E-piphany

A potentially important development in the market for pharmaceuticals – and in the pricing of the 1.6% of the Consumer Price Index that Medicinal Drugs represents – is the President’s move towards a “most-favored-nation” clause in the pricing of pharmaceuticals. The concept of a favored-nations clause is not new, although this is the first time it has been applied broadly to the pharmaceutical industry. In the investment management industry, it is not uncommon for very large investors (state pension funds, for example) to demand such a clause in their investment management agreements. Essentially, what such a clause does is guarantee to the customer that no other customer will get better pricing.[1] In the context of pharmaceuticals, the “problem” that the President is addressing is the fact that Americans buying a drug will often pay many times what a customer in another country will pay the pharmaceutical company for that same drug.

The optics are terrific for the President, but the economics not as much so. The argument is that demanding such a clause will force pharmaceutical companies to lower prices for American consumers drastically, to something approximating the price of those same products purchased abroad. The reality, though, is not so clear.

This is a story about price elasticity of demand. As I do often, I pause here and give thanks that I studied economics at a university that had a fantastic econ faculty. Economics is a great field of study, because done right it teaches a person to ask the right questions rather than jumping to what seems to be the apparent answer. (Incidentally, I feel the same way about Street research: done right, the value of that research is in guiding the questions, rather than handing us the answers.)

So let’s start at the ‘free market’ version of the pharmaceutical company’s profit-maximization problem. Let’s start by assuming that the marginal cost of production of a little pill is close to zero, or at least that it’s no different for the pill sold in one country versus the pill sold in another country. Then, the firm’s profit-maximizing linear programming problem is to maximize, independently for each country, the price where the marginal revenue is essentially zero – where in order to sell additional units, the price must be lowered enough that selling those additional units costs more in lost profit on the other units than it does on the incremental units. (If I sell 10 units at $10, and in order to sell the 11^th unit I have to lower the price to $9, then I go from $100 in revenue to $99 in revenue and so if I am a profit-maximizer I won’t do this).

This point will be different in each country, and depends on the demand elasticity for that drug in that country. If the demand for a drug is very elastic, then that market will tend to clear at a lower price since each incremental decline in price will produce a relatively large increase in incremental quantity demanded. On the other hand, if the demand for the drug is very inelastic, then that market will tend to clear at a higher price since each additional increase in price will result in the loss of relatively few units of quantity sold. Now, every country and every drug will have different price elasticities. A lifestyle drug like the little blue pill will face fairly elastic demand in a Third World country, while a malarial drug probably does not.[2]

As an aside, one of the things which creates a more-elastic demand curve is the availability of substitutes. So, if the FDA makes it more difficult for a new statin drug to be approved than does the equivalent agency in Italy, then demand for a particular statin drug (all else equal) will be more elastic in Italy, where it faces more competition, than in the US. If you want lower prices, promote competition. But back to our story:

Now the Trump Administration adds a constraint to the drug company’s linear programming problem, such that the maximization is now joint; the problems are no longer independent maximization problems but the company must find the price that maximizes revenue across all markets collectively. If the free market has found a perfect and efficient equilibrium, then any such constraint must lower the value of the revenue stream to the drug company because if it did not, then it implies the company would already have be operating at that single-price solution. Constrained solutions can never be more valuable than unconstrained solutions, if both are in equilibrium.

What the drug company most assuredly will not do, though, is immediately lower the price to the American consumer to the lowest price charged to any other country. What it will do instead is take the highest price, and then add the incremental market that has the most inelastic demand, and see how much total revenue will increase if they have to lower the universal price to induce demand in that market. Note that this outcome may lower the price in the high-priced country, but it will also raise the price in the low-priced country. Since the lower-priced countries probably have more-elastic demand than the high-priced country…which is suggested by the fact that they had lower prices when they were being separately optimized…it is easy to imagine a scenario where the drug company ends up only supplying the high-priced country because the large increases in price for other countries essentially eliminate that demand. And that outcome, or indeed as I said any constrained outcome, is likely to be bad for the drug company. But what it will almost certainly not do is cause drug prices in the USA to drop 70%, or a massive decline in the Medicinal Drugs portion of CPI.

It may cause a decline in US drug prices, but that is not as certain as it appears. If the optimal strategy is to supply the drug only in the United States, then prices need not change at all (the US would then be the Most Favored Nation because it’s the only customer). In fact, the drug company might need to increase prices in the US. That happens because when you allow price discrimination, any customer who pays more than the variable cost of the product (which we assume here is close to zero) contributes something to the fixed overhead of the company;[3] therefore, a company that understands cost accounting will sometimes sell a product below the total cost per unit as long as it is above the variable cost per unit. When a US company, then, sells a pill to Norway at a really low price but above the cost of production, it defrays some of its overhead. If a most-favored-nation clause prevents a company from doing this, it will need to raise the price of the product in its remaining markets in order to cover the overhead that is not being covered any longer by those customers.

OK, so that’s just one iteration. I suspect that most pharmaceutical companies will end up lowering prices a little bit in the US and in other countries where prices are similar, and only selling them in countries that now pay a very low price to the extent that those foreign countries and/or international charities subsidize those purchases. But then we get into the financial and legal engineering part of this: what happens if Pfizer now licenses the formula for a particular drug to an Indian company that is legally distinct and doesn’t sell to the United States? Does the licensing agreement also fall under the MFN clause? What if Pfizer spins off its South American operations, sharing the intellectual property with its spinoff? For that matter, it might be the case that for some drugs, it is optimal to sell it everywhere in the world except the US, because the value of the unconstrained-non-US portion of the business is greater than the constrained-US portion of the business.

Now wouldn’t that be a kick in the head, to see pharmaceutical companies leave the US and refuse to sell to the US consumer because it makes them subject to the MFN clause? In the end, it seems to me that this is a great political gesture but it will be very difficult to get the results the President and his team wants.

[1] As an aside, in investment management this has caused the universe of strategies available to institutions demanding this clause to be reduced, hurting their investors. There are many circumstances in which an investment manager will offer outstanding, and sometimes outlandish, terms to investors who are the first in a new strategy, or who are low-touch easy/sophisticated customers, etc; a later entry by a large, high-maintenance customer may not be economic under the same terms.

[2] I am not at all an expert on how drug price elasticity behaves in this riot of market/product combinations, so readers who are should give me a break! I’m just illustrating a point.

[3] Cleverly called “variable contribution.”