I posted a new paper under the philosophy section of my blog in response to George Soros’s new book - The New Paradigm for Financial Markets: the Credit Crisis of 2008 and What it Means. I reproduce it below as a blog post.
Reflecting on Reflexivity
In 1962, a (now famous) philosopher of science named Thomas Kuhn drew much attention for his controversial claim that, when scientific theories of the world change, the world itself changes, too.
Since 1987, financier and philanthropist (and self-admitted amateur philosopher) George Soros has been making a similar claim, but with a focus not on the natural sciences that Kuhn spoke of but instead the social sciences – specifically, economics.
Soros’s dictum – the theory of reflexivity – asks that we set apart sciences which have thinking and non-thinking participants and those which have only thinking participants. A chemist studying a molecule can make independently valid statements about the molecule that may qualify as knowledge. An economist studying the behavior of crowds to make judgments about supply and demand can do no such thing: his theories, in the Kuhnian sense, will impact the world of his participants and thus put knowledge out of reach.
If Soros’ theory is true, it cuts a wide swath. It implies at once a need for a new ethics, a new epistemology and – of course – the need to tear up our classical economics textbooks.
Yet, as Soros confesses in his latest treatise on the subject - The New Paradigm for Financial Markets: the Credit Crisis of 2008 and What it Means – published a year ago in response to the financial crisis, philosophers, economists and financiers alike have failed to give his theory serious consideration.
Naturally, this begs the question: why? Is Soros wrong or is the establishment wrong?
When was an undergrad at Penn, I never thought there would be much synergy from being a philosophy major and a business student. Here, I found my challenge.
What follows is a critique of Soros’s philosophy that I hope can shed some light on where his theory of reflexivity excels and where it falters.
My main claim is very simple: Soros commits the same mistake as Kuhn did in his magnum opus The Structure of Scientific Revolutions: he carries a good idea to an unjustifiable conclusion.
Just as one can read Kuhn and be convinced by his account of how new scientific theories become prevailing paradigms, one can ready Soros and be convinced about his account of how markets influence and are influenced by our attempts to understand them. But when he concludes from this that market participants “cannot base their decisions on knowledge,” Soros makes the same mistake as Kuhn did when he said that our scientific theories change the physical world: he carries his theory a step too far.
Part I of this post outlines Soros’s theory in greater detail to make sure that anyone who has not read it can be on equal footing. Part II outlines my criticism of his theory. Part III, in good faith, contains some advice for how he can strengthen his theory.
Part I: Soros’s theory explained
First, some background: Soros’s theory of reflexivity is based on another theory – the correspondence theory of truth – without which his whole paradigm could not work.
In its simplest form, correspondence holds that a proposition is true when it corresponds to reality, and false when it does not (cocktail party tidbit for you: this is what philosophers term “altheic realism” – the notion that truth hinges not on us, but on the world in which we live).
Soros argues that correspondence works well in the world of natural science but breaks down when applied to social science:
I contend that social events have a different structure from natural phenomena. In natural phenomena there is a causal chain that links one set of facts directly with the next. In human affairs the course of events is more complicated. Not only are facts involved but also the participants’ views and the interplay between them enter into the causal chain (New Paradigm, pg. 7 – emphasis added).
Another way to view this is in mathematical terms: a function is uniquely determined (and therefore true) if it has one independent variable which determines the value of the dependent variable. In natural science, there is just one variable – the natural world – but in social science, there are two variables: our understanding of social phenomena and the impact that our actions have on them.
Soros terms the first variable the “cognitive function” and the second variable the “manipulative function”. In other words, we seek to understand social phenomena but, as a result of being participants in them, we can simultaneously change, or manipulate them. And as we manipulate them, our understanding of them changes. As our understanding changes, so does our interaction with them. So changes in one function are reflected in the other, and vice versa. Soros calls this “reflexivity” and hence the name of the theory
“Reflexive situations are characterized by a lack of correspondence between the participants and the state of affairs,” Soros writes (emphasis his). Participants are always manipulating the state of affairs, so as soon as they make a statement about it that purports to be true, their actions can render the statement false.
