Denmark’s Niels Bohr considered himself more of a philosopher than a physicist, yet he contributed so much to the groundbreaking approach that became the basis for quantum physics. At the same time, Germany’s Werner Heisenberg was writing the famous paper on “his” uncertainty principle, Bohr was purportedly on vacation thinking up the deeper consequences and meaning of all its implications.

Upon returning from his holiday, the Dane wrote to Heisenberg urging the German to include some appreciation for what he thought must’ve been the greater fundamental truth going beyond strictly addressing uncertainty. Convinced, Heisenberg obliged and included the following as a note to his groundbreaking paper:

Bohr has brought to my attention the uncertainty in our observation does not arise exclusively from the occurrence of discontinuities, but is tied directly to the demand that we ascribe equal validity to the quite different experiments which show up in the theory on one hand, and in the wave theory on the other hand.

Indeed, Bohr’s idea came to be known as complementarity which he also applied to a great many more settings beyond strictly quantum physics or physics of any kind. In the latter, it was and is related to the uncertainty principle whereby Heisenberg had shown the limitations of even the best, most sophisticated experiments.

Is knowledge itself constrained by very real physical boundaries?

Niels Bohr made a stab at why these limits may have existed in the first place, and, more importantly, what they appear to mean for not just those who design and carry out the experiments but for those also trying to frame reality by human thought perhaps unsuited or at least ill-suited to the specific task.

Grossly oversimplifying, complementarity aligns with a great unknown; that “reality” itself may be literally indescribable in very big things as well as the very small purported building blocks of nature. We take what we know of fundamental particles, for example, from what we can observe of their specific interactions with each experiment, inferring a great deal from the instruments measuring only those specific interactions.

Thus, we have no choice (formalized in physics by Heisenberg’s uncertainty principle, as a start) but to keep in mind that each experiment only serves limited use. Individual experiments cannot produce a “single picture”, as Bohr said, of physical reality. By and large, complementarity demands we acknowledge the “totality of the phenomena.”

In Economics, though the differences between it and physical sciences are legion and immense, perhaps much of its inability to answer any larger questions stems at least in some large part from how it seeks out any answers at any scale: ceteris paribus. In holding “all else equal”, Economics then seeks to assure its interpretations of the wider reality of all those things artificially being held equal.

We do this all the time; in formalizing our understanding of bank reserves, for example, how do we describe them and their function? We get out our T-diagrams and conduct a thought experiment where one single bank might undertake a single transact with a single central bank, swapping directly one asset for another. Technically correct, this doesn’t actually tell us much or maybe anything important because in the real world the monetary “reality” which only begins with a QE transaction just might be equally indescribable and therefore leaving our limited thought experiment as valid upon only its own very narrow terms.

Mechanically true, otherwise potentially meaningless.

Such for the repo market, too. To understand this main point of modern money emphasis, we might conduct a limited thought experiment because to attempt anything more complex leaves us in a haze of misconception and within a tangle of seemingly potentially contradictory outcomes and reasoning. As Socrates in Plato’s Republic, we presume that whatever we might come up with for our narrow thought experiment must therefore apply to the system as a whole.

Bohr would be furious.

An example. Most people attempt to reconcile repo in this ceteris paribus sort of framework: as a bank which wants to conduct a repo transaction, singular. Already, public perception is at a disadvantage because the public is conditioned to think along the lines of financial transactions with which it is already familiar. If I want to buy a Treasury note, for instance, I put money in a brokerage account and my broker goes into the marketplace to buy the specified instrument.

Was Aristotle An Idiot? Welcome Back August 9, The Answer To That Question The Key To The Next Golden Age

That’s not how it works for banking. A modern financial institution has a myriad of different parameters to manage, only starting with a complex maze of asset choices which then must be harmonized with a wide variety of liability options. And then time itself is another key factor.

