Or: How Philosophy Can Make You a Billion Dollars
(sorry, I couldn’t completely resist the click-bait-y title)
This essay began as a series of notes on the George Soros’s book, “The Alchemy of Finance” (1994). The book itself is one that I lightly recommend. Soros proposes his philosophy, then uses it to drive investments that make him a billionaire by the end of the book. His wealth doesn’t validate his philosophy, but it does mean it passed through some filter that surely burns many others. However, his core philosophy often frustrated me with various inconsistencies and broken naming structures. As I kept struggling to form and form his ideas, in some places I eventually moved past his language altogether. As such, much of this essay is based on some bones he leaves lying around, but the skeleton is not entirely his.
What follows is about 5,000 words exploring the seed of an idea called reflexivity: the ability for predictions to affect outcomes, and vice versa.
Overview:
In Part I, we define stable and unstable systems, and how the combination of uncertainty + reflexivity can render a system unstable.
In Part II, we apply this knowledge to the world of finance, first by creating a model of a market, then by showing where the classical model of equilibrium in supply-demand curves breaks down, and ending with an examination of how reflexivity can create boom-busts cycles in the stock market and beyond.
In Part III, we begin to explore some of the larger conclusions that one might draw about the current economic model: that free markets and floating exchange rates might be inherently unstable, and that regulation is unable to prevent this in its current form, although it continually thinks it can.
Part IV ends with some closing thoughts and remaining questions.
Part I: Stable and Unstable Systems
All systems can be divided into two types: stable and unstable.
In a stable system, what is true at one time will be true at other times. Most of nature is (simplistically) a stable system, where the laws of physics hold constant over time, such Newton’s first law is true today as it is tomorrow. In such systems, the impersonal laws mean that man cannot achieve a desired result without understanding these laws (even if only implicitly, and not explicitly - you don’t need to define gravity in order to understand roughly how it works). Operational success - the bringing about of a desired state - is then equivalent to science, what I will call stable science. The goal is to understand Truth of a permanent nature.
Many natural phenomenon allow for stable science:
- Starting with initial conditions and hypothesized scientific laws, one can make a prediction.
- Based on final conditions and hypothesized scientific laws, one can posit an explanation.
- By comparing predictions against final conditions, one can do testing.
Note, predictions and explanations are thus reversible, the mirror images of each other, and scientific laws cannot be proven, only disproven.
In an unstable system, there may be hidden underlying fundamentals, but the movements of the system are ever-changing, with no set pattern. A causality might hold true for a short period, but not for all of time. Truth is inherently temporary. (“The Three Body Problem” opens with a description of scientists trying to wrestle with a world where nature is, itself, unstable. For that alone, I highly recommend the book). As such, set initial conditions will not always yield the same final conditions. A prediction or explanation may hold true sometimes, but not always.
(Note: one useful explanation that Soros misses completely is that of dynamical systems, and as much as I’d like to delve into those here, I’ll have to save it for a future essay.)
In stable systems, one must understand the permanent truths in order to achieve operational excellence. In unstable ones, there are no permanent truths, so pure scientific study will never suffice, and one must settle for what works (operational excellence) and accept temporary truth. The question is then, to paraphrase Jim Paul, “would you rather be right, or make money?” (Soros calls these two methods science and alchemy, but I find this categorization woefully lacking, disrespectful both to science and the ancient effort of alchemy).
Unstable systems are on average harder for the modern Western person to grasp, since it requires a radical shift in thinking to instead contemplate a process of change, even if that process is closer to the truth. To reference Ian McGilchrist, this in economics maps almost perfectly to left hemispheric bias: left brain favors static models, while the right sees the enduring identity of a moving river.
The human uncertainty principle
Human affairs always have 3 sources of uncertainty:
- Inability of participants to know what other participants are thinking
- Participants can be guided by differing and sometimes conflicting interests
- Participants can have in themselves conflicting interests
Each of these elements is inherently immune to observation, and so cannot be fully known with full certainty. it means that participants views are always partial and distorted. However, this uncertainty is not enough to affect a system’s stability - there needs to be a mechanism that propagates that distortion throughout the system.