Soros provides the following example to make this distinction clear:
Consider a statement about the objective aspect: “It is raining.” That is either true or false; it is not reflexive. But take a statement like: “Your are my enemy.” That may be true or false, depending on how you react to it. That is reflexive (New Paradigm, pg. 28).
Reflexivity is best studied and demonstrated in the financial markets, Soros argues, because market developments are not independently given by natural phenomena but instead are driven by human expectations. Classical economics, he argues, ignores this insight:
Demand and supply curves are presented in textbooks as though they were grounded in empirical evidence. But there is scant evidence for independently given demand and supply curves. Anyone who trades in markets where prices are continuously changing knows that participants are very much influenced by market developments. Rising pries attract buyers and vice versa. How could self-reinforcing trends [bubbles] persist if supply and demand curves were independent of market prices? (New Paradigm, pg. 55).
Our expectations (cognitive function) are reflected in the market when we buy or sell (manipulative function), which in turn impacts our expectations (cognitive function), and so on, ad infinitum. One reinforces the other. This, Soros argues, is how market bubbles are born.
More importantly, he argues that reflexivity “introduces an element of contingency or uncertainty into the course of events, and it prevents the participants’ views from qualifying as knowledge” (New Paradigm, pg. 4). In other words, market decision markers cannot base their decisions on knowledge.
Part II: A critique of reflexivity
Given this background, there are two possible readings of Soros’s philosophy.
One, I would argue, is uncharitable and the other is charitable (in philosophy, these terms typically mean taking the philosopher’s argument at its most extreme meaning versus giving it the benefit of the doubt).
Let’s start with the uncharitable interpretation:
In social situations, reflexivity produces uncertainty and thus prevents our views from qualifying as knowledge. Social science can therefore only succeed in producing opinion, whereas natural science can succeed in producing facts about the world.
This implies that we cannot make any claim to knowledge from observing our behavior and creating Newtonian rules such as the law of supply and demand. And so much, if not all, of economics, sociology, psychology and other social sciences are futile exercises.
Carry this line of thought further and soon you end up being the caveman in Plato’s Allegory of the Cave, mistaking shadows on a wall for a true account of the world. As with the allegory, where concealed persons chatting and carrying various objects before a fire result in the false impression that they are the real world, studying social situations inevitably results in a false view of the world in this Sorosian paradigm.
But is that what Soros is really advocating? I think not; one very easy way to disprove this reading is just to assume that it is correct. If Soros’s observations about the human position are true and reflexivity obtains, then social science can indeed produce some sort of knowledge. After all, what is The New Paradigm for Financial Markets if not a work of social science?
The real question is therefore not the black-and-white, “can we as participants in the social world order come to gain knowledge about it?” but rather, “how perfect is that knowledge?”
Ponder that question vis-à-vis his philosophy, and a much more charitable interpretation results:
In social situations, reflexivity produces uncertainty and thus prevents us from having perfect knowledge. Social science, like natural science, can therefore help us better understand the world, but it cannot reveal the ultimate truth about the world we inhabit.
To his credit, I think Soros believes as much. In later parts of The New Paradigm, he seems to soften his stance, arguing that “people are participants, not just observers, and the knowledge they can acquire is not sufficient to guide them in their actions. They cannot base their decision on knowledge alone” (New Paradigm, pg. 26, emphasis mine).
To make up for this shortcoming, Soros argues that we use various techniques – generalizations, similes, metaphors, habits, rituals – to organize information and make decisions. In doing so, though, various biases, emotions and other non-factual elements can enter our reasoning.
Now think back to Soros’s discussion of financial markets and what results is a pretty convincing case against the efficient market hypothesis.
This hypothesis (which Soros loathes) is predicated on the notion of perfect information: stock prices always reflect all relevant information and therefore trade at their fair value.
