Efficientmarket hypothesis  Wikipedia
The efficientmarket hypothesis ..
A Null Hypothesis for the New Year  Flirting with Models
Arthur D. Roy (1952) demonstrates that the would not choose to hold the closedend weight of an asset in a meanvariance efficient fund's NAV in a portfolio that contains the ten portfolio is equal to the expected excess return openend funds. In this case, the ratio of the of the asset conditioned on the excess return intercept to the residual variance will be zero. of all other available assets being zero, divided From the regression estimation used to pro by the variance of the asset's return condi duce Table 1, the portfolio weight measure was tioned on all other available asset returns. constructed for all 52 funds. The average (me This reasoning can be used to ascertain dian) measure is 0.009 (0.025). The tvalue whether closedend fund NAVs are redundant under the null hypothesis that the average measure is zero, is 0.67, implying that the average closedend fund NAV is redundant given the ten openend funds. A test of inter cepts is similar. The average (median) inter cept is 0.0004 (0.0010). The tvalue under the null hypothesis that the average intercept is zero, is 0.44.
Both John H. Cochrane (1991) and G. William Schwert ( 1991), in reviews of Shiller's work, stress the importance of the joint hypothesis of excess volatility and a timevarying discount rate model. For example, many variancebounds papers assume constant expected returns and, hence, constant discount rates through time. Even if a test allows time variation in expected returns, a variancebounds test might reject the null hypothesis if the expected return process is misspecified. The test conducted in my paper assumes that the discount rate of the fund is equal to the discount rate of the fund's portfolio. Thus, a model of timevarying discount rates is unnecessary.
MUTUAL FUNDS; QUBE PORTFOLIOS; THE WEIRD ..
Inferences from variancebounds tests also are sensitive to the inclusion of different types of shareholder distributions. For example, most variancebounds tests use dividends as the relevant cash distribution. Thus. other dis tributions to shareholders, such as cash gen erated from share repurchases and takeovers, are not considered. Although variancebounds tests usually reject the null hypothesis of no excess volatility when dividends are used as the relevant distribution, Ackert and Smith (1993) fail to reject the null when a broader definition of distribution is used. Problems with distributions are avoided with volatility tests that use closedend funds, since they fa cilitate the comparison of two marketbased prices.
Performance hypothesis testing with the Sharpe ratio: …
Statistical Modelling of the Returns of Mutual Funds

excess returns as a proxy for explaining ..
Efficient Market Hypothesis ..

What is Efficient Market Hypothesis (EMH)
It concludes that excess returns cannot be achieved using technical analysis

The efficient market hypothesis is also known by its acronym EMH
Importance of Efficiency Market Hypothesis  UK Essays