Institutional Repository of Vilnius University
The comparison of the performance of macroeconomic variables in forecasting fuel price growth in lithuania /
Abstract
dc:descriptionThe aim of this scientific paper is determining how effectively 9 possible indicators can estimate fuel price growth in Lithuania. The study evaluates predicting performance using two types of fuel and three predicting periods, in total yielding six model parameters (three gasoline, and three diesel). In addition, the research investigates the efficiency of three combination forecasting strategies. The study article opens with a review of previous research on the topic, continued by a detailed explanation of the technique utilized, including information on collected data. Finally, the analytical findings are presented in the study. The review of previous research on indicators and predicting is presented, which serves as the basis for selecting potential predictive variable, analysing forecasting methods, and forming expected outcomes. To obtain results, the study employees commonly used forecasting technique, where the considered indicators were incorporated as regressors in a simple ARDL model. This model was used to anticipate the fuel price growth, which was then compared to benchmark autoregressive (AR) projection employing the mean squared forecast error. Additionally, combination forecasts were created using methods such as media, weighted average, and simple mean. The study shows that five variables demonstrates superior forecasting performance compared to benchmark forecasts. The predictive value of these indicators in forecasting the growth of fuel prices was found to increase with longer predicting horizon. This can be attributed to the diminishing influence of prior prices on price in the future over long time periods. The consistency of these findings was observed across fuel prices and indicators groups. Across all model parameters, none of the indicators significantly decreased forecast error. Most of the forecast combination method decreased predicting error. Combination forecasts should be used in practical scenarios to address the challenge of identifying the best-performing indicators in all situations.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Belickaitė, Aušrinė,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192842041&prefLang=en_US