In this paper, we aim to assess the gender and ethnic earnings gaps in Vietnam by using the Labour Force Survey
in 2007. We show that in general, ethnic earnings gap is higher than gender earnings gap and the gaps are very
heterogeneous amongst different ethnic minority groups. Then, results from the multinomial logit model used to
correct for selectivity indicate that the gaps depend significantly on the sector choice: the gender earnings gap is
observed low in public sector while the ethnic earnings gap is widened in agriculture sector and reduces
significantly in public and formal sectors. Quantile regression result suggests that the gender earnings gap is
lowest at the bottom quantiles but does not vary much across the other quantiles while ethnic earning gap is
higher at the lower quantiles of the distribution and getting narrower at the higher level of the outcome
distribution. Our next aim is to assess how much of these gaps can be explained by individual characteristics and
their allocation between different sectors. We use the well-known Oaxaca Blinder and Neumark method to
decompose the gaps and the extended Appleton’s full decomposition approach to take into account selectivity in
sectoral allocation. Our findings show that differences in characteristics and sectoral allocation play an important
role in explaining the ethnic earnings gap but it is not the case for gender earnings gap.