Thanks for the kind words and for a fun discussion.

You might enjoy Condy Raguet’s 1839 treatise on currency and banking. Raguet scours some strange corners of the world for evidence that hard-money economies are not subject to the business cycle. From p.37:

“Such being the theory of this branch of my subject, I have the satisfaction to state in regard to the practice under it, upon the testimony of a respectable American merchant, who resided and carried on extensive operations for near twenty years at Gibraltar, where there has never been any but a metallic currency, that he never knew during that whole period, such a thing as a general pressure for money. He has known individuals to fail from incautious speculations, or indiscreet advances, or expensive living; but he never saw a time that money was not readily available, at the ordinary rate of interest, by any merchant in good credit. He assured me, that no such thing as a general rise or fall in the prices of commodities or property was known there; and that so satisfied were the inhabitants of the advantages they enjoyed from a metallic currency, although attended by the inconveniences of keeping in iron chests, and of counting large sums in Spanish dollars and doubloons, that several attempts to establish a bank there were put down by almost common consent.”

The GDP measure of savings is picked up from income statements, not balance sheets. The transaction in your example wouldn’t appear in savings statistics, and wouldn’t be interpreted as savings. And it shouldn’t be.

Your analysis is very good; you’re clearer on information asymmetry than I was.

I need more time to consider your ideas on fixed money supply and matched maturities, among others.

Thanks for an excellent discussion – hopefully we can recommence at a later date.

I am vehemently in agreement with the opinion that attempts to “stabilize” markets represent time-delayed feedback. While this is not the only conceivable cause of cyclical behavior, surely it is the most obvious!

In a free financial market, third-party accounting is the treatment for information asymmetry. The trust relationships work like this: the bank trusts the accountant not to reveal its proprietary information, and the bank’s creditor trusts the accountant to evaluate its balance sheet as effectively as if the creditor itself could see everything under the kimono. In principle, the accountant’s estimate of the bank’s financial position should produce all the information the creditor needs to evaluate its prospects of default, while retaining all the information the bank needs to keep a secret. While in practice this may not work perfectly, in practice nothing works perfectly.

As far as cascading waves of default, one way to think of it is to use the analogy of a snowball. For a snowball to get bigger as it rolls downhill, you need two things: snow, and a hill.

Ordinary risk is the snow. The inherent instability of maturity mismatching is the hill. If you have a financial system that does not create demand for long-term money by alchemically synthesizing it from demand for short-term money, an inherently unstable reaction as we’ve seen, you have snow but no hill. In other words, there is no automatic direction in which the system tends to panic. Keep a couple percent as a reserve buffer, and you have a tremendously resilient design.

Today’s financial system has no parallel in history in its ability to diversify risk. In this department, at least, it is like a tough kid who grew up in the ghetto. Perhaps there shouldn’t be a ghetto, but it’s not so bad to be tough.

The key to avoiding Knightian uncertainty is to keep patterns of cause and effect repeatable and analyzable. To overuse a metaphor, no market can predict the path of the hurricane from the butterfly’s wing. If you can get the feedback loops out of the system, you will still see real effects from real events – the doomsday asteroid, and so forth. But even asteroids are insurable risks, and there are no feedback loops involved. An asteroid can change the return on an investment, but changing the return on an investment can’t create an asteroid.

In order to figure out whether or not to sell the fish and exchange it for a more durable object or let it rot, you need to do an economic calculation, and this economic calculation is impossible without a unit of measure. That unit of measure could be Euros, dollars, ounces of gold, or anything else. To keep this unit of measure useful so that you can do calculations across time, it is useful to keep the value constant and predictable.

Again this word “value.” How can you hold something constant when you can’t even construct an objective definition of it?

Every definition of “real” (as in “inflation-adjusted”) that can be constructed – that is, every comparison of prices across time, or any other absolute barrier to exchange – depends on someone’s subjective, aesthetic assessment. These indexes are pure fudge. And who wants fudge in their equation? When you mix numbers and fudge, the result is fudge.

Look again at the calculation of saving I describe above. The goods are actual goods, the prices are actual prices, the numerators and denominators have the same units, etc, etc. And there is no fudge.

Thanks for the kind response. I’ll keep the anon tag temporarily for continuity …

If we compare non-market forces in an administered (apple) system with the free market system (orange), it makes intuitive sense that the former will be subject to more Knightian uncertainty, and the latter will be more efficient. The presence of the Fed and the PboC for example are surely factors of Knightian uncertainty.

A free market system should exhibit more continuity and lower volatility over time, as there would be no ongoing gaming of the ‘apple puts’ and so forth. This seems consistent with a more efficient market.

Conversely, when monetary authorities aim to reduce economic and inflation volatilities, monetary volatility inevitably follows. This includes cycles of overshooting. So they impede efficiency with a sort of feedback loop in their own actions. The more general administered “apple put” is a complex of pricing asymmetry (an option) that interrupts otherwise free market probability beliefs.

Another larger category of asymmetry – “information asymmetry” – has been in vogue recently. This is an old idea that is fundamental to the operation of a bank. Originally, a bank’s advantage in information asymmetry constituted value inherent in the intermediation services it provided. Now it’s become a beast of opacity, hidden from creditors and investors. This evolution was channeled largely by securitization. Securitization strips banks of their former integrating role in capitalizing information asymmetry. It replaces this with a partial role (origination/servicing) in a chain of relatively independent operators who capitalize different subsets of the same total bundle of information asymmetry. The result is a more complex chain of information asymmetries that was previously the case with a bank.

While the first kind of asymmetry – the “apple” put – would not be present in a purely free market, I’m not sure about the second. Perhaps the natural evolution of free markets toward more granularity and differentiation produces waves of information asymmetry that inevitably interfere with otherwise smoothly interacting risk probabilities. Both types are inherently hidden from the broad market, thereby inducing Knightian uncertainty.

Taleb’s theme is true. Too many risk management systems are dense black holes of stunted statistical analysis – necessary to a degree, but in no way sufficient, and to the degree that their insufficiency is unappreciated and their necessity overemphasized, extremely dangerous.

I particularly like one Mendelbrot quote on this subject:

“Value at risk is a theory of ocean waves whose swells are forbidden to exceed six feet.”

That says it all.

This isn’t true. Lots of people want two year old rotting salmon. Farmers. Marine biologists. If you have ten tons of rotting salmon, you might be able to sell it to a farmer for fertilizer in exchange for an iPhone. Or not.

In order to figure out whether or not to sell the fish and exchange it for a more durable object or let it rot, you need to do an economic calculation, and this economic calculation is impossible without a unit of measure. That unit of measure could be Euros, dollars, ounces of gold, or anything else. To keep this unit of measure useful so that you can do calculations across time, it is useful to keep the value constant and predictable.

Once you have a ruler (lets call them dollars), it becomes possible to do calculations regarding value. I can calculate how many ounces of gold to trade for salmon or to make calculations regarding time value.

moldbug: It is because a large number of market participants undergo a Darwinian selection process which tends to reward and promote those who correctly identify patterns, and punish and dismiss those who incorrectly misidentify them.

The people that tend to survive the markets are people who don’t try very hard to make firm predictions and who respect the fundamental uncertainty and unpredictability of reality.

“This is a great opportunity for Europe …. Europe will be far richer as a result – unlike a certain neighbor across the pond.

“History has a way of repeating itself, soon a loaf of bread will be the equilvalent monetary value to US dollars carried in a wheelbarrel.”

I come back to this otherwise excellent blog about once every few weeks to read Brad’s insight, and marvel at the ignorant jingoism of its similarly otherwise informed (with some exception) commenters.