B. Cannas, S. Cincotti, A. Fanni, M. Marchesi, F. Pilo and M. Usai
Many practical applications of neural networks require the identification of non‐linear deterministic systems or chaotic systems. In these cases the use of a network architecture…
Abstract
Many practical applications of neural networks require the identification of non‐linear deterministic systems or chaotic systems. In these cases the use of a network architecture known as locally recurrent neural network (LRNN) is often preferable in place of standard feedforward multi‐layer perceptron (MLP) networks, or of globally recurrent neural network. In this paper locally recurrent networks are used to simulate the behaviour of the Chua’s circuit that can be considered a paradigm for studying chaos. It is shown that such networks are able to identify the underlying link among the state variables of the Chua’s circuit. Moreover, they are able to behave like an autonomous Chua’s double scroll, showing a chaotic behaviour of the state variables obtainable through a suitable circuit elements choice.
Details
Keywords
Stefania Testa, Thaer Atawna, Gino Baldi and Silvano Cincotti
This paper aims at explaining variances in the contribution of Islamic crowdfunding platforms (ICFPs) to sustainable development (SD), by adopting an institutional logic…
Abstract
Purpose
This paper aims at explaining variances in the contribution of Islamic crowdfunding platforms (ICFPs) to sustainable development (SD), by adopting an institutional logic perspective (ILP). ICFPs represent a dual institutional overlap between two logics (the Western-mainstream and the Islamic logic) which have an impact on corporate social responsibility (CSR) interpretations, practices, and decisions and whose conflicts are mitigated by choosing different resolution strategies. The authors aim at showing that this choice affects SD differently.
Design/methodology/approach
The authors develop a conceptual typology through the following steps: (1) choice of variables and identification of corresponding variable domains, through literature review. Variables chosen are the elemental CSR dimensions related to various social and environmental corporate responsibilities to whom diverse meaning and emphasis are given under the Western-mainstream and Islamic logics. (2) Identification of three distinct ideal types of ICFPs, building on different resolution strategies to mitigate conflicts between logics; (3) development, for each ideal type, of a set of implications related to SD; (4) implementation of a first test aiming at assigning real cases to each ideal type.
Findings
The authors identify Western-mimicking (platforms adopting as resolution strategy decoupling or compartmentalizing strategies), Islamic-driven (platforms focusing on one prevailing logic) and Syncretism-inspired (platforms adopting hybridizing practices) ideal-types.
Originality/value
It is the first paper suggesting ILP to explain variances in crowdfunding platforms' role in addressing SD. It focuses on a specific type of CF platforms till now neglected.
Details
Keywords
Kalugala Vidanalage Aruna Shantha
The purpose of this paper is to examine the evolutionary nature of herding phenomenon in the context of a frontier stock market, the Colombo Stock Exchange of Sri Lanka.
Abstract
Purpose
The purpose of this paper is to examine the evolutionary nature of herding phenomenon in the context of a frontier stock market, the Colombo Stock Exchange of Sri Lanka.
Design/methodology/approach
This study applies the cross-sectional absolute deviation methodology for daily frequencies of data of all the common stocks listed during the period from April 2000 to March 2018. The regression coefficients are estimated by using both the ordinary least square and the quantile regression procedures.
Findings
The findings reveal significant changes to the pattern of herding over different market periods, each with specific characteristics. Herding is strongly evident in up and down market days in the 2000-2009 period, during which the market was highly uncertain with the impact of the political instability of the country due to the Civil War on the stock trading. Even after this Civil War period, herd tendency is strongly manifested toward the up market direction as a result of the investors’ optimism about the country’s economy and political stability, which caused to a speculative bubble in the market. After that, it is turned into negative herding due to the panic selling occurred in view of the uncertainty of the inflated prices, which led to a market crash. Notably, herding appears to be consistently absent over the period after the crash, despite the presence of herd motives such as high market uncertainties triggered by political instability and economic crisis during that period.
Research limitations/implications
The findings suggest that herd behavior is an evolving phenomenon in financial markets. Consistent with the adaptive market hypothesis, the absence of herding evident after the market crash could be attributed to the investors’ learning of the irrationality of herding/negative herding for adapting to market conditions. As a result, herding and negative herding tendencies declined and disappeared at the aggregate market level.
Originality/value
This study contributes to the literature by providing novel evidence on the evolutionary nature of behavioral biases, particularly herding, as predicted by the adaptive market hypothesis. With the application of the quantile regression procedure, in addition to customary used ordinary least squares approach, it also provides robust evidence on this phenomenon.
