In a recent WashingtonMonthly post I’m once again seeing people who don’t understand why pharmaceutical companies need to earn more profit than average companies in order to attract investment. As a service I thought I would offer a quick illustration of why they need to. The non-math answer is that pharmaceutical companies are riskier (lots of them go out business providing a -100% return to their investors. Those that don’t go out of business lose and gain money at wild intervals. Investors can invest in other things and have lower amounts of risk. This risk has to be compensated by higher up-side returns. For those not math-phobic lets look at that a little more closely.
For purposes of this example we will say that you invest for exactly one year and have exactly $1,000. We will assume the money is not capitalized at intervals during the year.
Investment Option 1
Government Bonds. It is very safe, with a lowish formal rate for a low return.
Lets say you can get a 1 year T-bill at 5.0%.
Your expected return is
1,000+1,000*0.05
That is $1,050 for an expected 5% return on your money.
Investment Option 2
McDonalds Franchise. Pretty safe with some chance of breaking even and some chance of losing everything.
Lets say you get a 95% chance of getting a 13.7% rate, a 3% chance of breaking even and a 2% chance of losing everything. If you don’t know which franchises are going to be the failures your expected return is .95(1,000+1,000[.137])+0.03(1,000)+.02(0).
That is $1,110.15 for an expected 11% return on your money. We will round that to $1,110.
Investment Option 3
Lets say we don’t know anything about it other than the risk profile. The risk profile is somewhat worse than McDonalds but not crazy (risk of losing money is 10% or less). The exact profile is a 90% chance of getting the up-side return (‘X’) a 5% chance of losing 10% of the value and a 5% chance of losing everything. The formula to find out what the up-side rate would have to be to break even with McDonalds is
.9(1,000+1,000X)+.05(1,000-1,000[0.1])+.05(0)=1,110
Solving for X we find that the rate to break even with McDonalds is .1833 or 18.33%.
What would happen if we gave Investment Option 3 a McDonalds level upside return?
Your expected return if you didn’t know which ones would lose money would be .9(1,000+1,000[.137]])+.05(1,000-1,000[0.1])+.05(0)=$1,068.3 for an expected 6.8% return. This is very close to the return rate on the government bonds (which are theoretically 100% safe).
So in order to have an expected ending value equal to the McDonalds investment you would need to earn 18.33% on the up side of the riskier investment compared to the 13.7% rate of the McDonalds investment. Of course you would be a fool to make the investment at only 18.33% because your chance of losing money is greater in the risky investment while your expected value is the same. So you need to earn even more if you are going to bother with the riskier investment. Now I don’t know exactly what McDonalds investment returns are, but I did set it up so that Investment Option 2 tracks the average annual return of the general US stock market, 11%. Pharmaceutical companies are much riskier than the general US stock market, so they need much higher returns on the upside to make the risk worthwhile.
Sorry about the font weirdness, I performed a cut and paste from word and apparently that didn't work well. I don't have time to fix it right now.
Pharmaceutical companies are much riskier than the general US stock market, so they need much higher returns on the upside to make the risk worthwhile.
Can you provide some evidence for this? How do you come to the conclusion that pharm companies are inherently riskier?
Specifically, I'd like some link that could quantitatively spell out how pharm companies engage in riskier business practices than the general stock market.
Posted by: manyoso | March 28, 2006 at 06:02 PM
Sebastian: I don't have time to think, really. That might be an argument against commenting, but hey! has it ever stopped me before?
So: wouldn't there be a difference, as far as your argument goes, between big pharma companies (Merck, Astra-Zeneca, etc.), and little ones (the ones that hope to develop a big drug and get bought out)? I would have thought that the chances of one of the big ones going out of business and thus returning -100% were no greater than those of any other large company; and that your point was restricted to the little fish in the pharma pond.
Posted by: hilzoy | March 28, 2006 at 06:02 PM
Ok, but as Ezra Klein points out, 36% of all medical research in this country is government-funded. I don't think 36% of all cheeseburgers are subsidized by the government.
In any event, I don't really see where Kevin is arguing that pharmaceutical companies are not entitled to higher profit margins than other types of companies. The question is whether their current profit margins, which derive in part from the inability of the government to negotiate lower prices the way private insurers do, are truly necessary.
Posted by: Steve | March 28, 2006 at 06:03 PM
Insofar as Sebastian is ascribing the need for greater profits to greater riskiness per se (as opposed to lower expected value), that seems like a factor weighing in favor of government-funded research, because the government can bear risk cheaply.
Posted by: Christopher M | March 28, 2006 at 06:14 PM
"I would have thought that the chances of one of the big ones going out of business and thus returning -100% were no greater than those of any other large company; and that your point was restricted to the little fish in the pharma pond."
The chance of negative returns on any one large pharmaceutical company are much larger than normal large companies.
I use bigcharts.com for stock charts because they are easy to play with.
Try a 5 year look at Merck (MRK). High (about 5 years ago) price, about $75. High just before the Vioxx problems, about $45. Price right now, about $35.
Try a 5 year look at Pfizer (PFE). High price (about 5 years ago), about $45. Current about $25. Pfizer had the worlds very best selling drug for all of the last five years.
Compare to the 5 year look at another company accused of monopoly-type powers, Microsoft (MSFT). Low (about 5 years ago) about $24. Current, about $27.
BristolMeyers (BMY) High (about 5 years ago) about $60. Current, about $25. (I'm skeptical of this one because I think there was a merger that messed with the price but I can't remember.)
Compare to the 5 year look at McDonalds, which has had some tough years lately. Low, about 3 years ago at around $13. Current, about $35.
Compare to the 5 year SP500 index. Low around 3 years ago--about 780. Current about 1290. In fact I compared all of the drug companies above to the SP500 on the comparison function and found that they rarely did better than the SP500 and have often done significantly worse.
The reason you get statistics like "The top 5 drug companies make ridiculous profit X" is because they aren't the same companies from year to year.
Posted by: Sebastian Holsclaw | March 28, 2006 at 06:35 PM
Hmm, where to start with this...
It's true that, as risk goes up, return must also go up. Why settle for a return equal to, or only slightly higher than, an extremely safe investment like a T-bill, if you're investing in something that has a risk of loss?
(2 optional geeky extensions follow; readers not particularly fascinated with decisions under risk may skip the next 2 paras.)
(Geeky extension #1: in fact, the relationship between return and risk must be better than linear, since you can always combine two assets in some combination to get a "basket" with the weighted average risk and return, i.e., a point on the line between them. So if someone else offers an asset they want to be attractive, it has to have either the same return for a higher risk, or lower risk for a higher return.)
(Geeky extension #2: moreover, because people are risk-averse, the return on any asset with multiple possible outcomes, such as your examples #2 and #3, must be higher than the "expected," i.e., probability-weighted average, return. For example: if I offer you an asset with a 50% chance of losing everything you invest and a 50% chance of grossing $1 million, then the expected value of the asset is 0.50 x 0 + 0.50 x $1M, or $500,000. But would you invest $500,000 in that asset? Wouldn't you prefer the $500,000 in cash, for certain? Therefore, in order to get you to invest in that asset, I'd have to price it lower, or equivalently, offer you a more attractive return.)
So because pharma is a very risky business indeed (another one with a similar profile is movie-making, of all things), it is true that investors demand high returns from investments in pharma projects.
BUT, while an individual pharma project has a very high risk of a total loss, pharmacos -- especially the big ones -- diversify away risk by undertaking many projects simultaneously. So the risk to investors is actually significantly reduced: they will almost certainly see a reasonable return on their investment, because some projects will be successful, and they will almost certainly not see a high return, because most projects will fail. (Geek translation: the variance in return is much lower, and since variance is the measure of risk, the required return is also much lower.)
Therefore, pharmacos actually need NOT see a much higher return than any business that faces risk. In fact, because a McDonald's franchise could actually fail COMPLETELY, since it's just one store, it could be the case that investment in a pharmaco with multiple projects is actually SAFER than investing in a (single) McDonald's, and therefore deserves LOWER return.
Now all of this is quite general, and a resolution of the issue depends on the details of costs and outcome possibilities.
And the simple accounting fact of the matter is, while exploratory and clinical development costs of a successful drug can be quite high -- order hundreds of millions of dollars -- the sales and marketing expenditures of big pharmacos are equally high; a typical "blockbuster" (annual sales >$1B) drug has M&S expenses of $100 million or more ANNUALLY. So a lot of the costs of drugs are due to the decisions of the pharmacos to push their drugs on TV, in magazines, and to doctors on junkets, and NOT to the inherent risk of drug development.
And digging even deeper, the underlying culprits are the pharmacos' poor (though improving) practices in controlling development risks (primarily their unwillingness to kill unpromising projects early), and the demands of the financial community for unrealistic rates of return (which in turn drives a dependence on the "blockbuster" model). In many ways, modern pharmacos resemble the US auto industry of the 1970s, and until they reform their practices -- and tell Wall Street to take a long walk off a short pier -- their financial picture will not improve.
And more to your point, none of those things -- poor risk management, financial greed, or bloated ad costs -- is an excuse for gouging people for health care. Even though the risks are high, drug costs can be reduced significantly.
Posted by: bleh | March 28, 2006 at 06:38 PM
"Insofar as Sebastian is ascribing the need for greater profits to greater riskiness per se (as opposed to lower expected value), that seems like a factor weighing in favor of government-funded research, because the government can bear risk cheaply."
But there is the choice problem. The government isn't nearly as good at figuring out what particularly to fund as the market as a whole is. If the government could figure out what to research that idea might work, but it can't process the information properly without the market. The government would be better off targeting specific market failure scenarios--like vaccines or treatments for low frequency diseases.
Posted by: Sebastian Holsclaw | March 28, 2006 at 06:39 PM
What bleh actually said is, I think related to what I was getting at with my big vs. little company comment -- the little companies are often banking everything on one or a few projects; the big companies are not. They let the little companies take a lot of their risks for them, and then swallow them up. Thus, if they are riskier, it can't be (I think) for the reasons Seb gave in the post.
Of course, since I just gestured vaguely at it, while bleh spelled it out clearly and eloquently, my point is "related" to his in sort of the same way as the idiot stepchild is "related to" his or her wildly successful, eloquent, poised, and gracious family.
Posted by: hilzoy | March 28, 2006 at 06:50 PM
"(Geek translation: the variance in return is much lower, and since variance is the measure of risk, the required return is also much lower.)"
But you have to attend to the numbers. The ratio of drugs explored to drugs taken to market isn't anything like 1:1. It isn't even anything like 10:1. It is much more like 40:1 (at the big phrama level, at the small pharma level I think it approaches 100:1). [Numbers not made up, but I haven't looked them up today]. Very few business models allow for even a 5:1 bad to good ratio for serious expenditures. Mine is horrified by the idea of even 1:1. When you say that the variance is much lower you are correct but only by comparison to the small drug companies. Comparing successful projects to unsuccessful in a drug context vs. your average business context provides a hugely different picture.
Also, marketing is not advertising. They aren't an identity. Marketing is a lot more than advertising. It includes such things as samples (and my understanding is that samples make up the largest part of the 'marketing' section.
"And digging even deeper, the underlying culprits are the pharmacos' poor (though improving) practices in controlling development risks (primarily their unwillingness to kill unpromising projects early), and the demands of the financial community for unrealistic rates of return (which in turn drives a dependence on the "blockbuster" model)."
Unrealistic? If they don't perform REALLY well people will just invest in McDonalds. Kill unpromising projects? If anything many scientists worry that they kill too many promising projects too quickly.
I really suggest that everyone browse through the archives at Derek Lowe's webpage. The right sidebar has topics which could be really useful. They include: Academia vs. Industry, Business and Marketing, Drug Development, Drug Prices and Who Discovers and Why. He isn't a full throated pharma company cheerleader, but he does debunk a number of common misconceptions--the main one being that pharma companies unfairly leech off of basic government research as if the basic research brought you within a stone's throw of a drug in most cases.