Introducing Reflexivity
Reflexivity is a circular mechanism, wherein expectations of the future affect future outcomes, and vice versa. The relationship between the “one doing the expecting” and the “world” is then described as “reflexive.”
Reflexivity as a term originates from the social sciences, attributed to father of psychology William James, expanded by sociologist Merton with his concept of “self-fulfilling prophecies,” and greatly explored by Karl Popper. Popper initially considered reflexivity a uniquely human element, but would later see reflexivity in certain aspects of molecular biology, and use it to hint at the limits of the scientific method.
A reflexive relationship opens the door to positive feedback loops, where feedback amplifies changes instead of dampening them. This pulls further from equilibrium, not towards. The result is an inherently unstable system, sometimes called a “complex adaptive system.”
Reflexivity + uncertainty creates instability
Due to the human uncertainty principles, participants make wrong guesses about the future decisions of others. Reflexivity means that these wrong guesses can affect the future decisions of participants: distorted views can influence the world, because false views lead to inappropriate actions (for ex, treating drug addicts as criminals, can increase their criminal behavior).
The combination of reflexivity and uncertainty will thus render any system unstable.
One can then begin to identify subsets of unstable systems by looking for those two elements.
History as unstable system
History, for example, is a process shaped by reflexivity and human uncertainty. Humans have predictions of the future, which then can determine the future. Similar to Darwin’s genetic mutations, one could say that the predictions - which are inherently flawed - shape the processes of history and are themselves encouraged or killed off by the process of history. Perhaps the ideas that make up history consist of fertile fallacies. These fallacies are originally conceived as an insight that, as they are translated into reality, reveal their shortcomings. These are successively replaced by new fertile fallacies antithetical to the old. Each fallacy provides a new experience, which engenders new lessons. To the extent that people learn from experience, the process can be described as progress.
Part II: Reflexivity in Finance
In the world of finance, one can find an interesting thesis in betting against the mis-categorization of systems. This is what George Soros does, by looking for reflexive mechanisms in systems that other investors treat as stable (and who thus treat reflexive relationships as unidirectional).
Specifically, Soros settles on the relationship of valuation and the fallacy of market equilibrium. He holds that money values do not simply mirror the state of an underlying reality, because valuation is a positive act that can itself impact the course of events. (As an aside, this maps to a deeper philosophical and possible quantum physical truth that the very act of paying attention is a moral act in and of itself, because attention changes both the attender and the object of their attention, a la McGilchrist).
Most common ‘mispricing’ involves the use of leverage (debt + equity leverage), but we’ll get back to this later.
Breaking down a market
A market is made up of many participants, whose views are bound to differ. Since a market is affected by (and essentially composed of) thinking minds, it is affected by a reflexive element, and thus an unstable system.
A market is a pool of participants taking different sides of a hypothesis about the future of the market. That hypothesis is submitted to the test of time. Hypotheses that survive the test are reinforced, and their backers are rewarded. The hypothesis is never fully validated or invalidated by the course of actual events - the only response is the future state of the market itself. Crucially, the markets don’t tell why a hypothesis was right or wrong (trial and error).
At any point in time, most of the individual biases of participants cancel each other out, leaving a “prevailing bias:'' the weighted average of the hypothesis of the participants. A positive bias raises the stock price, a negative one lowers it. The prevailing bias is thus observable, all other things being equal. Since all a market does is reflect the prevailing bias, a market cannot be said to be always right, nor always wrong.
To reiterate, operational success in markets (which translates to financial success) is then achieved by anticipating changes in the prevailing bias, i.e. the future state of the market.
Economic Theory and Equilibria
Economics is broken, in that it attempts to apply “stable science” to an unstable system.
Economy theory is devoted to the study of an axiomatic system: equilibrium position. The concept of equilibrium is undoubtedly very useful, but it is based on three assumptions: perfect knowledge, homogenous and divisible products, and a large enough number of participants that no single participant can influence the market price.
For long periods, financial markets follow the random walks mandated by the efficient market theory. In near-from-equilibrium conditions, corrective mechanisms prevent reality and perceptions from drifting too far apart. However, due to the double-feedback mechanism of reflexivity, there sometimes arise intermittent far-from-equilibrium conditions. In such cases, there is no tendency for perceptions and reality to come close together without a change of regime (Popper’s ‘paradigm shift’). This means that economic theory is intermittently false.
