The Costly Myths about Pharmaceutical R&D

Posted By


The following is an unsolicited guest post submitted by Donald W. Light, University of Medicine & Dentistry of New Jersey, and Rebecca Warburton, University of Victoria, Canada.  It is followed by some comments in response from Matthew Herper, the writer of the Forbes article in question. We are pleased to provide this forum for this discussion.


A recent article in Forbes sets a new record for the “staggering cost” of developing new drugs. It estimates R&D costs as $4-11 billion per new drug, 3-5 times more than the highest estimates by the industry-supported economists at the Tufts University Center for the Study of Drug Development.

This record estimate sets the stage for far more government subsidies and protections, paid by taxayers, than even Big Pharma ever hoped for.

As authors of an article on the costs of R&D cited by Forbes, we would like to straighten a few errors in these estimates that pharmaceutical reps, executives, and lobbyists use to justify higher prices and costs to employers and governments that pay for them.

First, the Forbes estimates accept company R&D figures uncritically and ignore evidence that what companies count as “R&D” may be much broader than the costs of bench, lab, and trial research that make up R&D. Drug companies work hard to hide any verifiable figures from independent review. And they never link their alleged costs to how quickly they earn them back at high prices.  See more on

Second, the Forbes estimates divide total reported costs by the number of “new drugs,” which is not defined. It probably means NMEs or new active ingredients; but most of the products coming out of industry R&D are newly patented variations on existing drugs that involve less risk, time and cost, even though much larger trials are needed for small differences to be proven statistically significant. These me-too are the main products of R&D, and they account for about 60 percent of the nation’s drug budget.

In other words, the Forbes estimates divide apples by oranges — total R&D for all research activities on all products, divided by the handful of NMEs.

Third, there is no evidence that the failure rate has increased, aside from industry claims. The industry-supported estimate by DiMasi and his colleagues is half the rate used here by Forbes. Both our and DiMasi’s cost estimates include all costs of failures as reported by companies to DiMasi’s shop at Tufts.

Fourth, no one knows the real cost of discovery, the “R” in R&D. DiMasi backed in a guesstimate that ended up being about 40 percent of total costs. Another way he inflated estimates was to claim that it takes on average12-15 years to develop a new drug, much more than independent evidence shows. See our articles on What matters to companies is what they actually spend, after taxpayers pay almost half if it.

Finally, half of the DiMasi estimate is not real R&D costs at all but highly inflated estimates of what companies would have made had they put their money in an index fund and not developed new drugs in the first place! Given the staggering cost estimates in Forbes, it looks as if drug companies should do just that and become investment banks.

Our own estimate of pharmaceutical R&D is often misquoted as an average of $43 million, which commentators reject as absurd without actually reading what we wrote. In fact, we make clear this estimate in 2000 does not include the cost of discovery (because no one has accurate figures and it varies from 3 months to 30 years), nor the “cost of capital” (for reasons explained in our article). It is also the net cost after taxpayers cover about 50% of corporate R&D expenses and the median cost, which is not biased up by 35% to the average cost by a few costly R&D projects. Any 2000 figure is 30% more in 2011.

In sum, our estimate is that the median, net, corporate cost to develop a new drug, based on the confidential cost data that companies reported to DiMasi and his Center, is $56 million in 2011, plus the unknown company costs of discovery and the artificial cost of capital, if you think it applies.

Our estimate is substantially higher than the only hard number we could find, audited tax returns in the late 1990s where companies reported that their average trial costs were only $22.4 million. DiMasi’s own high estimate applies only to the most costly 21.8% of new approvals, which his article indicates are 15 times more costly than R&D for new variations on existing molecules. If you actually read our deconstruction of how DiMasi et al inflated their estimate, you will learn a lot. (Search for “demythologizing the high costs of pharmaceutical research” )  See also

What do we get for all that money?

Forbes is right that staggering costs should not be “a badge of honor.” What matters is how much better new drugs are for patients against clinical criteria and how many people can afford them. Sadly, 85-90 percent of all newly approved products of pharmaceutical R&D are judged by independent review groups to be little better than existing ones. See THE RISKS OF PRESCRIPTION DRUGS on Amazon. These are the drugs that brilliant marketing turns into huge sales that make up that 60 percent of the nation’s drug costs.

But drug companies know what they are doing. Between 1995 and 2010 (the era of the so-called “innovation crisis”), they reported spending $34.3 billion more in R&D and generating $200.4 billion more revenues – not a bad return. Pharmaceutical companies average several times more profit than the Fortune 500. By contrast, if the Forbes figures and business arguments are correct, then nearly all the global companies would have gone bankrupt between 1997 and 2011.


I sent the article to the Forbes writer, Matthew Herper, to let him know that I had received this submission from Light/Warburton, and to see if he wished to respond.  He sent me some comments via email, which appear below:

“It’s certainly true that R&D budgets contain more than just the cost of inventing drugs and getting them to market. Postmarketing studies, though, are not always just mere marketing tools. They proved, for instance, that statins prevent second heart attacks and that Avastin works in lung cancer.

Some of my industry sources have pushed back by that same logic, saying that the average cost is closer to the $1.8 billion a group of Eli Lilly researchers published in Nature Reviews Drug Discovery in 2008. I think what this approach loses in depth it makes up for in simplicity: it’s a back-of-the envelope calculation where a lot of the arguments on both sides balance out.

