/or cost efficiency in alternative ways. Delaying the point of the proliferation helps manufacturers improve their responsiveness by postponing the ordering decisions of the final products until there is partial or full resolution of the demand uncertainty. This strategy can be implemented in two different ways: (1) redesigning the operations so that the point of proliferation is swapped with a downstream operation or (2) reducing the lead times. To establish cost efficiency, manufacturers can systematically reduce their operational costs or postpone the high-cost operations. We consider a multi-echelon and multi-product newsvendor problem with demand forecast evolution to analyze the value of each operational lever of the responsiveness and the efficiency. We use a generalized forecast-evolution model to characterize the demand-updating process, and develop a dynamic optimization model to determine the optimal order quantities at different echelons. Using anonymized data of Kordsa Inc., a global manufacturer of advanced composites and reinforcement materials, we show that our model outperforms a theoretical benchmark of the repetitive newsvendor model. We demonstrate that reducing the lead time of a downstream operation is more beneficial to manufacturers than reducing the lead time of an upstream operation by the same amount, whereas reducing the upstream operational costs is more favorable than reducing the downstream operational costs. We also indicate that delaying the proliferation may cause a loss of profit, even if it can be achieved with no additional costs. Finally, a decision typology is developed, which shows effective operational strategies depending on product/market characteristics and process flexibility.
Pages
52
ISSN (Print)
1866–3494
Working Paper
CESifo Working Paper Series No. 9227
Ingrid Huitfeldt, Andreas R. Kostol, Jan Sebastian Nimczik, Andrea Weber (2021)
Subject(s)
Economics, politics and business environment
Keyword(s)
internal labor markets, organization of labor, wage setting
JEL Code(s)
J310, J620, M500
This paper develops a new method to study how workers’ career and wage profiles are shaped by internal labor markets (ILM) and job hierarchies in firms. Our paper tackles the conceptual challenge of organizing jobs within firms into hierarchy levels by proposing a data-driven ranking method based on observed worker flows between occupations within firms. We apply our method to linked employer-employee data from Norway that records fine-grained occupational codes and tracks contract changes within firms. Our findings confirm existing evidence that is primarily based on case studies for single firms. We expand on this by documenting substantial heterogeneity in the structure and hierarchy of ILMs across a broad range of large firms. Our findings on wage and promotion dynamics in ILMs are consistent with models of careers in organizations.
ESMT Working Paper
ESMT Working Paper No. 21-01
Wolfgang Briglauer, Michał Grajek (2021)
Subject(s)
Economics, politics and business environment; Information technology and systems; Technology, R&D management
Keyword(s)
Fiber optic technology, state aid, ex-post evaluation, efficiency, OECD countries
JEL Code(s)
C51, C54, H25, L52, O38
The deployment of new broadband networks (NBNs) based on fiber-optic transmission technologies promises high gains in terms of productivity and economic growth, and has attracted subsidies worth billions from governments around the world in the form of various state aid programs. Yet, the effectiveness and the efficiency of such programs remains largely unstudied. We employ panel data from 32 OECD countries during 2002-2019 to provide robust empirical evidence of both. We find that state aid significantly increases NBNs by facilitating the deployment of new connections to 22% of households in the short term and 39.2% in the long term. By comparing the actual amounts of state aid support to the estimated impact on GDP growth, we also find it to be highly cost efficient, as the programs break even after three years on average.
Pages
35
ISSN (Print)
1866–3494
ESMT Working Paper
ESMT Working Paper No. 20-01 (R3)
Forthcoming in Management Science.
Subject(s)
Management sciences, decision sciences and quantitative methods; Product and operations management; Technology, R&D management
Keyword(s)
Data, machine learning, data product, pricing, incentives, contracting
This paper explores how firms that lack expertise in machine learning (ML) can leverage the so-called AI Flywheel effect. This effect designates a virtuous cycle by which, as an ML product is adopted and new user data are fed back to the algorithm, the product improves, enabling further adoptions. However, managing this feedback loop is difficult, especially when the algorithm is contracted out. Indeed, the additional data that the AI Flywheel effect generates may change the provider's incentives to improve the algorithm over time. We formalize this problem in a simple two-period moral hazard framework that captures the main dynamics among ML, data acquisition, pricing, and contracting. We find that the firm's decisions crucially depend on how the amount of data on which the machine is trained interacts with the provider's effort. If this effort has a more (less) significant impact on accuracy for larger volumes of data, the firm underprices (overprices) the product. Interestingly, these distortions sometimes improve social welfare, which accounts for the customer surplus and profits of both the firm and provider. Further, the interaction between incentive issues and the positive externalities of the AI Flywheel effect has important implications for the firm's data collection strategy. In particular, the firm can boost its profit by increasing the product's capacity to acquire usage data only up to a certain level. If the product collects too much data per user, the firm's profit may actually decrease, i.e., more data is not necessarily better.
Pages
48
ISSN (Print)
1866–3494
Working Paper
CEPR Discussion Paper No. DP16243
Simon P. Anderson, Özlem Bedre-Defolie (2021)
Subject(s)
Economics, politics and business environment
Keyword(s)
Trade platform, hybrid business model, antitrust policy, tax policy
JEL Code(s)
D42, L12, L13, L40, H25
We provide a canonical and tractable model of a trade platform enabling buyers and sellers to transact. The platform charges a percentage fee on third-party product sales and decides whether to be "hybrid", like Amazon, by selling its own product. It thereby controls the number of differentiated products (variety) it hosts and their prices. Using the mixed market demand system, we capture interactions between monopolistically competitive sellers and a sizeable platform product. Using long-run aggregative games with free entry, we endogenize seller participation through an aggregate variable manipulated by the platform's fee. We show that a higher quality (or lower cost) of the platform's product increases its market share and the seller fee, and lowers consumer surplus. Banning hybrid mode benefits consumers. The hybrid platform might favor its product and debase third-party products if the own product advantage is sufficiently high. We also provide some tax policy implications.
With permission of CEPR