The Linder hypothesis has attracted substantial empirical research over decades. However, the evidence has failed to provide consistent support for it. This paper explains the failure. Building a theoretical framework in which, as in Linder's theory, product quality plays the central role, I show that the Linder hypothesis is formally derived but holds only when formulated as a sector-level prediction. This prediction is then estimated using a sample of 64 countries in 1995. The results support the sectoral Linder hypothesis: controlling for the effect of intersectoral determinants of trade, countries of similar income per capita trade more intensely with one another.