Computational Economics最新文献

Understanding why extreme events occur is crucial in many fields, particularly in managing financial market risk. In order to explain such occurrences, it is necessary to use explanatory variables. However, flexible models with explanatory variables are severely lacking in financial market risk management, particularly when the variables are sampled at different frequencies. To address this gap, this article proposes a novel dynamic tail index regression model based on mixed-frequency data, which enables the high-frequency variable of interest to depend on both high- and low-frequency variables within the framework of extreme value regression. Specifically, it concurrently leverages information from low-frequency macroeconomic variables and high-frequency market variables to model the tail distribution of high-frequency returns, consequently enabling the computation of high-frequency Value at Risk and Expected Shortfall. Monte Carlo simulations and empirical studies show that the proposed method effectively models stock market tail risk and produces satisfactory forecasts. Moreover, including macroeconomic variables in the model provides insights for macroprudential regulation.

In today’s dynamic financial landscape, the integration of environmental, social, and governance (ESG) principles into investment strategies has gained great significance. Investors and financial advisors are increasingly confronted with the crucial question of whether their dedication to ESG values enhances or hampers their pursuit of financial performance. Addressing this crucial issue, our research delves into the impact of ESG ratings on financial performance, exploring a cutting-edge machine learning approach powered by the Extreme Gradient algorithm. Our study centers on US-registered equity funds with a global investment scope, and performs a cross-sectional data analysis for annualized fund returns for a five-year period (2017–2021). To fortify our analysis, we synergistically amalgamate data from three prominent mutual fund databases, thereby bolstering data completeness, accuracy, and consistency. Through thorough examination, our findings substantiate the positive correlation between ESG ratings and fund performance. In fact, our investigation identifies ESG score as one of the dominant variables, ranking among the top five with the highest predictive capacity for mutual fund performance. As sustainable investing continues to ascend as a central force within financial markets, our study underscores the pivotal role that ESG factors play in shaping investment outcomes. Our research provides socially responsible investors and financial advisors with valuable insights, empowering them to make informed decisions that align their financial objectives with their commitment to ESG values.

The Greymodels package presents an interactive interface in R for the statistical modelling and forecasting of incomplete or small datasets using grey models. The package, based on the Shiny framework, has been designed to work with univariate and multivariate datasets having different properties and characteristics. The functionality of the package is demonstrated with a few examples and in particular, the user-friendly interface is shown to allow users to easily compare the performance of different models for prediction and among others, visualize graphical plots of predicted values within a user chosen confidence interval. The built-in algorithms in the Greymodels package are extensions or hybrids of the GM((1,,1)) model, and this article covers an overview of the theoretical background of the basic grey model and we also propose a PSO-GM((1,,1)) algorithm in this package.

A recent literature within quantitative macroeconomics has advocated the use of continuous-time methods for dynamic programming problems. In this paper we explore the relative merits of continuous-time and discrete-time methods within the context of stationary and nonstationary income fluctuation problems. For stationary problems in two dimensions, the continuous-time approach is both more stable and typically faster than the discrete-time approach for any given level of accuracy. In contrast, for concave lifecycle problems (in which age or time enters explicitly), simply iterating backwards from the terminal date in discrete time is superior to any continuous-time algorithm. However, we also show that the continuous-time framework can easily incorporate nonconvexities and multiple controls—complications that often require either problem-specific ingenuity or nonlinear root-finding in the discrete-time context. In general, neither approach unequivocally dominates the other, making the choice of one over the other an art, rather than an exact science.

In recent years, artificial intelligence has helped to improve processes and performance in many different areas: in the field of portfolio optimization, the inputs play a crucial role, and the use of machine learning algorithms can improve the estimation of the inputs to create robust portfolios able to generate returns consistently. This paper combines classical mean–variance optimization and machine learning techniques, concretely long short-term memory neural networks to provide more accurate predicted returns and generate profitable portfolios for 10 holding periods that present different financial contexts. The proposed algorithm is trained and tested with historical EURO STOXX 50® Index data from January 2015 to December 2020, and from January 2021 to June 2022, respectively. Empirical results show that our LSTM neural networks are able to achieve minor predictive errors since the average of the MSE of the 10 holding periods is 0.00047, the average of the MAE is 0.01634, and predict the direction of returns with an average accuracy over the 10 investment periods of 95.8%. Our prediction-based portfolios consistently beat the EURO STOXX 50® Index, achieving superior positive results even during bear markets.