We do have a class of drug companies whose R&D spending is dedicated almost entirely to getting products to market: these are biotechnology firms. And looking at them reveals pretty clearly that Light and Warburton’s $55 million figure is amazingly low. If companies like Plexxikon and Optimer, which developed new medicines on the cheap, didn’t come in under that figure it’s unlikely to be a real cost.

Light and Warburton fall into the trap of viewing the purpose of the high estimate of the cost of R&D as a justification for high drug costs. This is one of many self-destructive arguments the drug industry has made over the years, but it no longer holds much sway. If AstraZeneca is spending triple what Novartis is on each new drug, that’s not a reason to pay AstraZeneca more. Some of the most expensive drugs, like Questcor’s Acthar Gel, have had the least amount of new research dollars put into them. The problems with drug pricing have nothing to do with research. Some of the medicines with the highest research costs, such as new blood thinners and new diabetes drug, launch with the lowest prices.

I’ve spent a decade watching huge sums of money spent on mergers and acquisitions for still unapproved drugs or even platforms for inventing new drugs. Surely we can agree that this was real money, and that the buyers would not have been spending it were their per-drug cost a mere $55 million.

The reality is that drug companies spend more on research to launch fewer new medicines than they did a decade ago. This is a problem, because medicines are generally a good thing. Trying to pretend that inventing and testing drugs is not a very expensive endeavor because one wants to try and keep drug prices down is every bit as silly and counterproductive as using the cost of inventing new medicines as the justification for making them expensive.”

 ADDENDUM on February 18:  Merrill Goozner also criticized the Forbes piece in The Fiscal Times.


You might also like


Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

Prof Donald Light

February 17, 2012 at 10:37 am

Matthew Herper makes some solid points but also perpetuates some mis-impressions. First, he still writes without distinguishing between the few NMEs or new molecular entities approved which are called “new drugs” and the 7-times more variations on existing molecules that constitute the main products of pharmaceutical R&D. Dividing all R&D costs by the few NMEs lies at the heart of how he arrived at such high estimates. As Merrill Goozner added in his column yesterday (, a large percent of R&D also goes to hundred of clinical trials to sign up physicians, try out other uses, and do post-marketing studies. If one could subtract all these other R&D activities, the costs for developing the few NMEs would be MUCH smaller.

Second, any average cost of R&D per new drug or per NME is an artificial number designed years ago by industry-supported economists to lobby for longer government protections against free-market price competition. It has little with how companies actually estimate and think about the costs of R&D. Their R&D costs vary greatly by which disease, which target, what kind of science, and what research strategy a given company or team is pursuing. Further, what matter to a large company is the net, median cost. The more it can get others to pay for early discovery and development and even trials (as in cancer drugs where NCI also pays for many trials), the better.

Third, M&A has been shown to not increase the development of NMEs. As Munos, a noted industry analyst put it, the evidence about M&As is that “1 plus 1 equals 1.”

Fourth, Herper is correct that the picture of small biotechs is entirely different. This whole rhetorical debate about “the average cost of R&D” has never had anything to do with them. Given that they are now the leading source of important innovation, it makes this “average cost of R&D” discussion even more irrelevant to anything.

Finally, Herper continues to misrepresent our estimated costs in this pointless debate to which he has just prominently contributed. Our estimate is the cost of pre-trial discover and development, a median net cost of $56 million, plus the cost of capital if you think it’s a justified cost. It looks like Herper has never read our article. If he did, he would see that we calculate the cost of development for the two kinds of NMEs and for the vast majority of new products that are not NMEs, with cost of capital and without. Isn’t it time that folks actually read what we wrote and writing critiques based on it?


February 17, 2012 at 11:22 am

This is actually a pretty easy question to answer.

A search of for the period January 1 2002 through today shows 16,834 industry-sponsored, interventional clinical trials that enrolled 4.42 million patients. 13,892 of these trials were pre-registrational (Phase I to III), enrolling 3.25 million patients, or 74% of the total. While it is difficult to manually assess all of the remaining 3,000 or so trials, if you rank them in order of number of patients, it is clear that the vast majority of the Phase IV trial population was enrolled in large cardiovascular outcome, vaccine efficacy, safety, or other scientifically valuable trials, many of which were required by the FDA as a condition of approval.

Certainly seeding trials should not occur at all.. But the data is out there to unambiguously refute accusations that they represent a significant fraction, let alone the majority, of industry R&D spending.

Harold DeMonaco

February 18, 2012 at 8:32 am

I think that there is another and perhaps more enlightening way of look at R&D costs of big Pharma. While touting the huge investment in R&D, Pharma routinely spends 2-3 times their budget on “Selling and Administrative Costs”

Pfizer (2010)
R&D 13.9%
S&A 29.0%

Astra Zeneca (2009)
R&D 13.4%
S&A 34.5%

GSK (2009)
R&D 14.4%
S&A 33.8%

The cost to bring a drug to market is important but sometimes a bit difficult to discern. How much a company spends on R&D as compared to administrative, marketing and sales costs is a better marker. One can hardly complain about the huge cost of research and development while spending 2-3 times that much on trying to convince physicians and consumers to purchase a product.

Steven J. Raphael

February 20, 2012 at 12:16 pm

I was struck but the unstated (and unsupported) claim in Mr. Herper’s note that all costs of pharmacy research and dev are reasonable and justified. I guess that what happens when economics becomes solely the realm of ideology and is ignorant of history.