I consider myself an economist-in-training, but I would not consider myself an Austrian economist-in-training in terms of my preferred methodologies, philosophy of economics, and research goals. I tend to be sympathetic to a view akin to that of Antony Davies that Austrian economic methods such as Praxeology are useful as a guide for model selection, and I love much of the argumentation employed by Ludwig von Mises, Friedrich Hayek, Carl Menger, and Murray Rothbard in particular, though I also disagree with parts of their work. In a sense, I view the best of the Austrian tradition as a properly consistent, (if sometimes logically ‘weak’) philosophical guide and sort of “handmaiden” to the rest of economics in a similar way to how many philosophers view philosophy as the handmaiden to narrowing the consideration and organization of science in general.
However, despite the limitations I occasionally see in Austrian economics, I truly do believe that Austrian Business Cycle Theory is and will continue to be the best model for understanding macroeconomic contraction and expansion in the coming two or even three or longer years following the Covid outbreak and spread.
For those who are unaware, a simplified version of Austrian Business Cycle Theory (though the originals and variations can be complex) is a macroeconomic model that seeks to explain the business cycle (booms and busts in the economy) as a consequence of artificially low interest rates set by central banks, causing investors and consumers to borrow money more than they otherwise would compared to a market-determined interest rate and often take on more risk than the market has priced in. Because they are investing or borrowing for consumption suboptimally, the theory asserts that this causes malinvestment from the ‘easy credit’ as people engage in ventures that appear more profitable or beneficial than they actually are, creating a situation of temporary high demand and growth or a ‘boom’. This process is thought to then end when the central banks either allow interest rates to return to higher levels (often in an effort to prevent high inflation), or financial and asset markets realize the real value of their investments is less than had been consensus (as possibly in the late 2000s mortgage and financial crisis). This signals the beginning of the ‘bust’. The values of assets then fall, and recession takes place as the value of malinvestments becomes apparently less than was thought. Thus, the theory argues for a narrative in which the traditional mainstays of central banking are pro-cyclical.
This issue could potentially be compounded, as some Austrians argue occurred during the late 2000s, by other forms of ‘easy credit’ outside of low interest rates alone that lead to similar malinvestment. Some have argued this was embodied in the late 2000s by the revelation of ‘toxic assets’ in the form of collateralized debt obligations and other instruments fueled by government-exacerbated subprime lending. The situation in that case was only resolved by massive federal reserve purchase of these assets through quantitative easing. The consequences moving forward as far as the theory is concerned are context-sensitive, but likely risky and unstable or even disastrous, as could be argued to have occurred in United States housing, for example, when lending practices such as adjustable rate mortgages led to mass foreclosures. This then had the consequence of increasing housing supply, devaluing homes, and encouraging more bankruptcies and foreclosures in a vicious cycle as people begin to owe more than their homes were worth. This aforementioned real-world example isn’t specifically predicted by the theory, necessarily, but provides an example of the how the situation may compound risk of structural failure.
I know that the model is not perfect or complete (to say the least), and many will turn their noses at it for not being “mainstream economics.” I have noticed problems, but some of the biggest problems in the minds of some critics of the theory are actually very able to be addressed and refined against, I’ve found.
For example, public choice theorist (and my all-time favorite economist) Bryan Caplan has argued the following on ABCT:
“What I deny is that the artificially stimulated investments [from artificially low interest rates set by central banking] have any tendency to become malinvestments. Supposedly, since the central bank’s inflation cannot continue indefinitely, it is eventually necessary to let interest rates rise back to the natural rate, which then reveals the underlying unprofitability of the artificially stimulated investments. The objection is simple: Given that interest rates are artificially and unsustainably low, why would any businessman make his profitability calculations based on the assumption that the low interest rates will prevail indefinitely? No, what would happen is that entrepreneurs would realize that interest rates are only temporarily low, and take this into account.”
My response to Caplan though is thus:
You ask, “Given that interest rates are artificially and unsustainably low, why would any businessman make his profitability calculations based on the assumption that the low rates will prevail indefinitely?”
I would rebut by mentioning that three important possibilities have been too hastily cast aside:
1.) The notion that agents in ABCT must act as though the low rates will prevail “indefinitely” is far too strong of a claim on what refined versions of ABCT assert. Perhaps if one were arguing with Mises or Hayek in the early 20th Century when it was first developed, that would be a very unrealistic thing to have to accept with the benefit of modern economic knowledge since then. However, there are formulations and even interpretations of the original work that make a weaker claim: essentially, “that the low rates must compel behavior that leads to malinvestment.” It’s weaker logically, but still powerfully compelling as it’s more believable, and accomplishes the same economic connection to malinvestment, just by more possible mechanisms. As long as the theory makes the case that people will act and cause malinvestment, even if believing interest rates are temporarily and artificially low, Caplan’s point loses water. Caplan does give the investors credit with a “rational expectations” framework, but it’s not as incompatible there as he has written it.
