nep-cmp New Economics Papers
on Computational Economics
Issue of 2023‒04‒24
fourteen papers chosen by

  1. Deep Calibration With Artificial Neural Network: A Performance Comparison on Option Pricing Models By Young Shin Kim; Hyangju Kim; Jaehyung Choi
  2. AI language models: Technological, socio-economic and policy considerations By OECD
  3. Stock Price Prediction Using Temporal Graph Model with Value Chain Data By Chang Liu; Sandra Paterlini
  4. A Macroscope of English Print Culture, 1530-1700, Applied to the Coevolution of Ideas on Religion, Science, and Institutions By Peter Grajzl; Peter Murrell
  5. Cryptocurrency Price Prediction using Twitter Sentiment Analysis By Haritha GB; Sahana N. B
  6. Artificial Intelligence and Dual Contract By Wuming Fu; Qian Qi
  7. Finding Regularized Competitive Equilibria of Heterogeneous Agent Macroeconomic Models with Reinforcement Learning By Ruitu Xu; Yifei Min; Tianhao Wang; Zhaoran Wang; Michael I. Jordan; Zhuoran Yang
  8. Accurate solution of the Index Tracking problem with a hybrid simulated annealing algorithm By \'Alvaro Rubio-Garc\'ia; Samuel Fern\'andez-Lorenzo; Juan Jos\'e Garc\'ia-Ripoll; Diego Porras
  9. Quantifying the Economic Effects of Land Reform Policy in South Africa: A Computable General Equilibrium Analysis By Khumbuzile C. Mosoma; Heinrich R. Bohlmann; Sifiso M. Ntombela; Renee van Eyden
  10. Quantifying SDG indicators for multiple SSPs up to 2050 with a focus on selected low and low-middle income countries and the bio-economy based on CGE analysis By Wilts, Rienne; Britz, Wolfgang
  11. Nutrition Indicators for CGE Models By Sands, Ronald; Beach, Robert
  12. Macroeconomic Impacts of Net Zero Pathway for Turkey By Dudu, Hasan; Beck, Hans Anand; Hallegatte, Stephane
  13. A Case for `Killer Robots': Why in the Long Run Martial AI May Be Good for Peace By Arandjelović, Ognjen
  14. Taureau: A Stock Market Movement Inference Framework Based on Twitter Sentiment Analysis By Nicholas Milikich; Joshua Johnson

  1. By: Young Shin Kim; Hyangju Kim; Jaehyung Choi
    Abstract: This paper explores Artificial Neural Network (ANN) as a model-free solution for a calibration algorithm of option pricing models. We construct ANNs to calibrate parameters for two well-known GARCH-type option pricing models: Duan's GARCH and the classical tempered stable GARCH that significantly improve upon the limitation of the Black-Scholes model but have suffered from computation complexity. To mitigate this technical difficulty, we train ANNs with a dataset generated by Monte Carlo Simulation (MCS) method and apply them to calibrate optimal parameters. The performance results indicate that the ANN approach consistently outperforms MCS and takes advantage of faster computation times once trained. The Greeks of options are also discussed.
    Date: 2023–03
  2. By: OECD
    Abstract: AI language models are a key component of natural language processing (NLP), a field of artificial intelligence (AI) focused on enabling computers to understand and generate human language. Language models and other NLP approaches involve developing algorithms and models that can process, analyse and generate natural language text or speech trained on vast amounts of data using techniques ranging from rule-based approaches to statistical models and deep learning. The application of language models is diverse and includes text completion, language translation, chatbots, virtual assistants and speech recognition. This report offers an overview of the AI language model and NLP landscape with current and emerging policy responses from around the world. It explores the basic building blocks of language models from a technical perspective using the OECD Framework for the Classification of AI Systems. The report also presents policy considerations through the lens of the OECD AI Principles.
