Assessing Computational Complexity in Selecting Periods for LMDI Techniques in Energy‐Related Carbon Dioxide Emissions: An Alternative Approach

ABSTRACT The Logarithmic Mean Divisia Index (LMDI) decomposition analysis is widely employed to examine the drivers behind changes in carbon dioxide emissions related to energy consumption. This analysis has been applied using single‐period, year‐by‐year, and multi‐period time frames worldwide. However, these time frames often overlook trend changes in carbon emission time series, which may lead to inaccurate and biased identification of driving factors. This study replicates previous findings and proposes a novel multi‐period methodology for defining time frames in decomposition analysis. The proposed approach addresses the limitations of traditional methods by accounting for trend changes in the time series and performing an exhaustive search to optimally identify the most suitable time frames for LMDI‐based decomposition. The methodology comprises two stages: the first generates an exhaustive list of possible time series partitions, and the second determines the optimal partition by minimizing the total mean square error (TMSE) using sequential linear models. The results, supported by computational performance tests, demonstrate that the proposed method effectively identifies optimal time frame definitions, making it particularly suitable for annualized case studies on carbon dioxide emissions decomposition in the context of the energy transition.