The Fatal Flaw of Supply-Demand Curves
The flaw in economic theory lies in one of its core assumptions: perfect knowledge is impossible for situations that involve human participants, since participants would need perfect knowledge of what other participants were thinking and thus going to do. Economists sidestep this by treating demand and supply curves as a given, produced by psychologists (demand) and engineers or management scientists (supply). Thus perfect knowledge is conditioned on the assumption that participants know all preferences and alternatives.
Economy claims to follow the scientific method, which makes it dangerously tempting. Natural science studies events that consist of a sequence of facts. A chain of causation leads directly from fact to fact, or in the case of economics, between information, human decisions, and economic outcomes. This is captured in the static supply-demand curve, and would be true in a stable system.
However, human decision-making (based on an imperfect understanding of the situation) creates a reflexive relationship that precludes this field of study from a scientific method. The sequence is not from fact to fact, but fact to perception and perception to fact. Thus modern economics and many of its adherents treat certain markets as stable systems, when they are not.
Paradigm shifters
Instead of a static supply-demand curve, we can use Soros’s splitting of functions that continually and dynamically feed into each other:
- The cognitive (passive) function, in which participants attempt to understand the situation
- The participating (active) function, in which the thinking of the participants impacts the real world.
We can classify all events into two categories vis-a-vis a market: humdrum, everyday events that are correctly anticipated by the participants and do not provoke a change in their perceptions, and unique, paradigm-shifting events that do (which Soros confusingly calls ‘historical’).
These events mean that information to human decision to human action is not a constant equation, and thus that the supply-demand curve is not constant over time. Information continually flows through the cognitive (passive) function to the participating (active) function, but occasionally information changes the cognitive function by shifting the paradigm of the participants. Equilibrium analysis reductively equates the participating (active) function as equivalent to the supply-demand curve, and assumes away the cognitive (passive) function through the sleight of hand of “perfect knowledge.”
There is another variable that is not captured in the supply-demand equation, which is the participants’ expectations about future prices (not just the price of today), which can sometimes be expressed outwardly in futures markets. Beyond the current price, it is these expectations that drive buy and sell decisions - and it is current buy and sell decisions that furthermore drive future prices.
(This doesn’t fully apply to commodities, where supply also depends on production, and demand on consumption. But still, the price that determines production and consumption is not necessarily the present price).
Those reared on classical economics struggle with the idea that events in the marketplace continually affect the shape of the supply-demand curve. The demand and supply curves are supposed to determine the market price. If they were themselves subject to market influences, prices would cease to be uniquely determined. Instead of equilibrium, we would be left with fluctuating prices. But of course, history has only ever given us fluctuating prices.
Fundamentals Aren’t Enough
Some argue that such fluctuations are short-term, and eventually ruled by fundamental value over a long enough time horizon. There is a one-way connection between companies and the prices of their traded stock. The fortunes of the company determines its relative value against other companies and, despite temporary delay due to speculative excess, eventually things smooth out. As such, stock market developments are unable to affect the fortunes of the companies themselves.
In actuality, we are faced with the ever present fluctuating prices and, in some cases, self-reinforcing trends. We must consider that speculation may itself alter the supposedly fundamental conditions of supply-demand. This opens the door to the possibility that free markets will always be inherently unstable, but we’ll get back to that in a bit.
In the meantime, we’ll focus on the fact that financial markets can actively impact underlying fundamentals. Therefore, market prices always distort underlying fundamentals.
Applied to the stock market
The stock market market, as described earlier, is the weighted average of all biases, which produces a numerical output for each listed company: the stock price. Stock prices move to match the prevailing bias.
The prevailing bias changes as paradigm-shifting events happen (usually external events). But a change in stock price can affect external happenings, which over time can cause a paradigm-shifting event to happen, which then affects the prevailing bias. This is where we see a cycle: the prevailing bias can cause a change in the prevailing bias. So when the price moves to anticipate a future event, it may in fact change the future course of events. This explains why markets can appear to anticipate events (i.e. recessions), while also precipitating them.
