The effect of ESCOs on carbon dioxide emissions

Abstract

Proponents of energy service companies (ESCOs) argue that these firms provide a crucial instrument for delivering improved energy efficiency in public and private sectors, thus contributing to carbon dioxide (CO₂) emission reduction around the world. Do ESCOs reduce CO₂ emissions? To answer this question, we develop an estimating equation, which approximates the IPAT model, from a simple model of production. We estimate a dynamic panel of 129 countries over the period 1980–2007 to show that the ESCOs effectively reduce CO₂ emissions and that this effect increases over time. These findings also prove robust to the inclusion of a set of control variables, different dates of the first ESCO and the Kyoto Protocol. Finally, we discuss energy policy implications.

Keywords:

• energy service companies
• carbon dioxide emissions
• IPAT
• dynamic panel model

JEL Classification:

• Q55
• Q56

Acknowledgements

Professor Fang gratefully acknowledges financial support from the National Science and Technology Program–Energy (NSTPE), National Science Council, Taiwan.

Notes

1. ¹ For Argentina and Philippines, we adopt the mid-point of the 1900s, 1995, as the year of the first ESCO. We adopt 1993, the mid-point of Germany’s range from 1990 to 1995, as the starting year. Italy’s range equals the early 1980s, which we translate into a starting year of 1983. Finally, for Hungary, we adopt 1990 as the starting year, since its range runs from the late 1980s to the early 1990s.

2. ² Data downloaded from http://unfccc.int.

3. ³ Data downloaded from http://data.worldbank.org/data-catalog The data on CO₂ emissions and total energy use originally come from the Carbon Dioxide Information Analysis Center of Oak Ridge National Laboratory and the International Energy Agency (IEA), respectively.

4. ⁴ For estimation, we use the Stata command xtabond2 (see Roodman, 2009).

5. ⁵ This model resembles the specification that Poumanyvong and Kaneko (2010) use in their study, except ours adopts a dynamic specification.

6. ⁶ Our data on ESCOs only identify the year or range of years in which ESCO operations initiated in each country. We do not know exactly the size of these operations and/or how these activities changed over time after the initial adoption. Thus, we interact the ESCO dummy variable, since the initial permitting of ESCO activities, to examine the ESCO effect over time.

7. ⁷ Models 8, 9 and 10 use the Vine (2005) dating of ESCO whereas Models 11, 12 and 13 correspond to Models 8, 9 and 10, respectively, but use the modified dates on ESCO as reported by Kiss et al. (2007).

8. ⁸ The insignificant estimate in Model 7 also appears in Martinez-Zarzoso et al. (2007) and Martinez-Zarzoso and Maruotti (2011) when they use the difference GMM estimator for European countries and developing countries, respectively.