Shift Happens On a New Paradigm of the

“Shift Happens” On a New Paradigm of the Markets as a Complex Adaptive System by Michael J. Mauboussin Lecture Notes for Finance 450 CSULB Dr. Ammermann

References: • “Shift Happens” – www. capatcolumbia. com • The Warren Buffett Portfolio, by Robert Hagstrom – Ch. 8, “The Market as a Complex Adaptive System” • Chaos and Order in the Capital Markets, by Edgar Peters

3 Forms of Market Efficiency • Weak form – technical analysis – Supported by studies on autocorrelation – Contradicted by seasonality, subsequent performance of “winners” and “losers” • Semi-strong form – fundamental analysis – Supported by typical (under)performance of fund managers, also by many event studies – Contradicted by P/E and M/B effects, consistent outperformance of some fund managers (e. g. , the “Superinvestors of Graham-and-Doddsville”)

3 Forms of Market Efficiency • Strong form – insiders – Contradicted by subsequent performance of stocks after insider share repurchases

Areas of Potential Exploitation • Two areas of potential exploitation under traditional theory – Better information • Day traders, arbitrageurs? – Better analysis • Warren Buffett?

Areas of Potential Exploitation • Tension exists between academics and practitioners – Academics – rational agents, random walk, efficient markets – Practitioners – outperformance of some fund managers, irrational investors, inefficiency – One model reconciles both, bringing theory together with practice: The Market as a Complex Adaptive System (CAS)

Paradigm Shifts • • Thomas Kuhn Steps in process: 1. Theory laid out 2. Scientists test theory, but some facts counter it 3. Original theory stretched (to encompass new facts) 4. New theory developed to supercede the old

Paradigm Shifts • Example: 1. Aristotle proposes geocentric universe with orbits as circles 2. Astronomers observe the orbits are elliptical, not circular 3. Ptolemy introduces “circles-upon-circles” 4. Copernicus, Kepler, and Galileo introduce heliocentric universe, elliptical orbits, and celestial imperfection

2 Tests for a New Paradigm • Jeremy Bernstein’s test of “Correspondence” • 2 tests: – New idea must explain why the old theory worked – Must also add some predictive (or at least explanatory) power

Classical Capital Market Theory • Economics still largely based on equilibrium systems: supply vs. demand, price vs. quantity, risk vs. reward • Stems from the view of economics as a science akin to Newtonian physics – direct cause and effect, with implied predictability • Many statistical tools can only be applied if equilibrium theory holds

Classical Theory 1. Stock market efficiency – prices reflect all relevant information when that info. is cheap & widely disseminated – – Purchasing stocks as a zero-NPV proposition Prices not always “correct, ” but not systematically wrong 2. Random walk – security price changes are independent of each other – – Lots of agents – current prices reflect all info. that is collectively known Normal distribution of stock returns typically assumed in conjunction with this 3. Rational agents – investors can assess and optimize risk / reward opportunities 4. Assumptions / prediction – modest trading activity, limited price fluctuations directly attributable to specific news “events”

Classical Theory Tested 1. Stock returns are not normal – – 2. “fat tails, ” or “Noah Effect” Similar to “punctuated equilibrium” in biology Random walk not supported by data – – 3. 4. Elements of persistence – “Joseph Effect” Campbell, Lo, & Mac. Kinley show prices are predictable Volume higher & price changes greater than predicted Risk & reward not linked via variance – 5. Fama – firm size and M/B more important, so interpreted as risk factors Investors are not rational – – Systematic judgment errors have been identified Humans operate inductively, not deductively

Theory Stretched 1. Non-normal distributions – – 2. Typically ignored, or “outliers” simply removed from data Mandelbrot and stable Paretian distributions for financial data Noise traders (Black, 1986) – – – 3. “Noise” fills theory / practice gap Should not trade “Noise theories were all derived originally as part of a broad effort to apply the logic of the [CAPM] to … behavior that does not fit conventional notions of optimization. ” Crashes – – 4. Fama – “I think the crash in ’ 87 was a mistake” Miller – recommends reading Mandelbrot Behavioral finance – – “as if” argument – if many investors whose errors are independent with no systematic biases, aggregate market should appear rational Dilutes theory

New Theory – Market as Complex Adaptive System • Three components of complex adaptive systems: 1. Decision rules 1. Lots of agents, each operating with their own decision rules, with the most effective surviving 2. Provides the “adaptive” part of CAS 2. Emergence 1. Emergence = complex, large-scale behaviors resulting from aggregate interaction of less complex agents 2. e. g. , ant colony – individual ants have simple tasks, but combine together to create a very complex colony 3. e. g. , Adam Smith’s “invisible hand” 3. Market as a “Meta-System”

Market as a “Meta-System” • • “Meta-System” = market has properties and characteristics distinct from the agents / investors who comprise it Two key characteristics: – Nonlinearity • – Output of the system will not necessarily be proportional to the input Critical points • Periodically, small-scale stimuli lead to large-scale effects – – • “the straw that broke the camel’s back” Sand piles Leads to booms and crashes in the market

Does CAS Conform to Reality? • Adaptive behavior – leads to high trading volumes • Nonlinearity – contributes to fat-tailed return distributions • Trend persistence, but with low levels of autocorrelation, found in many complex adaptive systems • Homogeneous vs. heterogeneous expectations – Heterogeneous expectations may lead to risk / reward inefficiency – Deductive vs. inductive decision making – “El Farol” problem - Independent errors O. K. , nonindependent errors can lead to self-reinforcing trends

Does CAS Conform to Reality? • Portfolio manager performance – Not much predictability with CAS’s – Nonetheless, some investors may be “hardwired” to be good investors • E. g. , Warren Buffett, George Soros • Artificial stock market – Santa Fe Institute – Replicate market activity – Generates realistic market behavior

Practical Investor Considerations • Risk & reward link may not be clear – CAPM a possible first approximation, but far from the whole story! • Cause-and-effect thinking is dangerous – Human nature to identify cause and effect – But, with nonlinearity, large-scale changes can come from small-scale inputs (cf. Haugen, “Beast on Wall Street” – Crash of ’ 87 • Traditional DCF analysis remains valuable – Sets out first principles for stock valuation – Good framework for sorting out key issues in valuation – Helps investors crystallize / quantify expectations • Strategy / micro-economics

Why does EMH still win out? • Still hard to beat the market consistently – Implies EMH is a good first approximation • Strategy for Buffett’s “know-nothing” investors – Market efficiency is effectively accurate • Haugen’s counter-argument – Market so IN-efficient that investors cannot count on rational values to resurface • Justification for Buffett’s focus on “margin of protection” over and above “margin of safety” – Near the “wrong 20 -yard line” – Cf. , Ch. 15, “The Inefficient Stock Market”