On the importance of clean accounting measures for the tests of stock market efficiency

Tests of the semi-strong form of the efficient market hypothesis (EMH) typically use earnings and book value of equity as benchmarks of fundamental value. Accounting earnings, however, are contaminated by noise due to their transient component and book value of equity tends to be biased downwards due to accounting conservatism. We investigate whether controlling for these effects impacts on the implications concerning the information efficiency of the Swedish stock market. We conclude that relevant adjustments increase both the magnitude and the consistency of the value premium earned on a contrarian investment strategy that buys (shorts) stocks with low (high) relative market valuation. The existence of the value premium cannot be explained by common risk proxies or transaction costs argument. Using cleaner accounting proxies thus strengthens the evidence on the imperfect efficiency of the Swedish stock market.