I've just read an interesting paper by Philip Pilkington (then of Kingston University, now of GMO), called "Endogenous Money and the Natural Rate of Interest: The Reemergence of Liquidity Preference and Animal Spirits in the Post-Keynesian Theory of Capital Markets". In it, he argues that the concept of "a" natural rate of interest implicitly assumes that the EMH holds.
This is a post Keynesian perspective, resting on Keynes' view that there was no single rate of interest that would bring the loanable funds market to equilbrium. Rather, there's a multitude of interest rates that exist throughout the market. One might think that the risk-free rate serves as a benchmark upon which other rates relate, however these rates require a risk adjustment. Pilkington's point is that in order for savings and investment to be equal, "every lender is pricing in the risk of the borrower correctly - i.e. they are lending to the borrower at the "correct" or "natural" rate of interest given this specific borrower's risk". He goes on to argue that this requires an assumption that lenders are perfectly rational and have perfect information. If this isn't the case, he says, there's no reason to expect the natural rate to "channel investment in a manner that ensures a stable equilibrium growth path".
What caught my eye was a footnote where Pilkington points out a perceived contradiction amongst Austrians:
One might note the superficial similarities between what we have just described and the boom-bust cycle of the Austrian Business Cycle Theory (ABCT). The key difference, however, is that the ABCT assumes that only central bank action can affect the money rate on interest. As we have seen, however, unless we assume perfect foresight on the part of savers/investors there is no logical reason to assume that they will set the money rate of interest in line with the natural rate. It would be interesting to consider how Austrian theorists, who generally recognize Knightian uncertainty as being operative in capital markets, would respond on this point. The only viable response to this so far as we can see is to advocate some form of the EMH and rational agents, but if Austrians were to do so it would no longer be clear what would distinguish them from, for example, New Classicals.
The Austrian point is that expectations don't need to be rational (in a RatEx sense) for there to be a tendancy towards equilibrium. The Austrian point is that the "saving" in the loanable funds market is - to paraphrase Roger Garrison's terminology - "saving for something". It thus bridges short run and long run models.
Keynesians emphasise the capacity for saving to not find its way into investment (i.e. the paradox of thrift can occur), whereas classical growth theorists argue that all unconsumed resources are necessarily channelled into investment uses. However the critical difference between these views is simply the time scale: in the short run Keynesians are right, in the long run the classicals are. But the Austrian approach finds a convenient middle ground. Higher saving (because it's driven by a purposeful reason) means greater future consumption and therefore greater profits for entrepreneurs that ready those resources. Rather than implicitly rely on an assumption of Rational Expectations, this fully captures the radical uncertainty that characterises entrepreneurial decision making. Indeed this is precisely why disruptions to the natural rate (i.e. signal extraction problems) matter. I think the Austrian position is robust on this. The middle ground is solid.
For more on my take on the differences (but also similarities) between Austrians and the New Classicals, see here.