CO2 emissions persistence: Evidence using fractional integration

Do bitcoin electricity consumption and carbon footprint exhibit random walk and bubbles? Analysis with policy implications

One of the main current focuses of global economies and decision-makers is the efficiency of energy utilization in cryptocurrency mining and trading, along with the reduction of associated carbon emissions. Understanding the pattern of Bitcoin's energy consumption and its bubble frequency can greatly enhance policy analysis and decision-making for energy efficiency and carbon emission reduction. This research aims to assess the validity of the random walk hypothesis for Bitcoin's electricity consumption and carbon footprint. We employed both traditional methods (ADF and KPSS) and recently proposed unit root techniques that account for structural breaks and non-linearity in the data series. Our analysis covers daily data from July 2010 to December 2021. The empirical results revealed that traditional unit root techniques did not confirm the stationarity of both bitcoin's electricity consumption and carbon footprint. However, novel structural break (SB) and linearity tests conducted enabled us to discover five SB episodes between 2012 and 2020 and non-linearity of the variables, which informed our application of the newly developed non-linear unit root tests with structural breaks. With the new methods, the results indicated stationarity after accommodating the SB and non-linearity. Furthermore, based on Phillips and Shi (2019)'s test, we identified certain bubble episodes in the bitcoin energy and carbon variables between 2013 and 2021. The major drivers of the bubbles in bitcoin energy consumption and carbon footprint are variables relating to the bitcoin and financial markets activities and risks, including the global economic and political risks. The study's conclusion based on the above findings informs several policy implications drawn for energy and environmental management including the encouragement of green investments in cryptocurrency mining and trading.

CO2 emissions in Latin America: a time series perspective based on fractional integration

This article deals with the analysis of [Formula: see text] emissions in Latin America by using a long memory process based on fractional integration. Using data of [Formula: see text] emission and [Formula: see text] emissions per capita, for 32 Latin American and Caribbean countries, the results show significant differences according to the variable examined, the model used, and the country under examination. In particular, for the [Formula: see text] emissions, mean reversion is found in Belize and also under some circumstances in Antigua and Barbuda, Colombia, Dominica, Dominican Republic, Ecuador, Grenada, Honduras, Nicaragua, Panama, Peru, and Uruguay. Thus, shocks in these series have a transitory effect. With respect to the time trends, only for some Caribbean countries, namely, Antigua and Barbuda, Aruba, Bahamas, Cuba, and Jamaica, the trend is insignificant; on the other hand, large countries like Brazil, Mexico, and Argentina display the highest time trend coefficients; for the [Formula: see text] emissions per capita, there are eleven countries where mean reversion is detected, and there are ten that share a lack of significance for the trend. The most significant trends now take place in Trinidad and Tobago, British Virgin Islands, Barbados, and Guyana. Policy implications of the results obtained are reported at the end of the paper.