nep-cmp New Economics Papers
on Computational Economics
Issue of 2022‒03‒28
sixteen papers chosen by
Stan Miles
Thompson Rivers University

  1. Machine Learning Models in Stock Market Prediction By Gurjeet Singh
  2. Pricing options on flow forwards by neural networks in Hilbert space By Fred Espen Benth; Nils Detering; Luca Galimberti
  3. Fairness constraint in Structural Econometrics and Application to fair estimation using Instrumental Variables By Samuele Centorrino; Jean-Pierre Florens; Jean-Michel Loubes
  4. Economists in the 2008 Financial Crisis: Slow to See, Fast to Act By Levy, Daniel; Mayer, Tamir; Raviv, Alon
  5. Nonparametric Adaptive Robust Control Under Model Uncertainty By Erhan Bayraktar; Tao Chen
  6. Control of a Conveyor Based on a Neural Network By Pihnastyi, Oleh; Kozhevnikov, Georgii
  7. Can LSTM outperform volatility-econometric models? By German Rodikov; Nino Antulov-Fantulin
  8. Neural Generalised AutoRegressive Conditional Heteroskedasticity By Zexuan Yin; Paolo Barucca
  9. Universal approximation of credit portfolio losses using Restricted Boltzmann Machines By Giuseppe Genovese; Ashkan Nikeghbali; Nicola Serra; Gabriele Visentin
  10. The Good Shepherd: An Oracle Agent for Mechanism Design By Jan Balaguer; Raphael Koster; Christopher Summerfield; Andrea Tacchetti
  11. Inverse Selection By Markus Brunnermeier; Rohit Lamba; Carlos Segura-Rodriguez
  12. Hierarchical Sensitivity Parity By Alejandro Rodriguez
  13. HCMD-zero: Learning Value Aligned Mechanisms from Data By Jan Balaguer; Raphael Koster; Ari Weinstein; Lucy Campbell-Gillingham; Christopher Summerfield; Matthew Botvinick; Andrea Tacchetti
  14. REPPlab: An R package for detecting clusters and outliers using exploratory projection pursuit By Daniel Fischer; Alain Berro; Klaus Nordhausen; Anne Ruiz-Gazen
  15. Political and Non-Political Officials in Local Government By Resce, Giuliano
  16. Foreign Doctorate Students in Europe By Laureti, Lucio; Costantiello, Alberto; Matarrese, Marco Maria; Leogrande, Angelo

  1. By: Gurjeet Singh
    Abstract: The paper focuses on predicting the Nifty 50 Index by using 8 Supervised Machine Learning Models. The techniques used for empirical study are Adaptive Boost (AdaBoost), k-Nearest Neighbors (kNN), Linear Regression (LR), Artificial Neural Network (ANN), Random Forest (RF), Stochastic Gradient Descent (SGD), Support Vector Machine (SVM) and Decision Trees (DT). Experiments are based on historical data of Nifty 50 Index of Indian Stock Market from 22nd April, 1996 to 16th April, 2021, which is time series data of around 25 years. During the period there were 6220 trading days excluding all the non trading days. The entire trading dataset was divided into 4 subsets of different size-25% of entire data, 50% of entire data, 75% of entire data and entire data. Each subset was further divided into 2 parts-training data and testing data. After applying 3 tests- Test on Training Data, Test on Testing Data and Cross Validation Test on each subset, the prediction performance of the used models were compared and after comparison, very interesting results were found. The evaluation results indicate that Adaptive Boost, k- Nearest Neighbors, Random Forest and Decision Trees under performed with increase in the size of data set. Linear Regression and Artificial Neural Network shown almost similar prediction results among all the models but Artificial Neural Network took more time in training and validating the model. Thereafter Support Vector Machine performed better among rest of the models but with increase in the size of data set, Stochastic Gradient Descent performed better than Support Vector Machine.
