Friday, 1 January 2010

An Austrian review

I’ve spent some time over Christmas reading a comprehensive review by Mario Rizzo of recent work in Austrian economic theory. Regular readers might have noticed that, though this is a school I broadly disagree with, I have given it a lot of attention. It does seem to hold a fascination for me.

I think the reason for this is that the founding thinkers of the Austrian school picked some really important issues to think about, and I instinctively agree with their choice of territory. Here is the list of topics Rizzo identifies:
“(1) the subjective, yet socially embedded, quality of human decision making; (2) the individual’s perception of the passage of time (‘real time’); (3) the radical uncertainty of expectations; (4) the decentralization of explicit and tacit knowledge in society; (5) the dynamic market processes generated by individual action, especially entrepreneurship; (6) the function of the price system in transmitting knowledge; (7) the supplementary role of cultural norms and other cultural products (‘institutions’) in conveying knowledge; and (8) the spontaneous - that is, not centrally directed - evolution of social institutions”
So the issues discussed are the right ones – but are the conclusions correct? My suspicion remains that Austrian economists start asking the right questions but often bring in some unacknowledged assumptions in deriving their results. I suspect that the original Austrians – Hayek and Schumpeter at least, and probably von Mises, were more explicit about stating their ideological assumptions than some more recent writers. And this criticism is certainly true of many other economic schools too. But either way, I remain unconvinced that all the typical conclusions flow logically from the premises.

[Since I wrote the above Rizzo has posted another item which says something similar - also worth reading]

But that doesn’t detract from the very interesting survey Rizzo has carried out. Here are some individual questions arising from my reading of that document. I hope these will be taken in the spirit of an interested but critical friend.
“...inappropriateness of the capital structure (malinvestment) generated by artificially low real interest rates (that is, interest rates that are lower than the real supply of savings would allow)”
The question is: does lending really require savings? Lending is about somebody spending their time now, in return for payment later (usually intermediated by a highly creditworthy institution such as a bank). There is no strict requirement for an existing stock of savings to enable this. In a simple example, three people could wash up on a desert island with no stock of wealth at all, and as long as two of them can find enough coconuts to support the third, she can get busy building a house for them to live in. This is lending with no associated savings. Rizzo mentions papers on 100% reserve banking, which implicitly acknowledges the possibility of zero reserve approach that I have described here.

