A NonTechnical Introduction To Benoit Mandelbrot’s MultiFractal View of Financial Markets Verses Modern Financial Models
By Clifford J. Short III, AAMS AWMA CPM
Scope
Accurately measuring and predicting risk and investment return has been the elusive ‘holy grail’ of financial market analysts and technicians since the beginning of capital markets. From diverse backgrounds such as astronomy, physics, and applied mathematics, to name a few, ‘students’ of the financial markets have developed and modified models to comprise the analytical tools used today. Based deeply in normal distribution, rational markets and rational participants, continuous probability, and smooth price changes, Modern Portfolio Theory is the foundation of today’s asset management and market rationale.
The current standard financial models for investing and market analysis suggest utilizing variance measurements and betas of different securities and classifying investment portfolios by their probability of risk we can build an “efficient” portfolio to target a specific return with a desired level of risk. Further, price changes are statistically independent and normally distributed.
To the contrary, Benoit Mandelbrot, the eightyyearold Polishborn mathematician awarded the Japan Prize in Science and Technology and the Wolf Prize in Physics, sees the markets and their movements much differently. The ‘inventor’ of fractal geometrya mathematic discipline that attempts to measure and explain “the roughness” of observationssubmits market prices do not follow a bell curve (normal distribution) and prices are very much dependent and have ‘memories’.
The Basis of Modern Financial Analysis
To start, the Efficient Market Hypothesis in an efficient market, all relevant information is already (or will adjust to new information quickly) priced into a security todayis the argument for Modern Portfolio Theory. Simply, equilibrium is current or corrects itself almost immediately to information, preventing one from consistently outperforming the overall market nullifying the idea of Alpha (valueadded long term active security selection). Today’s modern risk measurements (Beta, Sharpe, Standard Deviation, CAPM, Arbitrage Pricing Theory) derive their formulae from this assumption.
Developed by Harry Markowitz in the 1950’s, Modern Portfolio Theory is derived from Brownian Motion Theory. This physics theory was ‘discovered’ by Robert Brown, a Scottish botanist, by observing under a microscope the “erratic way that tiny pollen grains jiggled in a sample of water.” After comprehensive study, he discovered this erratic motion over time developed into a defined patternlater defined as normal distribution. From this came the universally applicable linear measurement the ‘bell curve’ and the observation that distribution will revert to the mean over a period of time. Subsequently, in 1900 a French mathematician Louis Bachelier began his efforts in the new field of probability theory and applying Brownian Motion Theory in financeFrench government bond prices. His approach foundation “Random Walk”postulated prices go up and down with equal probability. Standard deviation was used to explain and quantify the scatter of data outside the mean. Hence, with the ‘bell curve’ the “abundant small changes cluster in the center of the bell while the infrequent big changes (outliers) dot the edges”. Bachelier’s key model hinged on price changes being statistically independent and normally distributed.
What is Risk?
Risk, in the financial markets, is viewed as price volatility and the possible loss of capital. The two types of risk pertaining to the financial markets are systematic and unsystematic. Systematic refers overall market risk, interest rate risk, exchange rate riskmacro event risks that effect the entire market with no way to diversify out the risk. Unsystematic risk (diversifiable risk) points to securityspecific riskrisks the specific company or entity accommodatee.g. liquidity risk, default risk, poor management.
A Sampling of Risk Measurements of Modern Financial Analysis

Probably the most widely recognized and used measurement for a security is Beta. Beta measures a security’s risk relative to the overall market based on past volatility (linear regression). For instance, a Beta of 1 is ‘perfectly’ correlated to the overall marketwill do no better or worse relative to the marketwhile a Beta of less than 1 displays less volatility to the market and a Beta of more than 1 will be much more volatile as the number increases above 1.

Standard Deviation is the mathematical apparatus used to measure the variability of a security’s return around its average’s (mean) total risk return “A simple mathematical yardstick for measuring the scatter of data”. Utilizing the standard deviation model and graphing the observations, a bell shape will occur. Again, “small clusters in the center and infrequent big changes at the edges”. As the German mathematician Carl Friedrich Gauss imposed , the bell shape is ‘normal’ and any other shape is an anomaly. Subsequently, standard deviation considers volatility above the mean just as harmful as volatility below the mean.

The Sharpe Ratio, developed by William Sharpe, to become known as the Capital Asset Pricing Model (CAPM) takes Markowitz’s “cumbersome portfolio calculations” and reduces them to just a handful. CAPM basically sets ‘a hurdle rate’ for a security’s expected return. Simply, the value of an asset is based on how it functions within a market portfolio and how its risk relates to the market’s risk. The returns are based on the relativity to returns of a riskless asset (usually identified as U.S. government Tbills). The return above the riskless security is labeled excess return.