    Date: 2022–02
  2. By: Fred Espen Benth; Nils Detering; Luca Galimberti
    Abstract: We propose a new methodology for pricing options on flow forwards by applying infinite-dimensional neural networks. We recast the pricing problem as an optimization problem in a Hilbert space of real-valued function on the positive real line, which is the state space for the term structure dynamics. This optimization problem is solved by facilitating a novel feedforward neural network architecture designed for approximating continuous functions on the state space. The proposed neural net is built upon the basis of the Hilbert space. We provide an extensive case study that shows excellent numerical efficiency, with superior performance over that of a classical neural net trained on sampling the term structure curves.
    Date: 2022–02
  3. By: Samuele Centorrino; Jean-Pierre Florens; Jean-Michel Loubes
    Abstract: A supervised machine learning algorithm determines a model from a learning sample that will be used to predict new observations. To this end, it aggregates individual characteristics of the observations of the learning sample. But this information aggregation does not consider any potential selection on unobservables and any status-quo biases which may be contained in the training sample. The latter bias has raised concerns around the so-called \textit{fairness} of machine learning algorithms, especially towards disadvantaged groups. In this chapter, we review the issue of fairness in machine learning through the lenses of structural econometrics models in which the unknown index is the solution of a functional equation and issues of endogeneity are explicitly accounted for. We model fairness as a linear operator whose null space contains the set of strictly {\it fair} indexes. A {\it fair} solution is obtained by projecting the unconstrained index into the null space of this operator or by directly finding the closest solution of the functional equation into this null space. We also acknowledge that policymakers may incur a cost when moving away from the status quo. Achieving \textit{approximate fairness} is obtained by introducing a fairness penalty in the learning procedure and balancing more or less heavily the influence between the status quo and a full fair solution.
    Date: 2022–02
  4. By: Levy, Daniel; Mayer, Tamir; Raviv, Alon
    Abstract: We study the economics and finance scholars’ reaction to the 2008 financial crisis using machine learning language analyses methods of Latent Dirichlet Allocation and dynamic topic modelling algorithms, to analyze the texts of 14,270 NBER working papers covering the 1999–2016 period. We find that academic scholars as a group were insufficiently engaged in crises’ studies before 2008. As the crisis unraveled, however, they switched their focus to studying the crisis, its causes, and consequences. Thus, the scholars were “slow-to-see,” but they were “fast-to-act.” Their initial response to the ongoing Covid-19 crisis is consistent with these conclusions.
    Keywords: 2008 Financial Crisis; Financial Crises; Economic Crisis; Great Recession; Textual Analysis; LDA Topic Modeling; Dynamic Topic Modeling; Machine Learning; Securitization; Repo; Sudden Stop
    JEL: A11 C38 C55 E32 E44 E52 E58 F30 G01 G20 G21 G28
    Date: 2022–02–13
  5. By: Erhan Bayraktar; Tao Chen
    Abstract: We consider a discrete time stochastic Markovian control problem under model uncertainty. Such uncertainty not only comes from the fact that the true probability law of the underlying stochastic process is unknown, but the parametric family of probability distributions which the true law belongs to is also unknown. We propose a nonparametric adaptive robust control methodology to deal with such problem. Our approach hinges on the following building concepts: first, using the adaptive robust paradigm to incorporate online learning and uncertainty reduction into the robust control problem; second, learning the unknown probability law through the empirical distribution, and representing uncertainty reduction in terms of a sequence of Wasserstein balls around the empirical distribution; third, using Lagrangian duality to convert the optimization over Wasserstein balls to a scalar optimization problem, and adopting a machine learning technique to achieve efficient computation of the optimal control. We illustrate our methodology by considering a utility maximization problem. Numerical comparisons show that the nonparametric adaptive robust control approach is preferable to the traditional robust frameworks.