Hilzoy, "They let the little companies take a lot of their risks for them, and then swallow them up. Thus, if they are riskier, it can't be (I think) for the reasons Seb gave in the post."
Same answer as above. The magnitude of 'a lot' in your sentence is vastly different in the drug company world than it would be in almost any other business model. Furthermore even with that, the point in the research at which they take over still involves enormous risk because they normally take over in phase II or certainly before phase III. The failure rate then is still much higher (and hugely more expensive) than would be normal for another business. Imagine 10 "New Coke" level money-suckers for every successful drug. One "New Coke" gave Coca-Cola a huge scare.
Posted by: Sebastian Holsclaw | March 28, 2006 at 06:58 PM
Also note how very little risk I added into the equation. I'm still giving the 'risky' company a 90% chance of going positive at a and a 5% chance of only being -10%. That gets an 18% rate necessary to catch the SP500.
That is certainly very generous to the other side of the argument if we are applying it to drug companies. Pfizer has been -40% over the last 5 years. It has also had the most popular drug in the world for the past 5 years.
Posted by: Sebastian Holsclaw | March 28, 2006 at 07:04 PM
That should be "at a(n) 18.33% rate, and..."
Posted by: Sebastian Holsclaw | March 28, 2006 at 07:05 PM
Sigh, remember to refresh before posting.
Hilzoy gets at the diversification point.
Regarding research (Christopher M, Sebastian #2), the big costs in pharma are not exploratory research but clinical development, i.e., the part where they're testing an already extensively tested drug on large numbers of patients. Properly done, that is a relatively low-risk exercise; the real risks come early and in much cheaper phases. Government can and does sponsor a lot of the early stuff, but could not and should not pay for the later, expensive stuff.
Regarding stock market prices (Sebastian #3 and #5), one must distinguish between investment in a drug company or project and investment in a liquid financial instrument. Most of this discussion is, I think, more appropriate to the former. As to the latter, and to stock prices of Merck and Pfizer, there is a lot more at work. Merck was hit by fear of lawsuits over Vioxx and Bextra (and imho by the loss of some talented executives). Pfizer has several blockbuster drugs coming off patent in the next few years. And both are subject to worries about the political climate, notably how quickly the US is moving toward single-payer healthcare. That's a whole 'nother set of risks that are largely or entirely independent of the risks inherent in drug development, and they're material enough that pharmaco stock price movements are, I think, not a very clear measure of anything except, well, pharmaco stock prices.
Re failure rates (Sebastian #4), the funnel is even tighter than you think: start with 10,000 compounds, end up with 1 commercialized drug. But those numbers are themselves dwarfed by the immense returns on that successful drug: markups are literally 10,000% or more. So in fact, attending to the numbers (and I do; it's my business), the failure rates are entirely acceptable and more than made up for if the risks are managed well. (As to phase III failures, the saying goes, a phase III failure is a management failure.)
And re marketing (Sebastian #4 again), I beg to differ; in pharma, marketing is by product, not by company. It's budgeted that way, it's managed that way, and it's accounted for that way. One does not see marketing touting the GSK "brand," for example, one sees marketing touting a particular GSK product. There is some value to corporate brands, of course (Johnson & Johnson being the premier example), but pharma simply does not spend much money marketing corporate brands.
Posted by: bleh | March 28, 2006 at 07:08 PM
"Regarding research (Christopher M, Sebastian #2), the big costs in pharma are not exploratory research but clinical development, i.e., the part where they're testing an already extensively tested drug on large numbers of patients. Properly done, that is a relatively low-risk exercise; the real risks come early and in much cheaper phases."
Phase II and III are relatively low-risk exercises? Compared to what? Compared to the chance of having a heart attack due to Vioxx? I think not. That was barely perceptible. Vast numbers of drugs fail Phase IIA, many of those that don't fail IIB, many of those that don't fail III. I'll look for precise numbers when I get a chance tomorrow but "relatively low-risk" leans very heavily on 'relatively'.
"But those numbers are themselves dwarfed by the immense returns on that successful drug: markups are literally 10,000% or more. So in fact, attending to the numbers (and I do; it's my business), the failure rates are entirely acceptable and more than made up for if the risks are managed well."
I'd love to see where you get your numbers, but even if you totally accept them, the companies aren't doing as well as you would predict. Are you really claiming that the risk profile is similar to McDonalds? If so just be clear and state it.
Posted by: Sebastian Holsclaw | March 28, 2006 at 07:28 PM
Sorry, just to round out my last point...
Marketing, by which I mean advertising to consumers, hosting seminars and conferences, and building product brands, incurs very high and increasing costs, on the order of tens to hundreds of millions of dollars per year for a leading brand. Samples and goodies (calendars, pens), are insignificant in comparison: the actual cost of a of sample of pills is in the range of pennies and nickels.
You might be thinking of sales discounts, which are the negotiated price reductions to large buyers like insurance plans. Those are indeed large, but they are accounted for separately.
Posted by: bleh | March 28, 2006 at 07:28 PM
But there is the choice problem. The government isn't nearly as good at figuring out what particularly to fund as the market as a whole is.
It's possible that replacing patents with some sort of prize system could fix this. Even if government isn't good at predicting success, it might do all right at rewarding it. I'm not arguing for this, though, just pointing out the possibility.
Posted by: Christopher M | March 28, 2006 at 07:45 PM
"As to the latter, and to stock prices of Merck and Pfizer, there is a lot more at work. Merck was hit by fear of lawsuits over Vioxx and Bextra (and imho by the loss of some talented executives). Pfizer has several blockbuster drugs coming off patent in the next few years."
Whoa there, you just made a huge dodge. You can't dismiss the risks inherent to being a drug company when we are talking about the risks of losing money investing in a drug company. Having your top drugs come off patent and not being able to replace them with new drugs is exactly why investing in drug companies is risky. Having you compnay sued for super-minimal risks is in fact a risk of running a drug company. Having a jury award $253 million for a heart problem that isn't even of the type at issue in the science is one of the risks of running a drug company. You can't just dismiss those risks in a discussion on drug company risk. Drug companies have to make enough profit to cover their riskiness.
Posted by: Sebastian Holsclaw | March 28, 2006 at 07:51 PM
from the lawyer in the crowd:
Isn't there a huge difference between ROI and stock price?
from my utterly naive viewpoint, I would expect (a) industry ROI to be higher than other industry ROIs; (b) any one company's ROI being more variable over time against the industry average; and (c) stock price having virtually no correlation.
(a) & (b) -- because of the speculative nature of drug development.
(c) -- because stock price captures a lot more than ROI, including litigation and political risks.
Posted by: Francis | March 28, 2006 at 07:51 PM
Rats, I DID refresh that time, but I just missed it.
Phases II and III are low-risk relative to earlier, exploratory phases, where the real squeeze in the funnel comes. By the time a drug gets to phase III, it typically has perhaps a 1-in-2 or 1-in-3 chance of successful commercialization, and of course data early in the phase can halt the program long before the full costs of the phase have been incurred. Phase IIA is, of course, higher-risk -- typically phase IIA is seen as more similar to phase I, since efficacy data usually are not available, while phase IIB, like phase III, is seen as more of a statistical exercise -- but phase IIA is at least an order of magnitude (sometimes two) cheaper than phase III. Again, this is a manageable trajectory, and it does not justify exorbitant prices.
I'm not sure how a comparison to the cardio risk of Vioxx is relevant: certainly some effects do not become apparent until many more patients have taken a drug than it is possible to test in clinical trials, but often that is simply due to the time required for them to become apparent rather than to statistical frequency, and in this case it seems increasingly clear that there WERE data concerning the cardio effects of Vioxx available to Merck during clinical trials.
As to data sources, I would suggest publications by the Tufts Center for the Study of Drug Development, or of course into the planning factors used by pharmaco portfolio groups if you have access to those.
Posted by: bleh | March 28, 2006 at 07:51 PM
I'm afraid I would have to say that not having enough products in your pipeline is very much NOT a risk particular to pharma development, and therefore does NOT justify higher returns to the pharma industry. It is a management failure, pure and simple. If it justified higher returns, then so should McDonalds have higher returns, since it is possible that they might not invest in enough trucks to move frozen burger-patties around.
And as to jury awards, as noted above, that increasingly looks like a management failure -- they knew about the problems. And even if it weren't, it does not support higher returns for pharma in particular, unless you wish to argue that ANY industry that MIGHT be subject to high jury awards at ANY time deserves higher returns; by that standard, we should be paying the building materials industry -- one of the LOWEST-return industries -- big premiums for future asbestos-related litigation.
Posted by: bleh | March 28, 2006 at 08:00 PM
"Phases II and III are low-risk relative to earlier, exploratory phases, where the real squeeze in the funnel comes. By the time a drug gets to phase III, it typically has perhaps a 1-in-2 or 1-in-3 chance of successful commercialization..."
Right, but phase IIB is riskier than III and IIA is riskier than IIB. All three are quite expensive to deal with.
"Again, this is a manageable trajectory, and it does not justify exorbitant prices."
Well of course it is manageable, people do actually invest in pharma. If everybody always lost it wouldn't be a big sector. Once again you are focusing on a limited subset of risks. Pharmaceutical companies also face huge down-side lawsuits (even when they are not at fault see Bendectin) they face short post-FDA patent periods compared to other companies, they face highly uncertain pipelines, they face unforseen post-market side effects. But frankly, the research alone is a big deal. Very few companies spend such a large percentage of their total inputs on developmental research.
Posted by: Sebastian Holsclaw | March 28, 2006 at 08:22 PM
Sebastian
But there is the choice problem. The government isn't nearly as good at figuring out what particularly to fund as the market as a whole is. If the government could figure out what to research that idea might work, but it can't process the information properly without the market. The government would be better off targeting specific market failure scenarios--like vaccines or treatments for low frequency diseases.
In terms of health care/pharma (am I casting too wide a net?) I think your stated exception largely swallows the rule. I don't think the market is especially rationale when it comes to new drug development because it really can't be. Is there another situation where the 'perfect information' assumption is less apt?
Also, if the 'market' fully drives research, we get viagra and rogaine, which are nice, I suppose (I say at 29 with a full head of hair...) but I can't help but think that better things could have been developed.
Posted by: Pooh | March 28, 2006 at 08:57 PM
"In terms of health care/pharma (am I casting too wide a net?) I think your stated exception largely swallows the rule."
How so? Heart disease, Alzheimers, diabetes, stroke therapies, all sorts of the things we normally think of as needing treatment wouldn't fall into the exception.
"but I can't help but think that better things could have been developed."
Viagra wasn't initially developed for erections. If was a side effect of another test. Unfortunately that is often how science works. You can't always predict what (or when) you will find.
Posted by: Sebastian Holsclaw | March 28, 2006 at 09:12 PM
"Very few companies spend such a large percentage of their total inputs on developmental research."
What's the percentage of "research" that's devoted to finding a patentable way to extend a claim on a variant of a current drug, and what's the percentage devoted to actual new drugs that treat serious disease?
Posted by: Gary Farber | March 28, 2006 at 09:39 PM
Without having much to add, I'd just like to say that this debate is one of the more interesting I've seen online for some time.
Posted by: Hodgepodge | March 28, 2006 at 09:40 PM
Sebastian:
I appreciate the post but, for pragmatic and ethical reasons, will avoid any further comment. As I've mentioned in the past, I represent several pharmaceutical companies (which are on the side of truth and justice, of course).
Posted by: von | March 28, 2006 at 10:56 PM
"As I've mentioned in the past, I represent several pharmaceutical companies (which are on the side of truth and justice, of course)."
RNC attack commercial version: "Well-known blogging lawyer Von defends drug dealers as his clients. Do you want to vote for a lawyer who defends drug dealers?"
;-)
Posted by: Gary Farber | March 28, 2006 at 11:02 PM
Well, there does seem to be consensus on the primary points. Pharma IS inherently riskier than many other industries, even if you take account only of what I think is appropriate in considering return to investors, i.e., the scientific, clinical, and commercial risks of drug development, and not the effects of poor management, e.g., not keeping your pipeline stocked, ignoring clinical trial data in a way that exposes you to lawsuits, or gambling by spending way too much on advertising in support of the "blockbuster" model.
