I use micro-level panel data on six rural communities in Ethiopia to study the impact of these reforms on growth and poverty over the period 1989 to 1995. Locally, growth outperformed the average growth rate in GDP. Using an absolute measure of poverty, I find that overall poverty decreased relatively strongly. Nevertheless, poverty remains high in most communities. A significant number of households saw their welfare decrease in this period as well. Some even experienced a move into poverty in this period. I ask whether the changes in welfare and poverty can be explained by reforminduced higher returns to physical and human capital. An alternative explanation would be that poverty decreases are superficial, for example, that poverty gains are simply related to better weather in some communities. To do this, I use a profit function framework to explain growth using prices and endowments of land, labour, human capital and location characteristics, while controlling for shocks. Next I develop a regression-based decomposition of the changes in the poverty gap. It is found that common and idiosyncratic shocks matter, but that the main factors driving consumption changes are relative price changes, resulting in changes in the returns to land, labour, human capital and location. A detailed discussion of the factors that may have contributed to these changes supports the view that reforms have helped. Producer prices increased in most areas, while incentives for market-based activities contributed as well, the latter most likely aided by increased security. Locally and against the national trend, the agro-climatic conditions were not much better in the five preceding years than before, while in some villages, the most recent rains in 1994 were rather bad. The poor have benefited on average more from the reforms than the non-poor households, suggesting that the reforms have been rather pro -poor. But the experience of the poor is mixed. One group of the poor in 1989, with relatively good land and labour, as well with access to roads and towns, strongly outperformed all other households, while experiencing the best weather as well. Another group with much poorer endowments also faced poor rains and smaller producer price increases, resulting in virtually unchanged and persistent poverty for this group.