Skip to main content

Authors: Prabesh Luitely, Piet Sercuz, and Tom Vinaimont

Consider a standard discounted-cash- ow problem. For underdiversifed entrepreneurs
i, the cost of capital (CC) should reflect both the project's relatively large size and its non-scalability, which rules out the CAPM and its `total-beta' version. A correct CC is derived in
the usual `return-on-value' form. Plugging-in observed returns from similar-looking listed  rms
introduces inconsistencies and bias; even ROIs do better, as they at least reveal the sign of the
NPV. An analytical valuation, relying on dollar inputs, is available for one-period problems, and
can either be used sequentially or complemented by a simple MonteCarlo procedure (internet
appendix). All this is done for both the single-asset case and provided the investor somehow
manages to unite traded-asset values with the project's PV-ed income into a joint distribution|
the multi-asset case. Empirically, cash flows from small, non-listed  rms are almost uncorrelated
with the market return, which induces an implausibly low CC under the multi-asset view. The
single-asset CC does deliver reasonable CCs.

Subscribe