|
on Computational Economics |
|
Issue of 2020‒07‒13
seventeen papers chosen by |
| By: | Bartram, Söhnke M; Branke, Jürgen; Motahari, Mehrshad |
| Abstract: | Artificial intelligence (AI) has a growing presence in asset management and has revolutionized the sector in many ways. It has improved portfolio management, trading, and risk management practices by increasing efficiency, accuracy, and compliance. In particular, AI techniques help construct portfolios based on more accurate risk and returns forecasts and under more complex constraints. Trading algorithms utilize AI to devise novel trading signals and execute trades with lower transaction costs, and AI improves risk modelling and forecasting by generating insights from new sources of data. Finally, robo-advisors owe a large part of their success to AI techniques. At the same time, the use of AI can create new risks and challenges, for instance as a result of model opacity, complexity, and reliance on data integrity. |
| Keywords: | Algorithmic trading; decision trees; deep learning; evolutionary algorithms; Lasso; Machine Learning; neural networks; NLP; random forests; SVM |
| JEL: | G11 G17 |
| Date: | 2020–03 |
| URL: | http://d.repec.org/n?u=RePEc:cpr:ceprdp:14525&r=all |
| By: | S M Raju; Ali Mohammad Tarif |
| Abstract: | Bitcoin is the first digital decentralized cryptocurrency that has shown a significant increase in market capitalization in recent years. The objective of this paper is to determine the predictable price direction of Bitcoin in USD by machine learning techniques and sentiment analysis. Twitter and Reddit have attracted a great deal of attention from researchers to study public sentiment. We have applied sentiment analysis and supervised machine learning principles to the extracted tweets from Twitter and Reddit posts, and we analyze the correlation between bitcoin price movements and sentiments in tweets. We explored several algorithms of machine learning using supervised learning to develop a prediction model and provide informative analysis of future market prices. Due to the difficulty of evaluating the exact nature of a Time Series(ARIMA) model, it is often very difficult to produce appropriate forecasts. Then we continue to implement Recurrent Neural Networks (RNN) with long short-term memory cells (LSTM). Thus, we analyzed the time series model prediction of bitcoin prices with greater efficiency using long short-term memory (LSTM) techniques and compared the predictability of bitcoin price and sentiment analysis of bitcoin tweets to the standard method (ARIMA). The RMSE (Root-mean-square error) of LSTM are 198.448 (single feature) and 197.515 (multi-feature) whereas the ARIMA model RMSE is 209.263 which shows that LSTM with multi feature shows the more accurate result. |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2006.14473&r=all |
| By: | Daniel J. Egger; Claudio Gambella; Jakub Marecek; Scott McFaddin; Martin Mevissen; Rudy Raymond; Andrea Simonetto; Stefan Woerner; Elena Yndurain |
| Abstract: | This paper outlines our point of view regarding the applicability, state of the art, and potential of quantum computing for problems in finance. We provide an introduction to quantum computing as well as a survey on problem classes in finance that are computationally challenging classically and for which quantum computing algorithms are promising. In the main part, we describe in detail quantum algorithms for specific applications arising in financial services, such as those involving simulation, optimization, and machine learning problems. In addition, we include demonstrations of quantum algorithms on IBM Quantum back-ends and discuss the potential benefits of quantum algorithms for problems in financial services. We conclude with a summary of technical challenges and future prospects. |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2006.14510&r=all |
| By: | Heni Boubaker (International University of Rabat, BEAR LAB, Technopolis Rabat-Shore Rocade Rabat-Sale, Morocco); Giorgio Canarella (Department of Economics, Lee Business School, University of Nevada, Las Vegas; Las Vegas, Nevada); Rangan Gupta (Department of Economics, University of Pretoria, Pretoria, 0002, South Africa); Stephen M. Miller (Department of Economics, Lee Business School, University of Nevada, Las Vegas; Las Vegas, Nevada) |
