Growth Accounting is the empirical methodology that allows for the breakdown of output growth into its sources which are the factors of production and technological progress, and provides estimates of the contribution of each source in output growth. The concept of total factor productivity growth (TFPG) which is central in growth accounting, measures the part of output growth which is attributed to technological progress, and which corresponds to the part of output growth not accounted for by factors of production such as capital or labour. Growth accounting still remains a central concept in growth theory, although there are still conceptual disputes about the subject, and Easterly and Levine (2001) state that "economists need to provide much more shape and substance to the amorphous term TFP". In this paper we try to provide some additional "shape" by considering the use of environment as a source of growth.
It was Solow in the late 1950's, (Solow, 1957) who provided an explicit integration of economic theory into the growth accounting calculations, which imply decomposing total output growth and measuring the contribution to growth of specific factors, including that of technological progress. During the last decades many di§erent approaches have been used to measure TFPG, which include dual approaches using mainly factor prices instead of factor quantities, and approaches which basically involve disaggregations and refinement of inputs in the production function.
In the early 1970's, a new dimension was given to the theory of economic growth with the introduction into growth models of environmental damages created by emissions. This new dimension which has generated a large volume of literature on "Growth and the Environment", implies a new way of looking at TFPG measurement. Brock (1973) stated that "received growth theory is biased because it neglects to take into account the pollution costs of economic growth". This is because in an unregulated market the cost of pollution is not internalized. Pollution in this case is an unpaid factor of production, with production becoming more costly if less pollution is allowed. In this context environment is used as a factor of production which is not fully compensated, and its use in the production process can be captured by introducing emissions as an input in an aggregate production function.
Following this methodological approach, the idea developed in this paper is that when emissions are introduced as an input in the production process and are properly measured, the contribution from the use of en- vironment in total output growth can also be measured. This contribution can be approximated even when emissions is an unpaid factor in the absence of environmental policy. In this sense, emissions can determine, along with other inputs and technological progress output growth in a growth accounting framework. Therefore, the present paper can be regarded as an attempt to explore systematically whether the use of the environment as an input in production contributes to output growth, and how this contribution can be measured.
We develop a growth accounting framework for measuring TFP growth by approximating the use of the environment by carbon dioxide (CO2) emissions. We argue that environment such as the atmosphere can be regarded as a component of social overhead capital (Uzawa, 2003), and that CO2 emissions can be thought of as a reduction of this social capital - a form of disinvestment. Thus, we use CO2 emissions as a proxy for the use of this component of social capital in the production process. Our purpose is to examine the contribution of CO2 emissions' growth, as a proxy for the use of environment, on economic growth and to show that since external pollution costs which are created during the production process are not taken into account in the measurement of total factor productivity growth, the current measurements of TFP growth, or the Solow "residual", could provide biased results. Our basic hypothesis, which has been tested empirically, is that environment is basically an unpaid source of output growth. If this source is not taken into account into the growth accounting framework, then output growth which should be attributed to the use of the environment will be incorrectly attributed to TFPG. Furthermore, if emissions saving technical change is present this could be another source of growth in addition to the conventional labor augmented technical change. This hypothesis is tested empirically in this paper by using data from a panel of 23 OECD countries.