2.) When the weaker claim is incorporated, more possibilities become evident. Austrian economists Anthony Carilli and Gregory Dempster argue, for example, “that a banker or firm loses market share if it does not borrow or loan at a magnitude consistent with current interest rates, regardless of whether rates are below their natural levels. Thus businesses are forced to operate as though rates were set appropriately, because the consequence of a single entity deviating would be a loss of business.” In highlighting this proposed mechanism, Carilli and and Dempster implicitly reveal a key difference between short and long-term behavior from the get-go even before the crux of their argument: that people and firms often prioritize the next couple of years over the time directly following that period due to time preference and the notion that sooner monetary benefit is ceteris paribus better than that same monetary benefit later on. That principle is part of why banks charge interest on loans to begin with and very easy to consistently model in.
Additionally and more importantly on this point: there is a great deal of uncertainty as far as timing and magnitude of future booms and busts as just a reality of being in a real economy and not a model. Even if an investor believes and totally understands ABCT and applies it to then-current low interest rates and credit, fulfilling Caplan’s rational expectations, this is not the same as knowing the future. As well, even if you believe and apply ABCT as an investor, there’s no reason to think that the contraction will necessarily be a larger hit for you individually than the boom that lasts who knows how long, due to heterogeneity. The possibility remains that taking part in the credit expansion may benefit you more than the contraction brings losses, perhaps on the net leading you a to a better outcome than if you had forgone unsustainable market activity for the cycle. This brings us to the third point:
3.) Caplan has assumed too much about the opportunity cost framework of investors. Many industries do compete for market share and stand between losing some more certain amount (from missing returns and being outcompeted) by not investing, on one hand, even if they expect an artificial credit expansion, and on the other hand, an expected generalized loss of investment value over the macroeconomy that is not only uncertain in timing or magnitude, but also takes place down the road, when returns and losses of otherwise equal monetary value matter less.
To demonstrate formally,
If we model ultimate Expected Value of X to an agent as:
1.) E[X] = P(e) * U(e)
where P(e) is the probability of an event given the choice input, and U(e) is real utility from that event given the choice input; this is very evident to see. For choosing to marginally step back from artificially low interest rate borrowing as Caplan insists would occur, (call its ultimate expected value to the agent E[S]) we can assume that the agent expects some amount of utility:
2.) V1 = – U(S1)
to be negative at some point near to the present if they exit the markets or halt investing, based on Carilli and Dempster combined with the simple fact of missing the returns from the boom, at “some” negative value. They also consider:
3.) W1 = + U(M1)
or the positive utility of the present from choosing to malinvest (M) and stay in the market in the present. Let’s suppose for simplicity that:
4.) |W1| = |V1|
And with parallel reasoning for S2 and M2 (indicating outcomes from the second phase of time for the different choices: Stepping Back (S) or Malinvesting (M), during the expected bust):
5.) V2 = + U(S2)
6.) W2 = – U(M2)
7.) Assumption for Simplicity: |W2| = |V2|
Let’s also suppose, per criticism from Carilli and Dempster, we have no reason to necessarily assume:
(V1 + V2) > (W1 + W2)
and indeed, the assumption of money time preference and the proposed additional consideration of losing market share with S, ceteris paribus, would both imply that:
8.) |V1| > |V2|
9.) |W1| > |W2|
Additionally, it is almost certainly harder to assign a higher probability to an event in any given time (2), at least, from the perspective of the agent, due to the greater uncertainty about when exactly a contraction may take place in the future, compared to the presently more evident loss of market share from preemptively rejecting credit and booming returns from easy credit. With there being a greater range of future possibilities due to uncertainty, the sample space of possibly significant time periods and points of possible probability value for the future increases. This contrasts with the very intutitively more evident task of assigning probabilities to more obvious and evident present or near-present outcomes, implying a relatively smaller sample space, and, thus, relatively higher probability assignments in time (1), ceteris paribus:
From this, in a given 2-identical-length points or periods from a sample space, one from present or near-future (1), and one from only roughly predictable farther future (2),
10.) P(W1 or S1) > P(W2 or S2)
Finally, bringing together steps 8.), 9.), and 10.) with the relative signs and relative magnitudes for each time and event, and again, assuming for simplicity that the absolute values of alternative loss or gain from each period are equal, we may then conclude:
11.) E[M] = (^P(M1) * +U(M1) + P(M2) * -U(M2)) > E[S] = (^P(S1) * – U(S1) + P(S2) * +U(S2))
With the signs and W1 being a larger relative positive, and V1 being a larger relative negative, along with probability favoring the present for accuracy, we may then conclude:
12.) E[M] > E[S]
Under this framework, thus, is it quite clear enough to admit a strong scenario when minor assumptions to the model imply the decisive conclusion that malinvestment is rational, and at least can be preferable to Caplan’s demanded alternative of preemptively exiting if you detect credit expansion based on rational expectations.
(Note: This is a pretty important moment, as I’ve only initially disagreed with maybe five things total that Caplan has written or said in economics and still seems to agree with and stand by today).