| Abstract: | This paper proposes a hybrid modelling approach for forecasting returns and volatilities of the stock market. The model, called ARFIMA-WLLWNN model, integrates the advantages of the ARFIMA model, the wavelet decomposition technique (namely, the discrete MODWT with Daubechies least asymmetric wavelet filter) and artificial neural network (namely, the LLWNN neural network). The model develops through a two-phase approach. In phase one, a wavelet decomposition improves the forecasting accuracy of the LLWNN neural network, resulting in the Wavelet Local Linear Wavelet Neural Network (WLLWNN) model. The Back Propagation (BP) and Particle Swarm Optimization (PSO) learning algorithms optimize the WLLWNN structure. In phase two, the residuals of an ARFIMA model of the conditional mean become the input to the WLLWNN model. The hybrid ARFIMA-WLLWNN model is evaluated using daily closing prices for the Dow Jones Industrial Average (DJIA) index over 01/01/2010 to 02/11/2020. The experimental results indicate that the PSO-optimized version of the hybrid ARFIMA-WLLWNN outperforms the LLWNN, WLLWNN, ARFIMA-LLWNN, and the ARFIMA-HYAPARCH models and provides more accurate out-of-sample forecasts over validation horizons of one, five and twenty-two days. |
| Keywords: | Wavelet decomposition, WLLWNN, Neural network, ARFIMA, HYGARCH |
| JEL: | C45 C58 G17 |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:pre:wpaper:202056&r=all |
| By: | Hans B\"uhler; Blanka Horvath; Terry Lyons; Imanol Perez Arribas; Ben Wood |
| Abstract: | Neural network based data-driven market simulation unveils a new and flexible way of modelling financial time series without imposing assumptions on the underlying stochastic dynamics. Though in this sense generative market simulation is model-free, the concrete modelling choices are nevertheless decisive for the features of the simulated paths. We give a brief overview of currently used generative modelling approaches and performance evaluation metrics for financial time series, and address some of the challenges to achieve good results in the latter. We also contrast some classical approaches of market simulation with simulation based on generative modelling and highlight some advantages and pitfalls of the new approach. While most generative models tend to rely on large amounts of training data, we present here a generative model that works reliably in environments where the amount of available training data is notoriously small. Furthermore, we show how a rough paths perspective combined with a parsimonious Variational Autoencoder framework provides a powerful way for encoding and evaluating financial time series in such environments where available training data is scarce. Finally, we also propose a suitable performance evaluation metric for financial time series and discuss some connections of our Market Generator to deep hedging. |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2006.14498&r=all |
| By: | Akash Doshi; Alexander Issa; Puneet Sachdeva; Sina Rafati; Somnath Rakshit |
| Abstract: | Forecasting stock prices can be interpreted as a time series prediction problem, for which Long Short Term Memory (LSTM) neural networks are often used due to their architecture specifically built to solve such problems. In this paper, we consider the design of a trading strategy that performs portfolio optimization using the LSTM stock price prediction for four different companies. We then customize the loss function used to train the LSTM to increase the profit earned. Moreover, we propose a data driven approach for optimal selection of window length and multi-step prediction length, and consider the addition of analyst calls as technical indicators to a multi-stack Bidirectional LSTM strengthened by the addition of Attention units. We find the LSTM model with the customized loss function to have an improved performance in the training bot over a regressive baseline such as ARIMA, while the addition of analyst call does improve the performance for certain datasets. |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2006.04992&r=all |
| By: | Mueller, H.; Rauh, C. |
| Abstract: | There is a growing interest in prevention in several policy areas and this provides a strong motivation for an improved integration of machine learning into models of decision making. In this article we propose a framework to tackle conflict prevention. A key problem of conflict forecasting for prevention is that predicting the start of conflict in previously peaceful countries needs to overcome a low baseline risk. To make progress in this hard problem this project combines unsupervised with supervised machine learning. Specifically, the latent Dirichlet allocation (LDA) model is used for feature extraction from 4.1 million newspaper articles and these features are then used in a random forest model to predict conflict. The output of the forecast model is then analyzed in a framework of cost minimization in which excessive intervention costs due to false positives can be traded off against the damages and destruction caused by conflict. News text is able provide a useful forecast for the hard problem even when evaluated in such a cost-benefit framework. The aggregation into topics allows the forecast to rely on subtle signals from news which are positively or negatively related to conflict risk. |
| Keywords: | Conflict prediction, Conflict trap, Topic models, LDA, Random forest, News text, Machine learning |
