Recession: The Painful 'Tail Event' We Desperately Need

Story Stream
recent articles

It is amazing what occurs by accident, in the moment in which we least expect something to happen greatness often follows. This is true of nearly every major discovery, though the "answers" were often laid out in the process of trying to peer inside the mythical veneer of whatever the human mind can endeavor to examine. It takes an accident or unintentional deviation to awaken the conscious senses to see and understand what was previously hidden or believed unimportant or erroneous.

Take the example of Edward Lorenz, an MIT professor of meteorology who stumbled upon what has been called, rightly in my mind, the third great scientific advance of the 20th century (along with relativity and quantum mechanics). In the winter of 1961, he was in the middle of running computer simulations trying to do nothing greater than more accurately predict the weather. Dr. Lorenz had been among the first to move into statistical modeling as a means to advance predictability over what was previously believed unattainably beyond comprehension. The first run of his theories produced regularity, an obvious cue that the models were badly calibrated.

There is no regularity in weather, as the natural world is unalterably complex despite all the progress of modern "science." So to see it as a result of computer programming was an obvious rebuke to the particular effort. To get back on track, Lorenz decided to run the program for a longer period of time in an effort to perhaps gain more insight, but to save time he started a second simulation from the middle. That meant a process of transcribing results and figures from the first run to incorporate in the second.

To his astonishment, the second deviated quickly from the path of the first simulation, something that should not have happened. After checking the computer for malfunction he noticed a small error in the initial conditions of the second run - his computer system stored numbers out to six decimal places, but the printout he used to port them from the first run to the next only went to three. Thus, what should have been 0.506127 was input instead as 0.506.

The difference was exceedingly minor, less than one-tenth of one percent, but the change in behavior of his model was enormous.

About two years later, Dr. Lorenz published a paper on the effects of small changes in deterministic systems, Deterministic Nonperiodic Flow. To say that it was not well received would be an enormous understatement, as it acquired exactly three citations during its first decade in print. Essentially, what he was stating was that sensitivity to initial conditions was what caused non-periodic behavior, or the lack of steady repetition as his first simulation had experienced. It was the increase in the number of variables and their close interactions that allowed for an infinite number of possible outcomes.

In terms of weather and meteorology, the implications became quite drastic. Making even small errors on large-scale phenomena would double in magnitude every three days. Smaller scale errors, mostly measurement imprecision, could produce large-scale errors in as little as a single day. Thus, weather forecasting is limited only to the shorter timescales because of what are really "exploding errors." There can be no confidence that enough care and precision has been added to forecast models to allow for such a narrow range of variation as to be useful and meaningful.

The mainstream introduction to this idea was the butterfly effect. Lorenz himself had used the analogy of a seagull causing a storm, but for his seminal 1972 paper, fellow meteorologist Philip Merilees came up with the title, Predictability: Does the Flap of a Butterfly's Wings in Brazil Set Off a Tornado in Texas. And so "chaos theory" was finally given a wide audience for the first time, though few people seem to understand it or its wider implications even to this day.

Of course it was Benoit Mandelbrot who introduced mathematical theories about complex systems and took to trying to better grasp the nature of "error" in forecasting and in complexity. What Mandelbrot understood, standing sharply against the trend in mathematics and statistics of that age, was that there was usefulness in the irregularity of nature. Just as Lorenz saw the dangers of overlooking it, Mandelbrot saw the benefits of defining it, even if only poorly or incompletely.

Modern statisticians, particularly economists, see the world as only smoothed contours. The reason for that is as simple as Dr. Lorenz's transposition error, in that the mathematics breaks down at a certain level of complexity. The modern tools of calculus in the statistical discipline assume smoothness in everything because there is no method to extract every little and possible detail, taking instead small leaps of assumption and leaving much behind in doing so.

Mandelbrot's initial focus was on coastlines and geography, where he noticed a high degree of self-similarity - that is, patterns that repeated regardless of scale. Where the calculus of probability-based estimations removed the jaggedness of nature in order to preserve the math, Mandelbrot realized how much information was being discarded for the sake of making a calculation feasible. Orthodox statisticians make these assumptions because they have withdrawn to the comfortable confines of normal distributions, which assume such irregularity as random variation.

But like meteorology, in economics and especially finance non-linearity puts particular emphasis on error terms. As with the butterfly, a small error for whatever reason can quickly get out of hand and render forecasts unbelievably wrong. Nowhere was that more evident than in 2007 and 2008, where the best and most sophisticated and elegant economic models were totally outclassed by the real world's complexity. What was believed impossible, really impossible since some conditions were modeled then as being ten-standard deviations or more, actually happened.

There is no way to address that deficiency without understanding the idea that the current strain of economic theory and statistics intentionally discards an enormous amount of relevant data, rationalizing the process of doing so as nothing more than randomness. But what in finance or economics is actually random? To believe in such mythology is to deny human reality, because what happens today has significant bearing on what happens tomorrow.

The most recent example of this, in a larger scale setting, was the first quarter's GDP estimate. Rather than embrace the information that might be relevant to what is actually taking place, the entire orthodox economics profession is busy trying to convince everyone that there was nothing useful in that result. It was, as has been repeated over and over, an aberration of no significant value; an error term to be denied a full place in the analysis of where the economy might actually be headed.

A GDP figure of nearly minus three percent is decidedly rare, however, so unusual that it has never taken place outside of a recession. But that is precisely what we are supposed to ignore when being counselled to take no notice of it. Actually, counsel is too slight and too soft of a word, as what is really occurring is nothing short of a demand. The surety at which the orthodox profession, especially those of monetary disposition, exercises such confidence about forecasting is very much descriptive of ideology rather than science.

