nep-cmp New Economics Papers
on Computational Economics
Issue of 2023‒03‒20
sixteen papers chosen by
Stan Miles
Thompson Rivers University

  1. Sentiment Spin: Attacking Financial Sentiment with GPT-3 By Markus Leippold
  2. The Transformation of Public Policy Analysis in Times of Crisis – A Microsimulation-Nowcasting Method Using Big Data By O'Donoghue, Cathal; Sologon, Denisa Maria
  3. Data Driven Contagion Risk Management in Low-Income Countries using Machine Learning Applications with COVID-19 in South Asia By Abu S. Shonchoy; Moogdho M. Mahzab; Towhid I. Mahmood; Manhal Ali
  4. Mass Valuation of Real Estate Using GIS-based Nominal Valuation and Machine Learning Methods By Muhammed Oguzhan Mete; Tahsin Yomralioglu
  5. Predicting Firm Exits with Machine Learning: Implications for Selection into COVID-19 Support and Productivity Growth By Lily Davies; Mark Kattenberg; Benedikt Vogt
  6. ddml: Double/Debiased Machine Learning in Stata By Ahrens, Achim; Hansen, Christian B.; Schaffer, Mark E; Wiemann, Thomas
  7. Preparing random state for quantum financing with quantum walks By Yen-Jui Chang; Wei-Ting Wang; Hao-Yuan Chen; Shih-Wei Liao; Ching-Ray Chang
  8. Simulation analysis of the impact of blockchain-orchestrated home sales in Sweden on housing price dynamics By Jaroslaw Morawski; Anetta Proskurovska
  9. Stock Broad-Index Trend Patterns Learning via Domain Knowledge Informed Generative Network By Jingyi Gu; Fadi P. Deek; Guiling Wang
  10. Fourth Industrial Revolution and Evolution of Data Science: Challenges for Official Statistics By Popoola, Osuolale Peter; Adeboye, Olawale Nureni
  11. Modelling Subjective Attractiveness By Konrad Lewszyk; Piotr Wójcik
  12. Forecasting realized volatility in turbulent times using temporal fusion transformers By Frank, Johannes
  13. The supply, demand and characteristics of the AI workforce across OECD countries By Andrew Green; Lucas Lamby
  14. Age, wealth, and the MPC in Europe: A supervised machine learning approach By Dutt, Satyajit; Radermacher, Jan W.
  15. Using supervised machine learning to scale human‐coded data: A method and dataset in the board leadership context By Harrison, Joseph S.; Josefy, Matthew A.; Kalm, Matias; Krause, Ryan
  16. PRUDEX-Compass: Towards Systematic Evaluation of Reinforcement Learning in Financial Markets By Shuo Sun; Molei Qin; Xinrun Wang; Bo An

  1. By: Markus Leippold (University of Zurich; Swiss Finance Institute)
    Abstract: The use of dictionaries in financial sentiment analysis and other financial and economic applications remains widespread because keyword-based methods appear more transparent and explainable than more advanced techniques commonly used in computer science. However, this paper demonstrates the vulnerability of using dictionaries by exploiting the eloquence of GPT-3, a sophisticated transformer model, to generate successful adversarial attacks on keyword-based approaches with a success rate close to 99% for negative sentences in the financial phrase base, a well-known human-annotated database for financial sentiment analysis. In contrast, more advanced methods, such as those using context-aware approaches like BERT, remain robust.
    Keywords: sentiment analysis in financial markets, keyword-based approach, FinBERT, GPT-3
    JEL: G2 G38 C8 M48
    Date: 2023–02
  2. By: O'Donoghue, Cathal (National University of Ireland, Galway); Sologon, Denisa Maria (LISER (CEPS/INSTEAD))
    Abstract: The urgency of the two crises, especially the COVID-19 pandemic, revealed the inadequacy of traditional statistical datasets and models to provide a timely support to the decision-making process in times of volatility. Drawing upon advances in data analytics for public policy and the increasing availability of real-time data, we develop and evaluate a method for real-time policy evaluations of tax and social protection policies. Our method goes beyond the state-of-the-art by implementing an aligned or calibrated microsimulation approach to generate a counterfactual income distribution as a function of more timely external data than the underlying income survey. We evaluate the simulation performance between our approach and the transition matrix approach by undertaking a nowcast for a historical crisis, judging against an actual change and each other. Nowcasting emerges as a useful methodology for examining up-to-date statistics on labour force participation, income distribution, prices, and income inequality. We find significant differences between approaches when the calibration involves structural heterogenous changes. The model replicates the changes in income distribution over one year; over the longer term, the model is able to capture the trend, but the precision of the levels weakens the further we get from the estimation year.