| JEL: | F51 C53 |
| Date: | 2020–03–10 |
| URL: | http://d.repec.org/n?u=RePEc:cam:camdae:2015&r=all |
| By: | Nawaf Almaskati (University of Waikato); Ron Bird (University of Waikato); Yue Lu (University of Waikato); Danny Leung (University of Technology Sydney) |
| Abstract: | In a sample covering bankruptcies in public US firms in the period 2000 to 2015, we find that the addition of governance variables significantly improves the classification power and prediction accuracy of various bankruptcy prediction models. We also find that while adding governance variables improves the performance of bankruptcy prediction models, the additional explanatory power provided by adding the governance measures improves the further we are from bankruptcy, which implies that governance variables tend to provide earlier and more accurate warnings of the firm’s bankruptcy potential. Our analysis of five commonly used statistical methods in the literature showed that regardless of the bankruptcy model used, hazard analysis provides the best classification and out-of-sample forecast accuracy among the parametric methods. Nevertheless, non-parametric methods such as neural networks or data envelopment analysis appear to provide better classification accuracy regardless of the model selected. |
| Keywords: | corporate governance; bankruptcy studies; default prediction; non-parametric methods |
| JEL: | D81 G10 G14 G30 G32 |
| Date: | 2019–07–31 |
| URL: | http://d.repec.org/n?u=RePEc:wai:econwp:19/16&r=all |
| By: | Twahir Khalfan; Elineema Kisanga; Vincent Leyaro; Faith Masekesa; Michael Noble; Gemma Wright |
| Abstract: | The provision of a universal old age pension is increasingly recognized as an important instrument for strengthening and extending social protection. A growing number of emerging economies, including East African countries, are introducing universal old age pensions to guarantee at least a basic level of social security. However, such a benefit has not been established in Tanzania Mainland, and a lack of adequate financing is viewed as one of the main constraints. |
| Keywords: | microsimulation, Old age, public pensions, Pensions, Social security, Tanzania |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:unu:wpaper:wp-2020-81&r=all |
| By: | Ariel Neufeld; Antonis Papapantoleon; Qikun Xiang |
| Abstract: | We consider derivatives written on multiple underlyings in a one-period financial market, and we are interested in the computation of model-free upper and lower bounds for their arbitrage-free prices. We work in a completely realistic setting, in that we only assume the knowledge of traded prices for other single- and multi-asset derivatives, and even allow for the presence of bid-ask spread in these prices. We provide a fundamental theorem of asset pricing for this market model, as well as a superhedging duality result, that allows to transform the abstract maximization problem over probability measures into a more tractable minimization problem over vectors, subject to certain constraints. Then, we recast this problem into a linear semi-infinite optimization problem, and provide two algorithms for its solution. These algorithms provide upper and lower bounds for the prices that are $\varepsilon$-optimal, as well as a characterization of the optimal pricing measures. Moreover, these algorithms are efficient and allow the computation of bounds in high-dimensional scenarios (e.g. when $d=60$). Numerical experiments using synthetic data showcase the efficiency of these algorithms, while they also allow to understand the reduction of model-risk by including additional information, in the form of known derivative prices. |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2006.14288&r=all |
| By: | Amewu, Sena; Asante, Seth; Pauw, Karl; Thurlow, James |
| Abstract: | The objective in this paper is to estimate the economic costs of COVID-19 policies and external shocks in a developing country context, with a focus on agri-food system impacts. Ghana is selected as a case study. Ghana recorded its first two cases of COVID-19 infection on 12 March 2020. The government responded by gradually introducing social distancing measures, travel restrictions, border closures, and eventually a partial, two-week “partial” lockdown in the country’s largest metropolitan areas of Accra and Kumasi. Social distancing measures have been enforced nationwide and include bans on conferences, workshops, and sporting and religious events, as well as the closure of bars and nightclubs. All educational institutions are also closed. The partial lockdown measures in urban areas directed all residents to remain home except for essential business, prohibited non-essential inter-city travel and transport, and only essential manufacturing and services operations were permitted to continue (The Presidency 2020). At the time the lockdown was announced, Ghana’s Ministry of Finance revised its GDP growth estimate for 2020 downwards from 6.8 to 1.5 percent (MoF 2020), although the Minister warned that growth could fall further if lockdown measures were extended. The lockdown was initially extended for a third week but was officially lifted on 20 April. Social distancing measures remain in place nationwide, although a gradual easing of restrictions commenced in June. Ghana’s borders remain closed at the time of writing. |