Why then are savings typically associated with lending? It could be to provide a discipline on lenders/borrowers/investors who have no obvious incentive to limit the amount of credit they extend (and thus risk destroying the credibility of the instruments they issue). There might be other ways to offer that discipline.
To take a different approach, investment can be made out of the income from previously invested assets; but there’s no particular reason it has to be. There might be some moral symmetry here, that those who have foregone consumption in the past are more deserving of the benefits from investment than those who will forego it in the future; but a Coasian argument shows that this need not hold; the returns to investment can be seen simply as a property right – which could be assigned to anyone and can still reach an efficient equilibrium.
“None of this implies that Hayek, Garrison or Horwitz are insensitive to the problems that would be induced by an aggregate increase in the demand to hold money (a fall in income velocity), which can accompany recessions. This ‘secondary deflation’ should be avoided by a concomitant increase in the supply of money by the relevant monetary institutions. Horwitz (2000) is the first to integrate Austrian macroeconomics with monetary disequilibrium theory to analyse deflationary processes. Nevertheless, recessions are not primarily deflationary phenomena (or at least need not be), but occasions for correction of the misdirection of resources. Some Austrians, however, argue that increases in the demand for money have significant negative consequences only in the presence of legal restraints on price flexibility (Salerno, 2003).”
Good to know that the Austrians can recognise insights from other schools. However, in their view does money creation in any way impede the process of recalculation? The legal restraints argument seems unrealistic – in practice there is lots of inertia in prices and wages regardless of the law. If unemployment increases the speed of recalculation is it worth it?
“One of the most important possible obstacles to recovery from recessions may be in the behaviour of ‘big players’. These are agents whose discretionary behaviour, insulated from the normal discipline of profit and loss, can significantly affect the course of economic effects (Koppl and Mramor, 2003; Koppl, 2002; Koppl and Yeager, 1996). Thus, discretionary behaviour on the part of monetary authorities (in the United States, the Fed), fiscal policy makers (Congress or the Executive), or even in some cases private monopolists, can increase uncertainty faced by most economic agents (‘small players’).”
This is a useful insight. However does it really just apply to ‘big players’ or is there something broader? For example if small private players have an inertia within a certain market or group of customers (due to brand loyalty, personal connections or lack of information about alternatives) do they become effectively a big player? Or if an unprofitable shop continues to play out their business model due to sunk costs (thus gaining an effective capital subsidy over new players), either because of a cognitive/emotional attachment to their investment or the transaction costs of reallocating resources, is this any different from a government thus acting? Is it an ideological position which leads Austrians to be more concerned about government action than the economically inefficient actions of private individuals and firms? This may be the key distinction between the Austrian and Keynesian views.
"...changes in the riskiness of investment decisions are linked to the ‘old Austrian’ concern with the degree of futurity or roundaboutness in investments. For example, in Cowen’s analysis, an increase in the acceptable level of risk will encourage undertaking more longer-term investments (as well as, of course, investments of any given length with more uncertain yields). These can be both investments in durable capital goods (that is, investments with a continuous flow of payoffs over a long period of time) and investments with a long period of gestation before the ultimate output is produced. Cowen associates less risky (‘safe’) investments with consumption and shorter-term investments.
Cowen’s analysis is more general than the traditional ABCT because it allows many factors besides a fall in real interest rates to generate a lengthening of the capital structure. These include exogenous risk-preference shifts, increases in savings, easing financial constraints, and reductions in uncertainty (so as to reduce ‘waiting’ for acceptable investment opportunities). Any of these changes can generate an increase in the riskiness of investment. None of these changes must necessarily cause a cyclical boom and bust, but they might do so. "
This seems a valid point. In a world where people criticise insufficient investment and excessive risk aversion, it highlights some of the deeper factors which might explain or contradict these ideas. It also hints at the asymmetric nature of risks – there are downside risks with one probability distribution, and upside risks with another. The way to treat each of these may be quite different; the patterns of each call for different investment and lending patterns, different relationships to past and future assets perhaps.
“Free-banking advocates argue that bank profit maximization, under sound institutional constraints, will lead banks to expand or contract deposits or currency pari passu with changes in the demand for money. Banks will receive signals about the demand for (their) money as their reserves expand or contract.”
Can it be demonstrated that, under market conditions, this process generates a stable equilibrium? Or is an increase in demand for money amplified by changes in reserves, the valuation of bank money and deposits (and the bankruptcy of banks?) The answer is not obvious. If it is stable, and if reserves and pricing can fluidly adjust to correct such problems, does this lead to either:

  1. Moral hazard/commons problems, where individual banks have an incentive to overlend on the basis that other banks share the risk and the total economy-wide reserve pool? OR...
  2. information problems and transaction costs, where the cost for individuals of understanding and distinguishing between the money of different banks acts as a brake on economic activity? This might be mitigated by technology today, though it must have been a real difficulty in previous free banking systems.