    Keywords: big data, policy analysis, nowcasting, microsimulation, COVID-19
    JEL: I31 I38 C54
    Date: 2023–02
  3. By: Abu S. Shonchoy (Department of Economics, Florida International University); Moogdho M. Mahzab (Stanford University); Towhid I. Mahmood (Texas Tech University); Manhal Ali (University of Leeds)
    Abstract: In the absence of real-time surveillance data, it is difficult to derive an early warning system and potential outbreak locations with the existing epidemiological models, especially in resource-constrained countries. We proposed a Contagion Risk Index (CR-Index) - based on publicly available national statistics – founded on communicable disease spreadability vectors. Utilizing the daily COVID-19 data (positive cases and deaths) from 2020-2022, we developed country-specific and sub-national CR-Index for South Asia (India, Pakistan, and Bangladesh) and identified potential infection hotspots-aiding policymakers with efficient mitigation planning. Across the study period, the week-by-week and fixed-effects regression estimates demonstrate a strong correlation between the proposed CR-Index and sub-national (district-level) COVID-19 statistics. We validated the CR-Index using machine learning methods by evaluating the out-of-sample predictive performance. Machine learning driven validation showed that the CR-Index can correctly predict districts with high incidents of COVID-19 cases and deaths more than 85% of the time. This proposed CR-Index is a simple, replicable, and easily interpretable tool that can help low-income countries prioritize resource mobilization to contain the disease spread and associated crisis management with global relevance and applicability. This index can also help to contain future pandemics (and epidemics) and manage their far-reaching adverse consequences.
    Date: 2023–02
  4. By: Muhammed Oguzhan Mete; Tahsin Yomralioglu
    Abstract: Geographic Information Systems (GIS) and Machine Learning methods are widely used in mass real estate valuation practices. Focusing on the physical attributes of properties, locational criteria are insufficiently used during the price prediction process. Whereas, locational criteria like proximity to important places, sea or forest views, flat topography are some of the spatial factors that extremely affect the real estate value. In this study, a hybrid approach is developed by integrating GIS and Machine Learning for automated mass valuation of residential properties in Turkey and the United Kingdom. GIS-based Nominal Valuation Method was applied to produce a land value map by carrying out proximity, terrain, and visibility analyses. Besides, ensemble regression methods like XGBoost, CatBoost, LightGBM, and Random Forest are built for price prediction. Spatial criteria scores obtained from GIS analyses were included in the price prediction data for feature enrichment purpose. Results showed that adding locational factors to the real estate price data increased the prediction accuracy dramatically. It also demonstrated that Random Forest was the most successful regression model compared to other ensemble methods.
    Keywords: GIS; Machine Learning; Mass Valuation; Real Estate Valuation
    JEL: R3
    Date: 2022–01–01
  5. By: Lily Davies (CPB Netherlands Bureau for Economic Policy Analysis); Mark Kattenberg (CPB Netherlands Bureau for Economic Policy Analysis); Benedikt Vogt (CPB Netherlands Bureau for Economic Policy Analysis)
    Abstract: Evaluations of support measures for companies often require a good assessment of the viability of firms or the probability that a firm will exit the market. On March 17, 2020, a lockdown and associated social-restriction measures were announced, which hit specific in the economy severely. To compensate companies and the self-employed for the loss of income, an extensive package of support measures has been designed. These support measures had hardly any restrictions, because they had to be paid out quickly. This raises the question whether unhealthy companies have made disproportionate use of support and to what extent these support measures have kept viable or non-viable companies afloat. In this paper, we use machine learning techniques to predict whether a company would have left the market in a world without corona. These predictions show that unhealthy companies applied for support less often than healthy companies. But we also show that the COVID-19 support has prevented most exits among unhealthy companies. This indicates that the corona support measures have had a negative impact on productivity growth.
    JEL: C18 E61 E65 G33
    Date: 2023–03
  6. By: Ahrens, Achim (Economic and Social Research Institute, Dublin); Hansen, Christian B. (University of Chicago); Schaffer, Mark E (Heriot-Watt University, Edinburgh); Wiemann, Thomas (University of Chicago)
    Abstract: We introduce the package ddml for Double/Debiased Machine Learning (DDML) in Stata. Estimators of causal parameters for five different econometric models are supported, allowing for flexible estimation of causal effects of endogenous variables in settings with unknown functional forms and/or many exogenous variables. ddml is compatible with many existing supervised machine learning programs in Stata. We recommend using DDML in combination with stacking estimation which combines multiple machine learners into a final predictor. We provide Monte Carlo evidence to support our recommendation.