Straight from Soros: “I do not accept the proposition that stock prices are a passive reflection of underlying values, nor do I accept the proposition that they correspond to the underlying value. I contend that market valuations are always distorted: moreover - and a crucial departure from equilibrium theory - the distortions can affect the underlying values.” There is then a reflexive relationship between flawed perceptions and the actual course of events.
Because there are no statically independent and dependent variables, there is no tendency towards equilibrium. Instead, the sequence of events is a process of historical change in which the variables are always changing. Its typical pattern is the boom and bust, where the three variables (price, prevailing bias, underlying trend) can reinforce each other first in one direction, then the other.
This omission is more glaring in the stock market than in other markets. Stock market valuations have a direct way of influencing underlying values: through the issuance and repurchase of shares and options and through corporate transactions of all kinds: mergers, acquisitions, going public, going private, etc. More subtly also through credit rating, consumer acceptance, management credibility, etc.
Reflexivity behind the boom/bust (self-reinforcing trend)
The reflexive relationship of the prevailing bias and the underlying trend can create a certain pattern. First, the prevailing bias reinforces a prevailing trend. The reinforcement accelerates the trend (self-reinforcing). As it extends, it becomes increasingly vulnerable because the fundamentals (ex: trade, interest payments) move against the trend, in accordance with classical analysis, such that the trend becomes increasingly dependent on the prevailing bias. When a bias is reinforced, it widens the gap between expectations and the future stock prices. Eventually, a turning point is reached and, in a full-fledged sequence, a self-reinforcing process starts operating in the opposite direction.
Typically, a self-reinforcing process undergoes small corrections in the early stages. If it survives them, the bias tends to be reinforced, and is less easily shaken. When the process is advanced, corrections become scarcer and the danger of a climactic reversal greater.
The Boom-Bust Sequence:
- the unrecognized trend
- the beginning of a self-reinforcing process
- the successful test
- the growing conviction, result in gin a widening divergence between reality and expectations
- the flaw in perceptions
- the climax
- a self-reinforcing process in the opposite direction
It is ingrained that stock prices are the passive reflection of an underlying reality, and not an active ingredient in a historical process (denial of the reflexivity). However, investors do intuitively recognize and respond to the boom-bust sequence. However, they do so later than someone who is actively looking out for the crucial features shaping the price curve. “This is what has given me my edge.”
Booms can arise whenever there is a two-way connection between values and the act of valuation. The act of valuation takes many forms. In the stock market, it is equity that is valued; in banking, it is collateral.
Participants are not in a position to prevent a boom from developing even if they recognize that it is bound to lead to a bust. The best that a participant can do is to cease to be a participant at the right time. But that may not always be possible. For instance, in a system of fluctuating exchange rates, holders of financial assets have an existential choice in deciding what currency to hold: they cannot avoid holding some currency, except by buying options.
Part III: What This Means
The Credit and Regulatory Cycle
There seems to be a special affinity between reflexivity and credit. That is hardly surprising: credit depends on expectations; expectation involves bias; hence credit is one of the main avenues that permit bias to play a causal role in the course of events.
Loans are based on the lender’s estimation of the borrower’s ability to service his debt. The valuation of the collateral is supposed to be independent of the act of lending; but in actual fact the act of lending can affect the value of the collateral. This is true of the individual case and of the economy as a whole. Credit expansion stimulates the economy and enhances collateral values; the repayment or contraction of credit has a depressing influence both on the economy and on the valuation of the collateral.
The reflexive interaction between the act of lending and collateral values has led me to postulate a pattern in which a period of gradual, slowly accelerating credit expansion is followed by a short period of credit contraction - the classic sequence of boom and bust. The bust is compressed in time because the attempt to liquidate loans cases a sudden implosion of collateral values.
By contrast, when credit is not an essential ingredient in a reflexive process, the pattern tends to be more symmetrical.
