This paper describes mean and volatility transmissions between the Phelix (Germany) and Nord Pool (Scandinavia) energy auction markets. The main contribution of the paper is improved knowledge of the two markets’ mean and volatility interferences. A multivariate semiparametric GARCH model (seminon-parametric [SNP]) is optimally and sequentially evolved and used to identify sources and magnitudes of mean and volatility in physically interconnected (limited capacity) electricity markets with separately determined auction market prices. The two mean equations show strong negative serial correlation and positive mutual cross-correlation. The cross mean correlations are, both in size and in significance, dominated by the Nordic market (NordPool). The multidimensionalSNPpolynoms report non-Gaussian features and interactions. The volatility reports mean reversion and strong serial correlation (clustering). The cross volatility correlation is mainly found in the leverage and level BEKK coefficients.