Computational Economics最新文献

This study examines the impact of investor sentiment on stock price crash risk from the perspective of news photo sentiment. First, the paper derives investor sentiment from news photos based on deep learning models. Second, we develop regression models analyzing the relationship between investor sentiment and stock price crash risk. The empirical analysis results show that news photo sentiment has a significantly positive effect on stock price crash risk and exhibits a stronger predictive power than sentiment embedded in news text. In addition, our study shows that positive news photo sentiment has a stronger impact on stock price crash risk in bull markets than in bearish markets. Our findings have great implications for investors, market analysts, and policy makers.

This paper investigates whether cryptocurrency exchanges exhibit greater liquidity than traditional financial markets. Utilizing four different liquidity measures, we evaluate the liquidity of six leading cryptocurrency exchanges and nine traditional small-cap stock indices across diverse geographies and rank the markets according to their liquidities. We investigate the Pre-Pandemic, First and Second-wave COVID-19, and post-pandemic economic periods. Multi-Criteria Decision Analysis, employing Borda and Keener Ranking techniques, is used to validate the robustness of our liquidity rankings. Our findings reveal that the Russel 2000 Small Cap is the most liquid among traditional markets, while Binance is the most liquid cryptocurrency exchange. Results show that Small-cap indices are generally more liquid than cryptocurrency exchanges. However, during the second wave of the COVID-19 pandemic, individual and institutional investors used cryptocurrencies as a safe haven, with Binance exhibiting better liquidity than traditional markets such as Nifty SC 100. In the post-pandemic period, cryptocurrency market liquidity significantly deteriorated compared to pre-pandemic levels. We argue that despite investors using cryptocurrencies as diversification tools during economic stress periods, cryptocurrencies fail to serve as a dependable asset allocation tool compared to small-cap equities. With contributions encompassing a pre and post-pandemic liquidity assessment, the development of a multifaceted liquidity framework utilizing Multi-Criteria Decision Analysis, and liquidity comparisons between traditional and cryptocurrency markets, this study delivers substantive enhancements to the analysis and understanding of global market liquidity for traders and researchers.

The consideration of the study on dynamic cluster flows in international tourists is an aspect that has been scarcely addressed in research despite its importance in economic development. Dynamic Time Warping is the methodology applied to identify alignments of common patterns in hotel demand time series within applied economics. The automatic determination of the number of clusters proposes an optimal number of groups for tourist destinations, and this proposition is confirmed through internal validation. Similarities among time series, including identifying outliers through boxplots, have been identified through the applied methodology. It has been employed for the primary tourist destinations in Spain for 106 international hotel demand time series. The effects of COVID-19 on the tourism sector and temporal similarities have been observed through clustering. The results that have been obtained reveal international tourist market flows that go beyond traditional analyses of seasonality or climatic factors, thus constituting a valuable tool for economic analysis in both direct and indirect markets.

In this paper, we raise a new method for numerically solving the partial differential equations (PDEs) of the Black-Scholes and Heston models, which play an important role in financial option pricing theory. Our proposed method is based on the rational spectral collocation method and the contour integral method. The presence of discontinuities in the first-order derivative of the initial condition of the PDEs prevents the spectral method from achieving high accuracy. However, the rational spectral method excels in overcoming this drawback. So we discretize the spatial variables of PDEs by rational spectral method, which yields a system of ordinary differential equations. Then we solve it by the numerical inverse Laplace transform using contour integral method. It is very important to select an appropriate parameters in the contour integral method, we revise the optimal parameters proposed by Trefethen and Weideman (Math Comput 76(259):1341–1356, 2007) in hyperbolic contour to control the effect of roundoff error. During solving the independent shifted linear systems, preconditioned Krylov subspace iteration is used to improve computational efficiency. We also compare the numerical results obtained from our proposed method with those obtained from the finite difference and spectral methods, showing its high accuracy and efficiency in pricing various financial options, including those mentioned above.

In recent years, the surge of unofficial digital currencies, often referred to as cryptocurrencies, has disrupted traditional financial landscapes. Bitcoin, being the most prominent among them in terms of market adoption and capitalization, presents unique modeling challenges. This study delves into the application of an autoregressive model of order one, incorporating a skew-normal mean-variance mixture of Birnbaum–Saunders innovations, to better capture the dynamic behavior of Bitcoin prices. The model’s robustness to atypical observations and its effectiveness in handling the inherent price volatility associated with Bitcoin make it a promising tool for financial analysis and prediction in this novel asset class.

