This paper estimates the real gross domestic product (GDP) and unrecorded economy for Turkey using the Kalman filter technique. Using different tests, most of the research articles on energy policy investigate the causal relationship between energy consumption and GDP for different countries. On the other hand, other studies on climate change try to show the effects of both energy consumption and GDP on carbon dioxide (CO2) emission. Since the unreported economy has an important weight in developing countries where the recorded (or official) GDP suffers from considerable measurement problems, investigation of the relationship between the recorded GDP and energy consumption may lead to biased results. In this paper, the economic variables (GDP, country population) as well as environmental variables (CO2 emission, forest area) are used in order to estimate GDP, which is an unobserved variable in our model. The results clearly indicate that: first, the true GDP in Turkey, that our model estimates, is higher than the observed (recorded) GDP in the whole period of observation (1973-2003) and the size of unrecorded economy varies between 12 and 30 percent of the observed GDP; second, the gap between the true GDP and the observed GDP has an increasing trend; third, if the change in GDP per primary energy supply is smaller than the change in CO, per primary energy supply, then there may exist unrecorded economy. (C) 2007 Elsevier Ltd. All rights reserved.