Drug Research and Its Discontents: Does It Really Cost $2.6 Billion to Research a New Medicine?

A similar version of this Health Alert appeared at Forbes.

My last Health Alert discussed the high cost of researching and developing a new pharmaceutical compound. I noted that the latest estimate, by the Tufts Center for the Study of Drug Development, is “controversial” and promised to address the controversy.

The Tufts group now estimates that it costs $2.6 billion to research and develop a new medicine, 2.5 times more than the previous estimate, which was published in 2003. The 2003 estimate provoked criticism, against which the Tufts group defended itself without qualification.

The Tufts group has not substantively changed its method. So, we can expect the same criticisms to be raised against the $2.6 billion figure. Aaron E. Carroll has most recently challenged the estimate at the New York Times‘ Upshot blog.

There have been five criticisms of the Tufts group’s estimate: Lack of peer review, use of proprietary data, excluding the value of R&D tax benefits, including the cost of capital as a real cost and the fact that the research-based pharmaceutical industry funds the Tufts group.

The new estimate has not yet been published in a peer-reviewed journal, although the 2003 estimate was published in the Journal of Health Economics. There is no reason to believe that the Tufts group has dropped its standards in 2014; and it can still use the new data for an academic article. However, peer-reviewed journals can take a long time to publish an article.

The 2003 article was received by the Journal of Health Economics over a year before publication. Researchers often release working papers before publication by journals. For example, working papers released by the National Bureau of Economic Research (NBER) frequently have great impact in policy debates, while the associated journal articles are published years later with little impact (except for academic housekeeping).

Second, the Tufts group uses proprietary data provided by research-based drug companies. This would be a sound criticism except that it is also a feature of peer-reviewed articles. Most journal articles are published in Adobe Acrobat format and contain research that cannot be replicated by independent researchers. It is a sign of the companies’ trust in the Tufts group that they are willing to hand over their data without fear that it will be leaked to their competitors.

Researchers can use public data to estimate pharmaceutical R&D costs, but this leads to a wide range of estimates. In 2013, Forbes senior editor Matthew Herper counted the number of drugs launched by 98 firms over the last decade and divided that number by each firm’s R&D spending over the same decade. That was pretty straightforward and transparent and resulted in an estimated R&D cost per new drug of $5 billion ― twice as much as the Tufts group’s estimate!

Third, critics note that R&D costs are expensed immediately, not depreciated, and that there is a tax credit for orphan drugs. In a 2011 article, critics insisted that these tax breaks should whittle down the cost estimate by around 39 percent to 50 percent. This criticism distracts from the policy question, which concerns the actual dollars spent on R&D. If there were zero corporate income tax, that would not really change what contract research organizations or companies’ R&D employees would charge for their services. To improve R&D productivity, those actual costs have to be addressed.

Fourth, some critics dismiss including the cost of capital as a true cost. However, it is very real, because investors demand a risk-adjusted return on capital. The Tufts group estimated a real cost of capital of 10.5 percent. Using more recent data, Wayne Winegarden of the Pacific Research Institute noted that the nominal weighted average cost of capital (WACC) for the research-based pharmaceutical industry is 8.33 percent, which is 1.39 percentage points higher than the S&P 500’s WACC of 6.94 percent.

This is a true economic cost. The Tufts group’s latest estimate of $2.6 billion includes $1.4 billion of negative cash flow (labeled “out-of-pocket” costs) and $1.2 billion of capital costs. Far from irrelevant, reporting the cost of capital distinct from the cash flow may be the Tufts group’s single most valuable contribution to our understanding of the total cost of R&D. This is because productivity improvements that shorten the R&D cycle can dramatically reduce the capital cost of R&D. This will increase the benefits of competition by reducing the prices of new drugs.

The final criticism, that the Tufts group is funded by the research-based pharmaceutical industry, barely merits response. It is an ad hominem argument, often resorted to by critics who cannot successfully defeat the conclusions of the research itself.

Nobody kicks a dead dog. The reason the Tufts group’s research attracts such heightened criticism is that it conclusively demonstrates that prescription drug prices and costs can only come down by dramatically increasing R&D productivity, not by railing against the pharmaceutical industry’s “greed.”

Print