In a recent article in the QJAE (

pdf), Edward W. Fuller explains the differences between the Austrian "Net Present Value" (NPV) approach to ranking investment projects and the Keynesian "Marginal Efficiency of Capital" approach. The difference basically goes like this: The NPV criterion suggests that we should calculate the "present value" of a particular investment by discounting its future cash flows using the going interest rate. The "net present value" takes the present value and subtracts off the current costs of starting the project. Investments are ranked according to their NPV. The MEC criterion suggests that you should use the same information to calculate, effectively, a "rate of return" on the investment, and use that to rank investment projects.

While I found the comparison interesting, I think the article ran into a number of somewhat troubling inconsistencies. Most importantly: the article was inconsistent regarding the Austrian view of imputation.

Here's the Austrian story as I understand it: The present value tells you what the resources needed to start a project are "worth" - so, it's what entrepreneurs should be willing to pay to start the project. As long as the market is at all competitive, the value of the resources will rise to the present value. What does that mean? The NPV of any project can be, at most, zero. If we take PV - cost of starting the project, we'll get exactly zero, since the prices of the resources will be bid up to the PV. So, when we rank investment projects, we're really just comparing zero NPV projects with projects that would use the same resources but have a sub-zero NPV. Since the resource prices are bid up to reflect the most valuable project, this all follows. So, what matters in the Austrian view isn't the NPV - it's the PV itself. It's the PV that determines which project "wins" the resources needed to complete it. Is this a big difference? Maybe not - but it is more than just semantics.

What makes this particularly disturbing is that the article itself criticizes Keynes (I'd say rightly) for acting like the cost of starting the project is fixed, rather than being the result of a competitive bidding process. Then, the article goes on to adopt the same assumption under the "Austrian NPV" view.

One point that the article makes is that the MEC and NPV criteria can lead to different rankings of investment projects. This is true, IF the startup costs are fixed. But, they aren't. Startup costs are the result of a competitive market bidding process. If we suppose that the startup costs rise to reflect the full present value of the most valuable project, then the MEC criterion would give the same result as the NPV criterion, I suspect. Here's my proof: for the most valuable project, the going interest rate will be the MEC (since the interest rate is used to discount the future payoffs, and the NPV is zero. Technically, the MEC is calculated as the interest rate that results in a zero NPV.). For any less valuable project, the NPV is negative. To increase the NPV, the interest rate must be lower. So, the MEC for any unpursued project will be less than for the pursued projects.

This technical problem hints at a larger conceptual problem with the Keynesian system. The Austrian system is built on an economic foundation of scarcity - that's precisely why prices adjust to reflect the present value of a project. The Keynesian system is built on an economic foundation of abundance (or "idle resources"), which is why prices don't adjust - there are just so many of them that there is no "competitive bidding" that goes on. This leaves unanswered the question of where prices come from (Keynes suggests that there's a lot of historical dependence on that account). So, the real, fundamental problem with Keynes remains: his price theory is largely absent. He basically just assumes that prices don't change, gives a few possibilities for why that might be, and then moves on. While it may be true that prices don't change to reflect every little wobble in supply and demand, that doesn't mean we should just assume that they don't move.

Yet, that assumption is in Keynes, and disturbingly, ported into the Austrian NPV criterion as used by Fuller.