Adam says that he is now open to the possibility that Apple’s vast cash holdings represent a Principle-Agent problem. Indeed, my claim is not even that thorough. My point is simply that many Principles misunderstand what’s going on here.

It could be a classic Principle-Agent problem but my current thinking runs more like this:

Apple’s cash hoard represents a trust fund that keeps its designers and engineers working together and creating spillover effects between themselves. These spillovers are the source of Apple’s profits.

A traditional firm works like the following: You get together some workers, some machines and some raw materials. You use them to create a product which you sell. The workers are paid their marginal product which is the wage. The machines are paid their marginal product which is the rental rate on capital. The materials are paid their marginal product which is the price of commodities.

The organizer of all of this gets a return for taking on the risk of coordinating all of this activity and that is profit.

Now, for princple-agent reasons the organizer often buys the capital outright instead of renting it, so mixed in with the profit is a return on capital.

That organizer is the residual claimant. That means if the organization fails she loses money. If the organization succeeds she makes money. Thus we often measure the wealth of the residual claimant by whether or not the the revenue from the product is exceeding the cost of paying the factors. That is by free-cash flow.

Some folks prefer to use net income rather than free-cash-flow but I think that this is a mistaken understanding of the difference between common equity and holders of other securities, which is more of degree than kind. However, for our purposes this is neither here nor there.

That is a traditional firm.

Now, what about a Tech firm. It works more like this:

We get a gather a bunch human capital together. Each element of human capital radiates ideas. I say radiates because I think the process is fundamentally different than labor supply.

The ideas radiating from one element of human capital hit another element of human capital and increase its size. The radiation from a element of human capital is proportional to (really a concave function of) its size.

As you gather together more elements of human capital the pile grows in size and radiative force. However, because the concave down nature of the growth process several related things are true

  1. The pile can never go fully critical. It can go critical over some small range but the expansion will be self-limiting because of the concave down function.
  2. The marginal radiative return to adding new elements to the pile hits some maximum where marginal congestion outweighs marginal spillover.
  3. The marginal radiative return to each element grows the longer its in the pile.
  4. The marginal radiative return to each element is an increasing function of the size of the elements in its pile.

For simplicity we assume that the radiated ideas are captured by a black-box Intellectual Property mechanism which turns them into output at a one-to-one ratio. One idea yields one unit of output. This is done so that all of the action happens in the pile. We could have just as easily assumed that the pile itself was radiating output but I think that would have confused the matter.

Now here is our core problem. For a mature pile the sum of all the discounted marginal products of all the elements in the pile is less than the discounted total product of the pile. This is because if any element leaves the pile it will decrease the productivity of the pile but it will increase the equilibrium size of the rest of the elements within the pile partially offsetting the loss in productivity.

However, if the all the elements of the pile up and leaves, then the entire discounted total product of the pile goes with it.

My guess is that if we simulated this constraining the payouts to be discounted marginal products that the elements will keep wanting to seek different long term arrangement schemes. The evolution of the process will be completely dependent on the iteration mechanism. That is, who moves when.

One could solve this by creating a pile trust fund. This fund pays out to the elements of the pile who stay in for the entire game, or in our case until bankruptcy.

The cash pile that tech companies amass serves as this type of trust fund. It pays out to the employees who stay until bankruptcy.

If this is what’s going on – and this is just my random musings with no evidence – then there is no principle-agent problem.

What there is, however, is a misunderstanding by some folks as to what the actual residual is. The large pile of cash is not, in fact, free cash flow that can be paid out to the providers of capital. It is a trust fund that must be paid out to the employees.

Likewise much of the projected profits are really best thought of as retained wages, not actual returns on equity.