And that risk commands a higher return.
It's also certainly true that pharma spends more (as a fraction) on research than almost any other industry, even accounting for the subsidy provided by government-funded BASIC research. (The latter -- please note emphasis -- is unequivocally a GOOD thing, BTW; I certainly hope we don't need to get into the Pareto-superiority of government funding of public goods or the wisdom of the Vannevar Bush -- oh irony -- policy of government-sponsored university leadership of R&D in the US.)
As to the question of the "market choosing" how to allocate pharma R&D dollars, that turns out to be a very interesting problem. I think Sebastian is correct in the larger sense that pharmacos, following "medical need" and market opportunity (and appropriately corralled by the FDA, I will say), have improved medical care and the quality of life ENORMOUSLY both in the US and abroad. Even a little bit of historical research will reveal how much we have improved health and reduced costs for what are now seen as run-of-the-mill drug-treatable diseases but used to be killers. (Note the qualification: overall costs have gone up, but that's because what we treat has changed and the new stuff is more expensive; if we treated today only what we did 50 years ago, we'd be spending a lot less, and we'd have a lot less for doctors to do, thanks to drugs produced by mostly market forces.)
However, I think Pooh is correct about a couple of phenomena "on the margin." One is the tendency in the industry to develop "lifestyle drugs" for the "worried well" because it's a lucrative market (especially if you can scare people with advertising -- Do you sometimes find yourself nervous in social situations? You might have Social Anxiety Disorder! Ask your doctor!) rather than focusing on more serious medical needs. A second one is that, when considering how to allocate R&D money over a pipeline, even a sophisticated risk-adjusted expected-value analysis always favors those later in the pipeline, both because they've overcome risks and because they're closer to commercialization. But that leads to "starved" pipelines, such as many companies are experiencing today (thank you, Wall Street earnings expectations). Fortunately, the pharmacos are learning how to deal with the latter problem, and for the former one there are government market corrections such as Orphan Drug designation (thank you government, for real).
(I still don't agree re lawsuit costs and stock prices, but those are secondary issues, and it's late. Perhaps another time...)
Posted by: bleh | March 28, 2006 at 11:07 PM
Yes, I should have prefaced my comments by saying that SH is correct as to the desirably of pharmcos earning bigger margins on 'succesful' drugs (as defined by those that actually reach the market). That only gets us halfway there, I think, as the magnitude of that excess margin is still open for debate.
My personal intuition is that they could probably afford negotiated pricing, but I am willing to be convinced otherwise.
Posted by: Pooh | March 28, 2006 at 11:17 PM
Wow, this is a REALLY good debate! I can feel my brain expanding as we speak.
Ok, I'm going to step in here and show my
ignorance/libertarian bias, but here it goes:
1. How does the crowd here feel to a private company like Underwriters Laboratories taking over the job of the FDA in the drug certification process as a potential way to reduce the barrier of entry (assuming that a private organization usually is more efficient than a government one) by potentially lowering the amount of money needed to go through the approval process? I know this is politically unfeasible and causes most people the heebee-jeebees but I just want to test the waters...
2. Should we be criticizing how much money a pharmco makes, even when it's use is in blockbuster drug ads, if it marginally raises it's profits and allows for more investments in research?
3. Why should we worry about the "lifestyle" drugs if they in turn help research into hard-to-cure ailments, like cancer?
I know these ideas aren't going to fly too high here (especially the FDA vs. UL idea), but since the conversation is so good, I figure any answer is better than none :)
Posted by: Frank_A | March 28, 2006 at 11:59 PM
Frank A: the usual argument for greater efficiency in the private sector turns on the availability of competition. If you outsourced to only one 'private FDA', there would be no competition. If you outsourced to a whole bunch, bear in mind that their 'customers', with whom they would be competing for business, would be the pharmaceutical companies, who, um, might not be interested in having the very best, most aggressive monitor of drug trials around.
One of the main reasons why, sometimes, government is best suited to run an agency is precisely because that agency ought to serve some public interest not well captured by market mechanisms. This would seem to be such a case. (The armed forces are another.)
Posted by: hilzoy | March 29, 2006 at 12:27 AM
How important to the cost-risk analysis for pharm companies are the various financial intermediaries that manage funds?
When I came into a little money a number of years ago, I stuck it into one of those "socially conscious" index funds that excludes so many sectors and companies that it ends up mostly investing in tech and drugs. Obviously, it hasn't done that great--but it hasn't lost me money either, and I haven't had to think about it at all.
While I doubt that there are very many individual investors who give up better returns this way, some state investor groups or wealthy universities might have restrictive requirements on investment that would move them towards pharma in preference to similarly risky industries.
It's worth nailing the math down, of course, but the math won't be the only factor in play for investors and investor groups, and that's worth acknowledging as well.
Posted by: Jackmormon | March 29, 2006 at 12:32 AM
A couple of comments. One, Sebastian is rigging the game by looking at returns over the last five years. Five years ago, pharmaceutical companies had taken over the role of Can't Miss Wall Street Darlings, replacing telecoms and internet stocks. Anyone who thought they would maintain their valuations (JMN sheepishly raises his hand) was out to lunch.
Two:
Viagra wasn't initially developed for erections. If was a side effect of another test. Unfortunately that is often how science works. You can't always predict what (or when) you will find.
Well, yes. Viagra isn't the only drug for which this is true. The totally-out-of-the-blue blockbuster is something that has to be considered when looking at drug companies. It is a wild card that increases the potential upside.
Three. Sebastian's basic argument mostly demonstrates that pharmaceuticals are a dysfunctional market. In general, we expect the market price of a product to converge to the marginal price of producing the next unit. Very clearly, drugs do not meet this expectation, at least while they are under patent. Thus, any discussion of them must work from the realization that classical market analysis simply can't be applied. That doesn't necessarily mean that Sebastian is wrong, but there is an unspoken assumption in his post that ought to be made explicit.
Posted by: J. Michael Neal | March 29, 2006 at 12:36 AM
hilzoy,
That's a pretty good point that if only one and only one UL-type company had government approval to certify drugs, then we would have the standard problems with government-monopolies.
However, if there were multiple "ULs," seemingly there certifcation would have to be a two-way street between working with pharmcos and keeping credibility with the consumer. I would think that if the Bextra-scandel thing occured under "Pharmco Researchers", that would smash their credibility and so another competitor would be able to take the stage in certification.
As for actual examples, in terms of home and electronic appliances I only know about UL doing their safety certification (however, if there are more companies or if the government is very involved in regulating UL this assumption is moot).
Since the profit motive is very strong in all for-profit organizations, wouldn't the same effect also occur in these areas that UL works in? I checked out Wikipedia (http://en.wikipedia.org/wiki/Underwriters_Laboratories), and I coulnd't find any criticisms in it's entry on UL practices, which may suggest that UL really does do a pretty good job at certifying safety. But then again I admit my ignorance of the field...
Posted by: Frank_A | March 29, 2006 at 12:49 AM
comments from the utopian would be techno dweeb. I think that companies should be able to reduce their future risk by discovering linkages to gene markers for determining which drugs work and being able to target those drugs as well as filling the pipeline by exploiting third and fourth world countries for compounds occurring in nature which may yield potential drugs (not good, just saying), not to mention the already advanced steps that have been taken to limit liability through add ons to legistlation. This suggests that the amount of "risk" involved can be reduced by technology and legal gamemanship, so if the drug companies' motive is purely profit, they would have every reason to dismiss or understate such steps, but if the possibility exists, they should be taken into account in making a market analysis. This is not to suggest that Sebastian is wrong (as I think he may be dealing with only the present situation and specifically be ruling out future trends), but if those other factors are not considered, we would not be objectively analyzing the situation.
I realize that I don't know what the situation for drug prices is in Japan, nor do I know if what japanese companies engage in drug research. I'll try and do some research, but if it is the case that Western pharmacetical companies are bearing the bulk of the risk and countries with an organized national health care program are able to negotiate for lower prices, then the drug companies are transferring the liability to the US market because it is not organized enough to negotiate for the discounts. This could be seen as a problem (though not for me, as I am on Japanese national health)
Posted by: liberal japonicus | March 29, 2006 at 12:57 AM
Just to be clear, the utopian techno dweeb is me, not anyone else.
Posted by: liberal japonicus | March 29, 2006 at 12:58 AM
I'll try and do some research, but if it is the case that Western pharmacetical companies are bearing the bulk of the risk and countries with an organized national health care program are able to negotiate for lower prices, then the drug companies are transferring the liability to the US market because it is not organized enough to negotiate for the discounts. This could be seen as a problem (though not for me, as I am on Japanese national health)
Posted by: liberal japonicus | March 29, 2006 at 12:57 AM
..............................
From what I understand, this is correct since Americans are better at paying for meds than nationalized medicine/can afford to buy their medicine, that the USA bears the burdern of higher costs.
For example, an AIDS triple-cocktail costs over $1000/month in the US (last time I checked the figures) but costs a fraction in Africa (which is still a BIG burden but better than $1000/month).
Here's what CBS had to say about them in 2003: http://www.cbsnews.com/stories/2003/04/28/health/main551324.shtml
AP) GlaxoSmithKline on Monday further reduced the prices of its AIDS medicines for the world's poorest countries by up to 47 percent — the fifth time since 1997 that the company has cut prices.
...........
The discounts are available to the world's least developed countries and all of sub-Saharan Africa — a total of 63 countries.
The latest reduction lowers the price of Combivir, central to most HIV/AIDS treatment regimens, by 47 percent to 90 cents a day. The price of Epivir fell 45 percent to 35 cents a day, while Retrovir's price declined 38 percent to 75 cents a day.
In the United States and Europe, an AIDS treatment regimen averages about $15,000 per year. Under pressure from activists, the companies that manufacture AIDS drugs have cut prices for developing countries, and now many, like Glaxo, sell them there at no profit.
........................................
Posted by: Frank_A | March 29, 2006 at 01:15 AM
Frank A,
"A couple of comments. One, Sebastian is rigging the game by looking at returns over the last five years. Five years ago, pharmaceutical companies had taken over the role of Can't Miss Wall Street Darlings, replacing telecoms and internet stocks. Anyone who thought they would maintain their valuations (JMN sheepishly raises his hand) was out to lunch."
I don't know what you mean be rigging, but you would be hard pressed to find consistent winners on a 3 year look, 7 year look, 10 year look or 12 year look. Closest might be Pfizer, which basically lucked in to two drugs and half lucked in to a third.
"Three. Sebastian's basic argument mostly demonstrates that pharmaceuticals are a dysfunctional market. In general, we expect the market price of a product to converge to the marginal price of producing the next unit."
Come on now. Did you really read this whole thread and somehow miss how expensive research is? It isn't generally considered a dysfunctional market to pay for the costs of something.
L_j,
"This suggests that the amount of "risk" involved can be reduced by technology and legal gamemanship, so if the drug companies' motive is purely profit, they would have every reason to dismiss or understate such steps, but if the possibility exists, they should be taken into account in making a market analysis."
I didn't quote your explanation of what you thought ought to work. I did that on purpose. Whatever it is that should work, call it 'X'. If there isn't a monopoly (and there certainly is not a monopoly on medical research) any known useful technique for making discovery more efficient would give the company that used it a huge edge--making it more profitable. It is like complaining about the water engine or a conspiracy to not cure AIDS. If I had a cure to AIDS, it doesn't matter that theoretically it would kill off a health sector because I could be fabulously rich and famous. Basically you are positing a perfect cartel where no one cheats to get ahead of the other ones even though it would make the first cheater fabulously wealthy.
"I'll try and do some research, but if it is the case that Western pharmacetical companies are bearing the bulk of the risk and countries with an organized national health care program are able to negotiate for lower prices, then the drug companies are transferring the liability to the US market because it is not organized enough to negotiate for the discounts."