Moving onto a chapter about entrepreneurship:
“Even better, eschewing the excessively constraining categories of necessity and sufficiency, we might say that the serendipity of discovery favours the searching mind.”
This is the kind of rhetoric that makes me worry about the Austrian mindset. Most people see necessity and sufficiency as simple logical constructs, not “excessively constraining categories”. It is indeed clear that there can be statistical correlations without strict necessity or sufficiency (one could count this in the domain of fuzzy logic or else provide a more detailed description of the causal chain). In the case being discussed, it is surely true that searching increases the probability of discovery without guaranteeing it. This can be stated in a less suspect way without describing logic as “constraining”.
“In most Austrian treatments entrepreneurial discovery is important because it drives the market process.”
This also in the entrepreneurship chapter: which is an interesting one, but it’s not clear where the concept fits into the broader theory. At any rate it does not feel like a distinction between Austrian, Keynesian or neoclassical theory, all of which are impacted broadly in the same way by entrepreneurship. Perhaps the Austrian model recognises an explicit externality cost of government action or recession-fighting policies in their inhibiting effect on enterprise; if so, it seems to offer no way to quantify it and thus determine the welfare-maximising amount of policy.
“...a theory of entrepreneurial judgment. This theory makes entrepreneurship inseparable from asset ownership. The entrepreneur’s judgement is about the control of heterogeneous capital assets under conditions of radical uncertainty”
This seems rather old-fashioned. Entrepreneurship surely is not primarily about control of assets; unless one considers the human capital of the entrepreneur and her employees as assets, but the formulation above specifically mentions capital assets. As before, this doesn’t feel like an important criticism as the subject seems tangential.
"The Austrian approach to market processes is distinctive in a number of respects...First, markets are in process and not continually in equilibrium...Second, market processes are not instantaneous but take time. In the passage of time (‘real time’), knowledge changes...Third, market processes take place in the context of radical uncertainty. This is to be distinguished from risk, in which all of the possibilities are known with objective probabilities...the fourth feature of market processes: they are relatively indeterminate...The fifth, and final, feature of market processes is the communication of decentralized or scattered knowledge. Markets enable individuals to act on more knowledge than they can ever hope to possess explicitly. They can do this through entrepreneurially produced market prices and through nonprice manifestations of market behaviour."
These all seem like useful insights but are hardly unique to Austrian theory. The recognition of non-equilibrium dynamics is very important but its mere acknowledgement provides only a limitation on standard theory, and no explanatory or predictive power.
“spontaneous order...‘the results of human action but not of human design’... is an organic or emergent form of coordination that manifests itself in social institutions, some organizations and clusters of individual plans”
Once again, a useful insight but is there any real analysis behind it? Ironically, this area has much in common with the work of recent Nobel-winner Elinor Ostrom, perhaps ideologically as far from Austrian economics as any mainstream economist.
"[in the area of] law and economics...The field’s uniquely Austrian features consist of attention to (1) the process of law and state intervention in markets; (2) the need for relatively stable law in a world of external change; (3) the influence of decentralized knowledge on the character and limits of law; and (4) the privatization of some of the basic functions of the state."
The discussion of (1) focuses on unintended consequences of regulation and wealth redistribution. Once again, good topics but the analysis may be incomplete.

Two arguments for redistribution from an economic efficiency standard are: that diminishing returns to wealth argue for money to be partially redistributed to those who lack it; and that a lack of redistribution may lead to negative social externalities such as a breakdown of public order or public health. I imagine that the theoretical apparatus of Austrian economics can take account of these factors, but I wonder whether the typical analysis carried out by the typical Austrian economist actually does.

On regulation, the author cites a couple of his interesting recent papers regarding behavioural economics and its interaction with regulation, and the impact in turn on the emergence of a “regulatory culture”. This is an argument against behavioural economics that I’ve heard before, and it deserves to be answered.
First, with an emphasis that good behavioural economics is inspired by many of the same motives – understanding human cognition and its economic impact – as von Mises’ Human Action and much of the Austrian work that followed it. This – which I prefer to call cognitive economics – is a positive, not a normative, discipline.

Second, an acknowledgment that many behavioural economists – like many Austrian economists – approach their subject with a common ideological framework and this does influence the type of results that are discovered, either through subjective interpretation of data or choice of research topics. I was interested recently to hear a conversation between Dan Klein and Russ Roberts in which Klein both highlighted the dangers of approaching the discipline from a subjective viewpoint, and admitted and demonstrated his own subjectivity. The more common view is that a scientist (or social scientist) should attempt to recognise and eliminate their subjectivity when approaching data; but it’s also worth remembering the distinction between lawyer and judge. Perhaps researchers are the lawyers of the scientific method, presenting their view and making the best case they can, while editors or even readers are expected to be judge or jury and decide between competing viewpoints.
(2) discusses " the legal framework and its adaptability...the level of abstraction of the relevant rules...Whitman shows that an intermediate level of abstraction is optimal from the perspective of generating rules with predictable consequences...Rizzo (1980a, 1980b, 1985) and Roy Cordato (2007) both criticize the cost-benefit framework in many conceptions of negligence law because it produces legal decisions that lack predictability to those for whom the particular law is relevant... The economic data upon which efficient legal decisions are to be made are often unavailable, complex or transient."
A valuable insight and (though I haven’t read the cited papers) potentially a powerful philosophical building block. Again there’s a way to represent this potential economic inefficiency as an externality...the cost imposed on future economic agents by additional uncertainty in the legal outcomes of their decisions. But this is an aside. Is there a way to price this externality or uncertainty? Rizzo argues for a “set of rules” rather than a standard, but in principle one of the rules could be “carry out a cost-benefit analysis based on the best available estimates of costs and benefits”. Is there a way to determine the concreteness or abstractness of a given rule?