Asymmetry of the credit boom-bust
Booms and busts are not symmetrical because, at the inception of a boom, both the volume of credit and the value of the collateral are at a minimum; at the time of the bust, both are at a maximum. But there is another factor at play. The liquidation of loans takes time; the faster it has to be accomplished, the greater the value of the collateral. In a bust, the reflexive interaction between loans and collateral becomes compressed within a very short time frame and the consequences can be catastrophic. It is the sudden liquidation of accumulated positions that gives a bust such a different shape from the preceding boom.
The role of central banks
The institution of central banking has evolved in a continuing attempt to prevent sudden, catastrophic contractions in credit.
There is a reflexive relationship between the act of lending and the value of collateral for those loans. Net new lending is initially a stimulant that enhances the borrowers’ ability to service their debt. As the amount of debt outstanding grows, an increasing portion of new lending goes to service outstanding debt and credit has to grow exponentially to maintain its stimulating effect. Eventually, growth of credit has to slow down with a negative effect on collateral values. If the collateral has been fully utilized, the decline in collateral values precipitates further liquidation of credit, giving rise to the typical boom/bust sequence.
Regulation as Lacking
Regulations are generally designed to prevent the last mishap, not the next one. Regulation always lags behind events. By the time regulators have caught up with excesses, the corrective action they insist on tends to exacerbate the situation in the opposite direction. Just as freely floating currencies tend to fluctuate between over- and under-valuation, so market economies tend to fluctuate between over- and under-regulation.
The history of central banking is a history of crises followed by institutional reform. It is truly surprising that the lessons of the international debt crisis have still not been learned.
In the US, to achieve any lasting solutions, the whole machinery of the administration and Congress needs to be involved. That means that needed reforms are rarely enacted in time.
There is a basic imbalance in deregulating deposit-taking institutions and guaranteeing depositors against loss. The guarantee enables financial institutions to attract additional deposits at will, and deregulation gives them wide latitude in putting those deposits to use. The combination of the two is an invitation to unrestrained credit expansion. The problem has been inherent in the system of the Federal deposit insurance since its inception but at the time the FDIC was founded banks were strictly regulated.
The Federal Reserve was forced to expand its role as lender of last resort and guarantee all depositors against whatever the size of their deposits. This removed the last vestige of the discipline that depositors are supposed to impose on banks.
Free Markets are Unstable
Now, as promised, we return to Soros’s hottest take: the fatal flaw of a free market system is its inherent instability. However, one cannot ignore its adaptable nature and ability to survive.
Markets clearly tend towards excesses, not equilibrium. But if markets don’t tend towards equilibrium, one loses the argument of optimum allocation of resources. The reflexivity inherent to markets will always create instability - and there is no limit to how far away both perceptions and events can move away from whatever might be considered ‘equilibrium.’
Instability is not inherently harmful. We might call it ‘dynamic adjustment,’ which sounds much more benign.
Every form of previous society was found wanting in something that could only be found in its opposite. Tribal society had traditional thinking, totalitarian society had dogma, and Open Society has critical mode of thought. But the totalitarian lacked freedom, and Open Society lacks stability. In matters of value, one must start from the position that they are rooted in fantasy rather than reality. As a consequence, every set of values has a flaw in it.
Since financial markets are inherently unstable, stability can only be maintained if it is the objective of public policy. But there is always the temptation that the government may help prevent the crash. If it truly did, it would be a historic first, the dawn of a new era in which financial markets are manipulated for the benefit of the public good.
Our economy will keep going to the brink, then recoiling. The danger of a recession prompts remedial action. But we cannot be sure that the danger merely increases with our confidence - through a series of self-defeating prophecies. Current models are based on the misconception that markets only foreshadow events, without shaping them. So it will keep appearing as if financial markets get too excited in anticipating events that seem quite harmless in retrospect. But financial crashes will then occur only when they are unexpected (for the most part - many events happen even when they are anticipated).
The more people believe that markets are always right, the less likely they are to adopt corrective processes, increasing instability. Therefore, the more the theory of efficient markets is believed, the less efficient the markets become.
Free Floating Exchange Rates
Specifically, Soros hypothesized that freely floating exchange rates are inherently unstable; moreover, the instability is cumulative so that the eventual breakdown of a freely floating exchange rate system is virtually assured. In a system of freely fluctuating exchange rates, reflexivity constitutes the rule.
