Big pharmaceutical houses have long relied on drugs developed by others—particularly biotechnology firms—to fill the huge gaps in their product pipelines. In-licensed drugs accounted for 30 percent of Big Pharma's revenues in 2001, and with the looming expiration of many key patents, to say nothing of lagging R&D productivity, that reliance is likely to increase. But this licensing strategy, our analysis shows, has a widely overlooked flaw: deals are often struck too late to generate maximum value.

To put it simply, pharmaceutical companies are overdiscounting for the uncertain prospects of deals made early in the development process. To reduce the risk of licensing a drug that ultimately fails to win approval from the US Food and Drug Administration (FDA), these companies make low offers to biotechnology firms during preclinical testing, when researchers complete the synthesis, purification, and animal tests of a drug. About one-third of all licensing deals occur at this stage.¹ For the rest of them, pharma companies often don't commit substantial resources until clinical trials demonstrate the drug's safety and efficacy in humans. While this delay is understandable, it can cost companies tens of millions of dollars in higher fees and royalty payments to the biotechs for every compound. Up-front and milestone payments for a single late-stage drug can exceed the cost of buying the rights to ten drugs in the early stages—and at least one of those drugs, industry statistics show, will pay off, frequently with annual revenues of $500 million or more.

To find the right time for licensing deals, we first calculated their expected completion phases if pricing had been "fair"—that is, if the risk-adjusted net present value of the deals made both parties indifferent to the phase when licensing occurred.² The results show that the expected distribution of deals is skewed heavily to the preclinical phase (Exhibit 1). We next carried out a Monte Carlo simulation, based on industry-average data, to determine the optimal time of licensing for 10,000 hypothetical compounds. With every simulated deal, we tested variables such as the success rate at each stage of development, the duration and cost of the phases, and the licensing terms by phase.³ The model was designed to simulate the prospects of a given drug at the start of preclinical research (that is, on a forward-looking rather than retrospective basis), to help managers decide on their course of action before uncertainties are resolved, as they would have to do in reality. The outcome was clear: under current conditions, pharma companies would capture the greatest expected value from preclinical licensing virtually 100 percent of the time because the greater risk of failure for preclinical compounds was more than offset by the low terms available early on. The problem is that biotech companies are usually reluctant to settle for them.

Our calculations, however, show that pharma companies could dramatically increase the amounts they pay for compounds in early development and still come out ahead. In most cases, they could pay 150 percent more at the preclinical stage for the rights to a drug—for instance, by increasing up-front payments to $5 million, from $2 million, and milestone payments to $40 million, from $15 million. Although a pharma company would be paying more money sooner and waiting longer for a return, drugs licensed at the preclinical phase would be expected to create the maximum amount of value for the company in 85 percent of all cases in which a deal could be negotiated (Exhibit 2).⁴ In effect, pharma companies would be making more bets sooner and, rather like venture capitalists, building broader portfolios in hopes of securing a few hits when the potential returns were highest.

Biotech firms, by contrast, prefer to wait. The booming capital markets of the late 1990s, which provided them with enough money to finance their early research without a large steady income, made them more willing to gamble that a drug would make it to clinical trials. Accordingly, our simulation showed that, in today's typical deal, these firms almost always reap maximum value in Phase II or Phase III. The preference of biotech firms for late-stage deals is obviously rational, but with richer incentives to make deals earlier, that could change (Exhibit 3). In our model of sweetened deal terms, a biotech firm that licensed its rights at Phase I would collect, on average, $60 million in extra up-front and milestone payments and retain a 20 percent royalty, giving the firm a continuing stake in the drug's success. In such a scenario, the firm would gain maximum value from preclinical or Phase I deals in over 60 percent of all cases in which a deal could be negotiated. And by licensing drugs sooner, biotech firms could also capture more quickly the benefits of the pharma companies' powerful sales and marketing capabilities and clinical-trials expertise.

Given the innovation challenge now facing the pharma industry, the competition for new drugs developed by biotech companies is bound to increase. Pharma companies that can assess the risks of early-stage compounds and secure the rights to the most promising among them will probably be rewarded—and so will these companies' biotech partners.