This article suggests how state enterprises can be incorporated into the theoretical and empirical growth literature. Specifically, it shows that if state enterprises are less efficient than private firms, invest less, employ less skilled labor, and are less eager to adopt new technology, then a large state enterprise sector tends to be associated with slow economic growth, all else remaining the same. The empirical evidence for 1978-92 indicates that, through a mixture of these channels, an increase in the share of state enterprises in employment by one standard deviation could reduce per capita growth by one to two percentage points a year from one country to another.