||The project focuses on common factors driving conditional quantiles of returns and volatility of assets in (possibly large) portfolios. We hypothesize that these factors depend not only on the sector and region, but also on the investment horizon. In that case, the standard analysis of risk-return relationship can be significantly improved by decomposing the risk into its short-, medium- and long-run components. We hypothesize that effects of these components on the conditional quantiles are not the same and that their absolute and relative importance vary across quantiles, industries and over time. On the other hand, we expect some common patterns within industries and, especially in times of increased systemic risk, even across industries. Identification of such patterns would help us to better understand the sources and dynamics of systemic risk, which could have important implications not only for effective portfolio diversification, but also for financial markets regulation assessment. To test our hypotheses, we develop a methodology based on panel quantile regressions of asset returns on volatility that is decomposed using wavelet transform. First, we analyze the results for single assets from several industries; then, we move to panel data, which should enable direct testing of the presence of common factors.