A broader, more comprehensive model has been developed, again, rooted in the original thesis of normal distribution. The Arbitrage Pricing Theory (APT) uses not one factor, like Beta, but utilizes multiple factors. In fact, APT uses as many factors as the analyst desires: Beta market/book effect, Beta price/earnings effect, Beta state of the economy, and Beta “for any element that could affect a security’s price”. Essentially, APT dictates that the return of a risky asset is a linear function of a number of macro economic and financial factors. The value of the return responds to the changing values of the factors.

And probably the most celebrated formula in the last thirty years: The groundbreaking BlackScholes model for pricing options. Devised by the Nobel Prize winning duo Fischer Black and Merton Scholes in the late 1960’s and in use by 1973, this model is based on the thesis “you do not need to know what the stock price will finally be when the option expires, but what the traders themselves know, the terms of the option and how volatile the stock is.” They followed Bachelier, Markowitz, and Sharpe on assuming a stock’s risk and volatility can be gauged by the ‘bell curve’ standard. It asserts volatility will remain the same through the life of the option, prices do not ‘jump’. Still widely used by the ‘street’ (with acknowledgement of its limitationsespecially in pricing exotic options and companyoffered options which are both privately arranged transactions), but with much less blind faith.
In review, our standard financial models which risk measurements are predicated on the assumption returns are normally distributed, price changes are statistically independent, and the markets and its participants are rational. Additionally, utilizing measurements of variance and betas of different securities and classifying investment portfolios by their probability of risk, one can build an ‘efficient’ portfolio to target a specific return with a desired level of risk.
A Sampling of Outcomes Based on Modern Financial Analysis
Mental note: “small clusters in the center and infrequent big changes on the edges” (?)

From 1916 to 2003 there should have been 58 days when the Dow Jones Index Average moved more than 3.4%. There were 1,001

There should have been 6 days the Dow swung beyond 4.5%. There were 366.

Dow swings of more than 7% should have come once every 300,000 years. In the 20^{th} century there were 48 days of swings exceeding 7%.

In August 1998 there were 3 trading days where the Dow dropped 3.5%, 4.4%, and 6.8%. According to our current standard models the 6.8% drop should only happen 1 in 20 million. “If you traded every day for 100,000 years, you would not have expected that to happen once. The odds of having those 3 declines in one month: 1 in 500 billion.

In 1997 the Dow cratered 7.7% in one day. The odds: 1 in 50 billion.

In July of 2002 the Dow declined sharply three time within 7 trading days. The odds: 1 in 4 trillion.

