Skip to main content

Drug Development

Knocking Opportunity Costs

I came across the news item that a representative from Ohio (Sherrod Brown, D-13th) has sent a letter to President Bush contesting the White House’s recent report on pharmaceutical costs. (A press release of his addressing the same topic is supposed to be available from this list, but all of Representative Brown’s links are broken.) Looking at his record, it appears that he’s been using my industry as a trampoline for some years now.
There’s one part of his complaint that I wanted to address. The over-800-million-dollars-to-develop-a-drug estimate that the Tufts center arrived at is a constant target. (I hate to tell people that it’s quite possibly a low estimate.) Rep. Brown specifically complains about the way that this figure includes the opportunity costs associated with the risk of drug development, and he’s not the only one. Why, critics ask, should the amount a drug company would have made by investing its research money be included as a cost?
Well, because that’s what it is. Most drug development projects don’t work, and all the money spent on them goes down the hole. If you’re going to risk your cash on one, you need to figure out what you would get if you parked your money in some more reliable investments instead, and you need to realize up front that this is what you’re forgoing.
“But I don’t do my expenses like that!” goes the cry. Don’t you, though? If you have some extra cash around, should you put it in a savings account, the stock market, pay down some principal on your mortgage, or put it all on your lucky number at a roulette table? Depends on the rate of return for each one, and the risk. (This would be a good time to mention that the odds of hitting your roulette number are almost identical to the odds of getting a clinical candidate to market for a central nervous system indication.)
Those first three options don’t involve muchrisk , but the last one sure does – and if you decide to go for it, you should realize what it’s really costing you: the use of the excess interest on your money, interest which you would have been able to earn. The same consideration applies to the other choices, which is why paying down your mortgage is the equivalent of earning that percentage interest on the cash. An appreciation of opportunity costs would do a lot of people good.
But I have a more direct demonstration, one that I’m prepared to offer to Representative Brown: give me a thousand dollars, Congressman, and in ten years I’ll give you a thousand back. If you don’t think that this deal is costing you money, I’d love to hear you explain why not – after you’ve written the check, naturally. How about it?

8 comments on “Knocking Opportunity Costs”

  1. Jake McGuire says:

    Your analogy about the roulette table is a good one, but I think it’d be more accurate if the last option was a roulette table where the croupier threw the ball so hard that it spun around the wheel for ten years before stopping. That’s obviously a worse deal than a roulette spin that ends in thirty seconds, and someone with an inquisitive attitude (not sure about the good Congressman) might try to figure out in just what way it’s a worse deal and then come to the correct conclusion on their own.

  2. Ben Robinson says:

    I have read your blog for a few months, and this is the first time I have felt compelled to comment. This is going to get a little long, as it is an issue that I am a little passionate about.
    The Tufts number is cited frequently, and miscited even more frequently. The idea behind the Tufts number is finding the economic cost of developing a new drug, not the accounting cost. An economic world is one with rational consumers, information asymmetries, and zero profits in competitive industries. An accounting world is closer to the real world where assets have various forms of liquidity, cash is king, and profits accumulate for shareholders.
    Thus, the $800 billion is not a cost that the pharmaceutical industry pays in cash, it is even unclear if these costs are accrued to a single organization or if this is the total cost for the development incurred by all parties involved. It is frequently cited by PhRMA as if it were a real cash expenditure for each drug company developing new drugs. It is also a number that is very sensitive to the discount rate (the percentage at which you discount future cash flows). I cannot find the reference, but at the announcement some financial reporters pressed the author for the discount rate and he stated that roughly 50% of the cost was due to financing costs (these opportunity costs). It is a decent estimate, but the concept of accuracy is not that relevant in an economic world. For the given set of assumptions, this is the economic cost. It is also intriguing that (to the best of my knowledge) this study was never published in a peer reviewed journal and the details of the actual analysis have never been made public.
    Having degrees in both economics and chemistry, I have bounced around the pharmaceutical industry in a variety of roles and I think the Tufts number has done more harm than good. It comes across sounding a bit like a rationalization from an industry that is facing maturation pains moving from a young upstart industry to a more stable middle age. It also has focused the debate on haggling about costs rather than social returns. What is the economic value delivered for the $800 billion? Are there competitive substitutes that give better returns? Are there lower cost structures of the industry (large pharmaceutical firms spend more on sales and administration than on R&D, and the opposite for smaller start-up firms)?

  3. Ben Robinson says:

    Does not help my long diatribe when I quote $800 billion rather than the true number of $800 million. Not good to impulsively rant on a blog.

  4. Derek Lowe says:

    These are all good points, and worth another blog entry or two on their own. You’re right that the Tufts figure is very sensitive to the discount rate that’s assumed, and I don’t know what it is, either. My argument is mainly with people who don’t know what a discount rate is, though.
    Your thought that this estimate may have done more harm than good is an intruiging one. In the end, I don’t think that the Tufts center is going to convince many people of much of anything. Pro-pharma types will cite their numbers as backup for their positions, and anti-pharma types will assume that the numbers have been cooked: status quo ante.

  5. dr says:

    I’m not happy about the Tufts number either, despite the fact that business cases built on it have paid my salary. Two big technical problems, which seemed serious even to a layman:
    1. Who pays. Say an academic research program consumed $20M in government funding over a decade, then produces (along with lots of papers and postdocs) a lead compound, which the university licenses out for $50k. The Tufts analysis uses not the actual cost to industry, the $50k at year 10, but instead takes the $20M spent by the government, inflates it forward (at the industrial discount rate!), and adds it to the total. Maybe valid to compute total cost to society, but not for total cost to industry. Ben alluded to this.
    2. I’m no expert, but it sure seemed like some double counting was involved. Eg., inflating everything by a discount rate, and then _also_ comparing return on investment to a bond rate. You should only have to pay a time value of money once.

    It’s hard to address the technical issues,
    since the raw data is all proprietary, and not available for independent analysis.
    But more importantly, I don’t think that that number actually expresses what’s difficult about drug discovery, nor contributes to the discussion about how this research ought to get paid for.

  6. Derek Lowe says:

    Dr, in my experience, universities very rarely develop a lead compound that’s fit to outlicense. (And when they do, they know to hold out for a lot more than 50 grand.)

  7. dr says:

    Derek, I think the good stuff gets wrapped in a spinoff first, not sold for cash. That’s one way to hold out for a lot more, I suppose, so I don’t think we’re disagreeing.

    This actually brings us closer to a more reasonable way of estimating. I was going to complain the $800M is not a cash number, in the sense that there is no market where I can pay $800M and get a new drug. But there is, actually: the market in pharmaceutical companies themselves, and their IP. Drug companies merge all the time, sell drug lines all the time, and many are publically traded. I’d be quite interested to see what this market imputs for the value of the new drugs appearing. Then we can see whether we’re making a profit if we’re spending $800M (or whatever the company’s share of this is) per new drug discovered. This is the value side of drug discovery, the value to the company. The value to society is larger yet.

  8. M Rosenberg says:

    I have helped develop several drugs that have made it to the market over the last 10 years. I am also familiar with the budgets for these programs. The out of pocket cash costs for drug development are on the order of $150-$350MM with about 10-15% covering research, 20-30% covering process development and manufacturing, 30-40% covering clinical trials costs and the remainder covering sales, marketing, and administrative costs. The Tufts number includes all of these costs plus the inflation rate and lost opportunity cost for failures.

Comments are closed.