Introduction
In the scientific community, the ‘valley of death’ refers to the funding gap commonly observed during the attempted transition between research and commercial development [1]. As many as 80-90% of research projects fail during this stage due primarily to the challenge of attaining funding [2]. Researchers often find it impossibly challenging to concurrently prove the viability of their scientific work outside the lab and demonstrate its potential impact.
To illustrate the difficulties of navigating this chasm, let's consider two case studies. The first involves Dr. Katalin Karikó, who spearheaded the development of messenger RNA (mRNA) technology, underpinning COVID-19 vaccines today. Starting in 1989, she held a professorship at the University of Pennsylvania and focused on mRNA research. However, she could not win scientific grants to fund her work and never made more than $60,000 a year [3], leading to her demotion in 1995. Luckily, in 1997, an immunologist named Dr. Weissman joined UPenn and, with his funding, the two began collaborating on mRNA research. And in 2005, Karikó and Weissman showed that mRNA could be used for vaccine development. This breakthrough led to the formation of Moderna and BioNTech, which have a total market capitalization over $65B today [4].
The second case involves my experience in developing ‘ElectroDAR’ between 2015 and 2019 at Hatch. As part of a small R&D group (about 300 employees) within a private company, we faced constant financing constraints. The firm’s shareholders were undiversified, resulting in aversion to risky projects. Consequently, funding came primarily from end customers, who were reluctant to be early adopters, and a smattering of grants. This prolonged development took over four years, when a more efficient timeline could have been achieved in under two years.
This article delves into the challenges that early-stage research encounters in crossing the 'valley of death.' Then, it compares valuation methods for research projects, using the previously mentioned case studies as examples. Finally, the article advocates for the adoption of real option-based methods as a potential solution to stimulate funding for impactful ideas and technologies.
Why is early-stage research difficult to fund?
Funding early-stage research projects proves particularly challenging due to the exceptionally high uncertainty (or in finance speak, idiosyncratic risk). This uncertainty, encompassing both technical and commercial feasibility, is a fundamental consideration for all three of the primary funding sources: grants, venture capital, and industry:
- Grant funding, typically provided for specific, well-defined projects, is predominantly allocated by large government institutions, including the National Institutes of Health ($45B), the National Science Foundation ($9B) and the Defense Advanced Research Projects Agency ($4B). In total, about $46B dollars was spent in 2020 by the US government specifically on “applied research” (research targeted at a practical aim or objective) [5]. The median National Science Foundation grant in 2023 was $300,000 over 3 years [6]. Grants deal with uncertainty primarily through peer review (80% to 90% of proposals) as well as staged awards. However, the need for consensus in peer review biases evaluations against unorthodox ideas [7]. As a result, scientists often apply for grants on the basis of what they believe is fundable, rather than with their best ideas [8].
- Venture capital (VC) funding is a crucial form of equity financing for startups and spinouts based on scientific research. About $15B was raised by angels and VCs for pre-seed and seed (early-stage) rounds in 2023, according to data from Crunchbase. VC mitigates startup uncertainty through due diligence, negotiated deal terms, and fund diversification, resulting in the prevalent 'spray and pray' approach. Despite fund investors being well-diversified and risk-mitigation deal terms like liquidation preferences (found in over 90% of rounds), follow-on participation rights (in 65% of rounds) and ratchets, evidence suggests that idiosyncratic risk leads to substantial price reductions in VC deals because of the increased due diligence needed and the compensation structure of venture fund partners [10] [11].
- Industry, or the corporate sector, contributes to research investment through various means such as internal R&D, grants, partnerships and venture capital. US industry funding of applied research was estimated to be $78.6B in 2020 [5]. Corporate research and development decision-making, however, exhibits bias against highly risky or uncertain projects. While traditional finance theory holds that when evaluating projects, expected cash flows should be discounted against the weighted average cost of capital, several studies reveal substantial upward adjustments to discount rates (more than 5 percentage points) in response to idiosyncratic risk [12] [13]. Additionally, corporate managers face budget constraints, limited diversification or hedging opportunities, and often incur more significant penalties for failures than rewards for successes.
How might research funding become more tolerant of project risk?
Proposed solutions vary according to funding source. For grant funding, ideas include changing the peer review committee voting system or allocating money to people rather than specific projects [7, 8]. For venture capital, it is recommended that fund investors set the carried interest benchmark appropriately based on the expected return of the portfolio to encourage the desired risk-return profile [11].
For industry, new approaches are needed to value early-stage research and better align the appetite for risk between corporate managers and the shareholders they represent. Specifically, I argue here for the adoption of real option-based valuation strategies in addition to the commonly used net present value (NPV) and internal rate of return (IRR) techniques.
What is real option valuation?
A real option provides the right, but not the obligation, to pursue specific business endeavors in the future, such as investing in the commercialization of a successful research project. Valuation techniques for real options resemble those for pricing financial options contracts but focus on "real" assets rather than financial derivatives [14]. The underlying idea is when a project involves uncertainty and multiple stages of investment, it is important to consider not only the expected value of the project, but also the probability that the value may exceed initial expectations. Consequently, the valuation of real options involves setting a price to the right to investment and capture the resulting cash flows, conditional on the likelihood that the cash flows may be expected later to exceed the investment.
As a result, a project's worth can be determined using discounted cash flow (DCF) or real option methods. Discounted cash flow techniques, such as net present value (NPV), concentrate on the expected return and the downside of risk. The real options approach is unique in considering the expected return while also recognizing the potential upside for risk [15]. In the early stages of an innovative project, the NPV of a project will be low because of the low expected likelihood of future cash flows but at the same time, the real-option value will be much higher due to that same uncertainty. A Harvard Business Review article has a simple diagram which illustrates this relationship.