On October 19, 1987 (“the worst day of trading in over 100 years”) the Dow fell 29%. The odds; 1 in 10 to the 50^{th} power “a number outside the scale of nature”.
Benoit Mandelbrot and His MultiFractal View
Multifractal geometry is the study and measurement of the irregular, convoluted or ‘roughness’ in our environment think the coast line of any water lined continent viewed from a thousand feet, or a Jackson Pollack painting, or the chart of a market index over the past five years. Utilizing dimension as a tool of measurement, rather than an ‘inherent property’, fractal mathematics ‘is about spotting repeating patterns, quantifying them, and manipulating them ‘a tool of both analysis and synthesis’. Subsequently, fractals are inherently visual starting with an initiator usually a traditional geometric shape. Following the initiator is the generator ‘a template from which the fractal will be made’. And the rule of recursion: the actual process of building the fractal basically, one small simple shape giving birth to the same shape hundreds, thousands, hundreds of thousands, and so on forming a complex reverberation: ‘an echo of the whole’ (think DNA). To the point, fractal geometry is a ‘proper theory of the irregular’ and measurement of the ‘uncontrollable element of life’. Life is comprehensively ‘jagged’. And, too, the financial markets.
Beginning in 1961 with the study of Cotton prices, wind turbulence, and powerlaw distributions, Mandelbrot discovered distributions have ‘fat tails’ and ‘prices can and do vary by leaps and bounds’ unlike the nice neat idea of normal distributions. He discovered applying his fractal theory to the financial markets was a supreme fit. Mandelbrot comments: ‘It is an extraordinary feature of science that the most diverse, seemingly unrelated phenomena can be described with the same mathematical tools. The same quadratic equations with which the ancients drew right angles to build their temples can be used today by a banker to calculate the yield to maturity of a new, twoyear bond’.
Mandelbrot, and the failure of our modern financial analysis, shows prices in fact do not follow the bell curve and rather than move in a ‘random walk’ seamlessly, frequently ‘jump’ (discontinuity) and a far more volatile and risky than we preach. Additionally, different kinds of price series ‘exhibit different degrees of memory’. To that, the power of fractal geometry comes from its ability to express a ‘great deal of complicated, irregular data¼’.
The Snapshot of Mandelbrot’s Rules of Market Behavior
Rule No. 1: Markets are risky. Riskier than currently given credit or effectively measured. As the examples of the previous Dow debacles indicates, extreme price swings are not the anomalies we believed them to be, but ‘are the norm in financial markets, not aberrations¼’ Price movements do not follow the wellmannered, smooth bell curve assumed by modern financial models. ‘They follow a more violent curve.’ Not unlike measuring wind turbulence, there are ‘eddies’ and frequent discontinuous movements to prices in the markets that have been and are miscategorized as anomalies and infrequent outliers. Fractal geometry indeed expresses clearly risk has and is being mismeasured. Our current understanding of cause and effect is murky at best and catastrophically wrong at worst.
Rule No. 2: Trouble Runs in Streaks. ‘Market turbulence tends to cluster.’ As professional traders experience, a wild volatile trading day on Day 1 may well be followed by a wilder trading day on Day 2. As Mandelbrot observes, ‘Their (traders) intuition not included in the standard model of efficient markets is entirely validated by the multifractal model’. Multifractals make turbulence a ‘fundamentally new way of analyzing finance’. Markets stop appearing rational, ‘wellbehaved’ patterns. They are in fact, dynamic, unpredictable and sometimes ‘dangerous systems for transferring wealth’. Price charts are commonly measured by their volatility. However, volatility itself is volatile (an ‘echo of the whole’). Mandelbrot’s fractal approach is leveraged on spotting the regularity inside the irregular ‘the pattern in the formless’.
Rule No. 3: Markets Have Personality. ‘Prices are not driven solely by realworld events, news, and people’. A new kind of dynamic surfaces when the ‘gathering’ of all the traders and investors of the universe converge on the markets daily. Literally thousands of different motives, ideas, strategies, tactics, ambitions, needs and wants rush the markets that affect them endogenously. The fundamental process by which prices react to news does not change (mathematically stationary) completely contradictory to the randomwalk theory.
Rule No. 4: Markets Mislead. ‘Patterns are the fool’s gold of financial markets’. They appear predictable and bankable under modern financial analysis. Hence, ‘bubbles and crashes are inherent to the markets’. They are the culmination and byproduct of the ‘human need to find patterns in the patternless’. People often think and behave in hierarchies ‘with repetition and scaling.’ As Mandelbrot submits, ‘the brain highlights what it imagines as patterns and disregards contradicting information’. Human nature strives to see order and will create it where it cannot find it. Fractals do not discount patterns, but, in fact patterns are the essence of fractal geometry. Fractal geometry states a theory of clearly measuring patterns, not subjectively and subsequently Gerry rigging observations to form patterns.
Rule No. 5: Market Time is Relative. There is a relativity of time in the financial markets ‘distinct from linear clock time.’ Simply, and as traders well understand, in periods of high volatility, time seems to speed up multifold while the inverse is true in periods of low volatility. The foundation of the common trading terms ‘fast’ and ‘slow’ markets is the compression and expansion of time (scaling) within the confines of the financial markets. It has an effect on how accurate modern financial analysis measures the observations, but, unfortunately, ignores the fact, disregarding relative time as a factor in analyzing the data observed not even as an outlier, but as nonexistent. Fractal mathematics has the opportunity to not only measure this component, but make it part of the overall analysis for comprehensive modeling.
Conclusion
The modern financial models’ theories and assumptions that prices move independently, smoothly, and markets are rational as well as its participants have been tested, if not throttled by Mandelbrot’s multifractal approach and observations. Clearly there are limitations and stressors that frequently surface to condemn the ‘conventional wisdom’ of modern portfolio theory and its modified incarnations. In Mandelbrot’s words, “We know more about how our car engines work than about how our global financial system functions¼we lurch from crisis to crisis”. Mandelbrot’s intriguing assumptions strongly suggest a need for invention and change, but admits ‘clearly, fractal investment analysis has more questions than answers today’. One thing that is clear to Mandelbrot without exception is ‘the very heart of finance is fractal’.
References
Dynamic Portfolio Theory and Management (Oberuc E., Richard) 2004
The Fractal Geometry of Nature (Mandelbrot, Benoit) 1982
The (Mis)Behavior of Markets (Mandelbrot, Benoit and Hudson L., Richard) 2004
Dostları ilə paylaş: 