Clean energyconsumption economic growth nexus in China: Fresh evidence from the rolling and recursive causality approaches (R&R at Studies in Nonlinear Dynamics & Econometrics).
The current study investigates the dynamic linkages among clean energy usage and economic growth in China for the timespan between 1965 and 2019. Dissimilar to previous studies that have assumed parameter stability in the causality analysis, we apply the rolling and recursive rolling Granger causality techniques that allow time-varying behavior of the model parameters and offer increased reliability and inclusive inference on the causal relationship among the variables. Utilizing the conventional linear Granger causality test, we fail to detect a causation between clean energy usage and growth of the economy. However, we show that the causal relationships experience substantial time variation. Therefore, the inferences according to the linear framework are vulnerable to unreliability. Nevertheless, our study detects a bi-directional and some unidirectional causality evidence over the sample period through the time-varying methodologies and their date-stamping feature. The findings of our study emphasize the significance of considering time-varying causality tests to avoid the risk of misleading inferences.
Revisiting the CO2-growth nexus: A symbolic transfer entropy analysis in the OECD.
This study revisits the causal linkages between CO2 emission level and economic growth in the selected OECD economies spanning the period 1870–2019. The study utilizes various panel-based granger causality methodologies to provide a more comprehensive picture of the variables’ dynamic patterns. Using a symbolic transfer entropy causality test that accounts for cross-sectional dependence, nonlinearities, and parameter heterogeneity, only the evidence of causality running from economic growth to CO2 emission is confirmed for the panel as a whole. This contradicts the findings of the other panel-based causality methods, in which the bivariate causality linkages are supported. The results imply the need for adopting more cost-effective environmental policies to prevent sustainable economic growth from being jeopardized while addressing environmental concerns
Bubble Detection in US Energy Market: Application of Log Periodic Power Law (LPPL) Models.
Causality analysis is an essential and sometimes challenging step that plays a vital role in detecting the dynamic interrelationships between variables. The standard approach of causality (Granger causality method) assumes a straightforward nature of the dependence relations (linear relationships) subjected to a broad spectrum of criticism. In this study, I develop an effective connectivity metric that uses a local linear model trees (LoLiMoT) framework in the context of causality analysis. The proposed technique benefits from LoLiMoT’s divide-and-conquer strategy, allowing for a more complex structural setting and detecting linear and nonlinear causal information flows. The performance of the LoLiMoT based Granger Causality measure (LoLiMoTGC) is evaluated and compared with those of the Linear Granger Causality (LGC), Multilayer perceptron neural network Granger Causality (MLPGC), and Adaptive Neuro-Fuzzy Inference System Granger Causality (ANFISGC) methods. Numerical simulation indicates that LoLiMoTGC outperforms its counterparts in detecting both linear and nonlinear linkages. Besides, it illustrates that this new approach performs decent size and power.
LoLiMoT Based framework in Causality Analysis.
This study aims to investigate the ability of LPPL models to identify bubble (s) and its corresponding termination point (s) [𝑡𝑐] in the US energy market using two US major energy prices, namely, the daily US dollar crude oil price of West Texas Intermediate (WTI), and the US dollar natural gas price covering the 1986:01:02-2018:03:19 periods daily. This study is the first comprehensive work in the literature showing the possibility to develop LPPL models in US energy market over the long horizon. To raise the reliability of the estimation process, the new constraint on the Augmented Dickey-Fuller and Philips–Perron values are introduced to the model to accept fits that are nonstationary. Our findings reinforce that energy market prices during bubble periods oscillate with decreasing amplitude around a faster-than-exponential growth. Besides that, our results indicate that applying the LPPL model in the US energy market is error-free, where the confidence interval of termination point encloses the actual regime shift dates.