The purpose of this study to see if there is a relationship heterocedasticity test and cointegration test on the use of economic variables such as GDP and Gov. Expenditure. In this study used regression analysis and cointegration to get the proof of their relationship to the occurrence heterocedasticity co integration of data in the long term. Of the studies that have been performed concluded that, with the form data that is homocedasticitas used regression models 1, 2 and 3 and not heterocedasticity on regressions 4 and 5, the data on economic variables GDP and Gov. Expenditure in 194 countries in 2014 cointegrated.