1,721,051 research outputs found
Alternative distribution based GARCH models for Bitcoin volatility estimation
No full textKatsiampa (2017) shows that, among different GARCH models, the optimal conditional heteroskedasticity model regarding the goodness-of-fit to Bitcoin price data is the AR-Component GARCH (AR-CGARCH) model. However, in that paper the author does not take into account for statistical proprieties of Bitcoin’s return distribution, and even showing both skewness and non-normality of the data, we consider a standardized normal distribution for all studied GARCH models. This paper represents an improvement of the previous literature about GARCH model for Bitcoin. In particular, this paper examines different distributional assumptions about innovations distribution for some GARCH models, showing that it is possible to obtain better estimates through the AR(1)-APARCH(1,1) model assuming that innovations follow a t-student distribution
Improved precision matrix estimation for mean-variance portfolio selection
No full textThis paper deals with the problem of reducing the estimation error of the precision matrix in mean-variance portfolio selection. A new shrinkage estimator for the precision matrix is proposed and the optimal shrinkage intensity is obtained by maximizing the investor's utility function. An oracle estimator is proposed and many feasible estimators are derived in the paper. Feasible estimators are easy to implement in practice. The performance of the proposed shrinkage estimator is evaluated with both simulations and empirical experiments
Forecasting High-Dimensional Portfolios
No full textIn this paper, we investigate the usefulness of forecasting in a high-dimensional framework where the number of assets is larger than the temporal observations. The benefit of forecasting lies in the concept of timing, which means anticipating future market conditions. We find that when high-dimensional econometric approaches are used, forecasting either the mean or the covariance is better than predicting both and then approaches based on static estimates. Moreover, we find that timing portfolios also perform better than the naive strategy. Considering the portfolio returns over time, we find that a possible explanation for the better performance of volatility-timing portfolios is that they better manage risk during periods of high uncertainty
Forecasting binary outcomes in soccer
No full textSeveral studies deal with the development of advanced statistical methods for predicting football match results. These predictions are then used to construct profitable betting strategies. Even if the most popular bets are based on whether one expects that a team will win, lose, or draw in the next game, nowadays a variety of other outcomes are available for betting purposes. While some of these events are binary in nature (e.g. the red cards occurrence), others can be seen as binary outcomes. In this paper we propose a simple framework, based on score-driven models, able to obtain accurate forecasts for binary outcomes in soccer matches. To show the usefulness of the proposed statistical approach, two experiments to the English Premier League and to the Italian Serie A are provided for predicting red cards occurrence, Under/Over and Goal/No Goal events
Measuring unit relevance and stability in hierarchical spatio-temporal clustering
No full textUnderstanding the significance of individual data points within clustering structures is critical to effective data analysis. Traditional stability methods, while valuable, often overlook the nuanced impact of individual units, particularly in spatial contexts. In this paper, we explore the concept of unit relevance in clustering analysis, emphasizing its importance in capturing the spatio-temporal nature of the clustering problem. We propose a simple measure of unit relevance, the Unit Relevance Index (URI), and define an overall measure of clustering stability based on the aggregation of computed URIs. Considering two experiments on real datasets with geo-referenced time series, we find that the use of spatial constraints in the clustering task yields more stable results. Therefore, the inclusion of the spatial dimension can be seen as a way to stabilize the clustering
Another Look into Tail Risk Connectedness Using Network Modelling: Evidence from European Stock Markets
No full textThis paper proposes an approach for measuring tail risk connectedness in financial networks by leveraging spatial methods, particularly the spatial autoregressive model, applied to Value-at-Risk (VaR) and Expected Shortfall (ES) estimates. Given the definition of the network structure, and departing from traditional methods reliant on volatility spillovers, our approach aims to capture the effect of the nature of the event on the tail dimension of market interconnectedness, offering insights beyond conventional metrics. Our results show that since the nature of the event substantially affects the convergence (divergence) of the response of the agents of the financial network to a shock, accounting for this effect is crucial to correctly measure its transmission. Since not all shocks are similar, events such as the Great Recession or the COVID-19 pandemic do not affect the tail risk of the financial network in a similar way, even when volatility spillovers increase in both cases. The relevance of considering the source of the shock is shown through an empirical analysis of the most important European stock markets, demonstrating the efficacy of the proposed approach in assessing systemic risks, therefore providing a valuable tool for policymakers, investors, and financial regulators
Option Pricing under Multifractional Process and Long-Range Dependence
No full textWe introduced a new method to compute the European Call (and Put) Option price under the assumption of multifractional Brownian motion (mBm). The reason why we need a procedure for estimating the Option price is due to the absence of a closed formula for this process. To compute the Option price, we first simulated the logarithmic price under mBm and, by using a discount factor, we computed the option's pay-off. Then, we fitted the best probability distribution associated to the discounted pay-off, computing the European Call Option price as its average
Gender and authorship in energy studies: Is there an impact?
No full textAlthough gender inequality has been examined and debated as one of the most prominent challenges within the scientific community, relatively little attention has been paid to gender differences with regard to authorship. The aim of this paper is to identify whether gender differences exist with respect to the impact factors of the journals in which male and female authors publish their research. Existing studies use machine-assisted tools to determine author gender. Given the limitations of this type of approach to data collection and coding, we opted for a manual approach that ensured both the inclusion of a high number of journals and greater precision in determining author gender. This paper focuses on authors who published articles on a concrete area of research (energy) in a specific region (Central and Eastern Europe) over a 14-year period (2004–2017). Our study identified a gender bias within energy-related research: male authors (or male-dominated teams) publish more often and on average in journals with higher impact factors than female authors
Entropy weighted model-based clustering of skewed and heavy tailed time series
No full textThe goal of clustering is to identify common structures in a data set by forming groups of homogeneous objects. The observed characteristics of many economic time series motivated the development of classes of distributions that can accommodate properties such as heavy tails and skewness. Thanks to its flexibility, the Skewed Exponential Power Distribution (also called Skewed Generalized Error Distribution) ensures a unified and general framework for clustering possibly skewed and heavy-tailed time series. A clustering procedure of model-based type is developed, assuming that the time series are generated by the same underlying probability distribution but with different parameters. Moreover, we propose to optimally combine the estimated parameters to form the clusters with an entropy weighing k-means approach. Moreover, by applying skewness or kurtosis-based projection pursuit, the resulting interesting projections can be used as the input of the clustering procedure with a different distributional assumption. The usefulness of the proposal is showed by means of application to financial time series
Model-based fuzzy time series clustering of conditional higher moments
No full textThis paper develops a new time series clustering procedure allowing for heteroskedasticity, non-normality and model's non-linearity. At this aim, we follow a fuzzy approach. Specifically, considering a Dynamic Conditional Score (DCS) model, we propose to cluster time series according to their estimated conditional moments via the Autocorrelation-based fuzzy C-means (A-FCM) algorithm. The DCS parametric modeling is appealing because of its generality and computational feasibility. The usefulness of the proposed procedure is illustrated using an experiment with simulated data and several empirical applications with financial time series assuming both linear and nonlinear models' specification and under several assumptions about time series density function
- …