You are close. The proper economic term is "free rider". Countries like Canada free ride off of US research--they get the benefit of it without paying the price of it. The free rider problem is always at its worst when enough people are free riding that the train is not getting its upkeep paid for. Right not the US is steeply subsidizing the pharmaceutical research of the rest of the world. Just because some people can get cheap prices by free riding doesn't mean that free riding will pay for the train if everyone does it.
As for the FDA, I don't mind a gatekeeper but I think that a better cost-benefit calculus needs to be implemented. Drugs have side effects. People need to realize that.
Posted by: Sebastian Holsclaw | March 29, 2006 at 01:43 AM
Bloody hell. Blogger is down, and won't let me post. There hasn't been an open thread here more than a week, and there's none on the sidebar. I don't want to post a completely extraneous comment here that has nothing to do with any current thread.
I'm going to post a long comment on the last open thread here. People can go look for it, or not.
It directly concerns you, Hilzoy.
Posted by: Gary Farber | March 29, 2006 at 01:45 AM
Oh, crap. There isn't even an open thread left on the front page here, any more. Now I really don't know where to post what I want to say.
Posted by: Gary Farber | March 29, 2006 at 01:46 AM
Okay, Blogger came back, so problem solved.
This is still off-topic, but short: Don't expect any links from Garance Franke-Ruta at Tapped, Hilzoy.
Here's why.
Posted by: Gary Farber | March 29, 2006 at 01:49 AM
Sebastian,
You got the wrong guy.
Those comments were: Posted by: J. Michael Neal | March 29, 2006 at 12:36 AM http://obsidianwings.blogs.com/obsidian_wings/2006/03/a_math_look_at_.html#comment-15522128
Mine were:
- Posted by: Frank_A | March 28, 2006 at 11:59 PM
http://obsidianwings.blogs.com/obsidian_wings/2006/03/a_math_look_at_.html#comment-15521479
- Posted by: Frank_A | March 29, 2006 at 12:49 AM
http://obsidianwings.blogs.com/obsidian_wings/2006/03/a_math_look_at_.html#comment-15522341
- Posted by: Frank_A | March 29, 2006 at 01:15 AM
http://obsidianwings.blogs.com/obsidian_wings/2006/03/a_math_look_at_.html#comment-15522341
Posted by: Frank_A | March 29, 2006 at 02:03 AM
I didn't quote your explanation of what you thought ought to work. I did that on purpose.
I don't mean to be picky, but I wonder if I missed a comment or if you had an open window and didn't send a comment in, because I don't see any comment of yours between mine and this. If you meant to write "I'm not quoting.." fine.
But moving onto the actual subject, if you just want to simply label it as X for the purposes of your analysis, it is important to realize that it is not imagining a time machine or transporter, but something that is quite far advanced. (my own interest is the intersection with property rights of 4th world people, just to explain how I came to read some stuff about it) I'm not positing a perfect cartel, I'm just suggesting that a discussion about the risks that pharma companies take should be informed by possible technologies on the horizon. Hence calling myself a 'utopian techno-dweeb'.
As for free-riding, this means that either the US government better get on the stick and organize national health care, or that the US find a way to keep drug discoveries within the US. Since the latter is untenable, the former is something that arises from your discussion of the risk that pharma companies face and simply presenting the math for why pharma companies need higher returns is too narrow a frame. If it was realized that the drug companies need higher returns from US consumers, the tenor of the debate would change dramatically. And we all know how American consumers don't like stuff taken out of their pockets.
Again, you may want to check if you have something you have written that is in another tab, it's something that I've found myself doing at HoCB recently.
Posted by: liberal japonicus | March 29, 2006 at 03:19 AM
"As for free-riding, this means that either the US government better get on the stick and organize national health care..."
How does that follow as an alternative? People already think things like "If Canada free rides why can't we?" National health care will lead to price setting like Canada, which is to say much closer to marginal cost without counting research costs. You already hear people talk about how it "can't" cost 5 dollars per pill.
When people talk about national health care saving money, this is what 'saves' money on drugs. Since pharma companies make back their research costs here, if we take that away they aren't making back their research costs. I would rather have new drugs all the time and have to pay a premium for ten years (which isn't what really happens because the dreaded 'me-too' drugs drive down the prices when they are truly equivalent) than pay less and reduce the research level.
Posted by: Sebastian Holsclaw | March 29, 2006 at 04:48 AM
Having a jury award $253 million for a heart problem that isn't even of the type at issue in the science is one of the risks of running a drug company.
That's an American risk though. In this part of the world there are not such high awards (also because health costs and income decrease are less in case of problems). And from a Public Citizen 2002 report I found:
About the marketing costs: There even is a blog that is dedicated to pharma marketing and my impression is that quite a lot of money is put into marketing. The International Herald Tribune: "According to Jon Hess of Cutting Edge Information, a research company, total ad spending by drug companies in Europe is only 13 percent lower than in the United States. On both sides of the Atlantic, drug companies spend about 50 percent more promoting product awareness than they do for research and development, according to Stewart Adkins of Lehman Brothers". And that is just product awareness...
In quite a few articles I read that lobbying costs the industry quite a few million per year. Doesn't that also have the negative side-effect that they might try to hamper competition from other sectors? Theoretically if you produce a drug that is essential for diabetici, an enourmous growth market, you would not like it if a surgical solution (stem cells, transplantations) might get rid of the disease, so you'd try to delay those kind of findings as long as possible.
As another negative side-effect from the privatisation I quote from Merrill Goozners blog:
This link I just have to put in to pester Sebastian. Can't..... resist.... temptation....
Posted by: dutchmarbel | March 29, 2006 at 04:57 AM
Sebastian, in 2001 Merck admitted booking $12bn of non-existent revenue. Do you think that might have had something to do with their share price drop? And doesn't it have more to do with corporate governance than the nature of the pharmaceutical industry. The fact remains that big pharma has consistently higher profit margins than any other industry. Yes share prices may fluctuate a lot, but that has more to do with the stock market than the fundamentals of the pharma industry - nobody is going to go broke holding on to Merck or Pfizer stock.
And besides, there's nothing stopping big pharma companies increasing their debt to equity ratio - their stable returns are exactly the sort of thing debt investors like, and most of them have more cash in hand than they do debt. But then the senior management's stock options wouldn't be so lucrative, would they?
Posted by: Ginger Yellow | March 29, 2006 at 05:39 AM
Since pharma companies make back their research costs here, if we take that away they aren't making back their research costs.
Sorry I wasn't clear, but that actually was my point, but I was afraid that I might be putting it too sharply. Taking that to be the case, do you think this system is sustainable? Or rational? (Though I admit, a lot of people have made a lot of money by betting against the American people being irrational)
I'm also a bit worried that this exacts a cost on the US. I do think that this has an impact on the competitiveness of the United States, so the 10 years that you are paying a premium, you (and the next generation) will find yourselves in a country that is not a global leader.
I've been trying to find some comparisons between international prices for drugs, and it's not as easy as I thought it would be. This pdf explains why (the url looks a little funky, so if you google
PharmacoEconomics 14 (Suppl. 1): 115-128, 1998
you can find it)
Posted by: liberaljaponicus | March 29, 2006 at 05:46 AM
So in order to have an expected ending value equal to the McDonalds investment you would need to earn 18.33% on the up side of the riskier investment compared to the 13.7% rate of the McDonalds investment. Of course you would be a fool to make the investment at only 18.33% because your chance of losing money is greater in the risky investment while your expected value is the same.
No. A rational investor shouldn't care which of two investments with the same expected value he picks. (Most investors do, for the psychological reasons already mentioned, but that is not the same thing.)
And in any case, even if p/e ratios have to be higher in pharma to offset higher earning volatility (which is what you are talking about) that does not justify government action to inflate pharma profits at taxpayers' expense!
Posted by: ajay | March 29, 2006 at 05:48 AM
Dutchmarbel,
that pharma marketing blog looks a lot like my daily spam folder!!
Posted by: liberaljaponicus | March 29, 2006 at 05:56 AM
Frank A,
As the individual involved pointed out, I'm not Frank.
One thing to consider is that, over the last five years, pretty much everything that could go wrong for pharmaceutical companies has gone wrong. Several of them have been caught playing fast and loose with clinical data. A number of them have let their pipelines go dry. And so on. Considering all that has happened, their performance since 2001 isn't that bad.
You do not seem to realize that you are arguing that expected returns should be higher than the market average, and also that realized returns should be higher than the market average, despite the fact that the higher levels of risk materialized. Sorry, but the higher expected returns means that you are assuming risk. If that assumed risk goes south on you, you aren't going to realize your expected returns.
You seem to be arguing that investors expected returns should be guaranteed. In that case, what risk are they assuming?
How does that follow as an alternative? People already think things like "If Canada free rides why can't we?" National health care will lead to price setting like Canada, which is to say much closer to marginal cost without counting research costs. You already hear people talk about how it "can't" cost 5 dollars per pill.
Well, the argument here is that the rest of the world shouldn't be free riding off of the US health care system. Without taking a position on whether this is what is actually happening, a real believer in markets ought to say that the US should bargain as hard as it can, and force the drug companies to raise prices to other health care systems. That would produce a system in which everyone is sharing the R&D costs. Why is it that you want to perpetuate distorted markets?
Posted by: J. Michael Neal | March 29, 2006 at 05:56 AM
liberal Japonicus: I'm biased, I LIKE marketing :)
People already think things like "If Canada free rides why can't we?" National health care will lead to price setting like Canada, which is to say much closer to marginal cost without counting research costs. You already hear people talk about how it "can't" cost 5 dollars per pill.
Sebastian: Marcia Angell disagrees with most of what you say, and on this subject she says:
And also:
Read the whole article I'd say, I thought it made some very valid points
Posted by: dutchmarbel | March 29, 2006 at 06:45 AM
Sebastian,
I just realised what you meant by 'I didn't write'. Duh. Sorry about that
Posted by: liberaljaponicus | March 29, 2006 at 06:45 AM
this pdf file is an article originally published in New Doctor 81,
Winter 2004. It states:
Posted by: dutchmarbel | March 29, 2006 at 06:59 AM
RNC attack commercial version: "Well-known blogging lawyer Von defends drug dealers as his clients. Do you want to vote for a lawyer who defends drug dealers?"
Hey, every word is accurate ;-)
(Though "Big Pharma" is usually considered a Repubican constituency.)
Posted by: von | March 29, 2006 at 07:14 AM
I work for a five-man startup company in an industry where our products sell once or twice a year for half a million dollars a pop. It's a roll-the-dice sort of business when you're our size; one client can set you up for a year, but losing a sale can ruin you.
So I definitely understand the need for better return on risky investments in companies that need millions to do their development.
The difficulty is that pharm companies are developing (in many cases, at least) drugs that people need to live, or to overcome deadly diseases. I have no trouble with someone charging a hundred bucks a pop for hair-growth drugs or viagra. Those optional items are where the free market works best.
How should we, as a society, approach concepts like treatments for cancer, though? Vaccines for AIDS? Is that the sort of 'how much is it worth to YOU' thing we want supply and demand regulating? I'm not suggesting subsidizing anything deemed 'important enough,' or screwing over companies that have put huge effort into curing deadly diseases. Many companies would simply stop developing anything but 'elective' drugs if that threat were there.
As a civilized culture, though, how should we approach those scenerios when return-on-investment demands that people die unless they can help the bottom line?
Insurance, obviously, is part of the equation, but obviously not the whole part.
Posted by: Jeff Eaton | March 29, 2006 at 09:14 AM
Going to answer a bunch of people quickly because I'm in a hurry.
"The difficulty is that pharm companies are developing (in many cases, at least) drugs that people need to live, or to overcome deadly diseases."
Yes. And if we don't let them make profits how precisely will that need get filled?
"Yes share prices may fluctuate a lot, but that has more to do with the stock market than the fundamentals of the pharma industry - nobody is going to go broke holding on to Merck or Pfizer stock."
