Rational irrationality

The comments of this post from Coordination Problem are interesting, especially - as usual - O'Driscoll's. The topic is Bryan Caplan's theory of rational irrationality. There is also a comment of mine, but O'Driscoll's comments make it outdated. I think I will think about the issue during the holidays.

Just a small summary of Caplan's argument (I bet my language is different: I've never studied public choice seriously, I'm Austrian, and I read the book "The myth of the rational voter" more than one year ago).

1. If I believe that gravity doesn't exist and I jump off a bridge, I die. That's the cost of being irrational, which is particularly strong in real "private" life.
   
2. Where's this feedback mechanism in politics? Each single voter has a 0.0000001% probability of being relevant at the margin at the elections. The benefits of informed and rational choices are public, the costs are private: reasonableness is a public good. Rationality is systematically underproduced, because rational politics from the point of view of the voters is a tragedy of the common.

The argument is a priori correct and a posteriori touches a very important fact I consider beyond dispute: people which are clever in their private lives become much dumber when it comes to politics, and I have never noticed the opposite. Besides, the theory also explains the success of Keynesian ideas.

The objections raised in the comments are:

1. Organized political groups are competitive and will invest in "rationalizing" the voters.
   
2. Economists don't invest in laymen's education: when they did it, voters were more rational.

These are great issues. Unfortunately I don't know much public choice, but I think I have something to say about it. Next week.

Is money a present or future good?

This morning I read "credit creation or financial intermediation?" by Cochran, Call and Glahe (QJAE).

The paper precisely hits one point of disagreement in the monetary disequilibrium vs injection effects debate, which divides the orthodox Austrian camp (Mises, Hayek, Rothbard,...) and a new brand of Austrian-monetarist synthesis whose most important proponent is Horwitz (Microfoundations for microeconomics), and is at the root of the works of Warburton, Yeager, Selgin, White, and Sechrest.

The two theories appear to differ on many aspects, some of them of purely ideological and noneconomical nature, and some of theoretical interest: is free banking feasible?, is money a present or future good? are injection effects or price rigidities responsible for monetary non-neutrality? is the law of reflux a feasible way to enable a proper use of voluntary saving?

Today I want to talk about the issue of presentness, because I consider both camps to be wrong (which means that the only right explanation is to be found in this blog, and this adds to the unlikeliness of its reliability):

• Cochran says that there is no cost in money holding, in terms of foregone consumption, because money can be used without any delay: money is a present good.
• Horwitz says that money holding is saving, because it is not consumption.
• I say that money holding is neither of the two.
  

When a bank receives 1$ of deposit and decides to lend a fraction 0 < X < 1 of it, it is immobilizing X$ worth of real resources (assuming constant prices) for an investment project. The moneyholder, however, doesn't lose the availability of his spending power, so that it has not yet decided whether he will consume or not. The bank is right in lending the deposit if the consumer won't exercise his option of using his money to buy consumers' goods before the investment is liquidated. The bank is wrong if the option is exercised before final liquidation, because it will have to ask for credit to pay for something that was not intended as credit. If all the banks at once immobilize more money than money holders are willing to consider as real saving, the banks will run out of liquidity and a crisis will ensue. This is not a liquidity crisis: it is a real crisis, because it is the command over real resources that is not available, not liquidity per se, at least if money doesn't depreciate (i.e., if a firm and a consumer want the same commodity, this is real scarcity, it is not a monetary problem). Of course, the banks can reduce the real value of their monetary liabilities in order to avoid running out of liquidity by printing more money. This is exactly what happens under central banking. However, without a public safety net I don't believe that this behaviour would be feasible, other than in the shortest run: there are limitations to monetary devaluation in that outside money cannot be printed, as it is commodity money.

This is not an analysis of the business cycle, becasue, to say the least:

• a distinction between real and nominal movements should be done: if prices were perfectly flexible, market coordination were instantaneous, and monetary movements had no distributive effects... money would only be a nominal issue, of no interest for economics (this is unlikely, of course),
• not all banking crises result in inflation, so that an analysis of deflationary crises is necessary (there is no mystery here: credit and monetary aggregates are countercyclical, so not all crises result in devaluation, a stop in the credit creation process suffices to create a crisis when the underlying market structure is unsustainable, and its reversal adds to it).

