In Magni [Eur. J. Operat. Res. 137 (2002) 206] I present some inconsistencies implicit in the net-present-value criterion, as currently used in finance. This paper shows that the standard use of CAPM for capital budgeting, based on disequilibrium values, is at odds with arbitrage theory, and that the corresponding CAPM-based NPV rule is meaningless even in the simplest case, because net present value is any number one wants it to be. Cognitively, this amounts to saying that the NPV procedure leaves decision makers subject to a framing bias; financially, this amounts to saying that additivity does not hold. De Reyck's [Eur. J. Operat. Res. 161 (2005) 499] objection to my thesis is invalid because he mistakes a project's expected rate of return for a project's cost of capital. De Reyck's Proposition, on the basis of which my thesis is criticized, leaves decision makers open to arbitrage losses, because it is an (admittedly interesting) reframing of the security market line and (as surprising as it might be) the use of the SML for project valuation is incompatible with the no-arbitrage principle. To use NPV and CAPM for capital budgeting is not a good idea.