I was quite surprised by this:
“Block argues that the Coasian cost-benefit approach effectively abolishes property rights”
Again without having read the paper, I’m surprised because I think of Coase as the ultimate defender of property rights. Coase doesn’t argue that legal structures should be chosen according to costs and benefits, but that whichever definition of property rights is chosen, private individuals will trade according to their own personal cost and benefits and an efficient outcome will occur.

No doubt Walter Block is playing against this conventional assumption in making the argument he makes, but still, I wonder if this is contrarian thinking for its own sake.

The exposition of (3) seems broadly similar to (1):
“The decentralization of factual knowledge is a critically important factor limiting the feasibility of many forms of intervention... Using the internal standards of three major ethical approaches - utilitarianism, natural law and Kantianism - Rizzo argues that the factual knowledge needed to determine just what the moral course of action is in concrete cases is not available to the paternalist.”
Of course that must be true in many cases...but is it always? At the extreme, there are cases where an individual clearly does not have the capacity to choose in their own interest: a 5-year-old child or a person in a coma. In other cases, an individual might explicitly opt into letting someone else make a decision on their behalf: a defendant in a criminal trial may ask their attorney to choose the right argument for them, or a buyer of a pension plan may ask their adviser to choose on their behalf. And somewhere in between, there are open questions such as: should the default for organ donation be opt-in or opt-out; should a defendant, whose mental state cannot be evaluated by the police officers treating them, be opted into receiving legal representation if they refuse to answer police questions about it? The pragmatic question must be about determining where the boundaries lie between these cases; it seems unhelpful to deny that there is a valid question to ask.

And finally, item (4):
“Most economic analysis proceeds on the assumption that the state exercises at least its minimum functions: that is, provision of protection, enforcement of property rights and contracts, and the adjudication of disputes. Nevertheless some economists...have argued that privatization of at least some of these functions is feasible and desirable... Powell and Stringham (forthcoming) survey a surprisingly large extant literature on the economics of a stateless society.”
Certainly an interesting area, and I’d be interested to read that Powell and Stringham paper. But is this an integrated and necessary consequence of basic Austrian principles, or is it another ideological add-on? To be fair, this point can certainly be seen as a corollary of the “spontaneous order” principle described earlier. I’d be fascinated to see a good marginal analysis – drawing on Ostrom, Caplan and Benson, say – of the right place to draw lines between state and non-state power.

Ultimately this might be, for me, the chief criticism of Austrian economics. Instead of using the key analytical strength of the economic method – marginal analysis and trade-offs – the field tends to end up at firm philosophical positions based on principle rather than maximum efficiency. Some of course would see this as a strength not a weakness, and perhaps this will be the toughest thing for me to reconcile between my philosophy and the Austrian approach.

Update: I note that Pete Boettke has renamed the "Austrian Economists" blog to "Coordination Problem". Some of his reasoning is similar to my arguments above, and I think it's a good move. I have added Coordination Problem to the blogroll on the right.


Tom Hickey said...

"I suspect that the original Austrians – Hayek and Schumpeter at least, and probably von Mises, were more explicit about stating their ideological assumptions than some more recent writers. And this criticism is certainly true of many other economic schools too. But either way, I remain unconvinced that all the typical conclusions flow logically from the premises."

As someone trained in philosophy rather than economics and having an interest in economics as it relates to social and political philosophy, it seems to me that most economists do not recognize the implications of their assumptions, either from a philosophical perspective or even from the perspective of the latest research finding is science, especially cognitive science.

One of the consequences is a tendency to generalize about human nature from their own experience. As a result, a lot of thinking betrays is social origins in culture and class, for example. Another consequence is a limited understanding of the dimensions of thought that underlie economics as a branch of social and political philosophy, which themselves rest on the foundation of logic, metaphysics, epistemology, rational psychology, and ethics, as informed by contemporary scientific findings in biology, psychology, and consciousness studies, sociology and anthropology.

The result is often a pseudo-scientific approach that rests on ill-founded and unexamined ideological assumptions. As heterodox economist and historian of economics Michael Hudson recently put it, the watch that such watchmakers make is expertly crafted, beautiful in appearance, and the gears mesh perfectly. but it doesn't keep accurate time.

Leigh Caldwell said...