I believe that some of the deposits are to be considered as saving, and that fractional reserve banking is stable to the extent that it can distinguish, at a systemic level, between money intended as saving and money intended as consumption.

The error in the monetary equilibrium analysis is to confuse a tautology (money holding is saving now, as there is no consumption) with the real economic issue (in order to avoid problems, money holding will have to be saving from now to the end of the underlying investment).

At this point, a question must be raised about how banks can distinguish between real saving (in the economically relevant sense, i.e., over an horizong linked to the time structure of production, which in monetary equilibrium analysis is never considered) and merely instantaneously foregone consumption. I know of no reliable answer to this question, maybe because, in the literature I know, both "presentists" and "futurists" don't see the problem in these terms.

Counterfactuals

Uh, this morning I've understood another small element of the edifice of economic theory.

Here's my comment (on this post by Prof. Rizzo) which shows the intuition:

Moral hazard is not only relevant because of deposit insurance: it is relevant whenever costs are socialized.

If I invest too much now, together with everybody else, in a private-costs world I would expect high interest rates in the future, when there will be a scramble for credit, i.e., investment that raises the demand for capital, as Hayek put it.

When there is the central bank following a countercyclical monetary policy, during the scramble for capital, normally, interest rates don't rise in the short run, and probably also in the long run (the policy intervention will be at least as persistent as the crisis, i.e., the scramble for credit).

Counterfactually, an interest rate that in a private-costs world would have risen to 7-8% at the onset of the crisis, becomes a 0% interest rate because of the central bank. This may be imply great cost socialization… as the counterfactual market interest rate is not observable, however, it is difficult to evaluate if the phenomeon is relevant.

When costs are socialized, the market system is uncapable of retrieving and spreading correct information and forcing correct incentives. All the coordination process goes astray.

Competitive market agents will take moral hazard into account and “optimize” taking the cost socialization technology into account, i.e., they won’t “optimize” to find a “social optimum”, whatever it means.

Systemic risk cannot be avoided because noone has private incentives to avoid it. Only monetary authorities can stop socializing risks, but countercyclical policies work on the opposite principle.

Ebeling on Mises

This post by Ebeling on Coordination Problem is very interesting and well worth reading.

Mises's epistemology was characterized by a dichotomy: Theory and History. The epistemic properties of the two realms of social inquiry were opposite. I think that it is better to use a Husserlian terminology and call them "research of essence" and "research of existence": the former, Theory, is about logical relations among concepts, the latter, History, is about all the complications which arise when it comes to apply Theory to the real world.

The hyperrationalist interpretation of Mises, which I think was due both to Rothbard (in which constructivism and apriorism play important ideological functions) and Hayek (for instance, when he said that Mises's "Socialism" was a great and influential work, but too rationalistic), is not compatible with what Mises writes, with or without the additional important insights which Prof. Ebeling has added with his post (which I'm looking forward to reading in the referenced books by the Liberty Fund).

Without History, what remains of Mises is a delusional hyperrationalist who thinks to be able to derive propositions about the real world out of purely logical constructs. This interpretation is not consistent with Mises's writings, in which the dichotomy between Theory and History hides all the uncertainty of the real world in the latter camp. Mises claimed explicitly that the apriori may have had an evolutional origin, that all theories are incomplete and subject to radical changes, that all the rational constructs are based on more or less arbitrarily chosen "ultimate givens". Where's the hyperrational? To have clear ideas and to state them clearly cannot be a symptom of hyperrationalism.

Let's make an example: socialism is impossible because there cannot be monetary calculations AND monetary calculations are necessary for coordinating a complex economy. The last hypothesis is not apriori: it is an empirical generalization of very vast applicability.