But apply the notion of reflexivity to the markets and it soon becomes clear that markets don’t always price in everything they should or exclude everything they shouldn’t price in. Like a river that collects sediment along the way toward its mouth, market biases, emotions, rumors and fictions get swept in with actual facts about a stock, all of which impact participants’ views (cognitive function) and impact the market (manipulative function) when they buy/sell.
Market information is thus not perfect (like the dirty river, it is filled with impurities), markets are not efficient, and they certainly do not tend toward equilibrium; rather, they tend toward whatever prevailing biases preoccupy investors any given point (ex. housing prices will continue to rise) until the bubble pops and the market corrects itself.
Viewed this way, I think Soros’s ideas are far more serious than many economists realize. Reflexivity is not an argument against the very viability of economics (in which case it would be understandable why social scientists would want to thumb their noses at him), but rather a call to arms to better understand the fallibility of our existing paradigms, like the efficient markets hypothesis.
It seems, though, that philosophers, economists and social scientists alike have turned a deaf ear to him for the past 20 years, perhaps opting for the uncharitable interpretation argument of his work and dismissing it accordingly.
So with Kuhn. People originally interpreted his book as a crazy argument that we can change the world with our thoughts: Copernicus comes along and all of a sudden the planets realign around the Sun? Surely not. Eventually, a more charitable reading of his work gained consensus and his account of how scientific revolutions occur is now a key consideration in any class on the philosophy of science.
Will the same thing happen to Soros? Only time will tell; but in the meanwhile, he can certainly do some things to make his theory more tenable.
Part III: Some advice
Here are some questions that I believe Soros should answer to strengthen his theory of reflexivity (with the important caveat that I’m not an expert; undergraduate degrees in philosophy and economics only – no fancy stuff).
Questions of interest to philosophers:
- Does reflexivity solely rely on the correspondence theory of truth, or could it co-exist with another theory of truth, such as coherence or identity
- Reflexivity implies that the Cartesian view of the world (separation of body and mind) is wrong, so what is the proper place of the mind in a reflexive world?
- What are the boundaries of reflexivity? The theory in its current form implies that all social situations are reflexive. But what’s a social situation and what is not? (Philosophers are notorious for thought experiments; sooner or later, someone will throw a remote control robot into this discussion).
Questions of interest to economists:
- One of the assumptions of the efficient market hypothesis is that no one market participant is big enough to influence the market on his/her own. Does reflexivity accept the same assumption? (Another way to phrase the question: does someone like Warren Buffet – who can say “the economy has fallen off a cliff” and slice 80 points off the Dow - have more “reflexive” power than some guy with an internet connection and a blog?)
- Are some markets more reflexive than others? If so, what distinguishes a reflexive market from a non-reflexive market – is it the number of participants, the type of security, or what? For example: I can bet on fed fund futures all I want, but at the end of the day, unless I break into an FOMC meeting and force the Fed Governors to raise or lower the fed funds rate, my bets can do little to move the target. But if I own a large stake in a company, become bearish and start selling, my outlook can influence the market much more. So where do we draw the line?
- If markets are indeed reflexive, does that mean that we are doomed to forever go through bubbles and busts and super-bubbles and super-busts, or does reflexivity allow us to forge a cure for this?
One last piece of advice for Soros: pick a discipline, because philosophers and social scientists often don’t see eye to eye. I learned the hard way: I once had to spend a whole summer re-writing a thesis because one of my philosophy professors thought the paper I handed in belonged more in the realm of political science than philosophy.
So as long as reflexivity is rooted in both disciplines, chances are that some philosophers and economists will inevitably choose to ignore it because it’s out of their realm.
Good luck, George.
I had just returned from Poland, where I was visiting my family for Easter. And I had just witnessed the excitement in Poland about the fact that Russian Prime minister would become the first Russian leader to stand in ceremony commemorating the Katyn massacre and that the shadow of this horrible crime would finally disappear.