    Date: 2022–02
  6. By: Pihnastyi, Oleh; Kozhevnikov, Georgii
    Abstract: The present study is devoted to the design of the main flow parameters of a conveyor control system with a large number of sections. For the design of the control system, a neural network is used. The architecture of the neural network is justified and the rules for the formation of nodes for the input and output layers are defined. The main parameters of the model are identified and analyzed. The data set for training the neural network is formed using the analytical model of the transport system. The criterion for the quality of the transport system is written. For the given criterion for the quality of the transport system, the Pontryagin function is defined and the adjoint system of equations is given. It allows calculating optimal control of the transport system. For calculation is used additional model of the transport system with output nodes which are controls. A graphical representation of the results of the study is given
    Keywords: PDE-model production; PiKh-model; distributed system; optimal control
    JEL: C02 C15 C25 C44 D24 L23 Q21
    Date: 2020–10–09
  7. By: German Rodikov; Nino Antulov-Fantulin
    Abstract: Volatility prediction for financial assets is one of the essential questions for understanding financial risks and quadratic price variation. However, although many novel deep learning models were recently proposed, they still have a "hard time" surpassing strong econometric volatility models. Why is this the case? The volatility prediction task is of non-trivial complexity due to noise, market microstructure, heteroscedasticity, exogenous and asymmetric effect of news, and the presence of different time scales, among others. In this paper, we analyze the class of long short-term memory (LSTM) recurrent neural networks for the task of volatility prediction and compare it with strong volatility-econometric models.
    Date: 2022–02
  8. By: Zexuan Yin; Paolo Barucca
    Abstract: We propose Neural GARCH, a class of methods to model conditional heteroskedasticity in financial time series. Neural GARCH is a neural network adaptation of the GARCH 1,1 model in the univariate case, and the diagonal BEKK 1,1 model in the multivariate case. We allow the coefficients of a GARCH model to be time varying in order to reflect the constantly changing dynamics of financial markets. The time varying coefficients are parameterised by a recurrent neural network that is trained with stochastic gradient variational Bayes. We propose two variants of our model, one with normal innovations and the other with Students t innovations. We test our models on a wide range of univariate and multivariate financial time series, and we find that the Neural Students t model consistently outperforms the others.
    Date: 2022–02
  9. By: Giuseppe Genovese; Ashkan Nikeghbali; Nicola Serra; Gabriele Visentin
    Abstract: We introduce a new portfolio credit risk model based on Restricted Boltzmann Machines (RBMs), which are stochastic neural networks capable of universal approximation of loss distributions. We test the model on an empirical dataset of default probabilities of 30 investment-grade US companies and we show that it outperforms commonly used parametric factor copula models -- such as the Gaussian or the t factor copula models -- across several credit risk management tasks. In particular, the model leads to better out-of-sample fits for the empirical loss distribution and more accurate risk measure estimations. We introduce an importance sampling procedure which allows risk measures to be estimated at high confidence levels in a computationally efficient way and which is a substantial improvement over the Monte Carlo techniques currently available for copula models. Furthermore, the statistical factors extracted by the model admit an interpretation in terms of the underlying portfolio sector structure and provide practitioners with quantitative tools for the management of concentration risk. Finally, we show how to use the model for stress testing by estimating stressed risk measures (e.g. stressed VaR) for our empirical portfolio under various macroeconomic stress test scenarios, such as those specified by the FRB's Dodd-Frank Act stress test.
    Date: 2022–02
  10. By: Jan Balaguer; Raphael Koster; Christopher Summerfield; Andrea Tacchetti
    Abstract: From social networks to traffic routing, artificial learning agents are playing a central role in modern institutions. We must therefore understand how to leverage these systems to foster outcomes and behaviors that align with our own values and aspirations. While multiagent learning has received considerable attention in recent years, artificial agents have been primarily evaluated when interacting with fixed, non-learning co-players. While this evaluation scheme has merit, it fails to capture the dynamics faced by institutions that must deal with adaptive and continually learning constituents. Here we address this limitation, and construct agents ("mechanisms") that perform well when evaluated over the learning trajectory of their adaptive co-players ("participants"). The algorithm we propose consists of two nested learning loops: an inner loop where participants learn to best respond to fixed mechanisms; and an outer loop where the mechanism agent updates its policy based on experience. We report the performance of our mechanism agents when paired with both artificial learning agents and humans as co-players. Our results show that our mechanisms are able to shepherd the participants strategies towards favorable outcomes, indicating a path for modern institutions to effectively and automatically influence the strategies and behaviors of their constituents.