Details
Keywords
Engin Boztepe, Fatma Akyüz and Selçuk Gülten
Purpose: To investigate the impact of integrating artificial intelligence (AI) techniques, particularly deep learning, into bankruptcy prediction models within the banking sector…
Abstract
Purpose: To investigate the impact of integrating artificial intelligence (AI) techniques, particularly deep learning, into bankruptcy prediction models within the banking sector.
Need for the study: With the historical development of bankruptcy prediction models, there is a growing recognition of the potential for AI to enhance the accuracy of these models. This study addresses the need to explore how AI can improve the prediction of financial failures in banks.
Methodology: Using data from banks spanning from 2020 to 2023, this study applies well-established bankruptcy prediction models including the Altman Z model, Altman Z’ model, Springate model, Zmijewski model, and Taffler model. Deep learning techniques are employed to teach these models to AI. Evaluation of the results is conducted using a majority voting decision-making system, incorporating algorithms such as KNN, naive Bayes, and decision trees.
Findings: Integrating AI techniques into bankruptcy prediction models has the potential to enhance the accuracy of forecasts. Evaluation criteria encompass both accuracy and precision, with promising results observed through the majority voting decision-making system. This study suggests a shift toward more sophisticated techniques for bankruptcy risk assessment within the banking sector.
Practical implications: Improved bankruptcy prediction models facilitated by AI techniques could enhance risk management strategies within banks, leading to more informed decision-making processes. This, in turn, could contribute to the overall stability and efficiency of the financial system. Moreover, the importance of considering contradictory results when applying AI-driven models in practice, highlighting areas for further research and refinement in the field of financial risk assessment.
Details
Keywords
An agent-based market simulation is utilized to examine the impact of high frequency trading (HFT) on various aspects of the stock market. This study aims to provide a baseline…
Abstract
Purpose
An agent-based market simulation is utilized to examine the impact of high frequency trading (HFT) on various aspects of the stock market. This study aims to provide a baseline understanding of the effect of HFT on markets by using a paradigm of zero-intelligence traders and examining the resulting structural changes.
Design/methodology/approach
A continuous double auction setting with zero-intelligence traders is used by adapting the model of Gode and Sunder (1993) to include algorithmic high frequency (HF) traders who retrade by marking up their shares by a fixed percentage. The simulation examines the effects of two independent factors, the number of HF traders and their markup percentage, on several dependent variables, principally volume, market efficiency, trader surplus and volatility. Results of the simulations are tested with two-way ANOVA and Tukey’s post hoc tests.
Findings
In the simulation results, trading volume, efficiency and total surplus vary directly with the number of traders employing HFT. Results also reveal that market volatility increased with the number of HF traders.
Research limitations/implications
Increases in volume, efficiency and total surplus represent market improvements due to the trading activities of HF traders. However, the increase in volatility is worrisome, and some of the surplus increase appears to come at the expense of long-term-oriented investors. However, the relatively recent development of HFT and dearth of appropriate data make direct calibration of any model difficult.
Originality/value
The simulation study focuses on the structural impact of HF traders on several aspects of the simulated market, with the effects isolated from other noise and problems with empirical data. A baseline for comparison and suggestions for future research are established.
Details
Keywords
Hongquan LI, Gang Cheng and Shouyang Wang
The securities transaction tax (STT) has been theoretically considered as an important regulation device for decades. However, its role and effectiveness in financial markets is…
Abstract
Purpose
The securities transaction tax (STT) has been theoretically considered as an important regulation device for decades. However, its role and effectiveness in financial markets is still not well understood both theoretically and empirically. By use of agent-based modeling method, the purpose of this paper is to present a new artificial stock market model with self-adaptive agents, which allows the assessment of the impacts from various levels of STTs in distinctive market environments and thus a comprehensive understanding of the effects of STTs is achieved.
Design/methodology/approach
In the model, agents are allowed to employ the strategies used by the following five types of investors: contrarians, random traders, momentum traders, fundamentalists and exit strategy holders. Specifically, the authors start with the investigation of the dynamics of a tax free benchmark market; then the patterns of market behaviors and the behaviors of various types of investors are discussed with different levels of STTs in markets with mild and high fluctuations.
Findings
The simulation results consistently show that a moderate transaction tax does contribute to market stabilization in terms of reducing market volatility while with a price of mild decrease of market efficiency and liquidity. The findings suggest that a balance between market stability and efficiency could be reached if regulatory authorities introduce STTs to markets discreetly.
Originality/value
This paper enriches the comprehensive understanding of the effects of STT, and gives good explanation about the controversy between Tobin’s proponents and anti-Tobin group.
