Biotech’s high risks have led to the conventional wisdom that it’s valuable to construct a diversified portfolio of products at the company level. Reinforced by the “show me multiple shots on goal” mantra of the capital markets, many product-focused biotech companies staple together multiple assets with the hope that one of them hits. This storyline has been essential for most IPOs in the past decade. However, while conceptually appealing on the surface, building this diverse portfolio inside a single corporate structure often creates real practical challenges down the road.
Before diving in, a quick review of classic portfolio diversification theory: bring together assets with differing risk/return profiles and liquidity attributes, and one can ride “efficient frontier” to attractive returns with lower volatility. This type of portfolio theory works because, at least during normal markets, the value of lower risk assets like fixed income often inversely correlates with the value of higher risk equities. This lack of correlation is a key principle.
But this benefit of portfolio diversification doesn’t apply very effectively to biotech product portfolios. The reason is that biotech portfolios frequently suffer from an asymmetry in the maturity of the underlying assets, and this leads to highly correlated outcomes regarding portfolio and therefore company value.
Most product-oriented biotech plays (i.e., those without a novel drug discovery engine behind them) are characterized by a single lead program and one or more pipeline projects several years behind it in R&D. This is in part due to capital allocation and burn rates. It’s difficult to fundraise around two different Phase 2a assets when the burn for each is $10-25M. But adding on $3-5M to advance a lead optimization program or two is often doable, and is perceived to leverage the “human capital” in the team effectively. And if you have a team and a lab you might as well utilize them, or so the belief goes. But because of the temporal staging of these programs, this type of portfolio construction runs squarely into the asymmetry of maturity challenge: these companies are tightly leveraged to the fate of their lead programs with little or no value in the rest.
Take the example (an all too frequent one) of a lead asset in Phase 2a and a second program near Candidate stage.
If the lead asset does well and drives an exit such as an M&A deal, it’s typically 95+% of the eventual deal value of the company, leaving no real value for any prior expenditures on the secondary programs. We all know of stories that fit this description, and there are lots of examples of this in our own portfolio: Novexel was bought by AZ for its beta lactamase program, not the earlier stage projects (e.g., aminocandin, PBP inhibitor). Prestwick was sold for its lead Huntington’s drug, tetrabenazine, and we got zero creditable value for either of our earlier stage clinical programs in Parkinson’s and Schizophrenia despite spending millions on them.
In the bad scenario, if the lead program fails, it wipes out 95% of the existing value and fundamentally handicaps the remaining entity. The supposed value of diversification to protect the downside disappears as the anchor from the now defunct lead product wipes out the portfolio’s overall value. And for private venture-backed companies, if it isn’t a total recap in this challenging scenario, then the second asset carries around the baggage of a bloated capitalization table and stacked liquidation preferences. This lead product blow-up is often terminal for most private companies. For public companies, this value destruction is rarely recoverable though can be done with patience (e.g., Jazz). Not a pretty outcome for most companies, their investors, or management teams. This “lead program brings down the house” outcome isn’t theoretical, it happens frequently. Dynogen’s lead Phase 2b program in IBS failed and effectively killed the company despite having two other earlier development programs that consumed considerable capital. Puretech’s Solace Pharma is another sad example: its lead Phase 2 failed (as hypotheses often do) but it dragged Cliff Wolfe’s cool discovery pain program down with it.
Beyond biotechs constructed around portfolios of product candidates, this asymmetry of maturity challenge also can afflict drug discovery platform companies. Over time, these companies frequently get valued predominantly for their lead program, despite having an innovative discovery engine behind them. Tolerx and Phenomix both recently threw in the towel because their lead programs either failed in Phase 3 or lost their luster, leaving a number of very interesting early stage assets and discovery platforms discarded for close to no value.
So if spending capital on pipeline projects is rarely if ever accretive in today’s capital markets, why bother? Because they can be accretive if done right.