This paper proposes a Bayesian varying coefficient model to estimate parameters exhibiting time-dependence in the Cobb–Douglas (CD) production function. We expand upon the classical CD production function by incorporating time-varying properties to enable more sophisticated modeling. We utilize a flexible and efficient Bayesian approach-based computational algorithm for statistical inference in the constrained parameter space, where the sum of model elasticities must be less than 1. The proposed model is applied to four real datasets from macroeconomics, as well as various social science issues broadly covered by the CD production function. The real data applications demonstrate the effectiveness of the proposed model in estimating underlying time-varying effects for parameters in the CD production function.

This research explores the use of machine learning to predict alpha in constructing portfolios, leveraging a broad array of environmental, social, and governance (ESG) factors within the S&P 500 index. Existing literature bases analyses on synthetic indicators, this work proposes an analytical deep dive based on a dataset containing the sub-indicators that give rise to the aforementioned synthetic indices. Since such dimensionality of variables requires specific processing, we deemed it necessary to use a machine learning algorithm, allowing us to study, with strong specificity, two types of relationships: the interaction between individual ESG variables and their effect on corporate performance.The results clearly show that ESG factors have a significant relationship with company performance. These findings emphasise the importance of integrating ESG indicators into quantitative investment strategies using Machine Learning methodologies.

Accurately estimating the size of unregistered economies is crucial for informed policymaking and economic analysis. However, many studies seem to overfit partial data as these use simple linear regression models. Recent studies adopted a more advanced approach, using non-linear models obtained using machine learning techniques. In this study, we take a step forward on the road of data-driven models for the unregistered economy activity’s (UEA) size prediction using a novel deep-learning approach. The proposed two-phase deep learning model combines an AutoEncoder for feature representation and a Long Short-Term Memory (LSTM) for time-series prediction. We show it outperforms traditional linear regression models and current state-of-the-art machine learning-based models, offering a more accurate and reliable estimation. Moreover, we show that the proposed model is better in generalizing UEA’s dynamics across countries and timeframes, providing policymakers with a more profound group to design socio-economic policies to tackle UEA.

The paper examines an opinion dynamics game in a social group with two active agents (influencers) based on the Friedkin–Johnsen model. In the game, we assume sequential announcements of influence efforts by the active agents on the opinions of other (passive) agents of the group. We characterize the Stackelberg solutions as proper solution concepts under sequential play. We then analyze the solutions with a number of measures that quantify them in different aspects: (i) the role of the information structure, i.e., open-loop vs. feedback, (ii) the advantage of sequential over simultaneous moves, and (iii) whether being a leader in the game is more cost-effective than being a follower. Finally, we perform numerical simulations for Zachary’s karate club network to understand how the Stackelberg solutions are sensitive to a change in a parameter characterizing the stubbornness of agents to their initial opinions. The results indicate that the information structure has minimal effect; however, the greatest advantage of the open-loop policy could be achieved with a fully conforming society. In such a society, the efforts of influencers become more efficient, reducing the spread of opinions. Additionally, we observe that the follower has an advantage in the game, which forces each influencer to delay their action until the other one acts.

Fuzzy Data Envelopment Analysis is a modeling technique that efficiently ranks decision-making units (DMUs) based on imprecise inputs and outputs. The method constructs an efficient frontier line that separates efficient and inefficient DMUs. The goal is to improve the efficiency score of each inefficient DMU by moving them to the efficient frontier. In this study, we introduce a new approach, called the Pythagorean approach, which considers both the input and the output aspects. The approach is applied to the CCR model, and a new version of the BCC model is introduced, known as the Pythagorean approach-based BCC model. To handle the vagueness of the data set, the Pythagorean approach-based BCC model is extended to a fuzzy environment using a new type of fuzzy number called a sine-shaped fuzzy number. Finally, the efficacy of the model is tested in Indian public sector banks.