Topic 8 ©R. Schwartz Equity Markets: Trading and Structure Slide 1 Introductory Remarks ©R. Schwartz Equity Markets: Trading and Structure Slide 2 The Dynamic Behavior of Prices The Effect of Trading Costs • In a frictionless environment, prices would follow random walks • Do they? • If so, how valuable would technical analysis and algorithmic trading be? • If not, would shares have unique fundamental values? ©R. Schwartz Equity Markets: Trading and Structure Slide 3 Do Shares Have Unique Values? Belief • Equilibrium values exist • Complex information translates into a single price • Shares have fundamental values But This is not as simple as elementary economics ©R. Schwartz Equity Markets: Trading and Structure Slide 4 Remember This One? PRICE SELL The perfectly liquid, frictionless market solution P is Discovered BUY 0 ©R. Schwartz Q is Discovered QUANTITY Equity Markets: Trading and Structure Slide 5 Life Really is Not So Simple ©R. Schwartz Equity Markets: Trading and Structure The Standard Microstructure Model • Information traders • Liquidity traders • Noise traders ©R. Schwartz Equity Markets: Trading and Structure Slide 7 TraderEx Orders Come From 3 Types of Participants Information traders P* Is p*>offer or p*<bid? Liquidity Traders Quotes, Prices, Volume Noise Traders Is there a trend/ pattern? Trading Mechanism ©R. Schwartz Equity Markets: Trading and Structure Slide 8 Thoughts of a Efficient Market ©R. Schwartz Equity Markets: Trading and Structure Slide 9 What is Behind Standard Finance Models? • Markets are informationally efficient (EMH) • Shares have unique fundamental values • Informed investors form identical expectations Homogeneous expectations ©R. Schwartz Equity Markets: Trading and Structure Slide 10 With Homogeneous Expectations • Information maps uniquely into security values (fundamental values) • If trades are triggered for liquidity reasons only, shares will trade at bid & ask quotes that are appropriate given the fundamentals • Aside from bid-ask bounce, prices will follow random walks Hmmm… What economic function is left for an exchange? ©R. Schwartz Equity Markets: Trading and Structure Slide 11 Traditional Support for Random Walk Burton Malkiel, A Random Walk Down Wall Street, 1973 “Technical analysis is anathema to the academic world. We love to pick on it. Our bullying tactics are prompted by two considerations: (1) the method is patently false; and (2) it's easy to pick on." ©R. Schwartz Equity Markets: Trading and Structure Slide 12 Of Course, We All Know That… • Investors are rational and prices reflect fundamental information • Systematic patterns can be arbitraged away • But… hmmm…wait a minute… ©R. Schwartz Equity Markets: Trading and Structure Slide 13 ©R. Schwartz Equity Markets: Trading and Structure Slide 14 A Second Opinion “The efficient market hypothesis is the most remarkable error in the history of economic theory” Lawrence Summers Subsequently U.S. Treasury Secretary The Wall Street Journal, 1987 ©R. Schwartz Equity Markets: Trading and Structure Slide 15 An Earlier Opinion Bernard M. Baruch, My Own Story, Henry Holt & Company, 1957, p. 84 The prices of stocks – and commodities and bonds as well – are affected by literally anything and everything that happens in our world, from new inventions and the changing value of the dollar to vagaries of the weather and the threat of war or the prospect of peace. • But these happenings do not make themselves felt in Wall Street in an impersonal way, like so many jigglings on a seismograph. • What registers in the stock market’s fluctuations are not the events themselves but the human reactions to these events, how millions of individual men and women feel these happenings may affect the future. ©R. Schwartz Equity Markets: Trading and Structure Slide 16 With Whom Do You Agree? 1. Burton Malkiel 2. Lawrence Summers 3. Bernard Baruch ??? ©R. Schwartz Equity Markets: Trading and Structure Slide 17 Another Perspective ©R. Schwartz Equity Markets: Trading and Structure Slide 18 A Deceptively Simple Question: What motivates individuals to trade? Accepted academic answer • Informed traders • Liquidity traders • Noise traders Perhaps we should add a fourth • Divergent expectations (people disagree…) ©R. Schwartz Equity Markets: Trading and Structure Slide 19 Divergent Expectations Has Implications For • • • • Understanding market structure & operations Assessing market quality Government regulatory policy Understanding volatlity ©R. Schwartz Equity Markets: Trading and Structure Slide 20 Edward M. Miller “Risk, Uncertainty, & Divergence of Opinion” Journal of Finance, Sept. 1977 “…it is implausible to assume that although the future is very uncertain, and the forecasts are very difficult to make, that somehow everyone makes identical estimates of the return and risk from every security. In practice, the very concept of uncertainty implies that reasonable men may differ in their forecasts.” ©R. Schwartz Equity Markets: Trading and Structure Slide 21 Complexity of Information • • • • Information sets are typically huge, complex, & imprecise Crudeness of our analytic tools Price & quantity discovery may be more complicated than academicians previously thought Technical analysis and algo trading may be valid Wow, did an academician say this? ©R. Schwartz Equity Markets: Trading and Structure Slide 22 Difficulty of Assessing Share Valuations With Precision Can a stock analyst or portfolio manager say with precision that the expected growth rate for XYZ is: 7.000%, not 7.545%? ©R. Schwartz Equity