Input–output tables provide important data for the analysis of economic states. Most regional input–output tables in Japan are not publicly available; therefore, they have to be estimated. Input coefficients are pivotal in constructing precise input–output tables; thus, accurately estimating these input coefficients is crucial. Non-survey methods have previously been used to estimate input coefficients of regions as they require fewer observations and computational resources. However, these methods discard information and require additional data. The aim of this study is to develop a method for estimating input coefficients using artificial neural networks with improved accuracy compared to conventional non-survey methods. To prevent overfitting owing to limited data availability, we introduced a data augmentation technique known as mixup. In this study, the vector sum of data from multiple regions was interpreted as the composition of the regions and the scalar product of regional data was interpreted as the scaling of the region. Based on these interpretations, the data were augmented by generating a virtual region from multiple regions using mixup. By comparing the estimates with the published values of the input coefficients for the whole of Japan, we found that our method was more accurate and stable than certain representative non-survey methods. The estimated input coefficients for three Japanese cities were considerably close to the published values for each city. This method is expected to enhance the precision of regional input–output table estimations and various quantitative regional analyses.

This paper illustrates the power of Julia language for the solution and estimation of Dynamic Stochastic General Equilibrium models. We document large gains of the Julia implementation of Perturbation solution (first and higher orders) and Bayesian estimation using two workhorse models in the literature: the Real Business Cycle Model and a medium scale New-Keynesian Model. We release a companion package that implements 1st, 2nd a 3rd order approximation of Dynamic Stochastic General Equilibrium models and allows for estimation of (log-)linearized models using Sequential Monte-Carlo Methods. Our examples highlight that Julia has low entry costs and it is a language where it is easy to deal with parallelization.

Recent advances in Machine Learning and, especially, Deep Learning, have led to applications of these areas in different fields of knowledge, with great emphasis on stock market prediction. There are two main approaches in the literature to predict future prices in the stock market: (1) considering historical stock prices; and (2) considering news or social media documents. Despite the recent efforts to combine these two approaches, the literature lacks works in which both strategies are performed with Deep Learning, which has led to state-of-art results in many regression and classification tasks. To overcome these limitations, in this work we proposed a new Deep Learning-based approach to predict the Brazilian stock market combining the use of historical stock prices, financial technical indicators, and financial news. The experiments were performed considering the period from 2010 to 2019 with the Ibovespa index and the historical prices of the following Brazilian companies: Banco do Brasil, Itaú, Ambev, and Gerdau, which have significant contribution to the Ibovespa index. Our results show that the combination of stock prices, technical indicators and news improves the stock market prediction considering both the prediction error and return-of-investment.

Every market leaves its fingerprint in prices time series. The Efficient Market Hypothesis (EMH), considers that prices behave as random walks, a property that has been tested on whole data sets of both formal and informal markets. Here we extend this idea studying the Cuban informal exchange market using two standard tests, the Wald-Wolfowitz runs test and the Variance ratio test. Moreover, while these tests are usually done in the whole data set, we check whether different intervals of the series and the series on different time scales fulfill the EMH. Therefore, we repeated the tests in the fast components of the market obtained from an Empirical Mode Decomposition of the data and on separated time intervals defined through a Hidden Markov Model with two latent variables. We concluded that in all cases the Efficient Market Hypothesis is violated. We finish our work discussing some possible causes and consequences of this inefficiency.

Cryptocurrencies have garnered significant attention recently due to widespread investments. Additionally, researchers have increasingly turned to social media, particularly in the context of financial markets, to harness its predictive capabilities. Investors rely on platforms like Twitter to analyze investments and detect trends, which can directly impact the future price movements of Bitcoin. Understanding and analyzing Twitter sentiments can potentially provide insights into future Bitcoin price movements and can shed light on how investor sentiment affects cryptocurrency markets. In this study, we explore the correlation between Twitter activity and Bitcoin prices by examining tweets related to Bitcoin price sentiments. Our proposed model consists of two distinct networks. The first network exclusively utilizes historical price data, which is further decomposed into various components using the Empirical Mode Decomposition method. This decomposition helps mitigate the impact of irregular fluctuations on Bitcoin price predictions. Each of these components is then separately processed by Long Short-Term Memory (LSTM) networks. The second network focuses on modeling user sentiments and emotions in conjunction with Bitcoin market data. User opinions are categorized into positive and negative classes and are integrated with historical data to predict the next-day price using LSTM networks. Finally, the outputs of each network are combined to form the ultimate prediction values. Experimental results demonstrate that Twitter sentiment can effectively helps us predict Bitcoin price trends. Furthermore, to validate our proposed model, we compared it with several state-of-the-art methods. The results indicate that our approach outperforms these existing models in terms of accuracy.