Using a CCAPM-based risk-adjustment model consistent with general asset pricing theory, we perform yearly valuations of a large sample of stocks listed on NYSE, AMEX and NASDAQ over a thirty-year period. The model di§ers from standard valuation models in the sense that it adjusts forecasted residual income for risk in the numerator rather than through a risk-adjusted cost of equity in the denominator. The risk-adjustments are derived based on assumptions about the time-series properties of residual income returns and aggregate consumption rather than on historical stock returns. We compare the performance of the model with several implementations of standard valuation models, both in terms of absolute valuation errors, and in terms of the returns on simple investment strategies based on the di§erences between model and market prices in the respective valuation models. The CCAPM-based valuation model performs substantially better than the best performing standard valuation model when comparing absolute valuation errors. Both types of models are able to identify investment strategies with subsequent excess returns but also in this setting, the CCAPM-based valuation model outperforms the standard valuation models in most dimensions. We further show that the standard CAPM and the Fama-French three-factor based approaches to risk-adjustment substantially overestimate the cost of risk. This ìerrorîmore than o§sets yet another ìerror,îwhich is committed when using analystsíforecasts of long-term growth, which are three to four times higher than what can be considered to be empirically reasonable. Using the CCAPM-based approach to risk-adjustment in the numerator, the results are consistent with investors being very conservative in their valuation of long-term value creation but also very conservative in risk-adjusting future value creation ; Using a CCAPM-based risk-adjustment model consistent with general asset pricing theory, we perform yearly valuations of a large sample of stocks listed on NYSE, AMEX and NASDAQ over a thirty-year period. The model differs from standard valuation models in the sense that it adjusts forecasted residual income for risk in the numerator rather than through a risk-adjusted cost of equity in the denominator. The risk-adjustments are derived based on assumptions about the time-series properties of residual income returns and aggregate consumption rather than on historical stock returns. We compare the performance of the model with several implementations of standard valuation models, both in terms of absolute valuation errors, and in terms of the returns on simple investment strategies based on the differences between model and market prices in the respective valuation models. The CCAPM-based valuation model performs substantially better than the best performing standard valuation model when comparing absolute valuation errors. Both types of models are able to identify investment strategies with subsequent excess returns but also in this setting, the CCAPM-based valuation model outperforms the standard valuation models in most dimensions. We further show that the standard CAPM and the Fama-French three-factor based approaches to risk-adjustment substantially overestimate the cost of risk. This error more than offsets yet another error, which is committed when using analysts' forecasts of long-term growth, which are three to four times higher than what can be considered to be empirically reasonable. Using the CCAPM-based approach to risk-adjustment in the numerator, the results are consistent with investors being very conservative in their valuation of long-term value creation but also very conservative in risk-adjusting future value creation.