I've already written a few time earlier (1, 2 and 3) of Dow Chemical's battle with activist investor Daniel Loeb. The last time I did so, I predicted that since Dow's plastics unit seems to be performing better (which was the main criticism that Loeb had against Dow), Loeb would go quietly into the night.
I was wrong.
While not the "open letter" that started this whole shoving match, Loeb has written extensively in his May 1st investor letter and is still on the attack. (Hat tip to The Chemical Notebook blog for pointing out this letter.) His new battle is about transparency in transfer costs, a term that you probably are not familiar with unless you've worked in a large corporation that is vertically integrated to some extent.
Dow is such a corporation. They are basic in ethylene and other feedstock monomers that they can sell to outside entities or keep and use themselves. If they use it themselves, the company establishes a transfer cost, which is the price that the downstream unit pays the upstream unit for the raw material. This cost may be below what the outside market pays or it may not. As you can imagine, transfer costs are considered trade secrets. If an external customer knew that the Dow product they were buying had a higher margin than normal because of a low transfer cost, they could play a harder negotiation on price knowing that Dow could afford the price cuts.
Loeb is suggesting that Dow has too low of a transfer price and that ethylene production should be making more money, but without knowing the transfer price, he can't be certain.
While it is clear that Loeb will never find out the transfer costs for Dow materials, I also think he is fundamentally barking up the wrong tree. According to Wikipedia, Loeb has a degree in economics so he should have taken one or more classes in multivariable calculus. Unfortunately, he has forgotten some of the important lessons of it. Specifically, for a multiple variable system, the global optimum is seldom found by optimizing each variable independently. If you have four variables w, x,y, and z, you don't hold x, y and z constant and maximize w and then repeat for the other 3 variables. You have to look across entire ranges of the variables at the same time. And even once you do find an optimum, you have no idea if it is just local or it is the global optimum.
But taking the former path is exactly what Loeb is seeking to do. He wants to optimize performance of each business unit separately rather than looking for a global optimum. While the VP's of each business unit do in fact seek to optimize their local performance, the job of the CEO is to find the global optimum. Whether or not the CEO is doing a good job of that is (always) open to debate, but pretending that a global optimum can be found by optimizing each individual business unit is amateur mathematics.
P.S., Mr. Loeb, Please keep up the attacks on Dow Chemical. It makes it easy for me to find something to write about.