* Corresponding author. Tel.: #91-33-4125-7382; Fax: #91-33-412- 7905. E-mail address: jroy@cal2.vsnl.net.in (J. Roy) Energy Policy 28 (2000) 433}438 The rebound e!ect: some empirical evidence from India Joyashree Roy* Department of Economics, Jadavpur University, Calcutta 700 032, India Abstract The present paper looks at the e!ect of technical e$ciency gains on energy use in three sectors in India, at varying levels of aggregation. High positive rebound implies that e$cient technology, unless supplemented by appropriate pricing policy, will not be successful in containing demand. However, the main purpose of the paper is to focus on the unique mechanism of rebound that may occur in developing countries with unmet demand.  2000 Elsevier Science Ltd. All rights reserved. Keywords: Rebound e!ect; Unmet demand; Developing country; Climate change; Rural lighting; Technology vs. price policy 1. The issue Rebound e!ect, take back or back"re, essentially re- cords the behavioral response to energy e$ciency im- provements. Since 1980 (Khazzoom, 1980), quite a large number of studies have focused on the theoretical logic and empirical evidence on the rebound e!ect especially in the context of developed countries (Schipper and Grubb, 1998; survey available in Greening and Green, 1998). The rebound issue has gained an additional dimension in the context of global climate change (Kainuma et al., 1999) because of the suggestion of reduction in energy use as a win}win option (Jepma and Munasinghe, 1998). The debate is relevant especially for policy matters: relative e!ectiveness of technological e$ciency improvement ver- sus energy price or tax policy in reducing energy use. The objective is to contain the otherwise growing pattern of energy demand, now a very important component of environmental policy. The e!ectiveness of this policy in meeting the targeted objective would ultimately depend on the rebound e!ect (Grubb, 1990, 1992; Brookes, 1990, 1992). If the rebound e!ect is positive, then the bene"ts from savings in energy demand arising out of the tech- nical e$ciency improvement, and hence reduction in GHG emissions, are going to be taken back through increase in demand, depending on the magnitude of the e!ect. The theoretical logic for a positive value of re- bound is now well established in the literature despite some need for re"nements (Greening and Green, 1998). In the context of climate change policy, the need is more for empirical evidence of rebound for various regions and sectors. These behavioral response parameters are essen- tial for the integrated assessment modelers also to derive the region-speci"c implication of mitigation options. From the policy perspective, the magnitude of the rebound e!ect is of interest (Greene, 1992). Technical e$ciency improvement in energy use would imply im- provement in productivity of the input energy. Improved technology can produce the same output with less en- ergy. It may thus lead one to conclude that the control variable needs to be the e$ciency of technology in use if the objective is to reduce energy consumption. However, the arithmetic is not that simple nor is the task of the policy makers. What complicates the policy-making pro- cess is the whole range of behavioral responses of the end-users that follow any technical e$ciency improve- ment all of which may, however, not be traced empiric- ally. If a large share of bene"t arising out of technical e$ciency improvement is taken back by behavioral re- sponse parameters then technological solutions `onlya cannot take us to the goal of reduction in fossil fuel consumption. Several supplementary policies would then become necessary. 2. Developing country perspective The issue is of no less if not more relevance from the developing country perspective. While developing coun- tries' share of fossil fuel consumption is presently small, 0301-4215/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved. PII: S 0 3 0 1 - 4 2 1 5 ( 0 0 ) 0 0 0 2 7 - 6