Details
Keywords
Baki Unal and Çagdas Hakan Aladag
Double auctions are widely used market mechanisms on the world. Communication technologies such as internet increased importance of this market institution. The purpose of this…
Abstract
Purpose
Double auctions are widely used market mechanisms on the world. Communication technologies such as internet increased importance of this market institution. The purpose of this study is to develop novel bidding strategies for dynamic double auction markets, explain price formation through interactions of buyers and sellers in decentralized fashion and compare macro market outputs of different micro bidding strategies.
Design/methodology/approach
In this study, two novel bidding strategies based on fuzzy logic are presented. Also, four new bidding strategies based on price targeting are introduced for the aim of comparison. The proposed bidding strategies are based on agent-based computational economics approach. The authors performed multi-agent simulations of double auction market for each suggested bidding strategy. For the aim of comparison, the zero intelligence strategy is also used in the simulation study. Various market outputs are obtained from these simulations. These outputs are market efficiencies, price means, price standard deviations, profits of sellers and buyers, transaction quantities, profit dispersions and Smith’s alpha statistics. All outputs are also compared to each other using t-tests and kernel density plots.
Findings
The results show that fuzzy logic-based bidding strategies are superior to price targeting strategies and the zero intelligence strategy. The authors also find that only small number of inputs such as the best bid, the best ask, reference price and trader valuations are sufficient to take right action and to attain higher efficiency in a fuzzy logic-based bidding strategy.
Originality/value
This paper presents novel bidding strategies for dynamic double auction markets. New bidding strategies based on fuzzy logic inference systems are developed, and their superior performances are shown. These strategies can be easily used in market-based control and automated bidding systems.
Details
Keywords
Albulena Shala and Vlora Berisha
Purpose: This chapter highlights the key findings about the development of green finance in Central and Eastern European (CEE) countries. The main purpose is to analyze the policy…
Abstract
Purpose: This chapter highlights the key findings about the development of green finance in Central and Eastern European (CEE) countries. The main purpose is to analyze the policy to promote green financing, the role of financial institutions and regulators in the green finance agenda, and the challenges of green financing.
Need for the study: Through this study, we will identify the level of development of green finance in CEE countries. It is also necessary to know the role and contribution of central banks in the development of green finance.
Methodology: Using the comparative methodology, we analyzed the following indexes: green economic opportunities, the Green Growth Index, the Global Climate Risk Index, and the Green Central Banking Scorecard in the G20. We included 17 countries from Central and Eastern Europe (CEE): Albania, Bulgaria, Bosnia and Herzegovina, Croatia, Czech Republic, Poland, Estonia, Romania, Hungary, Latvia, Lithuania, Kosovo, Montenegro, North Macedonia, Slovenia, Slovakia, and Serbia.
Findings: The results show that the countries of the Western Balkans do not have a satisfactory performance compared to other countries, although they stand well in terms of legislation.
Practical implications: Studying green finance in a group of countries would bring new insights to their economies and to all other countries wishing to develop green finance markets. Their exposition, especially for the countries of the Western Balkans, will bring new knowledge and practices to the Central Bank and relevant institutions.
Details
Keywords
Linda Ponta, Gloria Puliga and Raffaella Manzini
The measure of companies' Innovation Performance is fundamental for enhancing the value and decision-making processes of firms. The purpose of this paper is to present a new…
Abstract
Purpose
The measure of companies' Innovation Performance is fundamental for enhancing the value and decision-making processes of firms. The purpose of this paper is to present a new measure of Innovation Performance, called Innovation Patent Index (IPI), which makes it possible to quantitatively summarize different aspects of firms' innovation.
Design/methodology/approach
In order to define the IPI, a secondary source, i.e. patent data, has been used. The five dimensions of IPI, i.e. efficiency, time, diversification, quality and internationalization have been defined both analyzing the literature and applying three different machine learning algorithms (regularized least squares, deep neural networks and decision trees), considering patent forward citations as a proxy of the innovation performance.
Findings
Results show that the IPI index is a very useful tool, simple to use and very promptly. In fact, it is possible to get important results without making time consuming analysis with primary sources. It is a tool that can be used by managers, businessmen, policymakers, organizations, patent experts and financiers to evaluate and plan future activities, to enhance the innovation capability, to find financing and to support and improve innovation.
Research limitations/implications
Patent data are not widely used in all the sectors. Moreover, the pure number of forward citations is not the only forward looking indicator suggested by the literature.
Originality/value
The demand for a useable Innovation Performance tool, as well as the lack of tools able to grasp different aspects of the innovation, highlight the need to develop new instruments. In fact, although previous studies provide several measures of Innovation Performance, these are often difficult for managers to use, do not appreciate different aspects of the innovation and are not forward looking.