No it doesn't and yes they could. Please look at their price shares and compare them to the SP500 if you want to see how the stock of those two companies has done vs. the market. A quick look (not calculating dividends because it is tedious will overstate my case. I'm sorry for that but the dividends aren't big enough to overcome the point.) shows 1 year--SP500 +10%, Pfizer -4%, Merck +11%; 2 year SP500 +16%, Pfizer -25%, Merck -18%; 3 year SP500 +50%, Pfizer -20%, Merck -30%; 4 year SP500 +15%, Pfizer -35%, Merck -33%; 5 year SP500 +13%, Pfizer -35%, Merck -50%; That is losing money and losing it big time. The interesting thing in looking at the charts is that both Pfizer and Merck had serious declines even before their recent bad news. This reinforces a point that also touches on about half of what dutchmarbel is gathering statistics on. The profitability of any year's top ten profitmakers doesn't speak to volitility if the top ten aren't the same year to year. Of course you would do well any year if you pick the pharma top ten for that year in advance. The problem is you usually can't. Hence, 'risk'.
Dutchmarbel, you will have to relink. But I suspect before looking at the data that European research will on inspection often be done in the US and making money in the US.
Posted by: Sebastian Holsclaw | March 29, 2006 at 10:35 AM
"The difficulty is that pharm companies are developing (in many cases, at least) drugs that people need to live, or to overcome deadly diseases."
Yes. And if we don't let them make profits how precisely will that need get filled?
Well, that's precisely the problem. It IS necessary for a corporation to be profitable at some level. But profit is not the only human motivation, and sometimes it is to the benefit of humanity to do unprofitable things.
I'm not arguing that monetary profit is bad -- just that many good and essential things are unprofitable from a monetary standpoint. Our culture's moral and ethical vocabulary seems to have collapsed to the language of economics. It's a fine and good language, but it only captures one part of the human equation.
Please don't think that I'm slagging on your core point. Rather, I'm trying to think about the core concern that causes many people to lash out at the 'big profits' pharm companies reap when they find a viable product.
Posted by: Jeff Eaton | March 29, 2006 at 11:08 AM
I’m not sure if I’ll be responding further today because things are shaping up to be busy but I want to make a point on this:
“No. A rational investor shouldn't care which of two investments with the same expected value he picks. (Most investors do, for the psychological reasons already mentioned, but that is not the same thing.)”
This is a formally accepted model in much of economics, but I don’t actually believe it is correct. You should understand that I believe people are simultaneously much too risk averse about things they don’t understand and embrace too much risk about things they feel are commonplace—see for example a smoker worrying about depleted uranium (I’m pretty sure that person exists). That said, I think the formal model doesn’t capture perfectly legitimate concerns.
Say you have $100,000 as a substantial component of your retirement seed money. You have only two investment options, a sure thing at 500% over 20 years and a 0.001% chance at $50Billion with a 99.999% chance at $0. Both have an expected return of $500,000. But you don’t get invest your retirement 100,000 times. You only have one retirement egg. Even though both investments have the same expected return, I don’t think you should favor both investments equally. In fact, I don’t think the question is even close. A rational investor would be a complete idiot to invest in the second investment with his main retirement seed money. What is he going to retire on the 99,999 times out of 100,000 that the risky investment misses? The fact that most currently used economic models have trouble distinguishing between those two cases is not a problem with the rational investor. (Yes I realize that some economists have been working on the problem).
Posted by: Sebastian Holsclaw | March 29, 2006 at 11:34 AM
"In a recent WashingtonMonthly post I’m once again seeing people who don’t understand why pharmaceutical companies need to earn more profit than average companies in order to attract investment."
[snip - very dumbed down basic finance]
Err, Kevin's an MBA, which means he's taken more finance courses than your good self.
But the answer is: So what? There are plenty of technology-orientated companies that have to take on risky investments, be it building an ethylene cracker or developing an aircraft for the Pentagon. If you aren't allowing Medicare to even use the lowest-bidder process that the Pentagon uses, you sure are allowing your supplier to be supremely fat and happy.
The issue isn't whether pharma is more risky than T-bills or the S&P index. The issue is whether the economic rents that the pharma companies are collecting, given their patents and the inelasticity of demand, is in excess of that *justified* by the underlying (not market) riskiness of the investment.
Studies suggest, yes, the pharma companies are collecting rents in excess of the riskiness of their R&D investment, never mind poor investments made because of the absence of price pressures on their product.
" The non-math answer is that pharmaceutical companies are riskier (lots of them go out business providing a -100% return to their investors."
First, let's distinguish between pharm companies and biotech.
Pharma companies - big, have established distribution chain, mostly East Coast of US (esp. New Jersey), British, or Swiss. Positive cashflows from existing products.
Biotech firms - small, VC-funded, mostly West Coast (SF and San Diego), Austin, Boston, Research Triangle, and some concentration in Israel, Britain and Germany. No distribution chain because they don't yet have products. Negative cashflow.
Pharma companies don't usually go out of business. VC-funded *biotech* firms go out of business, 'cos their Phase II or Phase III results aren't good enough to raise another round of funding. Even then, there's usually sufficient intellectual property to get some penniesrecovery of the investment.(As a matter of fact, VC funds that invest in pharma return less than hi-tech VC funds, but still manage to get funding from their limited partners because the volatility of their returns is less.)
"Pharmaceutical companies are much riskier than the general US stock market, so they need much higher returns on the upside to make the risk worthwhile."
But this is a circular argument, Sebastian. The market determines the riskiness of the stock in terms of its price volatility as well as its price. You have to both refer back to the volatility of actual earnings as well as the volatility of the stock price.
IIRC a study of the actual fluctuations of earnings versus returns for the pharma sector concluded that returns for the pharma sector were 2% higher than actual fluctuation of earnings justified.
"And digging even deeper, the underlying culprits are the pharmacos' poor (though improving) practices in controlling development risks (primarily their unwillingness to kill unpromising projects early),"
Spot on. Also, at least five years ago, the pharma companies were unwilling to use available tools to reduce their risk, such as pharmacogenomics, because it might reduce their market to only those with a particular SNP. There's the strong desire for the big blockbuster, rather than a string of small winners. (That oversight has led to some successful bottom feeders, like Shire and Cambrex, who take a 'failed' drug for a big market and exploit it for a small market.)
Posted by: Urinated State of America | March 29, 2006 at 11:48 AM
It also does not hurt to remember how rarely a drug company lands a saleable product. It takes an average of 20 years between the discovery/development of a compound and FDA approval of drug, and only about 1 in a thousand compounds make it that far (at best one in a hundred even make it out of the lab to proof of concept).
I work in the industry and I agree wholeheartedly that pharma needs to plow revenue back into R&D and stop spending it on hyperaggressive marketing that is dangerous for patients and overwhelming for physicians. If they did so they might actually find themselves with new drugs for currently untreatable conditions to sell, instead of betting their lives on copycats, which seems to be the predominant strategy in the industry right now. They would also do well to divert $$ from marketing to Phase IV trials, that is, trials of approved drugs to see how physicians are actually prescribing them in the real world (as opposed to the rigidly controlled world of the pre-approval clinical trial) to real patients. This is where the REAL safety data is found but very few companies spend the money necessary to build a drug's safety profile in an uncontrolled population.
Posted by: Constance Reader | March 29, 2006 at 11:50 AM
Jeez, I hate to repeat myself, but it seems that a couple of the points I made earlier have evaporated from the discussion.
Pharmacos DO NOT NEED the level of profits they currently make in order to cover their research costs. They are WASTING a LOT of money, notably on advertising, but also due to their mismanagement of risk and of their own pipelines.
Pharmacos are making a LOT of money, although admittedly not as much as they used to per the above. Stock prices are NOT a good measure of this profitability, because they are influenced by so many other factors, e.g., herd mentality, analysts' formal "expectations," macro and political concerns. Earnings are a much better measure, and pharmacos' earnings (before extraordinary items) have been consistently high.
The conclusion is that US drug prices COULD come down to levels closer to those of the rest of the world WITHOUT materially affecting pharmaco research. (That is, in fact, the operative assumption at pharmacos right now, and it's causing no end of angst both in pharmaco boardrooms and on Wall Street; witness the flurry of huge pharmaco mergers that has taken place over the past decade or so -- always a favorite tactic, and NEVER effective at improving much anything -- as well as the recent wild gyrations in the prices of what used to be among the bluest of blue-chip stocks.)
Expected value is not an accurate measure for use in decisions under uncertainty except for those involving outcomes that are small in comparison to the decision-maker's wealth. This is because using EV assumes risk-neutrality, which is obviously counterfactual, as my example of yeterday and Sebastian's of today make clear.
Posted by: bleh | March 29, 2006 at 11:58 AM
And not to be snarky, but risk attitude has been incorporated into decision theory since at least the 1970s, although for some reason there don't seem to be many -- or even any -- good textbooks on it. The central conclusion is that expectations should be calculated with respect to utility, not dollars, and the utility of dollar-denominated payoffs is subject to diminishing returns. (Geeks: utility is not linear in the amount of the payoff but is instead concave, like a log curve). In other words, you value a dollar more when you have only a few of them than when you have a lot. You therefore are less willing to gamble on, say, a big win versus a loss rather than on a couple of modest wins, even if the expected value of the alternatives is the same.
Posted by: bleh | March 29, 2006 at 12:28 PM
Sebastian: this is the pdf file.
I read elswhere that more European firms are bringing part of their R&D to the US, but that is not because of market mechanism, but because R&D gives better tax advantages in the US and because of the connections with universities. The first is government regulation, the second is partly due to the commercialisation of universities. But I'm not sure that the latter development is a good thing, as I stated in my post of 06.45.
I bet the spending on marketing is a lot higher in the US too; there are more regulations about it in Europe. And spending on marketing and sales is more than twice what they spend on R&D.
Posted by: dutchmarbel | March 29, 2006 at 12:37 PM
""Pharmaceutical companies are much riskier than the general US stock market, so they need much higher returns on the upside to make the risk worthwhile."
But this is a circular argument, Sebastian. The market determines the riskiness of the stock in terms of its price volatility as well as its price. You have to both refer back to the volatility of actual earnings as well as the volatility of the stock price."
Indeed. I'm not denying that the stock market investors consider big pharma a somewhat risky investment. Or that small pharma is indeed a very risky investment. But investors' risk assessment for big pharma is not rooted in real earnings or dividend volatility, it's because other investors overreact to things like trials failing or denials of indication. As I said above, a rational capital structure for the industry would account for the fact that despite the high risk of individual drugs, big pharma companies have vast portfolios of products that average out these risks and create extremely high, relatively stable returns.
Posted by: Ginger Yellow | March 29, 2006 at 12:44 PM
The marketing rules for pharmaceuticals changed in 1997. If the marketing-to-consumer issues were a big portion of the expenses you should expect things to look different before that. They aren't.
1997 Pfizer 10-K See page 59 and 60. The relevant expenditure is Selling, Informational and Adminstrative. It is found directly above the research and development line item (easiest way to find it is to do a search for "informational") This helpfully provides both numbers from 1996-1986. You will note that the SI&A expense ranges from 2.5 times research to 4.2 times research.
The current 10-K shows on page 35 an SI&A amount of 2.2 times research--the low end of the pre-rule change range.
Furthermore the SI&A expense includes Administration. This would include all sorts of totally run-of-the-mill non advertising business expenses. In fact it looks to me like all non-research staff salaries are included in that expense. Other companies have to pay secretaries too. It definitely includes all legal expenses.
In short, the SI&A spending is more than the research spending argument does not seem to mean nearly as much as people think.
Posted by: Sebastian Holsclaw | March 29, 2006 at 01:45 PM
SI&A includes overhead? Did I interpret that correctly?
Last time I looked (which was quite a while ago, admittedly) my labor got billed to contract at something like 2.5 times my salary. Translating that into the context of Sebastian's comment, would it be fair to say that my teenie little piece of the R&D budget has included into it SI&A equivalent cost of 1.5 times R&D? Just trying to understand, here.