    Date: 2023–04–17
  3. By: Chang Liu; Sandra Paterlini
    Abstract: Stock price prediction is a crucial element in financial trading as it allows traders to make informed decisions about buying, selling, and holding stocks. Accurate predictions of future stock prices can help traders optimize their trading strategies and maximize their profits. In this paper, we introduce a neural network-based stock return prediction method, the Long Short-Term Memory Graph Convolutional Neural Network (LSTM-GCN) model, which combines the Graph Convolutional Network (GCN) and Long Short-Term Memory (LSTM) Cells. Specifically, the GCN is used to capture complex topological structures and spatial dependence from value chain data, while the LSTM captures temporal dependence and dynamic changes in stock returns data. We evaluated the LSTM-GCN model on two datasets consisting of constituents of Eurostoxx 600 and S&P 500. Our experiments demonstrate that the LSTM-GCN model can capture additional information from value chain data that are not fully reflected in price data, and the predictions outperform baseline models on both datasets.
    Date: 2023–03
  4. By: Peter Grajzl; Peter Murrell
    Abstract: We combine unsupervised machine-learning and econometric methods to examine cultural change in 16th- and 17th-century England. A machine-learning digest synthesizes the content of 57, 863 texts comprising 83 million words into 110 topics. The topics include the expected, such as Natural Philosophy, and the unexpected, such as Baconian Theology. Using the data generated via machine-learning we then study facets of England's cultural history. Timelines suggest that religious and political discourse gradually became more scholarly over time and economic topics more prominent. The epistemology associated with Bacon was present in theological debates already in the 16th century. Estimating a VAR, we explore the coevolution of ideas on religion, science, and institutions. Innovations in religious ideas induced strong responses in the other two domains. Revolutions did not spur debates on institutions nor did the founding of the Royal Society markedly elevate attention to science.
    Keywords: cultural history, England, machine-learning, text-as-data, coevolution, VAR
    JEL: C80 Z10 N00 P10 C30
    Date: 2023
  5. By: Haritha GB; Sahana N. B
    Abstract: The cryptocurrency ecosystem has been the centre of discussion on many social media platforms, following its noted volatility and varied opinions. Twitter is rapidly being utilised as a news source and a medium for bitcoin discussion. Our algorithm seeks to use historical prices and sentiment of tweets to forecast the price of Bitcoin. In this study, we develop an end-to-end model that can forecast the sentiment of a set of tweets (using a Bidirectional Encoder Representations from Transformers - based Neural Network Model) and forecast the price of Bitcoin (using Gated Recurrent Unit) using the predicted sentiment and other metrics like historical cryptocurrency price data, tweet volume, a user's following, and whether or not a user is verified. The sentiment prediction gave a Mean Absolute Percentage Error of 9.45%, an average of real-time data, and test data. The mean absolute percent error for the price prediction was 3.6%.
    Date: 2023–03
  6. By: Wuming Fu; Qian Qi
    Abstract: With the dramatic progress of artificial intelligence algorithms in recent times, it is hoped that algorithms will soon supplant human decision-makers in various fields, such as contract design. We analyze the possible consequences by experimentally studying the behavior of algorithms powered by Artificial Intelligence (Multi-agent Q-learning) in a workhorse \emph{dual contract} model for dual-principal-agent problems. We find that the AI algorithms autonomously learn to design incentive-compatible contracts without external guidance or communication among themselves. We emphasize that the principal, powered by distinct AI algorithms, can play mixed-sum behavior such as collusion and competition. We find that the more intelligent principals tend to become cooperative, and the less intelligent principals are endogenizing myopia and tend to become competitive. Under the optimal contract, the lower contract incentive to the agent is sustained by collusive strategies between the principals. This finding is robust to principal heterogeneity, changes in the number of players involved in the contract, and various forms of uncertainty.
    Date: 2023–03
  7. By: Ruitu Xu; Yifei Min; Tianhao Wang; Zhaoran Wang; Michael I. Jordan; Zhuoran Yang
    Abstract: We study a heterogeneous agent macroeconomic model with an infinite number of households and firms competing in a labor market. Each household earns income and engages in consumption at each time step while aiming to maximize a concave utility subject to the underlying market conditions. The households aim to find the optimal saving strategy that maximizes their discounted cumulative utility given the market condition, while the firms determine the market conditions through maximizing corporate profit based on the household population behavior. The model captures a wide range of applications in macroeconomic studies, and we propose a data-driven reinforcement learning framework that finds the regularized competitive equilibrium of the model. The proposed algorithm enjoys theoretical guarantees in converging to the equilibrium of the market at a sub-linear rate.