There are cases in which empirical hypotheses, which are logically necessary to apply Theory to History (essence does not beget existence, Saint Anselm notwithstanding), are not as clearcut. While it can be claimed that reality is complex, production takes time, capital is heterogeneous and information is scarce without fear of being falsified on a posteriori grounds, it cannot be said with the same level of certainty that the structure of production is sustainable or not, if a boom is fundamentally sound or will be turned into a bust, whether economic agents are conned into believing, or acting as if, something is true when it is not...

Relevance is never a priori concept: it is an a posteriori judgement about the relative importance of several conceivable causal factors in the interpretation of a specific historical event.

My reading skills fail me

I was asking a question to Prof. O'Driscoll on this Prof. Boettke's post on the public choice dangers of monetary policy (post and comments strongly suggested). Then I realized that I didn't understand its comment, as I thought he was saying that the Fed has sterilized monetary interventions in Fall 2008 (instead of BEFORE Fall 2008).

However, this is a 15m work on Fed FRED data, and I post it as a summary.

Fed Policies in Fall 2008.

Target interest rates: in Fall 2008 it fell from 2 to 0. This is expansionary.

Monetary base: in Fall 2008 it came from 850 to 1700b$. This is expansionary, too. Quantitative easing implies that monetization is not sterilized, otherwise is a portfolio rebalance (buy toxic waste, sell t-bills). The Fed would have run out of good collateral.

Extra facilities like TAF, TALF, MMIFF, AMLF, CPFF, created from Summer 2008 to Fall 2008, created reserves. I'm quite sure about that regarding TAF (and the already existing, in Fall 2008, PDCF), the others are a little bit to complicated for me. The already existing TSLF was not a monetary facility, but a credit transformation facility.

One element remains out of sync with the other policies: interest payments on deposits at the Fed, started in mid Oct 2008.

However, this rate fell to zero (dot something) after two months of existence, so I wouldn't consider it relevant at the margin: it's a 0.0025 * M0 gift for the banks, i.e., less than 5b$/y.

If return rates were not (risk adjusted) zero elsewhere in the economy, the substitution effect would have been quite nil. If they were zero elsewhere in the economy (like in New Keynesian ZIRP models with negative real natural rates), this is what has to be explained.

Besides, the CPI started falling before interest payments on reserve (July 2007), and deflation accelerated in September 2008, before interest on reserves. Inflation turned back in early 2009, so the eventual effect of this policy on the CPI was very short lived, if existent.

Interest on reserves is somehow a mystery, and I can only explain it as fear of inflation, but with 0.25% interest paid, I think it's also practically irrelevant.

Why did they do it? The standard answer is "to enlarge the Fed balance sheet without losing high quality collateral, i.e. t-bills", but why did they feared this? Just expand M0 as if there is no tomorrow. Japan already did it, and with zero effect (it makes the public debt marginally less expensive and it helps overvaluing toxic waste).

Little fragments that do not add up to a monetary theory

I'm amassing material on several issues in Austrian business cycle theory which do not seem to me well founded from a theoretical perspective. I hope noone has noticed these problems yet, so that when I solve them I will get in the Olympus together with Mises, but I often find papers which have treated some part of the problem and give important hints (not necessarily a satisfactory treatment) of the very same problems I was investigating.

One of these problems is how (in detail) money influences real activity. Objections that can be found in the literature are: small interest rate movements cannot explain much variation in investment decisions, the risk-free rate is dwarfed by risk premia, money is effective only in the short run and thus cannot affect long-term investment, rational entrepreneurs will make random (unbiased) errors and not biased errors favoring long-term with respect to short-term investment. Then, there are debates on whether the money stock (M) or the money flow (MV) set the stage for the boom-bust cycle, whether only M1 is relevant or also Mx, with x = 2, 3, ZM, whether price rigidities or injections effects are relevant, whether money is to be considered an information-carrying signal or (also) a means of exchange (in the former case, the Lucasian framework, monetary economics would be a mere subset of statistical estimation theory), whether rules are really preferable to discretion and which kind of rules (NGDP, M0, M1, M2 targets...) is preferable.

At the moment, I'm just collecting pieces. I've no full-fledged monetary theory in my head, and I only have a mass of papers on a shelf over my bed, which is somehow dangerous for my safety (and I hope this will incentivize me to read them as soon as possible).