Posted by: Slartibartfast | March 29, 2006 at 02:17 PM
1997 Merck 10-k
Here the relevant listing is "marketing and administrative". It is on page 37. It shows similar pre-rule-change expenditure ratios between "marketing and administrative" and research. While I read these nasty 10-Ks for other reasons as part of my job, I'm not sure about this issue--Look at something like the above 10-K. The costs/expenses areas are pretty much "Materials and Production", "Marketing and Administrative", "Research and Development" and "Other". What category do salaries and employment costs go under? I suspect Marketing and Administrative but I'm not sure. What category does non-manufacturing building leases go under? Are research scientist salaries in the same place as everyone else's salaries or do they go in research and development? I know that legal expenses and settlements go under Marketing and Administrative because that is where Merck reported the Vioxx stuff in their most recent 10-K.
Posted by: Sebastian Holsclaw | March 29, 2006 at 02:30 PM
Sebastian I guess I'm wondering what the topic sentence is here. Are you just arguing that the pharmaceutical industry has a particular risk profile (high product development costs; ubiquitous failure punctuated by occasional disproportionate ROI spike; generally swallowed by giants if successful...)? If so, that doesn't seem at all controversial and I commend you for getting a high turnout on a non-sexy issue.
Or are you actually arguing in favor of the current regulatory landscape as regards drug comapnies? Because there seems to be a consensus (unless I missed some comment where you suggest otherwise) that the current system is, de facto, one in which the government subsidizes the pharmaceutical industry (by refusing to negotiate, by funding academic research, and by using non-tariff mechanisms to enforce international price discrepancies).
If you're arguing that the system in place is more or less appropriate and satisfactory then I wish you would say so in so many words for those of us who already acknowledge that drug development is a high risk venture. Especially if what you mean is that you're opposed to allowing any government entities to negotiate their purchase contracts with drug companies.
Posted by: radish | March 29, 2006 at 03:00 PM
Sebastian,
The fact that most currently used economic models have trouble distinguishing between those two cases is not a problem with the rational investor.
What "economic models" are you talking about? The risk-aversion you're talking about is a commonplace in economic and financial literature. See., e,g., this "tutorial". Or am I missing your point?
Posted by: Christopher M | March 29, 2006 at 03:11 PM
I talked to a finance friend and he confirmed that Marketing and Administration includes: salaries and employment expenses (research scientists probably in the research area), the cost of leasing non-manufacturing facilities, legal expenses, non-research regulatory compliance (think warning labels) and general overhead expenses. As such, the fact that marketing and administrative expenses are a bit more than double research costs shouldn't be surprising.
Posted by: Sebastian Holsclaw | March 29, 2006 at 03:50 PM
"The risk-aversion you're talking about is a commonplace in economic and financial literature."
I was responding to "No. A rational investor shouldn't care which of two investments with the same expected value he picks. (Most investors do, for the psychological reasons already mentioned, but that is not the same thing.)"
"Sebastian I guess I'm wondering what the topic sentence is here."
It kind of a preface for another post I'm working on. I find it frustrating to argue about something for a while only to realize that the disagreement was entirely somewhere else. The classic case is the time I spent days arguing about the particulars of a school reform issue only to have the discussion end with "I don't think US schools are all that bad off anyway". Of course all the reforms are going to look silly if you think the school system is fine.
So before I make my health post, I wanted to have a quick discussion about why pharmaceutical companies have to have higher profit to attract investment.
Posted by: Sebastian Holsclaw | March 29, 2006 at 04:03 PM
Also the comments have caused me to do some research which pretty much debunks an often repeated anti-pharma talking point. So that is great news.
Posted by: Sebastian Holsclaw | March 29, 2006 at 04:05 PM
I talked to a finance friend and he confirmed that Marketing and Administration includes: salaries and employment expenses (research scientists probably in the research area), the cost of leasing non-manufacturing facilities, legal expenses, non-research regulatory compliance (think warning labels) and general overhead expenses. As such, the fact that marketing and administrative expenses are a bit more than double research costs shouldn't be surprising.
I am very suprised that all of that falls under marketing and administration. It works different in the companies I worked for, but I was a marketeer not an accountant. But all the costs you described would still be costs mostly aimed at and spend in the US - and might be a lot less in Europe.
The non-transparancy (there is a better word, I'm sure) of the financial situation of the industry is actually making me kind of suspicious - and the fact that you than assume all sort of things to make them fit your perception is also not really convincing me.
You never reacted to the articles I linked to btw. My post of 6.45 links to an interesting article and a lot of the arguments in it were given by other knowledgeable posters (at least that is my impression) here.
Posted by: dutchmarbel | March 29, 2006 at 05:11 PM
"salaries and employment expenses (research scientists probably in the research area)"
???
Just to clarify ('cos the above is a bit ambiguous) - the bit in parentheses is saying that research scientist salaries are in the research area.
"the cost of leasing non-manufacturing facilities, legal expenses,"
Wait, though - what about the lease costs of R&D labs, support staff for R&D scientists & facilities, utilities costs for R&D facilities and costs for patenting?
Like Slarti said, there's a big chunk of overhead associated with R&D staff, and not all of it is indirect corporate-level.
Posted by: Urinated State of America | March 29, 2006 at 05:42 PM
Actually, I didn't say that.
What I said (or attempted to say) was there's a good chunk of cash that's charged to the end user for jobs like mine that's NOT associated with giving me a place to sit, a computer, and such. Basically the cost of managing the company is passed on to the customer in one form or another; I was just asking if what Sebastian was saying was (in essence) that corporate overhead (which as I've noted can be fairly substantial) is included in the SI&A category, where pharma is concerned.
Whether pharma expenses are divvied up in this way, I have no idea.
Posted by: Slartibartfast | March 29, 2006 at 06:06 PM
Dutchmarbel, your pdf still points to nowhere as far as I can tell.
RE: The Angell article. It has lots of problems. I always reccommend reading Derek Lowe on pharmaceutical companies. At least half of his blog has dealt with these issues, but he deals with Angell especially here and here.
In other areas, she makes this classic mistake which we have dispensed with above:
She misunderstands how the NIH and small biotechs work here: "Second, the pharmaceutical industry is not especially innovative. As hard as it is to believe, only a handful of truly important drugs have been brought to market in recent years, and they were mostly based on taxpayer-funded research at academic institutions, small biotechnology companies, or the National Institutes of Health (NIH)." She is completely wrong about 'brought to market' for most academic institutions and the NIH. (See Lowe's discussion). She also completely misses the boat on the small biotechnology companies. They make some initial discoveries and get bought out by the big ones. They are getting paid with the money the big ones later make IF the prospective drug makes it through Phase II and III. Those payments have to be made to the small biotechs or the staggering risk of being a small biotech wouldn't be funded.
She than makes a super common and incredibly bad argument that most drugs are 'me-toos' and that drugs with different side effects don't really count as important research. That should be self-evidently wrong to anyone who has dealt with serious side effects, but in case it isn't you can read Alex Tabarrok. (This link contains links to his two other posts on the matter, which are both very good). But the basic point is that improvements on side effects are often really good things. Furthermore if you don't have the side effects in the original you can pay less than you would have otherwise because of the competition from the me-too drug. If you don't need the new drug, you don't have to take it.
She has this odd thought:
First, the research output of American universities and the NIH are available in English all over the world. Welcome to peer review. And if the point is that the best scientists are in the US and that they can collaborate personally, one might be a little more interested in finding out why the best locus of scientists is here.
Second, for purposes of spurring more and more research, it doesn't matter if the company is owned by Europeans. What matters is the innovation is being prodded by available profit in the US.
Anyone see a strange rhetorical turn in the bolded sections? Also note she is relying on the faulty Marketing and Administrative issue again. She alludes to a serious problem, but she doesn't seriously address it in the "drugs are different" section. If the drug wasn't invented, people would still 'need' it, but they couldn't get it. They are no worse off not being able to afford it than they would be if we cripple the development system by not incentivizing it. But far worse, when the drug would come off patent it would be available. It isn't available 10 years later if it is never invented.
"Similar legislation permitted the NIH itself to enter into deals with drug companies that would directly transfer NIH discoveries to industry."
In the analogous situation with small pharmaceutical companies deals with drug companies that transfer discoveries to industry take place because getting from the 'discovery' to the drug is still hugely expensive. The small pharmaceutical company can't do it. So the make a deal to transfer discoveries to industry. The normal name for that is "selling". If Angell doesn't believe the NIH is selling for the proper price, that is a concern to take up with the NIH. But considering the time, expense and failure rate I'm not convinced. (See Derek Lowe's side bar for a large number of articles on the how the academic world and the pharmaceutical company world interact.)
There we go again. Everyone (including the NIH or the small biotech) profits! Horrifying! She somehow doesn't realize that this doesn't help her argument for gutting prices by big pharma. Who exactly will pay the small biotechs off for their "most innovative" discoveries if the drugs don't get to have high prices?
So non-profit organizations are now able to get more funds than they used to by licensing. I'm sorry is this a bad thing? And what exactly is pro-industry bias in this context?
Unmentioned is the fact that this effectively puts drug companies back into the more normal useful patent lifetime. Very few successful consumer products spend almost half their patent lifetime without ever being sold.
She repeatedly makes mistakes about the benefits of drugs with different side effects. She repeatedly makes the Marketing and Administrative mistake. She repeatedly suggests that prices for discoveries made by the super-risky small biotechs shouldn't be reflected the price of drugs made from those discoveries if they are purchased by large companies--as if premium paid vanishes in the purchase.
I don't find her rhetoric impressive.
Posted by: Sebastian Holsclaw | March 29, 2006 at 06:24 PM
Slarti, coporate overhead is typically found in the Administrative component of SEC filings.
Posted by: Sebastian Holsclaw | March 29, 2006 at 06:25 PM
That SA&I costs are between 2 and 4 times R&D costs may or may not be "surprising," but it's also nearly irrelevant as a measure. For one thing, R&D costs have gone up very manyfold, almost exponentially, over the past decade and more. SA&I costs therefore have also done likewise, and a difference of 2 in the ratio is therefore not particularly significant compared to the actual numbers.
Moreover, because R&D costs include many of the most expensive aspects of drug development, notably including CRO and CMO costs, and also the salaries of the most expensive caste of employees (MDs and scientists), the fact that SA&I includes some salaries, e.g., of administrative personnel, plus facilities and G&A, is underwhelming. (I would also note that, for many reasons, companies routinely allocate as many costs as possible to R&D rather than to G&A -- including regulatory and administrative staff, for example -- so the breakdown is not as clean as it might seem from a few lines in a 10-K.)
And in any case, these are very gross numbers, with many "moving parts," not unlike stock prices.
The fact is, advertising costs, which include but are by NO means limited to DTC ads, are very large and escalating rapidly -- as can easily be found by reading any of a number of online articles about DTC ads, for example. Those costs, along with financial earnings pressures, have a material upward effect on the price of drugs, and they could be reduced, allowing reductions in the price of drugs, without compromising R&D efforts in the slightest.
Now this is not to say that advertising costs are the main culprit, merely that they are a significant one of a handful of reasons that drug prices are unnecessarily high in the US. Others include, as noted above, mismanagement of risk and undue fealty to the blockbuster model created in part by Wall Street earnings expectations. And still another, that I haven't mentioned, is a historically rooted bloat at Big Pharma, although that too is beginning to change.
But just so the main point is not lost: pharmacos do NOT need the kinds of prices currently charged in the US to adequately fund R&D AND fairly compensate them for its relatively high risk.
Posted by: bleh | March 29, 2006 at 07:29 PM
And just because it really is almost a separate thread, it is simply not the case that a "rational" decision-maker should be indifferent between two prospects with equal expected values. A risk attitude that is different from neutrality is perfectly rational, and there are simple, well-validated models of risk-aversion that are entirely consistent with the usual (Von Neumann-Morgenstern) principles of rational choice under uncertainty. While risk-averse people may behave in a way approximating risk-neutrality in certain cases (namely, ones in which the possible outcomes are not particularly material relative to their wealth), an argument that asserts equality in value of two prospects merely because they have equal expected values would not be taken seriously by a sophisticated reader.
