1 / 25

E NTE PER LE N UOVE TECNOLOGIE L’ E NERGIA E L’ A MBIENTE

E NTE PER LE N UOVE TECNOLOGIE L’ E NERGIA E L’ A MBIENTE. The causality between energy consumption and economic growth: A multi-sectoral analysis using non-stationary cointegrated panel data. Valeria Costantini, Chiara Martini Department of Economics, University Roma Tre

adamsdana
Download Presentation

E NTE PER LE N UOVE TECNOLOGIE L’ E NERGIA E L’ A MBIENTE

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. ENTE PER LE NUOVE TECNOLOGIE L’ENERGIA E L’AMBIENTE The causality between energy consumption and economic growth: A multi-sectoral analysis using non-stationary cointegrated panel data Valeria Costantini, Chiara Martini Department of Economics, University Roma Tre ENEA, National Institute for New Technologies, Energy and Environment

  2. Outline • Introduction • Aim of the paper • Theoretical background • Methodology • Empirical results • Conclusions

  3. Introduction • Toman and Jenelkova (2003):economic growth as the main driver for energy demand • Stern and Cleveland (2004): the shift from lower-quality to higher-quality energy services as impulse to economic growth • Brookes (1990) and Khazzoom (1980): rebound effect; Binswanger (2001) and Barker (2007):energy efficiency leads to a reduction in energy prices which could result in an increase in energy services demand

  4. Medlock and Soligo (2001) energy consumption (EC) at time t for each j-th end-use sector is a function of economic output (Y), energy prices (p) and technology (τ) Lee and Chang (2008); Stern (2000) economic output (Y) is a function of the capital stock (K), labour (L) and energy inputs (EC) Introduction

  5. Aim of the paper • To investigate the strength of the structural linkage between energy demand and economic growth using a panel dataset • To analyse energy-economic growth causality at the sectoral level (industry, service, transport and residential) • To develop a multivariate model that accounts for energy prices of each end-use energy sector (only for OECD countries) • To estimate the sector-specific elasticity parameters of energy demand related to economic variables and energy prices

  6. Theoretical background • First-generation: time series assumed to be stationary, based on a traditional VAR methodology. • Second-generation: recognized the non-stationarity of the data series, tested for cointegration and used ECM to test for Granger causality. • Third-generation: used multivariate estimators. • Fourth-generation:panel methods to test for unit roots, cointegration and Granger causality (Al-Iriani, 2006; Lee, 2007, 2008; Mahadevan and Asafu-Adaye, 2007). Sectoral studies on single time series (Jumbe, 2004; Zachariadis, 2007). Multivariate studies on panels (Stern,2000; Oh and Lee, 2004; Al-Rabbaie and Hunt, 2006; Lee and Chang, 2008).

  7. Methodology Three steps for testing the causal relationship between energy consumption and economic growth in a panel context: • Panel unit root tests are used to investigate the order of integration in the economic and energy time series variables. • Panel cointegration tests are used to examine the long-run relationships between the variables when they are I(1). • If a cointegrating relationship is found, a Vector Error Correction Model (VECM) is used to investigate the direction of the causal relationships in a dynamic context.

  8. Bivariate models: ECTs estimated as the residuals (itand it) from the two following equations Multivariate models (with prices): ECTs (it,it,it) given by Methodology

  9. Bivariate model lagged residuals from the long-run cointegrating relationship short-run adjustment coefficients Multivariate model Methodology

  10. We have collected data for 71 countries: 26 OECD (homogeneous) and 45 non-OECD (heterogeneous) Time series for 1960-2005 (OECD) and 1970-2005 (non-OECD) We have collected data separately for four end-use energy sectors: industry, services, transport, residential Dataset description

  11. Dataset description

  12. Dataset description

  13. Dataset description

  14. Dataset description Trends in energy consumption for OECD countries (ktoe)

  15. Dataset description Trends in energy consumption for non-OECD countries (ktoe)

  16. Results Having estimated the VECM for all the sectors and distinct sub-samples, we performed a Wald test with a Chi-squared statistic distribution on the significance of the coefficients, evaluating three different Granger causality relationships. 1) Short-run causality, testing the significance of the coefficients related to the lagged economic and energy variables: and 2) Long-run causality related to the coefficient for the ECT term.

  17. Results 3) Strong causality to test whether the sources of causation are jointly significant: The strong Granger causality test can be interpreted as a test of weak exogeneity (Engle et al., 1983) of the dependent variable. Only when both the t and Wald Chi-sq statistics in the VECM reveal the absence of causality nexus, this will imply that the dependent variable is weakly exogenous.

  18. Results

  19. Results

  20. Results

  21. Results Large differences when the econometric estimations are carried on the full sample or on sub-samples. Wider differences when we analyze the causality relationships on the basis of disaggregated energy end-use sectors. Results largely change when the role of energy prices is introduced in a multivariate model for OECD sample. Industrial sector:converging trend in the short-run, but causality directions diverge when strong causality is tested separately for the two sub-samples.

  22. Results Transport sector: all three kinds of causality show contrasting results for OECD and non-OECD countries the application of similar energy policies in structurally divergent countries could bring to different impacts. Residential sector: no univocal causality relationships in both developed and developing countries  policy evaluations and model settings should be performed with caution accounting for endogeneity and mutual causality.

  23. Conclusions Our results cast some doubt on the capacity of bivariate models to shape causal relationships in the energy-economy binomial especially when different sectors are investigated. Working with specific sectors allows the existence of divergent trends to emerge, even in a quite homogeneous sample such as OECD. Looking at the industry and transport sectors, it is worth noting that the causality direction changes when different time horizons are accounted for.

  24. Conclusions In the short-run, the economic growth process determines the energy consumption trend so that energy consumption is mainly driven by production demand, and policies oriented towards promoting energy saving do not seem to affect economic development negatively. On the contrary, long-run causalities show that energy consumption and economic performances could be mutually influenced by each other, reducing the neutrality of energy policies on the development path.

  25. Thanks for your attention! Chiara Martini chiara.martini@enea.it

More Related