This study re-examines the issue of causality between investment shares and economic growth. A methodology is applied based on Arellano and Bond (1991), and Holtz-Eakin, Newey and Rosen (1988) to quinquennial panel data on growth and investment shares for the post war period and shows that, contrary to previous results in the literature, causality between fixed investment and growth runs in both directions. Investment shares Granger-cause growth rates and growth rates Granger-cause investment shares. Granger causality from investment shares to growth rates is found to be negative. The result is in contrast with a capital fundamentalist view which sees fixed investment as the key to long run growth, but is fully consistent with the predictions of Solow-type growth models.