We introduce a single measure for testing market integration, contagion and malfunction simultaneously based on an efficient market hypothesis (EMH) network. The U-statistic whose value and distribution vary depending on the underlying market conditions can be consistently utilized for calculating p-values for given market conditions under the ideal EMH network (or random walk model) as benchmark null. An application of the U-statistic to globally integrated stock markets finds that emerging markets follow the ideal EMH more faithfully than developed markets. Our network approach based on the U-statistic resolves the intrinsic problem of analysing correlation increase due to volatility increase during financial stress periods when testing contagion.​

Keywords: U-statistic, market integration, contagion, efficient market hypothesis network.