We derive and test q-theory implications for cross-sectional stock returns. Under constant returns to scale, stock returns equal levered investment returns, which are tied directly to firm characteristics. When we use generalized method of moments to match average levered investment returns to average observed stock returns, the model captures the average stock returns of portfolios sorted by earnings surprises, book-to-market equity, and capital investment. When we try to match expected returns and return variances simultaneously, the variances predicted in the model are largely comparable to those observed in the data. However, the resulting expected return errors are large.