We used unit root and cointegration techniques to determine the long run relationship between GDP and investment for 90 countries using data from World Bank for the period 1960-1992. In the first step of our analysis we found GDP and investment integrated of different orders for 33 countries. Second step of our analysis shows no cointegration between GDP and investment for 25 countries and cointegration for 25 countries with both variables of order I(1). The other 7 countries with both variables of order I(0) are in long run relation and do not need cointegration test. To determine the direction of causal effect between GDP and investment we used Granger causality test as the third step of our analysis. We found causality in the short run for 15 countries and in the long run for 23 countries. Bi-directional causality is found for 10, unidirectional causality from GDP to investment for 18 and from investment to GDP for 10 countries. The causality from GDP to investment is positive for 11 countries and from investment to GDP for 6 countries. Bi-directional causality is mostly positive between the two variables.