Focusing on nonstationarity of the time series of CO2 emissions, weanalyze the relationship between CO2 emissions and macroeconomicvariables in the framework of an error correction model. To mitigate lowpower and size distortion of unit root tests and cointegration tests in asmall sample, various testing methods are employed. Cointegration testresults suggest a relatively robust long-run relation among CO2emissions, income, technology and energy price. We also find thatenergy mix and final energy consumption mix play important roles inexplaining short-run CO2 emission changes. The nuclear energy and therenewable energy contributes to mitigation of CO2 emissions, whereastransportation sector’s energy consumption is associated with anincrease of CO2 emissions.