Purpose: The present work has proposed and employed an innovative hybrid method based on the combination of factor analysis and an artificial neural network (ANN) model to forecast customer satisfaction from the identified dimensions of service quality in India, a developing country.Design/methodology/approach: The qualitative study is conducted with Internet banking users to understand e-banking clients' perceptions. The data is collected with the help of a questionnaire from randomly selected 208 customers in India. Firstly, factor analysis was performed to determine the influential factors of customer satisfaction, and four factors i.e. efficiency, reliability, security and privacy, and issue and problem handling were extracted accordingly. The neural network model is then applied to the factor scores to validate the key elements. Lastly, the comparative analysis of the actual ANN and the regression predicted result is done.Findings: The success ability of the linear regression model is challenged when approximated to nonlinear problems such as customer satisfaction. It is concluded that the ANN model is a better fit than the linear regression model, and it can recognise the complex connections between the exogenous and endogenous variables. The results also show that reliability, security and privacy are the most influencing factors; however, problem handling and efficiency have the slightest effect on bank client satisfaction.Research limitations/implications: This research is conducted in India, and the sample is chosen from the urban area. The limitation of the purposeful sampling technique and the cross-sectional nature of the data may hamper the generalisation of the results.Originality/value: The conclusions from the study will be helpful for policymakers, bankers and academicians. To our knowledge, few studies used ANN modelling to predict customer satisfaction in the service sector
Purpose: The paper aims to analyse the extent of volatility and generating forecasts of exchange rates of British pound and Indian rupees in US terms. Design/methodology/approach: This study applies different combinations of GARCH and EGARCH models suggested in the Econometric literature to capture the extent of volatility. The forecast of exchange rates of British Pound and Indian Rupees in US terms are generated applying artificial neural network (ANN) technique using different combination of networks with hyperbolic tangent function at hidden and output stage of the model. Findings: The presence of volatility depicts that there is noise and chaos in the forex market. Prediction of exchange rate of the respective currencies underscores that exchange rates will increase marginally in near future. Practical Implications: The results proposed in this study will be benchmark for the hedgers, investors, bankers, practitioners and economists to foresee the exchange rate in the presence of volatility and design policies accordingly. Originality/value: In literature, no study has applied ANN for forecasting exchange rate after measuring the extent of volatility. The present study is a unique contribution in the existing pool of literature to forecasts the concerned variable(s) after ascertaining the noise and chaos in the data by applying GARCH family models.
13 November 2024 in Environment, Development and Sustainability
Purpose: The Quadrilateral Group for Security Dialogues (QUAD), comprising Japan, Australia, India, and the USA, has experienced a noteworthy escalation in both environmental depletion and economic performance over the preceding 2 decades. Consequently, it becomes crucial to outline strategies through which the QUAD can attain a harmonious equilibrium between sustainable economic growth and ecological well-being. This research endeavors to dissect the intricate relationships among trade globalization, regulatory quality, and environmental sustainability within the QUAD for the period 1990–2021. The study checks for the Pollution Haven Hypothesis (PHH), and Pollution Halo Effect Hypothesis (PHEH).Design/methodology/approach: The long-run association is based on the autoregressive distributed lag (ARDL) model and bounds test approach to cointegration while divergence or convergence is studied with the help of decoupling index (DI). Results have been verified by applying serial correlation LM test, autoregressive conditional heteroskedasticity (ARCH), and cumulative sum of recursive residuals (CUSUM) tests to ensure the robustness and stability of the model.Findings: The empirical results of this study affirm the applicability of the PHEH in the contexts of India, Japan, and the USA, whereas the PHH is validated in the case of Australia. Furthermore, the analysis reveals the existence of relative decoupling solely in the case of India. This testifies that the rate of growth of the Indian economy surpasses the rate of growth in ecological footprint (EF), indicating a relative reduction in the intensity of environmental impact per unit of economic growth.Research limitations/implications: The empirical findings of our study suggest that countries with effective regulatory systems are better positioned to control and mitigate the potential adverse environmental effects resulting from increased global trade. Thus, policymakers are prompted to reassess the development policies for sustainable economic growth that will minimize adverse environmental repercussions. The implication of the negative relation between urbanization and EF is paramount for policymakers in developing countries seeking strategies for balanced urban development that aligns with environmental sustainability.Originality/value: The present study is a unique exploration of the impact of trade globalization and regulatory quality on EF, specifically on PHH/PHEH in the context of QUAD.