The paper studies the persistence of the high-growth phenomenon in Austria using social security data for the years 1985 to 2007. The Eurostat-OECD definition is used to identify high-growth firms and a modified Birch Index to identify high-impact firms. Bringing the definitions to the data confirms that the two definitions lead to the selection of different firms. We use matching as non-parametric preprocessing and estimate survival, persistence and growth regression on balanced datasets. The results show that being a high-growth firm does not improve the likelihood of survival in future periods in excess of the size effect induced by the high-growth event. For persistence and high growth we find a high-growth firm treatment effect. For high-impact firms we find a significant treatment effect for survival, persistence and growth. High-impact firms show a much higher persistence than high-growth firms. The average growth rate after the high-growth episode is quite modest for both high-impact and high-growth firms. Policy implications of the findings are discussed.