    Date: 2023–02
  8. By: \'Alvaro Rubio-Garc\'ia; Samuel Fern\'andez-Lorenzo; Juan Jos\'e Garc\'ia-Ripoll; Diego Porras
    Abstract: An actively managed portfolio almost never beats the market in the long term. Thus, many investors often resort to passively managed portfolios whose aim is to follow a certain financial index. The task of building such passive portfolios aiming also to minimize the transaction costs is called Index Tracking (IT), where the goal is to track the index by holding only a small subset of assets in the index. As such, it is an NP-hard problem and becomes unfeasible to solve exactly for indices with more than 100 assets. In this work, we present a novel hybrid simulated annealing method that can efficiently solve the IT problem for large indices and is flexible enough to adapt to financially relevant constraints. By tracking the S&P-500 index between the years 2011 and 2018 we show that our algorithm is capable of finding optimal solutions in the in-sample period of past returns and can be tuned to provide optimal returns in the out-of-sample period of future returns. Finally, we focus on the task of holding an IT portfolio during one year and rebalancing the portfolio every month. Here, our hybrid simulated annealing algorithm is capable of producing financially optimal portfolios already for small subsets of assets and using reasonable computational resources, making it an appropriate tool for financial managers.
    Date: 2023–03
  9. By: Khumbuzile C. Mosoma (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa); Heinrich R. Bohlmann (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa); Sifiso M. Ntombela (National Agricultural Marketing Council); Renee van Eyden (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa)
    Abstract: Although South Africa has implemented several land reform policies and farmer development support programmes, little progress has been achieved in bridging the inequality gap between mainly black emerging and mainly white established farmers. This study seeks to quantify the effects of land reform policy in South Africa using a dynamic computable general equilibrium (CGE) model. The study simulates two policy scenarios. The first policy scenario assumes that the current land reform approach will continue where government transfers land, but without transitional farmer support to improve the productivity of new farmers. The second policy scenario assumes that the state will additionally allocate transitional farmer support to new farmers, including those operating on land-reform farms and in the former homeland areas. The results reveal that the effects of land reform policy are minimal but positive at the aggregate economic level across the two scenarios. Achieving a land reform target of 30 percent will benefit the real GDP by R242.4 million under scenario 1 and R608.6 million under scenario 2 by the end of the simulation period. There is also a positive effect on selected macroeconomic indicators such as imports, employment, and investment, notably when comprehensive support services are provided. Primary industries like field crops, horticulture, and livestock experience significant output gains. Similar to industrial output, exports for the primary agricultural industries are impacted positively in the long term as new land is made available, making more output available for the export market. Although the implementation of land reform might be a costly exercise initially, it can be achieved at a lower cost than what is assumed or expected. Simulation results suggest that land redistribution will not harm the economy if accompanied by comprehensive farmer support. The study recommends that the government and the private sector work together to create a just and inclusive agricultural landscape.
    Keywords: Productivity, Land reform policy, Computable General Equilibrium modelling
    JEL: Q15 C68
    Date: 2023–03
  10. By: Wilts, Rienne; Britz, Wolfgang
    Abstract: A wide range of indicators beyond GDP growth is necessary to measure progress towards more sustainability as reflected by the indicator frameworks developed by the United Nations (2021). Still, such progress builds on its core on economic growth and related structural change. Given its multi-sector and global perspective, dynamic CGE analysis depicts these key processes and thus offers a starting point to quantify various SDG indicators. Multiple scholars have therefore developed SDG indicator frameworks which fit their CGE models, such as Philippidis et al. (2020) and Lui et al. (2021). Existing auxiliary data available from GTAP, such as CO2 (Peters, 2016), non-CO2 (Chepeliev, 2020a) and air emissions (Chepeliev, 2020b) already help to access important aspects of environmental sustainability and to relate emissions to human health. Further indicators require partly sector and product detail beyond the GTAP Data Base which motivates the development of more detail data base in this study. Distributional aspects of economic growth, also beyond income distribution, remain a challenge in CGE analysis, and are addressed in this study by micro-simulations. We propose to quantify 75 indicators relating to 13 of the 17 SDGs in order to assess SDG developments up to 2050 for different Socio-Economic Pathways to extend existing work in this field.