Hi Tom

Agreed. I do think the more mathematical economists are better at stating their assumptions, because it's part of the discipline of mathematical proof; and from what I've seen, Austrian economics rarely follows that discipline.

But even when economists do make the assumptions explicit, they often don't examine how realistic those assumptions really are.

Some do, and the field is not submerged in such a complete fantasy as some critics suggest; but there definitely isn't enough questioning of the foundations.

Jule Herbert said...

You quote Rizzo and then follow with a comment on the relationship between saving and investment (and lending):

“...inappropriateness of the capital structure (malinvestment) generated by artificially low real interest rates (that is, interest rates that are lower than the real supply of savings would allow)”

The question is: does lending really require savings? Lending is about somebody spending their time now, in return for payment later (usually intermediated by a highly creditworthy institution such as a bank). There is no strict requirement for an existing stock of savings to enable this. In a simple example, three people could wash up on a desert island with no stock of wealth at all, and as long as two of them can find enough coconuts to support the third, she can get busy building a house for them to live in. This is lending with no associated savings."

But surely the unconsumed coconuts found but not eaten by the two gatherers are "savings" loaned to the builder, and the factors of production (grasses, sticks, mud, whatever) used by the third person, the builder, not for immediate consumption, but as capital inputs, are also "savings."
No; in a non-monetary economy one cannot have investment without savings.

Leigh Caldwell said...

Hi Jule

Thanks for the comment. Yes indeed - investment is by definition the same as savings.

But my point is slightly different: in order to enable lending, you don't need to have a pre-existing stock of savings.

For brevity I did not include any monetary instruments in my example. But I realise that this point would be clearer if we invented a bank which issues a fixed amount of money. This money would be lent to the builder in order to buy coconuts during the construction of the house. Interest on the money would be payable to the central bank and (presumably) shared out among all three citizens.

In a non-monetary economy it is not really meaningful to talk about lending, because a loan is an instrument denominated in a fixed nominal amount of currency. On the other hand investment, as you point out, can happen whether or not there is money.

Jule Herbert said...

Three islanders -- and we already are modeling a central bank.

I think Menger would argue that it would be more likely that coconuts -- or whatever the most liquid commodity in the community might be -- would begin to serve as a medium of exchange or money. And, yes, one could have lending during the period when this commodity was becoming accepted as the monetary commodity.

As for banking, deposit banking might at some time evolve. But a non-deposit-holding bank issuing chits or a currency numeraire but which had no real capital assets (saved coconuts here) would find no home on the island (unless perhaps one of the inhabitants was one shrewd operator, a real John Law type).

It is, according to Mises et al., the modern (and not so modern) banking system which allows the disconnect between savings and investment and leads to the business cycle -- through credit expansion and interest rate manipulation. Rather than serving as honest brokers or intermediaries between savers and those seeking loanable funds, bankers, through the political means of a state-imposed monetary system of legal tender and such, enrich themselves and set into play the process which constitutes the business cycle.

Tom Hickey said...

I'm not following you here, Julie. Bank deposits are liabilities (payables) of the bank, not assets. Loans (receivables) are the corresponding assets on the bank's books. Conversely, the deposit is the asset of the borrower, and the loan is the liability of the borrower.

Deposits don't create loans; loans create deposits. The bank loans against its capital (equity), not against its deposit liabilities. Banks leverage their equity, and charge interest commensurate with risk.

Banks cannot increase net financial assets through lending, since loan (asset) and deposit (liability) net to zero. Banks provide credit money temporarily until the loan is extinguished. Meanwhile, the credit money can be used to increase the real assets of the borrower through employment as capital investment.

Jule Herbert said...

Tom: You are correct: I should have said a "non-capitalized-bank" (a bank not having any real coconuts to lend out). But banks also lend out their deposits.

Tom Hickey said...

"But banks also lend out their deposits."

Winterspeak speaks to this issue in a blog post that is brief and worth reading.

Jule Herbert said...

Not so fast there!

The key sentence in the Winterspeak analysis is:

"So, at an individual bank level, banks make loans as they like, and then either borrow the reserves they need at the discount window or overnight market, OR they grow their deposit base to meet their reserve requirements. But the banks make the loans first, and then do whatever to hit their reserve target at the end of the day." (Emphasis supplied.)

And he is speaking of bank operating the the current central bank system, where reserves are not held by the bank itself but on account at the FED. Our island bank is a stand alone bank.