    Date: 2022–02
  11. By: Markus Brunnermeier (Princeton University); Rohit Lamba (Pennsylvania State University); Carlos Segura-Rodriguez (Banco Central de Costa Rica)
    Abstract: Big data, machine learning and AI inverts adverse selection problems. It allows insurers to infer statistical information and thereby reverses information advantage from the insuree to the insurer. In a setting with two-dimensional type space whose correlation can be inferred with big data we derive three results: First, a novel tradeoff between a belief gap and price discrimination emerges. The insurer tries to protect its statistical information by offering only a few screening contracts. Second, we show that forcing the insurance company to reveal its statistical information can be welfare improving. Third, we show in a setting with naive agents that do not perfectly infer statistical information from the price of offered contracts, price discrimination significantly boosts insurer’s profits. We also discuss the significance our analysis through three stylized facts: the rise of data brokers, the importance of consumer activism and regulatory forbearance, and merits of a public data repository.
    Keywords: Insurance, Big Data, Informed Principal, Belief Gap, Price Discrimination
    JEL: G22 D82 D86 C55
    Date: 2020–04
  12. By: Alejandro Rodriguez
    Abstract: In this work we present a new framework for modelling portfolio dynamics and how to incorporate this information in the portfolio selection process. We define drivers for asset and portfolio dynamics, and their optimal selection. We introduce the new Commonality Principle, which gives a solution for the optimal selection of portfolio drivers as being the common drivers. Asset dynamics are modelled by PDEs and approximated with Neural Networks, and sensitivities of portfolio constituents with respect to portfolio common drivers are obtained via Automatic Adjoint Differentiation (AAD). Information of asset dynamics is incorporated via sensitivities into the portfolio selection process. Portfolio constituents are projected into a hypersurface, from a vector space formed by the returns of common drivers of the portfolio. The commonality principle allows for the necessary geometric link between the hyperplane formed by portfolio constituents in a traditional setup with no exogenous information, and the hypersurface formed by the vector space of common portfolio drivers, so that when portfolio constituents are projected into this hypersurface, the representations of idiosyncratic risks from the hyperplane are kept at most in this new subspace, while systematic risks representations are added via exogenous information as part of this common drivers vector space. We build a sensitivity matrix, which is a similarity matrix of the projections in this hypersurface, and can be used to optimize for diversification on both, idiosyncratic and systematic risks, which is not contemplated on the literature. Finally, we solve the convex optimization problem for optimal diversification by applying a hierarchical clustering to the sensitivity matrix, avoiding quadratic optimizers for the matrix properties, and we reach over-performance in all experiments with respect to all other out-of-sample methods.
    Date: 2022–02
  13. By: Jan Balaguer; Raphael Koster; Ari Weinstein; Lucy Campbell-Gillingham; Christopher Summerfield; Matthew Botvinick; Andrea Tacchetti
    Abstract: Artificial learning agents are mediating a larger and larger number of interactions among humans, firms, and organizations, and the intersection between mechanism design and machine learning has been heavily investigated in recent years. However, mechanism design methods make strong assumptions on how participants behave (e.g. rationality), or on the kind of knowledge designers have access to a priori (e.g. access to strong baseline mechanisms). Here we introduce HCMD-zero, a general purpose method to construct mechanism agents. HCMD-zero learns by mediating interactions among participants, while remaining engaged in an electoral contest with copies of itself, thereby accessing direct feedback from participants. Our results on the Public Investment Game, a stylized resource allocation game that highlights the tension between productivity, equality and the temptation to free-ride, show that HCMD-zero produces competitive mechanism agents that are consistently preferred by human participants over baseline alternatives, and does so automatically, without requiring human knowledge, and by using human data sparingly and effectively Our detailed analysis shows HCMD-zero elicits consistent improvements over the course of training, and that it results in a mechanism with an interpretable and intuitive policy.