Posted by: bleh | March 29, 2006 at 07:48 PM
"The fact is, advertising costs, which include but are by NO means limited to DTC ads, are very large and escalating rapidly -- as can easily be found by reading any of a number of online articles about DTC ads"
The Marketing and Administrative costs don't seem to be growing faster than the business. You can see that by looking at the Pfizer numbers--between the two links you have 14 years you can look at. They seem to be growing at about the same rate as the research costs which are also growing with the business.
Posted by: Sebastian Holsclaw | March 29, 2006 at 07:50 PM
" an argument that asserts equality in value of two prospects merely because they have equal expected values would not be taken seriously by a sophisticated reader."
Well at least we agree on something.
Posted by: Sebastian Holsclaw | March 29, 2006 at 07:51 PM
The Marketing and Administrative costs don't seem to be growing faster than the business. You can see that by looking at the Pfizer numbers--between the two links you have 14 years you can look at. They seem to be growing at about the same rate as the research costs which are also growing with the business.
But as pointed out above, the RATIO is not particularly meaningful. And "growing with the business" is a very loose construction indeed, since, for example, costs have been gorwing nearly exponentially, but the number of new drugs produced has been basically flat.
The fact is, marketing expenditures have been growing exponentially, and that is unrelated (as pre-1997, and especially pre-1984, experience makes crystal clear) to the inherent risks of developing drugs. It is a business strategy, but it is only that; it is not necessary to developing drugs, and it is not a consequence of the risks of developing drugs, and therefore neither are its very material effects on drug prices.
Moreover, as also pointed out above, those numbers are very gross, and it would be methodologically suspect at best to assume that their constituents, or even the accounting practices that produced them, were substantially static across the period. Pharma isn't Hollywood (although as noted far above, there are interesting and non-obvious parallels), but there are powerful incentives for doing, erm, aggressive things with gross financial measures, and pharma is far from immune to them.
And regardless of arguments over methodology, it is a simple fact, which can easily be verified by Googling the popular business literature or trade press, that even DTC advertising alone tops $100M a year for leading brands, and that is about the same as even a moderately large phase-III trial would cost. In short, what big pharma spends on ads is of the same order as what they spend on R&D.
Given that, the proposition that pharma deserves even the earnings they enjoy today, because of the risks and costs of their R&D, simply falls apart.
But happily, as noted above, the writing has been on the wall for some time, and things are changing for the better, in many ways, both tactical and strategic. Structural change -- including the inevitable reduction in US drug prices -- is painful, but the industry is being improved by it.
Posted by: bleh | March 29, 2006 at 10:10 PM
Since the major issue here is why are prices higher here than in Europe, then how much of that is the effect of regulation put on the prices of drugs and how much is it the marketplace?
More-or-less, is there any legislation in Europe that directly or indirectly lowers the prices of the drugs?
Posted by: Frank_A | March 29, 2006 at 10:25 PM
"But as pointed out above, the RATIO is not particularly meaningful. And "growing with the business" is a very loose construction indeed, since, for example, costs have been gorwing nearly exponentially, but the number of new drugs produced has been basically flat."
It isn't a very loose construction. For the two very large companies I looked up the categories "Marketing and Administrative" and "Research and Development" have been stable as a percentage of total revenues for a very long time--including the period BEFORE the advertising changes were passed. If you want to argue that Pfizer and Merck aren't representative of big pharma companies you can do so, but you are going to need at least a good couple of counterexamples at this point.
"The fact is, marketing expenditures have been growing exponentially, and that is unrelated (as pre-1997, and especially pre-1984, experience makes crystal clear) to the inherent risks of developing drugs."
I've heard that story, and it sounded plausible. But now that we have looked, it is clear that marketing expenditures have not been growing exponentially for Merck and Pfizer unless they have been lying about it for almost 20 years.
"Moreover, as also pointed out above, those numbers are very gross, and it would be methodologically suspect at best to assume that their constituents, or even the accounting practices that produced them, were substantially static across the period. Pharma isn't Hollywood (although as noted far above, there are interesting and non-obvious parallels), but there are powerful incentives for doing, erm, aggressive things with gross financial measures, and pharma is far from immune to them."
Now you just sound like you don't want to talk about facts when they cut against your storyline. I'm amenable to factoring in other facts. I've spent quite a bit of time sifting through 10-Ks and at this point I haven't seen what you are talking about. You have access to the SEC filings too. Show me a large company where your contention is true.
"And regardless of arguments over methodology, it is a simple fact, which can easily be verified by Googling the popular business literature or trade press, that even DTC advertising alone tops $100M a year for leading brands, and that is about the same as even a moderately large phase-III trial would cost. In short, what big pharma spends on ads is of the same order as what they spend on R&D."
Now your methodology is falling apart. Surely you aren't claiming that the ratio of leading brands to drug trials is 1:1. Lets go back to the data. According to Pfizers most recent 10-K (Page 35), their research costs are $7,442 million. In order for your contention to be true (even if they were spending $100 million on all of their top-selling drugs) they would have to have 74 blockbuster drugs that they are selling at that level. By my count they have maybe 8. Those are Lipitor, Norvasc, Zoloft, Celebrex, Viagra, Zyrtec, Xalatan and perhaps Bextra. I find it very unlikely that you are correct when you say that what they spend on ads is on the order of $7 billion.
"Given that, the proposition that pharma deserves even the earnings they enjoy today, because of the risks and costs of their R&D, simply falls apart."
I can't give you 'that' becuase it actively contradicts the evidence before me. I'm open to other evidence but I think I've done my due diligence in the discussion.
Posted by: Sebastian Holsclaw | March 29, 2006 at 11:11 PM
According to Pfizers most recent 10-K (Page 35), their research costs are $7,442 million.
Right above that, it has "Selling, informational, and administrative expenses" as $16,977 million. How much of that would plausibly considered 'advertising' as much of it goes to trying to convince doctors to prescribe their products?
I'm also not convinced of selecting one firm's statements and using it as a stand in for the industry gives a true picture of the industry. The PDF given on this webpage notes that a company that doesn't appear in the top 10 (Schering-Plough) spent $136 million publically purchased advertising Claritin in 1998, which, as the page notes, was more than advertising for either the Coke brand or the Budweiser brand. It also notes that with the relaxing of restrictions on DTC (direct to consumer) advertising, DTC advertising spending went from 791 million in 1996 to an estimated 2.5 billion in 2000, which probably accounts for the 2x figure above research costs. If American consumers are paying not for research costs, but for more and more elaborate advertising campaigns (much of which is hidden because it goes to doctors who are prescribing the drugs), that would negate any idea of you paying 10 years of higher prices in exchange for getting cutting edge drugs.
Also, if I understand correctly, many R&D costs can be written off as tax losses (around 30%?), so it is not as clear cut as the numbers on ond company's balance sheet.
having said that, I certainly acknowledge that it is the case that one company cannot simply forgo the advertising as we seem to have the equivalent of MAD there. Also, with the mergers in the industry, a lot of mileage is to get a 'brand' awareness, and that's unavoidable business practice. I also don't have any particular answers for dealing with the problems that would not require some serious changes in the philosophy of health care delivery, so you may rightly feel that I am picking at your point while not offering any proposal on my own. But my feeling is, despite the fact that I can not offer a proposal, I think things are going to change for the worst if people try to cling to the status quo.
Posted by: liberal japonicus | March 30, 2006 at 12:18 AM
"It also notes that with the relaxing of restrictions on DTC (direct to consumer) advertising, DTC advertising spending went from 791 million in 1996 to an estimated 2.5 billion in 2000, which probably accounts for the 2x figure above research costs."
As I said above, it definitely does not explain the 2X figure because that ratio (and higher) existed BEFORE the rule change in 1997.
"Right above that, it has "Selling, informational, and administrative expenses" as $16,977 million. How much of that would plausibly considered 'advertising' as much of it goes to trying to convince doctors to prescribe their products?"
We've just had a discussion about the fact that administrative expenses include nearly all normal overhead expenses for a company. Even if you posit that advertising is equal to overhead, you then get advertising equal to research costs--which is far better than most companies which do little to no research at all.
I frankly don't understand the obsession with advertising costs. Almost all companies spend huge amounts on advertising costs. Coke doesn't have to advertise as much as some because it has one of the most recognizeable brands in the world and can make money off of it for decades. Pharma companies have an extremely limited window to earn lots of money on their products. It isn't shocking that they try to make lots of money during that small window. One way to do that is good marketing. Unless you want to make a general anti-advertising argument, but that is whole different world.
Posted by: Sebastian Holsclaw | March 30, 2006 at 04:47 AM
you then get advertising equal to research costs--which is far better than most companies which do little to no research at all.
Actually, you don't because 34% of research fees is tax deductible, which means that you only have about $4,600 million dollars actually going out of the companies pocket into research. At any rate, with my last paragraph, I tried to acknowledge your point about advertising, so on that note of quasi-agreement, I'll bow out of this.
Posted by: liberaljaponicus | March 30, 2006 at 06:04 AM
"Also, if I understand correctly, many R&D costs can be written off as tax losses (around 30%?), so it is not as clear cut as the numbers on ond company's balance sheet."
I'm not a tax guy, so what does that mean? If you don't make a profit you can write off losses. Are you saying that even if they make a profit they can write off research as a loss?
Posted by: Sebastian Holsclaw | March 30, 2006 at 10:26 AM
I'm not a tax guy either, but I believe that a multinational corporation like Pfizer can shift its earnings outside of the US and thereby not pay taxes on them, making it appear as if they are running at a loss. If that is the case, then the US government is returning 1/3rd of the money that is spent in research.
Googling finds this, which says
A New York Times report notes that, in their public filings, companies are often unclear about what percentage of their profits comes from domestic operations as opposed to foreign operations, and they almost never discuss profit-shifting. For example, Pfizer, the pharmaceutical giant, said in its 2003 annual report that as of the end of last year, it had not made a United States tax provision on what it called $38 billion of unremitted earnings at its international subsidiaries. It was not clear whether that money was actually earned by the international subsidiaries or by Pfizer's operations in the United States and later shifted to those subsidiaries for tax purposes, and a Pfizer spokesman declined to provide any details or comments to the Times.
This page has this
Likewise, Abbot Laboratories and Pfizer are both in the drug business, but the former paid almost 29 percent of its profits in taxes from 1996 to 1998, while the latter paid only 3.1 percent.
This Times article is interesting in that regard.
A new tax break for corporations is allowing the biggest American drug makers to return as much as $75 billion in profits from international havens to the United States while paying a fraction of the normal tax rate.
The break is part of the American Jobs Creation Act, signed into law by President Bush in October, which allows companies a one-year window to return foreign profits to the United States at a 5.25 percent tax rate, compared with the standard 35 percent rate.
This link has this
In 1998, 24 companies received tax rebates. These companies reported U.S. profits before taxes in 1998 of $12 billion, yet received tax rebates totaling $1.3 billion. Among the companies receiving tax rebates in 1998: Texaco, Chevron, CSX, PepsiCo, Pfizer, J.P. Morgan, Goodyear, Enron, General Motors, Phillips Petroleum and Northrop Grumman.
As a lawyer, this might be more interesting to you, the section on 2003, where the group 'Responsible Wealth' asked for more complete tax reporting on the part of Pfizer.
Now, this may be simply because Pfizer is one bad apple, and we can't assume that their behavior is industry wide (in fact, Merck evidently pays a much larger percentage of taxes and is often used as a contrasting example to Pfizer) so I apologize for the threadjack, but this would seem to suggest that Pfizer is not the best example to use.
Posted by: liberal japonicus | March 30, 2006 at 11:04 AM
“No. A rational investor shouldn't care which of two investments with the same expected value he picks. (Most investors do, for the psychological reasons already mentioned, but that is not the same thing.)”