    Keywords: st0001, causal inference, machine learning, doubly-robust estimation
    JEL: C14 C21 C87
    Date: 2023–02
  7. By: Yen-Jui Chang; Wei-Ting Wang; Hao-Yuan Chen; Shih-Wei Liao; Ching-Ray Chang
    Abstract: In recent years, there has been an emerging trend of combining two innovations in computer science and physics to achieve better computation capability. Exploring the potential of quantum computation to achieve highly efficient performance in various tasks is a vital development in engineering and a valuable question in sciences, as it has a significant potential to provide exponential speedups for technologically complex problems that are specifically advantageous to quantum computers. However, one key issue in unleashing this potential is constructing an efficient approach to load classical data into quantum states that can be executed by quantum computers or quantum simulators on classical hardware. Therefore, the split-step quantum walks (SSQW) algorithm was proposed to address this limitation. We facilitate SSQW to design parameterized quantum circuits (PQC) that can generate probability distributions and optimize the parameters to achieve the desired distribution using a variational solver. A practical example of implementing SSQW using Qiskit has been released as open-source software. Showing its potential as a promising method for generating desired probability amplitude distributions highlights the potential application of SSQW in option pricing through quantum simulation.
    Date: 2023–02
  8. By: Jaroslaw Morawski; Anetta Proskurovska
    Abstract: The introduction of blockchain-based solutions to the real estate transaction process has received significant attention in the recent years. However, there is still little clarity as to what the impact of such step might be on the dynamics of the real estate market. We build an agent-based model calibrated on the Swedish market and use it in a simulation analysis to study how redesigning of the house sale process with blockchain can affect housing price dynamics. First, the model looks at the traditional transaction setup using the sequence of interactions between parties as is typical in Sweden today. In the second step, we simulate a streamlined transactional workflow orchestrated by a hypothetical blockchain smart contract and compare its outcome with that of the traditional process. Our main goal is to investigate whether streamlining housing transactions via permissioned blockchain application exacerbates market volatility and generates flash crashes. At this point, the analysis is still ongoing and the results are outstanding.
    Keywords: blockchain; housing market; Simulation
    JEL: R3
    Date: 2022–01–01
  9. By: Jingyi Gu; Fadi P. Deek; Guiling Wang
    Abstract: Predicting the Stock movement attracts much attention from both industry and academia. Despite such significant efforts, the results remain unsatisfactory due to the inherently complicated nature of the stock market driven by factors including supply and demand, the state of the economy, the political climate, and even irrational human behavior. Recently, Generative Adversarial Networks (GAN) have been extended for time series data; however, robust methods are primarily for synthetic series generation, which fall short for appropriate stock prediction. This is because existing GANs for stock applications suffer from mode collapse and only consider one-step prediction, thus underutilizing the potential of GAN. Furthermore, merging news and market volatility are neglected in current GANs. To address these issues, we exploit expert domain knowledge in finance and, for the first time, attempt to formulate stock movement prediction into a Wasserstein GAN framework for multi-step prediction. We propose IndexGAN, which includes deliberate designs for the inherent characteristics of the stock market, leverages news context learning to thoroughly investigate textual information and develop an attentive seq2seq learning network that captures the temporal dependency among stock prices, news, and market sentiment. We also utilize the critic to approximate the Wasserstein distance between actual and predicted sequences and develop a rolling strategy for deployment that mitigates noise from the financial market. Extensive experiments are conducted on real-world broad-based indices, demonstrating the superior performance of our architecture over other state-of-the-art baselines, also validating all its contributing components.
    Date: 2023–02
  10. By: Popoola, Osuolale Peter; Adeboye, Olawale Nureni
    Abstract: Fourth Industrial Revolution is describes as exponential growth of several key technological fields’ concepts, such as intelligent materials, cloud computing, cyber-physical systems, data exchange, the Internet of things and blockchain technology. At its core, data represents a post-industrial opportunity. The effects of technologies have provided new avenues of data for official statistics, which can then be harnessed through the power of data science. However, as data continue to grow in size and complexity; new algorithms need to be developed so as to learn from diverse data sources. The limitation of conventional statistics in managing and analyzing big data has inspired data analysts to venture into data science. Data Science is a combination of multiple disciplines that use statistics, data analysis, and machine learning to analyze data, and extract knowledge and insights from it. These swathes of new digital data are valuable for official statistics. This paper links industrial eras to the evolution of statistics and data; it examines the emergence of big data and data science, what it means, it benefits and challenges for official statistics
    Keywords: Industrial Eras, Data Evolution, Big Data Revolution, Data Science, Official Statistics
    Date: 2023
  11. By: Konrad Lewszyk (University of Warsaw, Faculty of Economic Sciences and Data Science Lab WNE UW); Piotr Wójcik (University of Warsaw, Faculty of Economic Sciences and Data Science Lab WNE UW)
    Abstract: Attractive people obtain greater economic and reproductive success. This article attempts to grasp individual preferences of facial attractiveness and create reliable models that will accurately predict a beauty score on a binary and quintary scale. Based on extensive research conducted on factors of attractiveness, we derive the most important facial features that have the highest impact in beauty perception. Based on a sample of 681 images of faces using facial a landmark detector. We derive various numerical features represented by face characteristics and. The application of various machine learning algorithms shows that attractiveness can be predicted accurately based on facial characteristics. In addition, we show that indeed the attractiveness is subjective as the same features have different importance for different subjects.
