Ensayos sobre la relación entre el crecimiento económico y la energíaaplicaciones matemáticas y econométricas

  1. Feria Gallardo, Julia
Supervised by:
  1. Emilio Congregado Ramírez de Aguilera Director
  2. Mónica Carmona Arango Director

Defence university: Universidad de Huelva

Fecha de defensa: 27 January 2016

  1. José Luis Galán González Chair
  2. José María Millán Tapia Secretary
  3. Antonio Jesús Sánchez Fuentes Committee member

Type: Thesis


This thesis focuses on the relationship between economic growth and some variables as energy consumption or C02 emissions. This relationship has been studied extensively in the empirical literature on energy economics. The articles of this thesis try to shed some light and new evidences on this relationship. Thus, in the first chapter, try to investigate the relationship between energy energy consumption and GDP, using Spanish quarterly data from 1980:1- 2010:4. The results show that the elasticity of energy consumption on income is positive. A second step, is to test if this elasticity is time invariant, to carry this task, we perfomr a test allowing the existence of structural breaks following the methodology proposed by Gregory and Hansen (1996) and Hatemi-J (2008). Estimates show that this relationship is time- variant. Finally, Granger causality is tested, finding mixed evidence. The second chapter, is an analysis to contrast causal relationship described above is performed, but in contrast to the time, if that energy is persistent, that is, if transient changes in such consumption become permanent. The methodology used is proposed by Sinclair (2010) using unobservable components model. The results support the existence of persistence (hysteresis) in energy consumption. The methodology permit us to estimate the cyclical component of both variables, and then, we can test the Granger causality between both variables. The results show bidirectional causality. Finally, the possible existence of structural breaks, is contrasted by the technique of Bai-Perron (1998.2003). This indicates the existence of a structural break in early 2009 (at the beginning of the recent economic crisis), and the existence of the two-way relationship but only for the last period, that is, only after confirming the current economic crisis, maintained throughout the period causality from economic cycle energy consumption. The third paper deals with the causality of GDP energy consumption, disaggregating by source (fossil, coal, oil, natural gas, nuclear, renewable, biomass and geothermal). For this study, using the methodology of Toda-Yamamoto (1996) and its extension proposed by Hatemi-J (2012) in the case of asymmetries. Basically, the results show the existence of "neutral hypothesis" in the case of renewable energies and "conservation hypothesis" in the case of fossil fuels energy. However, when asymmetry is allowed, it is seen as the emerging "growth hypothesis" for the case of petroleum, this important result, given the relative importance of petroleum to the total energy consumption. Finally, the last chapter of this thesis, study the relationship between GDP and C02 emissions testing the existence of the "Environmental Kuznets Curve" (EKC hereafter). Initial results show the non-existence of the EKC, with the positive elasticity, and therefore, showing that higher economic growth will be higher C02 emissions. However, given the existence of structural breaks, using the methodology propose by Kerjiwal-Perron (2008, 2010) to find the breaks and Arai-Kurozumi (2009) technique allowing breaks in cointegrating relationship. Three breaks are found. Reestimating elasticity, shows that the relationship is dependent on time, and, what is more important, as considering the existence of such breaks, the EKC hypothesis is confirmed in the United States. In this thesis we have tried to give new evidence about the relationship between energy consumption, C02 emissions and GDP. Results show that these relationships are time variant, depending of the source and according to the phases of expansion and recession. Finally, it confirms the existence of EKC in the United States, but pointing out the need to consider the long-term relationships between the variables.