Prof. Boettke has recently written a post on his blog regarding his public choice doubts about the policy of stabilizing the flow of monetary services (MV), also known as nominal GDP. I too have doubts, although nothing I can write anything of interest upon: I just started reading about monetary disequilibrium, and although I'm not convinced by this theory, I don't think I'm already grasping enough details. However, I will try to sketch an argument in favour of freezing the monetary base. I will divide the argument in steps.

1. A decentralized market system can work provided that the price system conveys correct information and gives correct incentives. The price system is endogenous to the market process, meaning that entrepreneurs improve upon the price system and the latter helps the former. Errors cannot be ruled out, but absent systematic distortions, it can be argued that entrepreneurs will act so as to push the economy as close as possible to some optimal equilibrium. For this point to be relevant, it is not necessary that the market system really reaches some optimal status, or even gets close to it, but only that monetary interventions cannot improve upon its behavior. It's a relative and not an absolute comparison.

2. Monetary policy has redistributive effects: it socializes part of the costs of investing and taking on risks. This implies that there are systematic externalities within the market process imposed upon by central bank policies. In these conditions, entrepreneurs are led astray and the price system is distorted. This is an induced market failure: it is not that the market process fails because there is something wrong with it (this could be contracted away, because any market failure is a profit opportunity, although its exploitation may be hampered by transaction costs), but the result of explitic policies imposed upon the market.

3. The key point here is that cost socialization devices shouldn't be used, otherwise entrepreneurs, instead of moving toward some sort of optimal equilibrium, move to a suboptimal one. Somehow, and in this somehow I should put something from a theoretical point of view, the market process and entepreneurship cannot be trusted upon to internalize monetary externalities (which are, despite their name, real externalities, not pecuniary ones). Here I give two examples of market distortion: the baseline (neoclassical) view of markets is that market agents optimize some utility function subject to constraints, and externalities hampers this process, which loses its optimality properties. A more realistic (austrian) view of markets is that markets are also made of contracts, institutions, bargaining, market/firm choices; externalities will lead all this complex entrepreneurial process astray, for instance by incentivizing insufficient screening in credit transactions. It money is the tailor which sews together the market process, whatever happens to money is something that is at least in part beyond the scope of entrepreneurship: coordination is bound to fail, and if monetary policies are systematic, failure will be systematic as well.

4. Any rule which eliminates or minimizes these redistributive effects will enable the market process to behave properly. If the rule is credible (although I think this hypothesis to be wholly unrealistic), which rule is followed matters much less than the fact that entrepreneurs no longer have any "externality producing" technology to exploit. This is a point that runs counter to both orthodox Austrian economics and monetary disequilibrium theory (although I only have a cursory understanding of the former).

5. In conventional Austrian economics, it is injection effects (increases in the money supply that are spent in credit markets) which start the business cycle. This process requires the cooperation of the banking system, which however will also pay the cost for these errors. Here's the point: if the monetary rule is followed, the banking system will be forced to internalize this cost, and will be less compelled to expand credit. The US banking system in the XIX century didn't need a central bank to cause a mess, but it can be argued that its weakness was due to limitations to branch banking, institutional uncertainty (silverites, greenbacks...), the forced use of securities as outside money (instead of plain gold, or, as I'm advocating, a fixed amount of base money), and moral hazard caused by banking holidays and specie conversion suspensions. Banking instability was not the rule in the world before central banks, so I guess there must be some specific US factor behind the misery of its monetary performance, both before and after the creation of the Fed.

6. In monetary disequilibrium theory, either in the aggregate Yeager's version or in the more Austrian Horwitz's version, it is monetary expansion in excess of money demand that starts the problem. This calls for some stabilization of MV, in which V is somehow taken as the inverse of money demand. The transmission mechanism relies on price rigidity: however, I believe that price rigidities, among all the frictions which create monetary non-neutrality, is unlikely to be the most important. Financial frictions, credit injection effects, moral hazard in the demand and supply of credit are likely to be much more relevant, although I can't justify this convinction in detail. The market being a complex system, it is unlikely that only one friction is relevant in causing monetary non-neutrality, and it must be noted that boom and bust policies are never a matter of mere money, but always involve credit, so that the secret of non-neutrality must lie in the relation between money and credit, and not in money per se (monetary disequilibrium, at least in the pre-Horwitz version, appears to be so aggregate that even the distinction between consumption and saving is irrelevant for its analysis of business downturns).

