A few days ago Alexander W. Salter wrote an interesting post on the problem of optimal resource allocation during an Austrian business cycle. His argument is that the Austrian business cycle theory (ABCT) can be understood within "a framework of rational expectations and ‘equilibrium always’ modeling conventions." The argument against the ABCT based on rational expectations is that if we assume a representative agent, then it is not plausible to assume that he persistently errs on investing in projects that might be too long or roundabout. If the agent truly is rational, then he should be able to sort out the “model” and correct his mistakes rather than always erring on the “too long or "roundabout” side. I want to add another angle (see here for a more complete analysis) to what was mentioned by Salter. The use of a representative economic agent constrains the analysis by doing away with heterogeneous individuals interacting with each other. The ABCT is not driven by a representative entrepreneur, but by a marginal entrepreneur. It should be noted that by marginal entrepreneur I mean those entrepreneurs that consider an interest rate that is below its natural level to be correct. Because the equilibrium conditions are unknown and have to be estimated by the economic agents, some of them will erroneously think that the equilibrium value is below the correct equilibrium value while others will think that the equilibrium value is above the correct value. On average, we might say that all economic agents are correct. The point, however, is that these mistakes do not cancel out. This is why it is important to move away from a representative agent and instead towards heterogeneous agents. Movements of interest rates that are used to discount investment projects affect the present value of a project’s cash-flow differently. Investment projects attached to a cash-flow with a higher duration see their present value increase more than investment projects attached to a cash-flow with a lower duration. There is a rise in the relative price of projects with more duration, which is the financial representation of the average period of production. In this scenario, then, those entrepreneurs that mistake a low interest rate for the correct one are willing and able to out-bid the entrepreneurs that err on the other side of the interest rate equilibrium from the market of factors of production and intermediate goods. These marginal entrepreneurs, not representative ones, are the ones driving the ABCT scenario. As time goes by, irreversible investment on projects with too much duration start to accumulate. When the interest rate rises towards its natural level then the present value of these projects fall or even become negative. The rational expectation criticism of the ABCT rests on both (1) a representative economic agent and (2) rational expectations. By so doing, the subtleties of the ABCT are ignored and the theory is criticized in the wrong terms. In our present monetary institutions, changes in money supply are usually channeled through the credit markets. This changes the relative price of time (interest rates) and produces a particular Cantillon effect among the present value of projects with different durations. The ABCT, in contrast to other business cycle theories, has been built to deal with this particular effect.