The traditional view of the currency market is that it tends toward equilibrium. An overvalued exchange rate encourages imports and discourages exports until equilibrium is reestablished. However, exchange rates have found a way of influencing the fundamentals. For instance, a strong exchange rate discourages inflation; wages remain stable and the price of imports falls. The fact is that the relationship between the domestic inflation rate and the international exchange rate is not unidirectional but circular. Changes in one may precede changes in the other, but because they are mutually reinforcing, there is no unidirectional causal chain.
Exchange rates are determined by the demand and supply of currencies. We can group the various factors that constitute demand and supply under three headings: trade, non speculative capital transactions, and speculative capital transactions. As a general rule: speculative capital is attracted by rising exchange rates and rising interest rates.
The cycle of fluctuating exchange rates
Each self-reinforcing circle is unique, but here is a universally valid generalization about freely fluctuating exchange rates:
- The relative importance of speculative transactions tends to increase during the lifetime of a self-reinforcing trend
- the prevailing bias is a trend-following one and the longer the trend persists, the stronger the bias becomes
- Once a trend is established, it tends to persist and to run its full course. When the turn finally comes, it tends to set into motion a self-reinforcing process in the opposite direction. In other words, currencies tend to move in large waves, each move lasts several years.
The longer a benign circle lasts, the more attractive it is to hold financial assets in the appreciating currency and the more important the exchange rate becomes in calculating total return. Those who are inclined to fight the trend are progressively eliminated and in the end only trend followers survive as active participants.
The greater the relative importance of speculation, the more unstable the system becomes: the total rate of return can flip-flop with every change in the prevailing bias.
(Another tentative generalization: when a long-term trend loses its momentum, short-term volatility tends to rise. It is easy to see why: the trend-following crowd is disoriented.)
If these generalizations are valid, the eventual demise of a system of freely fluctuating exchange rates is inevitable. Fluctuations become so wild that either the system has to be modified by some kind of government intervention or it is bound to break down. Currency markets thus provide the best support for the contention that financial markets are inherently unstable.
Part IV: Final Thoughts
As a quick recap of what we’ve covered:
Overview:
In Part I, we define stable and unstable systems, and how the combination of uncertainty + reflexivity can render a system unstable.
In Part II, we apply this knowledge to the world of finance, first by creating a model of a market, then by showing where the classical model of equilibrium in supply-demand curves breaks down, and ending with an examination of how reflexivity can create boom-busts cycles in the stock market and beyond.
In Part III, we begin to explore some of the larger conclusions that one might draw about the current economic model: that free markets and floating exchange rates might be inherently unstable, and that regulation is unable to prevent this in its current form, although it continually thinks it can.
Part IV ends with some closing thoughts and remaining questions.
Some closing thoughts:
Soros’s books and lectures have held a disproportionate amount of my mind-space over the past few months. That doesn’t mean I agree with all of Soros’s more philosophical conclusions. I still don’t understand why he is the target of so many conspiracy theories (although I suppose this isn’t uncommon for the Jewish financiers of history), I wouldn’t say I always align with his philanthropic goals or visions for an Open Society.
I see the world in terms of ideas, and each idea is like a seed or the mouth of a river. Some seeds are ‘bigger’ than others, not in size but in the size of what can emerge from them (consider the mustard seed…). In Popper’s reflexivity and Soros’s application of it to finance, I feel I have stumbled upon a fairly interesting seed that keeps showing new branches in places I don’t expect.
Even now, I can feel some of the branches that will probably keep developing and overlapping with other seeds. There is the exploration of other boom-bust cycles through the lens of unseen reflexive relationships, either historically or still ongoing (tech dot com bubble or tech bubble of 2010s. There is also the description of financial markets and history as dynamical systems and state-space (see Van Gelder on computation and cognition). In both of these are three of the core fascinations I cannot get away from: boom-bust cycles, the value creation (or destruction) of good (or bad) debt, and the chaos that emerges in complex adaptive systems.
For more on those topics, check back here as I’ll probably be adding some more essays. And if you’re intrigued by Soros’s ideas, I suggest you go straight to the source. I’d recommend starting with his lectures (I and II cover reflexivity as a concept and its application in finance).