Electricity Consumption, Labor Force and GDP in Turkey: Evidence From Multivariate Granger Causality

Abstract

This article attempts to examine the dynamic relationship between gross domestic product (GDP), electricity consumption, and labor force in Turkey over the period 1980–2008. Applying the bounds test approach to cointegration developed by Pesaran et al. (2001), we found that there was a long-run relationship between electricity consumption, GDP, and labor force when electricity consumption and labor force are the dependent variable. Using four long-run estimators, we also concluded that electricity consumption has a positive and a statistically significant impact on Turkey's economic growth. The coefficient on labor force is positive and statistically significant at the 1% level where a 1% increase in labor force generates a 1.22% and 2.52% increase in electricity consumption, respectively. Turning to the model where labor force is taken as a dependent variable, the coefficient on income is positive and statistically significant for all estimators with a 1% increase in real income generating between 0.23 and 2.77% increase in employment. We concluded that there exists bidirectional Granger causality between electricity consumption and GDP which confirms the feedback hypothesis. On the other hand, there is neutrality between labor force and electricity consumption. When we examine the long-run results, the coefficient on the lagged error correction term which indicates the speed of the adjustment process to restore equilibrium following a disturbance in the long-run equilibrium relationship is significant in the electricity consumption equation at the 5% level. We concluded that almost 5% of the disequilibrium of the previous year's shocks adjusts back to the long run equilibrium—an indication that once shocked convergence to equilibrium is slow.

Keywords:

• electricity consumption
• GDP
• Granger-causality
• labor force

Notes

Note: * denotes a 99% confidence level. Optimal lag determination is based on SBIC.

Note: ***, **, and * denote statistical significance at the 1%, 5% and %10 level, respectively.