    Date: 2022–02
  14. By: Daniel Fischer (LUKE - Natural Resources Institute Finland); Alain Berro (IRIT-REVA - Real Expression Artificial Life - IRIT - Institut de recherche en informatique de Toulouse - UT1 - Université Toulouse 1 Capitole - Université Fédérale Toulouse Midi-Pyrénées - UT2J - Université Toulouse - Jean Jaurès - UT3 - Université Toulouse III - Paul Sabatier - Université Fédérale Toulouse Midi-Pyrénées - CNRS - Centre National de la Recherche Scientifique - Toulouse INP - Institut National Polytechnique (Toulouse) - Université Fédérale Toulouse Midi-Pyrénées, UT1 - Université Toulouse 1 Capitole - Université Fédérale Toulouse Midi-Pyrénées); Klaus Nordhausen (TU Wien - Vienna University of Technology); Anne Ruiz-Gazen (TSE - Toulouse School of Economics - UT1 - Université Toulouse 1 Capitole - Université Fédérale Toulouse Midi-Pyrénées - EHESS - École des hautes études en sciences sociales - CNRS - Centre National de la Recherche Scientifique - INRAE - Institut National de Recherche pour l’Agriculture, l’Alimentation et l’Environnement, UT1 - Université Toulouse 1 Capitole - Université Fédérale Toulouse Midi-Pyrénées)
    Abstract: The R-package REPPlab is designed to explore multivariate data sets using one-dimensional unsupervised projection pursuit. It is useful as a preprocessing step to find clusters or as an outlier detection tool for multivariate data. Except from the packages tourr and rggobi, there is no implementation of exploratory projection pursuit tools available in R. REPPlab is an R interface for the Java program EPP-lab that implements four projection indices and three biologically inspired optimization algorithms. It also proposes new tools for plotting and combining the results and specific tools for outlier detection. The functionality of the package is illustrated through some simulations and using some real data.
    Keywords: Genetic algorithms,Kurtosis,Java,Particle swarm optimization,Projection index,Tribes,Projection matrix,Unsupervised data analysis
    Date: 2021
  15. By: Resce, Giuliano
    Abstract: This paper investigates the impact of non-political administrators on the financial management of local governments. The activity of prefectorial officials is compared with the activity of elected mayors exploiting data extracted from a panel of 7826 Italian municipalities from 2007 to 2018. To address the potential confounding effects and selection biases, we combine a Difference in Difference strategy with machine learning methods for counterfactual analysis. Results show that non-political administrators bring higher financial autonomy and higher collection capacity, raising more revenues at local level. This is consistent with the hypothesis that, since they do not respond to electoral incentives, non-political administrators have lower motivations to behave strategically, not taking their own interests about electoral successes into account when they have to choose the proportion of local versus external revenues for financing local expenditure.
    Keywords: Local Government, Electoral Incentives, Accountability
    JEL: D7 H2 H77
    Date: 2022–03–16
  16. By: Laureti, Lucio; Costantiello, Alberto; Matarrese, Marco Maria; Leogrande, Angelo
    Abstract: The determinants of the presence of “Foreign Doctorate Students” among 36 European Countries for the period 2010-2019 are analyzed in this article. Panel Data with Fixed Effects, Random Effects, WLS, Pooled OLS, and Dynamic Panel are used to investigate the data. We found that the presence of Foreign Doctorate Students is positively associated to “Attractive Research Systems”, “Finance and Support”, “Rule of Law”, “Sales Impacts”, “New Doctorate Graduates”, “Basic School Entrepreneurial Education and Training”, “Tertiary Education” and negatively associated to “Innovative Sales Share”, “Innovation Friendly Environment”, “Linkages”, “Trademark Applications”, “Government Procurement of Advanced Technology Products”, “R&D Expenditure Public Sectors”. A cluster analysis was then carried out through the application of the unsupervised k-Means algorithm optimized using the Silhouette coefficient with the identification of 5 clusters. Finally, eight different machine learning algorithms were used to predict the value of the "Foreign Doctorate Students" variable. The results show that the best predictor algorithm is the "Tree Ensemble Regression" with a predicted value growing at a rate of 114.03%.
    Keywords: Innovation, and Invention: Processes and Incentives; Management of Technological Innovation and R&D; Diffusion Processes; Open Innovation.
    JEL: O30 O31 O32 O33 O34
    Date: 2022–02–11

This nep-cmp issue is ©2022 by Stan Miles. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.