The World Bank, the United Nations, and virtually all economists agree that economic growth is essential to reduce poverty in developing countries (UNDP 2001; World Bank 2001b). Yet economic growth reduces poverty only if it is broad based; economic growth that does not change the incomes of the poor cannot reduce poverty. More generally, increased inequality implies that the incomes of the poor grow at a slower rate than the incomes of the rest of the population. Some economists claim that economic growth sometimes leaves the poor behind (see a discussion of this in Winters, McCulloch, and McKay ). Such claims can be resolved only by empirical analysis. This article investigates this claim for the case of Vietnam. Vietnam's rate of economic growth in the 1990s was one of the highest in the world. The average annual rate of economic growth from 1990 to 2000 was 7.9% (World Bank 2002). This high rate of economic growth led to a sharp reduction in poverty; estimates based on the 1992-93 Vietnam Living Standards Survey (VLSS) show a poverty rate of 58%, while 5 years later estimates based on the 1997-98 VLSS indicate that poverty had dropped to 37% (World Bank 1999). Such a sharp drop in poverty in only 5 years is an achievement that is rarely seen in any developing country, and Vietnam's continued economic growth since 2000 has continued to reduce the poverty rate, which had fallen to 16% by 2006 (General Statistics Office 2007).by increased inequality. The Gini coefficient for household per capita expenditures rose from 0.33 in 1992-93 to 0.35 in 1997-98. This increase may seem small, but it suggests that the poor in Vietnam benefited less from economic growth than did other households. This article uses panel data from the 1992-93 and 1997-98 VLSS to investigate whether Vietnam's growth has been "pro-poor." It divides the population into five groups-the poorest 20%, the next poorest 20%, and so on, up to the wealthiest 20%-and estimates growth rates for each group. One important issue is whether one should compare households who constituted the poorest 20% of the population in 1992-93 with the poorest 20% of the population in 1997-98, some of whom may not have been among the poorest 20% in 1992-93, or whether one should compare the poorest 20% in 1992-93 with the same households in 1997-98, some of whom are no longer among the poorest 20% of the population. A second important issue is more methodological in nature: household expenditures are likely to be measured with error, which can lead to serious biases, especially for the second type of comparison. The rest of this article is organized as follows. Section II discusses Vietnam and the data, Section III explains the methodology, and Section IV gives estimates of changes in per capita expenditures among different groups of Vietnamese households, using both types of comparisons and presenting estimates that ignore measurement error and estimates that correct for measurement error. A final section summarizes the findings.