    Keywords: International Relations/Trade, Research Methods/ Statistical Methods
    Date: 2022
  11. By: Sands, Ronald; Beach, Robert
    Abstract: Computable general equilibrium (CGE) models have proven useful for simulating future economic activity and environmental indicators, especially in response to global drivers such as population, income, technology, and dietary preference. The focus of this paper is to show how output from CGE models can also be converted to nutritional indicators such as calories, carbohydrates, protein, fats, and micro-nutrients. This paper covers post-simulation analysis of food demand, rather than how to specify food demand within a general equilibrium model. There are strong links between the specification of food demand in a model, and how that is calibrated, to the realism possible for reporting calories and other nutritional indicators. It turns out that modification to the underlying social accounting matrix (SAM) can improve the realism of projections of food demand, by increasing the consistency between monetary units in the SAM and physical units (metric tons) in food balance sheets such as those published by the Food and Agricultural Organization (FAO) of the United Nations. If model output by food commodity can be expressed by weight (e.g., consumption in terms of grams per person per day), then food conversion tables can be applied to obtain a comprehensive list of nutrient consumption, including macro- and micro-nutrients. This information can be summarized in a variety of nutritional indicators. We cover two key steps: (1) pre-processing of the SAM and food balance sheets; and (2) post-processing of CGE model output.
    Keywords: Food Security and Poverty
    Date: 2022
  12. By: Dudu, Hasan; Beck, Hans Anand; Hallegatte, Stephane
    Abstract: In this paper we are analyzing the impacts of reaching to net zero by 2053 on Turkey’s economy. We use a CGE model that is calibrated to 2018 Social Accounting Matrix of Turkey. Our scenarios incorporate the results of sectoral analysis from Turkey Country Climate and Development Reports published by the World Bank (2022). We take the results of land use change, energy, transport, and buildings sectors and translate them into shocks in the CGE model. Our results suggest that high levels of electrification of buildings and transport are likely to pose challenges for the net zero pathway of Turkey, although the energy efficiency gains thanks to the mitigation policies are likely to compensate the adverse effects of increasing electricity prices in the short to medium term. Hence Turkey needs to revise the energy sector policies to increase the production capacity of renewables further to ease the transition to a net zero economy. Mitigation policies are progressive in the sense that they do not harm poorer households as much as richer households but still the lower income groups would need to be compensated especially in the early years of the transition. Increase in government revenues thanks to a carbon tax and removing subsidies on fossil fuels would create enough fiscal space for social protection programs required for a just transition. Last, a well-managed transition to a net zero economy offers significant growth benefits for Turkey.
    Keywords: Environmental Economics and Policy
    Date: 2022
  13. By: Arandjelović, Ognjen
    Abstract: Purpose: The remarkable increase of sophistication of artificial intelligence in recent years has already led to its widespread use in martial applications, the potential of so-called ‘killer robots’ ceasing to be a subject of fiction. Approach: Virtually without exception, this potential has generated fear, as evidenced by a mounting number of academic articles calling for the ban on the development and deployment of lethal autonomous robots (LARs). In the present paper I start with an analysis of the existing ethical objections to LARs. Findings: My analysis shows the contemporary thought to be deficient in philosophical rigour, these deficiencies leading to an alternative thesis. Value: I advance a thesis that LARs can in fact be a force for peace, leading to fewer and less deadly wars.
    Date: 2023–03–23
  14. By: Nicholas Milikich; Joshua Johnson
    Abstract: With the advent of fast-paced information dissemination and retrieval, it has become inherently important to resort to automated means of predicting stock market prices. In this paper, we propose Taureau, a framework that leverages Twitter sentiment analysis for predicting stock market movement. The aim of our research is to determine whether Twitter, which is assumed to be representative of the general public, can give insight into the public perception of a particular company and has any correlation to that company's stock price movement. We intend to utilize this correlation to predict stock price movement. We first utilize Tweepy and getOldTweets to obtain historical tweets indicating public opinions for a set of top companies during periods of major events. We filter and label the tweets using standard programming libraries. We then vectorize and generate word embedding from the obtained tweets. Afterward, we leverage TextBlob, a state-of-the-art sentiment analytics engine, to assess and quantify the users' moods based on the tweets. Next, we correlate the temporal dimensions of the obtained sentiment scores with monthly stock price movement data. Finally, we design and evaluate a predictive model to forecast stock price movement from lagged sentiment scores. We evaluate our framework using actual stock price movement data to assess its ability to predict movement direction.
    Date: 2023–03

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.