Say my name! (Walter White - Episode 7, Season 5) “Heisenberg”
The uncertainty principle states that the position of a particle and its momentum cannot both be described accurately.
The rollercoaster above serves as an analogy for how the uncertainty principle works at scales much smaller than this. Left: When the rollercoaster car reaches the peak of the hill, we could take a snapshot and know its location. But the snapshot alone would not give us enough information about its speed. Right: As the rollercoaster car descends the hill, we can measure its speed over time but would be less certain about its position. The uncertainty principle is a trade-off between two complementary variables, such as position and speed.
Credit: Lance Hayashida/Caltech
I don’t pretend to be able to follow the math or the physics, but it does seem to me to have something to say about financial markets and macroeconomics. We can measure any number of things related to the financial markets. We can measure correlations and prices, trend lines and volatility. Intuitively, it would seem that measuring more ought to give us more information, more useful information about not only the individual inputs that comprise the market, but also the direction of travel and behavior of the market overall. What actually happens, the more we dig in, analyze and absorb, is that we become way better informed but less certain about outcomes. The more we know, the more we understand that we don’t know enough. I touched on this here. If historical trends and regressions perfectly predicted future outcomes, investing would be easy. They don’t. It isn’t.
It seems ironic that, notwithstanding this conclusion is reasonably obvious, that the currency of market commentary is conveying certainty. Not the kind of commentary that counts as advice, in a regulated, SEC, FINRA way, but elsewhere, on X on Reddit and other social media platforms where there is no accountability. I listened today to a discussion between Brent Johnson (Milkshake Theory) and Grant Williams on the excellent Grant Williams Podcast. They discussed this apparent contradiction. Johnson is a money manager and also an active participant on X, so is aware of the conflict. Johnson freely admits that, early in his career, he had a tendency (not unusual) to form knowledge based on the most recent thing read and to parrot that knowledge as truth with apparent understanding. As he has aged in the business, he has become more thoughtful, less quick to form and dispense opinions and more skeptical of inbound information, questioning everything: measure 10x, opine/act once.
As I work my way through Michael Howell’s Capital Wars, it is becoming ever clearer that the under-appreciated aspect of financial markets behavior is global liquidity. Fixed exchange rates, the gold standard, regulated capital flows have all been swept away (largely) in an era of global capital markets. Capital is supposed to flow freely across borders finding its highest and best use wherever that might be. Government is supposed to exercise a light touch, making sure that playing fields are level, competition unbridled and legal systems functioning and fair. They are supposed to regulate where markets themselves have difficulty doing so and provide backstops and liquidity at times of stress where collective reactions may become irrational.
Oddly, though, this has not immunized the global economy from shocks and crises. In fact, as Howell observes:
“...even though financial crises were not unknown under the gold standard, the last 30 years have been among the most tumultuous in monetary history...
...Whereas in the 1960s, the World economy mostly suffered labour cost shocks and, in the 1970s and 1980s, oil and commodity-price shocks, it is more often buffeted these days by Global Liquidity shocks. Financial markets spin on fragile axes and this common driver emphasises that modern financial crises tend to be neither purely national, nor simply isolated events. Moreover, the Global Liquidity shocks are typically bigger, longer-lasting and more pervasive than the calibre of shocks studied by economists and Central Banks when using their so-called DSGE12 (Dynamic Stochastic General Equilibrium) models of the economy. In fact, since 1980 well over sixty countries have experienced asset booms followed by banking crises, with at least six episodes of major asset price bubbles: (1) 1980s Japan; (2) early 1990s in Sweden and across much of Scandinavia; (3) Thailand and the neighbouring South-East Asian econo-mies in the mid-1990s and (4) US in the late 1990s and, so far, twice again in the 2000s. The social and economic costs have been high, with national banking systems in many of these countries subsequently collapsing, after facing loan losses from these bubbles that on occasions exceeded a staggering one-quarter of their GDPs. As 007 agent, James Bond, keenly observed in Goldfinger: “Once is happenstance. Twice is coincidence. Three times is enemy action ”.
The data we have available to analyze and measure using these models and continue to generate is so vast that it is beginning to overwhelm the ability of the electrical grid to fuel the data centers needed to store and crunch it
Data is becoming such a pressing issue that Microsoft, among other cloud service data center providers, is actively considering powering its data centers with small modular nuclear reactors.
Not only do we have to bring humility to the analysis of the data we measure, we also have address how to store the raw material of measurement. The more tools we have available to us, the more skills and resources we need to mobilize to use them. Hardly surprising that we needed to develop large language (and data) models - AI - to make something of all this. The chalenge is whether we will understand or be able to adequately explain the output/conclusions or just act on them.