Cost of capital

A common misperception is that drug development companies should be viewed as risky

investments due to the high risk rooted in each single project’s ability to get successfully through the development phases. And from looking at the often high failure rates discussed in section 2.1.4 it is easy to believe this and discount projects with a higher discount rate. However, the risk associated with a specific project or company is diversifiable and according to financial theory should not be accounted for in the cost of capital (Ross, Westerfield & Jordan, 2006, p.408). The reason is that an investor can hold a number of different projects or companies in his portfolio and the asset-specific risks should in theory cancel each other out. Hence an investor should not be paid for taking unsystematic risk.

In this thesis we use the opportunity cost of capital as the discount factor when discounting the cash flows. We use two different costs of capital, one relating to the R&D costs and another relating to the expected revenue of the project to obtain results that are more theoretically

correct. With the use of two different costs of capital it becomes possible to deal more accurately with the specific risks when discounting different kind of cash flows.

In terms of the R&D costs of a specific project continual investments have to be made

throughout its entire development process. The cash flows of these costs have a relatively low risk as they will only occur if there are good prospects of a successful market launch. They are

actively managed by the management and therefore have a relatively low risk. With the cash flows relating to the R&D costs being known and secure they should be discounted with a rate reflecting such cash flows. We argue for the use of the risk free interest rate, which we will determine in section 6.1.1.

When discounting the expected revenue we use another cost of capital. For practical reasons the company’s WACC is chosen (Brandão & Dyer, 2005, p.23). The WACC resembles the rate of return that the company needs in order to give the owners their required return subject to the embedded risk. The operating cash flows are subject to market risk and these cash flows should therefore be adjusted by an appropriate discount rate that incorporates the same risks as the WACC does.

Using the WACC implies the use of the CAPM-model as mentioned in section 1.4.2. Hence a number of assumptions are accepted that are fundamental to the model such as government bonds being risk-free and lending and borrowing at the same interest rate. These assumptions are not completely fulfilled as no government bonds are 100% risk-free and borrowing rates are typically higher than lending rates (Brealey, Myers & Allen, 2006, p.197). Many assumptions of this model are questionable but as there are few alternatives, we intend to use the model bearing in mind the above-mentioned.

As discussed in section 2.1.5 biotech companies generally have the same capital structure, which is close to being 100% equity financed. It is clear that with such a capital structure the cost of equity has a significant influence on the value of the final WACC. This distinct capital structure makes the job of computing the WACC easier as a 100% equity ratio, which is used by most industry analyst²⁰, implies that there is no need to estimate the return on debt as the debt ratio is 0.

The components used in the CAPM-model and consequently to estimate the WACC will be discussed and estimated in section 6.1.

As the WACC is a company-specific discount rate one could argue whether it is suitable to evaluate individual projects that might have another risk profile than that of the entire company

(Brandão & Dyer, 2005, p.23). Especially the estimation of the beta to compute the return on equity requires some discussion about the theoretical correctness of this application.

The classic approach to finding a beta is to do a peer analysis where peers with the same

systematic risk characteristics are used to estimate the beta. However, it is practically impossible to find beta values for a single project as no or very few listed companies only have one project in their pipeline portfolio. So to be able to find a beta value we must relax our assumptions a bit and look at companies which resemble the characteristics (the systematic risk) of our project but are not exact matches. This restriction on the benchmarking material leads to the use of biotech companies with no or only little sales as the peer group. Including pharmaceuticals in the peer group would be incorrect as the production and sales activities would lower the beta of the company compared to higher beta R&D activities just as the more mature state of a

pharmaceutical would imply a more debt-heavy capital structure.

To obtain an even more precise estimate of the beta we could choose to only look at biotech companies that research in the same therapeutic area as our project. Also the use of an equity index benchmark including the relevant sales regions could enhance the exactness of our beta estimate. However as the main focus is on the valuation methods and not on the value of the project, we will conduct a simpler estimation process of the beta which will be discussed in section 6.1.2.

From the discussion above it is clear that the WACC is not optimal for finding a discount rate as it will only be an approximation. But with hardly any alternatives we accept the error margin involved in using the company WACC for a single project. For internal decision makers in a biotech company the estimation of the cost of capital might be more accurate as they have access to more information than an external analyst.

The use of two different costs of capital is recommended by several authors dealing with the valuation of development projects. For instance this practice is applied by Datar and Mathews in their article A Practical Method for Valuing Real Options: The Boeing Approach (Datar &

Mathews, 2007). They argue strongly for the use of two costs of capital. The somewhat same argumentation is seen in the work of Collan, Fullér and Mezei (Collan, Fullér & Mezei (1), 2009) who also make use of two different costs of capital.