One way to ensure value gets ascribed to earlier stage assets is to spin them out of transactions that don’t value them (e.g., Sequel’s spin-out from Novacardia when Merck only wanted the lead program; Conforma’s spin-out of Cabrellis when Biogen only wanted the hsp90 programs; Serenex spun-off its oral mucositis product because Pfizer didn’t want it). More recently, Amira spun-out a number of assets that BMS didn’t want (or they didn’t want BMS to have, like the previously partnered programs’ milestones). But these post hoc spin-out approaches aren’t economically advantageous due to tax liabilities, aren’t easy to do, and are frequently done as a sideshow to the main deal (without upsetting the deal process).
There are better approaches to structuring that can manage the asymmetry of maturity problem than spinouts: smarter corporate structures upfront can create a number of options, but they require breaking with conventional multi-asset company constructs.
- Single-program corporate entities: Creating one legal entity per program prevents losing the value of an earlier stage pipeline project in a sale or a blow-up. If you’ve got a couple programs of various stages worth funding, why not capitalize them individually? Create a “family” portfolio of these entities. This keeps the bar high for each program (i.e., is it worth funding on a standalone basis). Also, for leveraging human capital, management teams can spread their time across multiple companies. We’re doing this now with Dave Grayzel and his team with our Atlas Venture Development Corp initiative. Furthermore, this single entity approach keeps the clarity of focus, reduces the aggregate invested capital (mgt teams should read “lower liquidation preference”), and enables straightforward exit negotiations. We’ve done this with several companies (e.g., Arteaus, Stromedix).
- Asset-centric holding companies: Where there’s a thematic need (like a drug discovery engine or platform) to keep all the candidate projects under one roof, this can be done by creating a holding company structure. In one permutation, there’s a passive, non-operating parent LLC which owns the set of underlying programs (or platforms) via dedicated C-corp subsidiaries. Through this structure, a firm can avoid having to sell the entire story for just one lead program, which could enable greater long-term company growth and monetization of discovery engines. Importantly, this structure possesses all the strategic options afforded conventional single-corporation models, but with the benefit of asset-by-asset flexible monetization for shareholders. It also offers alternatives for capitalizing subsidiary programs differently. Our flagship company with this model is Nimbus Discovery, LLC; in the past couple years, several other firms have adopted related LLC models.
In both of these alternatives, the value of earlier stage programs can be detached from the fate of the lead program – solving the asymmetry of maturity challenge. This also then allows for real portfolio diversification benefits, at both the venture fund and shared management team level, as the discrete assets create uncorrelated outcomes.
Importantly, these types of structures are best done at the time of venture creation; trying to retrofit them later in a company’s life ranges from a headache to the impossible.
UPDATE – December 8, 2011
I just got a great email from an investment banker I have a lot of respect for, who wants to remain anonymous:
I usually agree with most of what you write, but I have to admit on this one, I take a slightly different view. I have had countless public company clients where the first product was either shelved or was less successful than was hoped, but who went on to build impressive organizations based on the second product. Just as examples, this was the case for Cephalon, Vertex, Alkermes, Onyx…..
I also think that the quality of executive you can recruit is higher if you are building a company rather than asking them to run an experiment.
Its nice to get the feedback, and certainly agree there are lots of historic datapoints about public companies who’ve remade themselves after lead assets failed. A few additional thoughts to this:
- Public companies are certainly in a different place than private companies in their ability to recover from a blow-up of their lead program if (the big if) they are “well financed”; Private companies have to deal with liquidation preferences and bloated cap tables, which can crush private companies when blow-ups happen (or lead to complete wipe-out recaps)
- In prior biotech vintages, great stories like the four examples above could be built because of cheap public capital. They got public in a much different public capital market than today (early 1990s IPOs) and so getting “well financed” at a reasonable cost of capital was still a possible outcome (hard to do today)
- Great teams can be recruited today into all of these types of structures (we have done so!). They aren’t project managers of “experiments”: they are building a lot of value via great assets. Spreading management talent across multiple programs, either in one holding company with subs or across individual legal entities, is an important part of leveraging “human capital” effectively and helps to support more efficient staffing (though this is clearly something to be managed)
- Lastly, I’m by no means suggesting companies shouldn’t have multiple products in their portfolios. They should just think carefully about their stage, their likely correlation with regard to company valuation, how to prevent the asymmetry of maturity issue from handicapping their ability to create value from secondary programs.