Recent studies have documented that various factors such as discretionary accounting accruals, underwriter reputation, venture capital backing, and firm size will affect the long-run performance of IPOs. However, it is not clear whether the return predictability of these attributes are the manifestation of one phenomenon, or independent results. In this study, we use univariate and multivariate analyses on these factors to trace the sources of return predictability. We find that these previously identified effects are not the same phenomenon, though correlated to some extent. The results show that a confluence of these determinants is more important than any individual factor in explaining the IPO long-run performance. We also identify a subset of IPOs that outperform their benchmark and another subset that consistently underperforms.