Institutional change has taken place gradually since 1978 for State-Owned Enterprises (SOEs) in the Industrial Sector of China. In this paper we estimate the effect of deep reform (the right to hire and fire labour, buy and sell capital and operate on international markets) on the productivity dynamics of enterprises. Using a unique balanced panel of 681 SOEs for the period 1980 to 1994, we find consistent production function estimates using an algorithm put forward in Olley and Pakes (1996), which corrects for simultaneity bias. Furthermore, we allow selection to reform to be endogenous, and correct for this selection bias by formulating an entry rule to reform similar to the Olley and Pakes (1996) exit rule. We show that exposure to deep reform have lead to higher productivity realisations while remaining under state ownership.