7. The "distorted market process" view is more robust because it does not rely on a single transmission mechanism. Whatever causes coordination problems by incentivizing excessive risk taking, capital consumption, overconsumption and malinvestment will more or less have the very same effect. Freezing the monetary base will eliminate the cost socialization technology that leads the market process astray. Thus, entrepreneurs will be forced to take into account the risk that their behavior will cause instability and, these costs being privatized, there will be an optimal risk-return choice. Externalities are likely to be the key building block in explaining monetary non-neutrality: tragedies of the commons that cannot be corrected by entrepreneurs, and impinge upon their ability to coordinate, may explain why injection effects are biased (without externalities, I have trouble imagining a tragedy of the commons in monetary matters). If Mises, answering in 1943 to Lachmann, said that the entrepreneurial cluster of error was not an apriori principle but an empirical generalization, systematic externalities created by monetary (and sometimes non-monetary) factors may give a robust microfoundation to this often criticized aspect of Austrian business cycle theory.

8. In the short run (at least a couple of years), this policy will be wildy contractionary because the banking system will be forced to step back, no longer being able to rely on the Fed's monetary safety net. Almost all of the risk-related aspects of the financial structure will be affected: there will be less maturity mismatch (monetary multipliers included), less financial leverage, less opacity and complexity, less reliance on originate-to-distribute business models, less currency carry trade, less short term foreign-currency borrowing... the enormous costs of this process are surely a very good reason to look for alternatives. But I see no alternatives that eliminate the root cause of the problem without having these adverse effects in the short run. Deleveraging is necessary to return to normalcy.

9. With respect to free banking, as described by Selgin and White, there is one major difference. Shifts in the currency/deposit mix will keep on having a destabilizing impact on the banking system: reserves turned circulating media will cause credit and monetary contraction, and banks will not be able to respond by printing banknotes convertible in outside money. Not being able to solve the problem by printing banknotes (offset by reduced deposits), banks will need to keep relatively high reserves in order to minimize the impact of these changes. With reserves in the range of 20%, it is likely that even large shifts in the demand of banknotes against deposits will have a minor impact. This war chest may not be required in free banking (with freedom of note issuance), but I don't see how it can be avoided in the system I'm advocating, which is surely easier to implement than an overall banking reform. I have no particular reasons not to eliminate this source of disturbance by also introducing freedom of note issuance.

10. My policy will not be able to eliminate all injection effects (banks may continue to issue money primarily through loans), price-stickiness effects (banks are constrained in their capacity of accomodating changes in the demand for money, because they need high reserve ratios to accomodate the risk of deposit/note shifts), and financial friction effects (there will always be moral hazard and adverse selection issues to take care of throughout the financial sector). The point is that there would no longer be an externality producing technology to shift eventual costs to third parties: so the entrepreneurial process itself is not hampered, and may be able to take into account and optimally adjust to all these factors.

11. The only remaining limitation, thus, is what limits entrepreneurship: transaction costs in credit markets and in monetary transactions (I think the former to be more relevant). Absent the cost socialization technology created by the central bank, all benefits and costs are private, or at least can be made private by entrepreneurial experimentation with the market structure. Only problems that are impossible to solve because they are too complex, because of too high transaction costs, or because they require too much information, will remain. But without systematic moral hazard agents have at least the incentive to avoid problems they cannot solve.

12 Finally, the monetary base is easy to check, and, thus, public choice concerns, to the extent that the rule is respected, are relatively low: there is nothing to estimate, like the neutral rate of interest and nothing to distinguish, like good (productivity driven) and bad (money side) deflation. Of course, there will be no political will to adopt this policy, which given the amount of confusion I have on these issues, may not be a bad thing.