Earning quality as an investment signal has been popular among equity portfolio managers for the last decade. The basic idea behind this accruals anomaly is that stocks with high and increasing accruals tend to have low earnings quality, while stocks with low and decreasing accruals tend to have high earnings quality. The earnings quality signal stopped working in the mid-2000s, but has staged a remarkable rebound since the end of 2008. The authors evaluate whether earnings quality is a true alpha signal, a risk factor, or both. They find that, in the periods when the signal worked, the strategy was largely driven by stock selection, suggesting that earnings quality is indeed an alpha signal. The authors also find that earnings quality is not a good risk factor, in that it does not have high statistical significance when regressed cross-sectionally on returns, along with other well-known risk factors, and is not very volatile over time. Overall, their results indicate that earnings quality may be that rare example of a pure alpha factor.