    Keywords: Attractiveness, beauty-premium, image processing, machine learning, predictive models
    JEL: C40 C53 J71
    Date: 2023
  12. By: Frank, Johannes
    Abstract: This paper analyzes the performance of temporal fusion transformers in forecasting realized volatilities of stocks listed in the S&P 500 in volatile periods by comparing the predictions with those of state-of-the-art machine learning methods as well as GARCH models. The models are trained on weekly and monthly data based on three different feature sets using varying training approaches including pooling methods. I find that temporal fusion transformers show very good results in predicting financial volatility and outperform long short-term memory networks and random forests when using pooling methods. The use of sectoral pooling substantially improves the predictive performance of all machine learning approaches used. The results are robust to different ways of training the models.
    Keywords: Realized volatility, temporal fusion transformer, long short-term memory network, random forest
    JEL: C45 C53 C58 E44
    Date: 2023
  13. By: Andrew Green; Lucas Lamby
    Abstract: This report provides representative, cross-country estimates of the artificial intelligence (AI) workforce across OECD countries. The AI workforce is defined as the subset of workers with skills in statistics, computer science and machine learning who could actively develop and maintain AI systems. For countries that wish to be at the forefront of AI development, understanding the AI workforce is crucial to building and nurturing a talent pipeline, and ensuring that those who create AI reflect the diversity of society. This report uses data from online job vacancies to measure the within-occupation intensity of AI skill demand. The within-occupation AI intensity is then weighted to employment by occupation in labour force surveys to provide estimates of the size and growth of the AI workforce over time.
    Keywords: Artificial Intelligence
    JEL: J21 J23 J24 J31 J44
    Date: 2023–02–23
  14. By: Dutt, Satyajit; Radermacher, Jan W.
    Abstract: We investigate consumption patterns in Europe with supervised machine learning methods and reveal differences in age and wealth impact across countries. Using data from the third wave (2017) of the Eurosystem's Household Finance and Consumption Survey (HFCS), we assess how age and (liquid) wealth affect the marginal propensity to consume (MPC) in the Netherlands, Germany, France, and Italy. Our regression analysis takes the specification by Christelis et al. (2019) as a starting point. Decision trees are used to suggest alternative variable splits to create categorical variables for customized regression specifications. The results suggest an impact of differing wealth distributions and retirement systems across the studied Eurozone members and are relevant to European policy makers due to joint Eurozone monetary policy and increasing supranational fiscal authority of the EU. The analysis is further substantiated by a supervised machine learning analysis using a random forest and XGBoost algorithm.
    Date: 2023
  15. By: Harrison, Joseph S.; Josefy, Matthew A.; Kalm, Matias (Tilburg University, School of Economics and Management); Krause, Ryan
    Date: 2022
  16. By: Shuo Sun; Molei Qin; Xinrun Wang; Bo An
    Abstract: The financial markets, which involve more than $90 trillion in market capitalization, attract the attention of innumerable investors around the world. Recently, reinforcement learning in financial markets (FinRL) emerges as a promising direction to train agents for making profitable investment decisions. However, the evaluation of most FinRL methods only focus on profit-related measures, which are far from satisfactory for practitioners to deploy these methods into real-world financial markets. Therefore, we introduce PRUDEX-Compass, which has 6 axes, i.e., Profitability, Risk-control, Universality, Diversity, rEliability, and eXplainability, with a total of 17 measures for a systematic evaluation. Specifically, i) we propose AlphaMix+ as a strong FinRL baseline, which leverages Mixture-of-Experts (MoE) and risk-10 sensitive approaches to make diversified risk-aware investment decisions, ii) we11 evaluate 8 widely used FinRL methods in 4 long-term real-world datasets of influential financial markets to demonstrate the usage of our PRUDEX-Compass, iii) PRUDEX-Compass1 together with 4 real-world datasets, standard implementation of 8 FinRL methods and a portfolio management RL environment is released as public resources to facilitate the design and comparison of new FinRL methods. We hope that PRUDEX-Compass can shed light on future FinRL research to prevent untrustworthy results from stagnating FinRL into successful industry deployment.
    Date: 2023–01

This nep-cmp issue is ©2023 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.