This is a formally accepted model in much of economics, but I don’t actually believe it is correct.
It is not correct, and it is not a formally accepted model either. Notions of risk aversion influencing investment behavior are crucial to understanding finance.
I haven't had a chance to look up numbers, but I would like to add one thing to bleh's comments: In evaluating the riskiness of pharmaceutical companies from an investment point of view it is simply wrong to look at individual companies. You have to evaluate how a portfolio of such stocks- preferably value-weighted - would perform in the aggregate. Such an approach "diversifies away" much of the risk of the individual companies.
The portfolio will have less variability for the same level of expected returns (to the degree the risks are based on individual product success or failure). Among other things, this gets around the "top ten stocks change every year" problem. To see this, take Sebastian's option 3 and imagine that you get to play it not once but many times. Then the risk drops dramatically. The chance of getting at least 8 successes is well over 90%.
It may well be that the industry is riskier than average - I promise to look later - but the you cannot evaluate this by looking at individual companies.
Posted by: Bernard Yomtov | March 30, 2006 at 11:38 AM
"Now, this may be simply because Pfizer is one bad apple, and we can't assume that their behavior is industry wide (in fact, Merck evidently pays a much larger percentage of taxes and is often used as a contrasting example to Pfizer) so I apologize for the threadjack, but this would seem to suggest that Pfizer is not the best example to use."
It may be that Pfizer is otherwise a tax cheat in the sense that they are hiding earnings somehow (I'm certainly not able to analyze that). But the scheme you outline doesn't really speak to research costs or operating costs so I don't see how it impacts the question. In terms of ratio of certain types of costs to other types of costs (which seem to preoccupy people in this context) Pfizer is right in line with Merck. If Pfizer is engaging in shady tax practices that is a much more general corporate governance issue. It wouldn't be specific to the pharma industry at all. So when you say "but this would seem to suggest that Pfizer is not the best example to use." I don't know what you mean.
Posted by: Sebastian Holsclaw | March 30, 2006 at 11:45 AM
"The portfolio will have less variability for the same level of expected returns (to the degree the risks are based on individual product success or failure). Among other things, this gets around the "top ten stocks change every year" problem. To see this, take Sebastian's option 3 and imagine that you get to play it not once but many times. Then the risk drops dramatically. The chance of getting at least 8 successes is well over 90%."
Maybe I'm confused about what you are saying here. You can make a portfolio of any type of company. If you have a set of companies all with low risk (A) and a set all with high risk (B) you mitigate the risk of either with a portfolio approach, but so far as I can tell if all or most of the individual constituents of A are less risky than all or most of the individual constituents of B, you aren’t going to get a Portfolio A which is more risky than Portfolio B unless some very weird things are happening. You can obviously create Portfolio B that is less risky overall than individual members of A, but if you use the same system on A it is difficult to see how you are going to get Portfolio B to be less risky than Portfolio A. (It isn’t a mathematical impossibility, but on average—which is what we are talking about—it should almost never happen and it certainly won’t be predictable.)
Also you are limited by your number of options. You can't invest in a portfolio of 1,000 companies that regularly do Phase III trials because there aren't that many companies that do Phase III trials on a regular basis.
Posted by: Sebastian Holsclaw | March 30, 2006 at 11:55 AM
"Basically the cost of managing the company is passed on to the customer in one form or another; I was just asking if what Sebastian was saying was (in essence) that corporate overhead (which as I've noted can be fairly substantial) is included in the SI&A category, where pharma is concerned."
Yeah, but there's (at least) three pools of overhead for you:
1) Benefit, payroll taxes, etc. that are immediately associated with you. This is really a variable cost, and would go away if you left the company.
2) Your immediate management, support staff (e.g. admin, HR), lease costs for your facility, staff associated with your facility but not with production and/or services to the customer (e.g. safety, environmental, maintenance). These costs would go away if the line of business or function you're in went away.
3) Corporate level management.
It'd seem to me bad accounting practice to include overhead costs in 2) associated with R&D staff in the same lump as 3). If, say, Elan pharmaceuticals (for example) decided to become a generics manufacturer and close down its R&D function, then the overhead costs associated with R&D staff would go away. (Mind you Elan has used very dodgy accounting practices in the past, so who knows.)
I'd know the answer to this if a certain very large, very profitable Bay Area Biotech company, after I convinced them to write a job description around my CV, hadn't changed its mind abruptly and decided not to hire me. (Good thing too - I wouldn't really have enjoyed corporate finance).
Posted by: Urinated State of America | March 30, 2006 at 11:58 AM
I sorta knew those things, USA, but what I was unclear about was how much of them get folded into SI&A.
"Sorta", because my wife does corporate finance for a living, but we don't discuss our respective specialties in all that much detail.
Posted by: Slartibartfast | March 30, 2006 at 12:14 PM
...you aren’t going to get a Portfolio A [of low-risk equities]which is more risky than Portfolio B [of high-risk equities] unless some very weird things are happening.
The happenings need not be so very weird. If the correlation between the individual returns of a pair of stocks in Portfolio B is low, better yet if it is negative, you can create a portfolio with minimal risk. If they are all pharma companies that's unlikely, but a lot depends on the kind of risk you are worried about. For example, if company B1 is racing company B2 to get a drug on the market, holders of one stock only are at risk but holders of both are not.
However, it seems to me that this focus on the risk facing equity holders is a bit beside the point. If you are concerned about rewarding innovation the sort of risk you really need to focus on is that facing bright students who must choose between careers in research and other prospects. A system which works by granting monopolies may improve the incentive to become lawyers!
Sebastian, if you are returning to this topic I suggest you forget finance. Focus on is the rules of the game: what they are, what you want them to be, what you fear your opponents would have them be. The best thing would be to pick on the ideas of some intelligent blogger whose views bother you; maybe John Quiggan or Brad DeLong. That way you aren't open to the charge of attacking a straw man.
That's just a suggestion, meant to be helpful; if it isn't please disregard.
Posted by: Kevin Donoghue | March 30, 2006 at 12:27 PM
Kevin makes the point well. The issue is whether the risks are independent or not. If the risks of the stocks in portfolio A (low-risk stocks) are strongly correlated and those in B (high-risk stocks) are independent then it is quite likely that B could be a lower risk portfolio than A.
Nor do you need a thousand stocks. A few will do. Just to give you an idea, a diversified portfolio of fifteen or twenty stocks will give you almost as much diversification as the whole S&P500.
This is the principle behind the whole notion of diversification in general. The idea is that if you invest in only one company you are assuming too much risk in exchange for the expected return - that is, you could get the same expected return at less risk with a more diversified portfolio. When evaluating the risk of a stock it is necessary to evaluate not the individual stock but the risk it adds or subtracts from the portfolio you already own, hence Kevin's point about negative correlation.
In this light, the way to evaluate the risks of pharmaceuticals is by judging the risk they add to a fully diversified portfolio.
Posted by: Bernard Yomtov | March 30, 2006 at 12:49 PM
My understanding from talking to the finance guy I know is that 1) goes to research, 2) much of 2 goes to research with some fuzziness on the edges and 3) goes to administrative.
The administrative costs of running any large company are huge. Many companies do little to no research and still have huge administrative costs. As such, and the reason I mention it, I think it is misleading to attack a company for having "Marketing and Administrative" costs of about double the research costs. And such attacks do in fact get made regularly (see the Angell article).
Posted by: Sebastian Holsclaw | March 30, 2006 at 12:49 PM
Seb,
Let me try again. I suggested that the figure given by Pfizer does not take into account the tax rebate and that Pfizer was able to utilize tax dodges so the figures on their balance sheet may not truly reflect the cost of research versus SIA. I backed it up with links detailing problematic tax behavior on the part of Pfizer (and there seemed to be a lot more, but that doesn't necessarily mean they are correct. But I was spoiled for choice). You have used Pfizer as your representative company, and so, given the lack of transparency (Pfizer refuses to give a breakdown of SIA) when you have stories like this, I wonder exactly which budget this is coming from. This link has lots of info from what seems to be a presentation on the acquisition of Pharmacia, so I post it here for anyone who has a better insight to what the figures represent. However, it seems that the first half of 2002 has about 1/4 rather than 1/2 of what would seems to be the budgeted spending for SI&A and R&D. I am sure this is probably aggravating and seems like a threadjack, but if your example is Pfizer (and I'm guessing that it is because they have a lot of their statements on the web), I'm just saying that what they seem to be doing is not as cut and dry as you make it out to be. The point that the research staff would go under R&D, and the sales staff would be working on commission (which is what the Cost of Sales line seems to be from my reading) suggests that the SI&A incorporates the corporate level management and the help, with two large groups of people not actually in SI&A.
Again, apologies for throwing out a lot of links which you may feel are unrelated to the point you are making. It is just that in googling about Pfizer, and trying to get some insight into how SI&A is broken down, or what exactly goes under the rubric of R&D, lots of other links turn up that I think are of interest to the discussion so I include them. I was interested in this discussion when you linked last year to Derek Lowe here, yet at that time, you were arguing that millions of dollars went down the tubes because "Lots and lots of (research) is done by privately-funded itty-bitty research firms." Now, you are arguing that the arguments that Derek Lowe makes there are equally applicable to one of the largest companies in the industry. That doesn't seem to be self-evident to me. YMMV
Posted by: liberal japonicus | March 30, 2006 at 12:50 PM
I'll also jump in and agree with you that far too many people take the SIA figure as equivalent with advertising, and that is sloppy, which is something that I wouldn't have realized had you not pointed it out. I've tried to go with the documents themselves, and avoid those arguments. I didn't see any similar mistakes in the arguments about Pfizer's tax burden though. And it would seem that if it were the case, it would be easier for Pfizer to give a breakdown of SI&A expenditures, but in that as well, YMMV
Posted by: liberal japonicus | March 30, 2006 at 12:54 PM
"Nor do you need a thousand stocks. A few will do. Just to give you an idea, a diversified portfolio of fifteen or twenty stocks will give you almost as much diversification as the whole S&P500."
Sure, if you pick a careful variety. The problem is that is very difficult to do in in pharmaceutical companies.
"For example, if company B1 is racing company B2 to get a drug on the market, holders of one stock only are at risk but holders of both are not."
Right, assuming at least one of them is successful at bring it to market--not a sure thing at all.
"If the risks of the stocks in portfolio A (low-risk stocks) are strongly correlated and those in B (high-risk stocks) are independent then it is quite likely that B could be a lower risk portfolio than A."
Right, but you are assuming constituents of A which don't make a fair comparison to B. Basically you are suggesting that you use one diversification model in creating B and a totally different model for creating A. If you use the same diversification model in both cases you won't get A riskier than B.
Posted by: Sebastian Holsclaw | March 30, 2006 at 12:57 PM
"I am sure this is probably aggravating and seems like a threadjack, but if your example is Pfizer (and I'm guessing that it is because they have a lot of their statements on the web), I'm just saying that what they seem to be doing is not as cut and dry as you make it out to be. "
Not really, I just chose Merck and Pfizer because someone else mentioned them specifically. All the public US companies have their SEC filings on the web. That is why I looked at both companies before 1997 (the advertising rules change) and current. You could choose any of the other companies and get just as much information. In fact you can choose non-pharma companies as well and look at their Marketing and Administrative lines. All you need to do is know their ticker symbol and plug it into the SEC site (under the EDGAR database). The 10-Ks are all there. Finding them is no problem. Reading them is the annoyance. I did it for both Merck and Pfizer in the present and in 1997. I found remarkable stability of those costs compared to the size of the company and the size of research over a 20 year period for Pfizer (they went back 10 years from 1997) and over 12 years for Merck (they only went back two years from 1997).
I figure 4 10-Ks is enough burden for me this week. :)
If you desperately want to look at other examples I would be happy to look at the data you find. The search page on the SEC site is here. I didn't choose Pfizer because it was easy, it was just under discussion. All public US companies should have their filings in that database.
Posted by: Sebastian Holsclaw | March 30, 2006 at 01:07 PM