It is also entirely contradictory - it seems that the degree and frequency in which economists' forecasts end up missing is proportional only to the certainty and conviction with which their new predictions are issued. It is not so much observational as it is, again, a command. To achieve it, policymakers have gone to extraordinary lengths, some perhaps beyond legality, to "ensure" their success. That is command, as it allows no other or contrary position.

In George Orwell's book, 1984, the animals on the farm were taught about equality in the same manner. Prediction was never about accuracy, but rather control over the system, the farm in which they inhabited. The idea of equality contributed to that system as an ideology on which to base and embrace the theory of what was really feudal function. Comrade Napoleon was equal like all animals there, but "more equal" in every meaningful sense. On one occasion, Napoleon's lieutenants explained,

"No one believes more firmly than Comrade Napoleon that all animals are equal. He would be only too happy to let you make your decisions for yourselves. But sometimes you might make the wrong decisions, comrades, and then where should we be?"

What Orwell was really talking about, for my reading of it, was the eternal struggle between freedom and order. The status quo is a very powerful thickener of intentions and actions. In purely economic terms, the Federal Reserve, like all modern central banks, is dedicated to order above all else. The governing premise which birthed the institution just over a century ago, currency elasticity, is nothing more than a tool to stamp out "tail risks", to push and prod the banking system into the fat part of the bell curve. In treating finance as a standard normal distribution, the tails are degraded into undesirability, even dangerous to the point of actively disallowing free thought about them.

You can see exactly that in the actions of quantitative easing here, all the massive intrusions from the Bank of Japan, and even Mario Draghi's lack of action through nothing more than promises. What all these policies are directed toward is nothing more than the idea that order takes full precedence over free expression of opinion - in this case financial markets and free discovery of prices.

The philosophic framework that asserts validity for that effort is that there is nothing good to come out of allowing "the downside." Whether that is panic or recession (or depression), the orthodox statisticians of the FOMC membership are dedicated to this idea, that "tail risks" have no meaning and are nothing more than random errors. By smoothing out the paradigm of projection and interpretation, the statistics (mostly linear) judge that free expression is secondary to order in the hierarchy of political intrusion into the economy.

That view is contrary to everything we know about nature, and there is no way to describe even human systems as anything but. Just as Dr. Lorenz's discovery about error terms was itself a "tail event" (and self-similar), there is much to be lost in trying to enforce order out of, forgive the usage here, chaos. Freedom is messy, and a full part of the advance of economic society and the wealth of nations is due to that messiness, or what are nothing more than tail events. These are not random errors, but rather the blooming of what was previously unpredictable, a flourishing of ingenuity without the constant intrusion of those "more equal" making sure that their views receive priority.

In the book Superfreakonomics, the authors recount dire predictions, from "experts" no less, about major cities then almost-drowning in horse manure. The density of people and horse had created a massive waste problem, so much so that apparently in 1894 the Times of London estimated by 1950 the entire city would be buried nine feet deep in "it." In New York City, it was thought that by 1930 horse apples would rise three stories all over Manhattan.

None of that ever happened because of the automobile, which, in the context of 19th century conveyance, would have been a "tail event." It would have made no sense, in broad economic terms, to try to maintain order over chaos as cars displaced horses, which we are fortunate took place so deliberately over time so as to be impenetrable to the affected powers that no doubt would have had it occurred to them.

Yet today, everything about such dynamism is drowned by fear. Sure, everyone wants the benefits of all the innovation that can take place, and especially the problems that are solved in doing so (no manure mountains ever came to pass), but we are now taught, scolded and demanded to countenance none of the messiness in which such dynamic opportunities are brought forward. Recession and financial rebalancing are exactly those kinds of events that release efficiency and propel "tail events" toward broad acceptance. Old firms die to make way for those new entrants with new ideas, a messy transition that is often painful but necessary. The march of actual employment, the most valuable kind, is created not by small businesses generically, but by exactly these kind of disruptive firms seeking not just to compete but to displace the old order.

That is as true in finance as the real economy, yet everything that central banks do is trained upon stamping out free expression in the name of order. We may not like the messy nature of free markets, but to reduce and restrain them as we have we might as well take account of the costs of doing so.

In thinking about butterflies and tornados, Edward Lorenz in his 1972 paper wrote, "If the flap of a butterfly's wings can be instrumental in generating a tornado, it can equally well be instrumental in preventing a tornado." That is what the modern central bank believes, but as Lorenz went on to demonstrate it is impossible to ever know which one is actually at fault. How can you possibly know ahead of time, no matter how elegant your math and how precise you think your instruments, which butterfly shall unleash so much destruction and which might cause enough breeze in a far flung locale as to accidentally change the terms of an experiment enough to allow accidental discovery of historic proportion?

The interdependent chains of cause and effect are far too complex to disentangle, yet that is exactly what we are led to believe can be done. The only way to maintain a consistency of that belief is to ignore the deep canyons and irregularities of nature, to deny the beneficial effects of dynamic change, messy and violent though it may be, and assume such power in narrow concentration. The power to pick winners and losers in the policy and process of redistribution is to say that all butterflies are dangerous and must be extinguished for the sake of an ordered world that can't possibly occur outside the math. Nature will not long be shoved into a single standard deviation; it will break out eventually and visit its own means of retribution.


Jeffrey Snider is the Chief Investment Strategist of Alhambra Investment Partners, a registered investment advisor. 

Show commentsHide Comments

Related Articles