| Keywords: | GHANA; WEST AFRICA; AFRICA SOUTH OF SAHARA; AFRICA; Coronavirus; coronavirus disease; Coronavirinae; economic impact; agrifood systems; models; pandemics; recovery; policies; Covid-19; Social Accounting Matrix (SAM); lockdown; Covid-19 policy responses; economic cost |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:fpr:gsspwp:52&r=all |
| By: | Castillo, Marco (Texas A&M University); Petrie, Ragan (Texas A&M University) |
| Abstract: | We examine optimal incentives for charitable giving with a large-scale field experiment involving 26 charities and over 112,000 unique individuals. The price of giving is varied by offering a fixed match if the donation meets a threshold amount (e.g. "give at least $25 and the charity receives a $25 match"). Responses are used to structurally estimate a model of charitable giving. The model estimates are employed to evaluate the effectiveness of various counterfactual match incentive schemes, taking into account the goals of the charity and donor preferences. Two of these optimal incentives were subsequently implemented in a follow-up field study. They were found to be effective at implementing the desired goals, as predicted by theory and our simulations. Our findings highlight the pitfalls of relying on a particular parameterization of a policy to evaluate effectiveness. The best-guess incentives in our initial field experiment turned out to be ineffective at increasing donations because optimal incentives should have been set higher. |
| Keywords: | charitable giving, mechanism design, field experiment |
| JEL: | D64 H41 C93 D91 |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:iza:izadps:dp13321&r=all |
| By: | Dupuy, Arnaud (University of Luxembourg); Galichon, Alfred (New York University); Jaffe, Sonia (University of Chicago); Kominers, Scott Duke (Harvard University) |
| Abstract: | We analyze the effects of taxation in two-sided matching markets where agents have heterogeneous preferences over potential partners. Our model provides a continuous link between models of matching with and without transfers. Taxes generate inefficiency on the allocative margin, by changing who matches with whom. This allocative inefficiency can be non-monotonic, but is weakly increasing in the tax rate under linear taxation if each worker has negative non-pecuniary utility of working. We adapt existing econometric methods for markets without taxes to our setting, and estimate preferences in the college-coach football market. We show through simulations that standard methods inaccurately measure deadweight loss. |
| Keywords: | matching, taxation |
| JEL: | C78 D3 H2 J3 |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:iza:izadps:dp13328&r=all |
| By: | Yuta Inoue (Faculty of Political Science and Economics, Waseda University); Koji Shirai (School of Economics, Kwansei Gakuin University) |
| Abstract: | This paper develops revealed preference analysis for limited consideration models. A revealed preference test is given for the decision model obeying two well-established hypotheses on a decision maker fs consideration: the attention filter property and competition filter property. We also provide a test for the two-step decision model called the (transitive) rational shortlist method. As an application, we conducted a simulation to compare the relative strength of observable restrictions across leading models, in addition to an experiment to compare models in terms of Selten fs index, which is a measure for plausibility of a model in explaining a given data set. |
| Keywords: | ; Revealed preference; Limited consideration; Limited attention; Rational short- listing; Bronars f test; Selten fs index |
| JEL: | C6 D1 D9 |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:kgu:wpaper:217&r=all |
| By: | Julien Blasco (THEMA - Théorie économique, modélisation et applications - CNRS - Centre National de la Recherche Scientifique - CY - CY Cergy Paris Université, Institut national de la statistique et des études économiques (INSEE), LIEPP - Laboratoire interdisciplinaire d'évaluation des politiques publiques [Sciences Po] - Sciences Po - Sciences Po); Elvire Guillaud (CES - Centre d'économie de la Sorbonne - CNRS - Centre National de la Recherche Scientifique - UP1 - Université Panthéon-Sorbonne, LIEPP - Laboratoire interdisciplinaire d'évaluation des politiques publiques [Sciences Po] - Sciences Po - Sciences Po); Michaël Zemmour (CES - Centre d'économie de la Sorbonne - CNRS - Centre National de la Recherche Scientifique - UP1 - Université Panthéon-Sorbonne, LIEPP - Laboratoire interdisciplinaire d'évaluation des politiques publiques [Sciences Po] - Sciences Po - Sciences Po) |