In our simple example, the bank doesn’t actually need any “money” whatsoever to come into possession of an asset. To begin with, we have to first define our terms: what we mean when we say “buy” is that whomever is doing the buying wishes to establish title, or full ownership, whereas other forms of possession such as securities borrowing confers more limited rights and uses (including the right to reuse). Specific situations might require only the latter rather than the former in order to accomplish the bank’s goals.

For a bank that wishes to establish title, buy, a Treasury note, there needn’t be cash on hand at the start, either. In the morning, it directs its market-making operation to go into the market and claim title on a note from a perspective seller who agrees to the terms of the transaction. The bank posts no cash, but the seller is credited anyway (or the seller’s agent).

Instead, our experimental bank has simply created and incurred what’s called a daylight overdraft. Having met the terms of the seller, the seller’s bank (if the seller isn’t a bank itself) is immediately credited the agreed-upon amount while the buyer is deducted.

Deducted from what, or from who?

In the old days (pre-90s), this was the Federal Reserve. As a matter of inadvertent policy, the Fed didn’t charge a single penny for banks running daylight overdrafts (Fedwire); penalties were only applied if they weren’t settled by the close of business. There came to be position limits and more scrutiny in the wake of 1985’s Bank of New York debacle, but it wasn’t until the next decade that policymakers began to realize securities dealers were using daylight overdrafts as sort of an intraday funding mechanism (and why wouldn’t dealers?)

Eurodollar University’s Making Sense; Episode 46; Part 1: Getting Fishy in the CHIPS

Yet, once the Fed started to charge and clamp down on them, the private marketplace (meaning JP Morgan and Bank of New York) stepped forward and recreated pretty much the same thing in what’s called triparty repo. Instead of daylight overdrafts being extended and as ultimately the responsibility of the Fed, these were then offered – for “free” – by the triparty repo system and its twin custodians which then developed and expanded (so much that it was repeatedly at each big shock epicenter during the first Global Financial Crisis).

So, our bank has bought (obtained title) a UST by borrowing a free intraday credit from its triparty repo custodian. This means that it has until the end of the day to settle what it might owe. This is where the menu of liabilities comes into play: in our fictional single purchase transaction, it could be “funded” by a scheduled incoming intrabank payment from another counterparty, or maybe as an offset to a collateral swap or short sale already being made.

Or, if nothing else, the bank can arrange a repo financing for this particular UST note. According to this other, this bank has put down no cash to buy the instrument and then closes out its daylight overdraft with the triparty repo custodian by arranging a secured financing transaction (SFT) with someone else in the repo market also using triparty repo (who, by the way, is granted limited seizure rights over the very UST which our bank has gained full title).

If there is any cash out-of-pocket at day’s end, it would only be for the small haircut difference.

But even that’s not really a thing because this isn’t really much closer to what must be shadow money reality, either. Complementarity: I keep writing that you have to look beyond these simplistic, very crude thought experiments if you’re ever going to more completely (maybe never fully) assess the totality of this modern money phenomena. What I mean is, these aren’t single transactions piled discretely one on top of the other in which we theoretically might be able to trace one-for-one, from buyer to seller, from full rights to limited reuse rights. Thinking about them in this way you do yourself a disservice.

Banks run entire portfolios of securities (assets) and at the same time seek to maximize their options among a truly wide array of funding options (liabilities); what they are not doing is putting on isolated trades one at a time, even if, for ours and regulatory purposes, that would make this whole mess so much easier to untangle and appreciate.

It’s instead been the job of regulators to have come up with some way of basic tracking for even the contours of all this, but, as I often have to still write, too, it’s just not part of the mainstream worldview – even after 2008 – which even today remains steadfastly focused on the wrong place. Any simple thought experimentation is limited, in the official sense, to whatever ends up with bank reserves as if these satisfy enough or even of a few permutations.

That they don’t, they continue to refuse all in the name of more QE’s and more bank reserves, has left us with developing a workable set of experiments and observations to do their job so that we can hopefully better do our job. Quite often, as with TIC data, we have to work with what’s available to us even (especially) if authorities won’t, can’t, or just don’t. 

Another example and set of experimentation makes up Part 2.