How can you model a research project with real option techniques?
This article introduces a set of model frameworks (password: ‘template’) built by both myself and Dr. Michael McIntyre, Professor of Finance at the Sprott School of Business. It is designed to value early-stage research projects using both NPV and real option techniques. The models, grounded in a common set of assumptions, considers a project with an extended pre-revenue research period following by a low probability of commercial success. There are two models, one in which commercialization is achieved through a second stage of capital expenditure and the other in which commercialization is achieved through licensing like at a university. I believe the assumptions used are generally conservative and, where possible, based on academic research. More details are shared in the Excel file.
To assess the model, I considered two cases: 1) mRNA research and 2) the development of a hardware device similar to ElectroDAR. In addition, I compared those results to that of a published thesis which values a small startup using real options [16]. These results are shown in the table below.
|
Case |
Key Assumptions |
DCF NPV |
Option Value |
Delta |
|
1) mRNA |
TAM: $10.8B (expectation of 0.5%) |
$-100,250 |
$3,532,986 |
$5,072,667 |
|
2) Hardware device |
TAM: $200M (expectation of 5%) |
$-635,632 |
$823,307 |
$1,458,939 |
|
3) Startup ‘Biketastic’ |
TAM: €3.2M (expectation of 12.5%) |
€ -157,296 |
€2,059,253 |
€2,216,549 |
For reference, the online copy of the model via the hyperlink above includes the mRNA and hardware device examples; TAM = total addressable market (estimated).
As can be seen, based on DCF NPV alone, it would not be recommended to fund any of these projects. In contrast, based on real option analysis, the research projects offer considerable value based on their potential upsides. The case for investing in these projects based on the option value is made even stronger when less conservative assumptions are used, especially by reducing the discount rate. The figure below shows the sensitivity analysis of the mRNA case based the discount rate and idiosyncratic volatility assumptions.
The sensitivity analysis highlights the option value's stronger sensitivity to discount rate changes than to volatility, a trait linked to the prolonged commercialization horizon. This prompts the question: can we anticipate funding organizations to lower their discount rates? Research indicates an average corporate perceived cost of capital at 8.4% (at which rate the mRNA option value becomes about $15M). Despite financial theory, at least 33% of managers justify higher rates for execution risk [16]. I argue that staging investment according to real option analysis could allow for a more modest discount rate buffer in corporate decision-making.
Why are real option approaches not used today?
In more than 90% of cases when selecting projects, corporate managers prefer more straightforward valuation strategies such as NPV and IRR, with little mention of the use of real options [12]. Several explanations have been offered, including the difficulty in finding good proxies for the input variables along with the complexity of the model [17].
However, I believe the primary reason for the limited adoption of real options lies in the practical challenges of structuring investments as options. For an investment to have options embedded, and be appropriate for real option valuation, three key questions need to be answered affirmatively [15]:
- Is the first investment a pre-requisite for the later investment/expansion?
- Does the firm have an advantage or form of exclusivity to the later investment/expansion?
- Are the competitive advantages sustainable?
In early-stage research, instances of option-based investment structures are sparse, with notable exceptions like the MIT Media Lab and Semiconductor Research Corporation. Generally, companies exhibit a preference for assuming the role of second movers in applied research, investing only after the technology or market has been sufficiently de-risked. This challenge is compounded by the time-consuming process of identifying research projects with promising commercial opportunities for the company.
Conclusion
The absence of option-based investment opportunities and limited adoption of real option valuation practices represent both a societal burden and economic opportunity. Research productivity is reported to be declining, and the ‘disruptiveness’ of papers and patents has seen an 80 to 90% decrease [7]. A recent study from the University of Chicago suggests that corporate budgeting practices, including aversion to idiosyncratic risk, can explain the puzzle of “missing investment” of over 20% of the $69T US capital stock [16].
Demonstrating the implications of missed opportunities through counterfactuals is challenging. Yet, one can consider the consequences of current funding practices by examining instances where projects narrowly averted failure. Imagine if Dr. Karikó secured earlier mRNA research funding, allowing for the timely development of vaccines for the swine flu pandemic. Conversely, picture the potential consequence if her research had received no funding, leading to an inability to leverage mRNA technology for COVID-19 vaccines.
I believe that better methods for evaluating and funding early-stage R&D are imperative to bridge the 'valley of death.' Corporations, in particular, have an opportunity to contribute for both their own economic benefit as well as that of society. In the future, I will be investigating how real option-based instruments are created and their applicability for applied research.
References
[1] Valley of Death - an overview | ScienceDirect Topics
[4] After being demoted and forced to retire, mRNA researcher wins Nobel | Ars Technica
[6] https://tableau.external.nsf.gov/views/NSFbyNumbers/Details
[7] New ways to pay for research could boost scientific progress (economist.com); To supercharge science, first experiment with how it is funded (economist.com)
[8] A Vision of Metascience (scienceplusplus.org)
[9] Cost of Experimentation and the Evolution of Venture Capital (nber.org)
[10] Volatility and Venture Capital by Ryan H. Peters :: SSRN
[11] The Price of Diversifiable Risk in Venture Capital and Private Equity - Ewens
[12] DecairePaul_CapitalBudgetingAndIdiosyncraticRisk_20191201_1.pdf (upenn.edu)
[13] Presidential Address: Corporate Finance and Reality by John R. Graham :: SSRN
[14] Real Option: Definition, Valuation Methods, Example (investopedia.com)
[15] Ch5 - REAL OPTION VALUATION (nyu.edu)