| Abstract: | Consumption taxes are often considered as the most anti-redistributive componentof the tax system. Yet, very few estimates, and fewer international comparisons of theredistributive impact of consumption taxes exist in the literature, due to scarce dataon household expenditures. We use household budget surveys and microsimulation toprovide consistent estimates of the regressivity of consumption taxes for a large panelof countries and years. We propose a new method for imputing consumption expen-diture across countries, using widely available data on income and socio-demographiccharacteristics of households. We show that including the distribution of housing rents,when data is available, to impute households' consumption greatly improves the pre-diction of the model. Our results are threefold. First, there is a 1 to 2 ratio betweenthe propensity to consume of the top decile (around 50% of their income) and thatof the bottom decile (100% of income). Second, consumption taxes entail a signifi-cant rise in the Gini coefficient of income (between 0.01 and 0.04 point), yet of muchsmaller magnitude than the positive redistribution operated by direct taxes and trans-fers. Third, cross-country differences in the distributive effect of consumption taxes aremainly explained by variations in the tax rate (from 7 to 24% in our sample), ratherthan variations in the distribution of consumption, since everywhere the propensity toconsume declines sharply with income. |
| Keywords: | Indirect taxes,Redistributive Effect,Consumption,Income,Microsimulation,Luxembourg Income Study |
| Date: | 2020–02 |
| URL: | http://d.repec.org/n?u=RePEc:hal:wpaper:hal-02735145&r=all |
| By: | Mandler, Martin; Scharnagl, Michael |
| Abstract: | We assess the macroeconomic effects of the Eurosystem's asset purchases on the four largest euro area economies using simulation exercises that combine unconventional monetary policy shocks with a fixed policy rate for the duration of the purchase programme. We identify unconventional monetary policy shocks in a large Bayesian vector autoregressive (BVAR) model as shocks to the term structure of interest rates using zero and sign restrictions. We propose a multi-country model in which we impose identification assumptions mainly on euro area aggregate financial variables and on country averages of output and price responses. Furthermore, the multi-country structure allows testing for cross-country differences in the effects of the asset purchase programme in a statistically rigorous way using the posterior of the difference between the country-specific effects. We estimate positive output effects in all countries as well as positive effects on bank lending to firms. Effects on HICP inflation, generally, are much weaker. We find substantial cross-country heterogeneity with the largest price level effects in Spain while output effects were smallest in France and inflation effects were smallest in Italy. |
| Keywords: | asset purchase programme,unconventional monetary policy,euro area,Bayesian vector autoregression,regional effects of monetary policy |
| JEL: | C32 E47 E52 E58 |
| Date: | 2020 |
| URL: | http://d.repec.org/n?u=RePEc:zbw:bubdps:292020&r=all |
| By: | Michael Christl (European Commission - JRC); Silvia De Poli (European Commission - JRC) |
| Abstract: | Financial incentives affect the labour supply decisions of households, but typically the impact of such incentives varies significantly across household types. While there is a substantial literature on the labour supply effects of tax reforms and in-work benefits, the impact of changes in social assistance benefits has received less attention. This paper analyses the impact of the Austrian reform proposal 'Neue Sozialhilfe' ("New Social Assistance"), which was introduced in 2019 and substantially cut social assistance benefits for migrants and families with children. We show that the labour supply effects of these changes in social assistance differ substantially across household types. While women exhibit higher labour supply elasticities in our estimates, the overall effects of the reform are especially strong for men and migrants. Couples with children and migrants, i.e. the groups which were hit the hardest by the reform's social assistance reductions, show the strongest labour supply reactions to the 'New Social Assistance'. Furthermore, we show that overall the reform has a positive, but small, effect on the intensive margin of labour supply. |
| Keywords: | social assistance, reform, labour supply, discrete choice, microsimulation, EUROMOD |
| JEL: | J08 H31 H53 |
| Date: | 2020–06 |
| URL: | http://d.repec.org/n?u=RePEc:ipt:taxref:202005&r=all |