Literature has frequently evaluated the proliferation of preferential trade agreement (PTAs) by counting the number of PTAs or the share of international trade covered by these agreements. But these indicators have been recently criticized. In this paper, using a CGE model, we attempt to quantify the economic impact (measured by its welfare effect) of the proliferation of PTAs. Results show that the welfare impact of PTAs is limited and often vanishes quickly because preferences are eroded by the proliferation of PTAs. This contrasts with the gains from non-discriminatory liberalization. Therefore we argue that the proliferation of PTAs is unlikely to continue. This paper also has methodological implications for CGE modeling: ignoring the proliferation of PTAs bias significantly the results; and the bias is not only potentially large but also its direction is a priori uncertain.