Markets: Trading and Structure Slide 23 Analyst Evaluation of XYZ Dividend one year from now = $1.35 Appropriate cost of eq. cap. = 10% (1) Growth rate (g) = 7.000% (2) Growth rate (g) = 7.545% Share price if g =7.000% Share price if g =7.545% = $45.00 = $55.00 ©R. Schwartz Equity Markets: Trading and Structure Slide 24 Evidence of Divergent Expectations • Private information • Analyst recommendations commonly differ • Prevalence of short selling • Two large institutions trading with each other on an ATS (e.g., Posit, Pipeline or Liquidnet) – Neither is likely to be a liquidity or noise trader – Neither may presume to have an informational edge – They are simply “agreeing to disagree” ©R. Schwartz Equity Markets: Trading and Structure Slide 25 Representing Divergent Expectations in TraderEx Informed Traders Liquidity Traders P* Is p*>offer or p*<bid? Do the informed Traders agree with each other? maybe not! ©R. Schwartz Quotes, Prices, Volume Noise Traders Is there a trend/ pattern? P* + 10% = VH (the bulls) P* - 10% = VL (the bears) Equity Markets: Trading and Structure Slide 26 Price Discovery ©R. Schwartz Equity Markets: Trading and Structure Slide 27 The Inside Scoop on Price Discovery • A complex, protracted process • Contributes to intra-day volatility • Equilibrium depends on the sequence of order arrivals & on how orders are handled → A coordination problem • The quality of price discovery depends on trader behavior & market structure • Divergent expectations underlie the complexity of price discovery ©R. Schwartz Equity Markets: Trading and Structure Slide 28 Divergent Expectations: A Simple Setting • • • A company is facing a jury trial – its share value will be affected appreciably by the outcome Investors can have 1 of 2 expectations Some believe pr(acquittal) = .80 Some believe pr(acquittal) = .35 Shares are valued at $55 by those who expect acquittal $45 by those who expect conviction ©R. Schwartz Equity Markets: Trading and Structure Slide 29 Lets Be More Generic • Bi-variate outcome: a decision will soon be made that will appreciably affect the value of a company Legal case: Acquit or convict Loan application: Grant or deny Takeover campaign: Win or loose shareholder votes • Investors disagree about probability of positive outcome • For bulls: positive expectation – stock is worth VH • For bears: negative expectation – stock is worth VL • The truth will soon be revealed ©R. Schwartz Equity Markets: Trading and Structure Slide 30 Price Determination in the Bi-variate Context VH = $55 (k percent of participants are bulls) – A* – B* Bid-Ask Spread for k = 0.6 VL = $45 (1-k percent of participants are bears) “Quote Setting and Price Formation in an Order Driven Market” Puneet Handa, Robert Schwartz, & Ashish Tiwari (HST) Journal of Financial Markets, August 2003 ©R. Schwartz Equity Markets: Trading and Structure Slide 31 From Divergent Expectations to… • • • We have not heard much about this It implies endogeneity of the trading decision A “Wisdom of the Crowds” reality A crowd is more likely to reach a correct decision than any single member of the crowd assuming independence ©R. Schwartz Equity Markets: Trading and Structure Slide 32 Picture It This Way • 800 observes are guessing the number of beans in a jar (the jar holds a lot, say 2500) • Each observer looks at the jar individually and forms an estimate • The observers come up one at a time and disclose their expectations • Each observer’s expectation depends on his initial estimate and on what he observes others guessing • As more observers arrive, each places less weight on his initial estimate ©R. Schwartz Equity Markets: Trading and Structure Slide 33 Adaptive Valuations (AV) Imply • • Random (multiple) equilibria Path dependency With random equilibria, shares do not have unique values ©R. Schwartz Equity Markets: Trading and Structure Slide 34 Our Starting Point: The Handa, Schwartz, Tiwari Model • • • • Risk neutral participants Participants arrive in random sequence Order driven, limit order book market There are just two valuations: VH & VL • Orders are placed w.r.t. VH, VL, and k → k percent are bulls (VH) → (1-k percent are bears (VL) ©R. Schwartz Equity Markets: Trading and Structure Slide 35 HST Model Cont. • • Market bid and offer prices for XYZ can be solved for if we know → VH, VL, & k If we know VH & VL only, → Price discovery is equivalent to k discovery → Remember… VH = $55 (k percent are bulls) – A* – B* Bid-Ask Spread for k = 0.6 VL = $45 (1-k percent are bears) ©R. Schwartz Equity Markets: Trading and Structure Slide 36 1 k 1 k k 2 HST’s Optimal Bid (B*) and Offer (A*) B* = γ VL + (1-γ) VH A* = VH + (1- ) VL where 1 k 2 1 k k k 1 k k 2 ©R. Schwartz Equity Markets: Trading and Structure Slide 37 What if k is Not Known? Orders are Based On • Each participant’s own assessment of information • Others’ assessments [ADAPTIVE VALUATIONS] • Others’ opinions are reflected in k, the % who are bulls ©R. Schwartz Equity Markets: Trading and Structure Slide 38 How Does Price Evolve When Everyone Uses This Basic Algorithm “The Dynamic Process of Price Discovery in an Equity Market,” J. Paroush, R. Schwartz & A. Wolf Working paper, 2008 ©R. Schwartz Equity Markets: Trading and Structure Slide 39 Representative Price Paths Alternative Equilibrium Prices 55 54 53 Price 52 51 50 49 48 47 46 45 0 5 10 15 20 25 30 No. Arrival ©R. Schwartz Equity Markets: Trading and Structure Slide 40 Volatility Consequences • RC(t) = PC(t)/PC(t-1) • Multiply and divide the RHS by VL(t)/ VL(t-1) • Rearranging gives: • RC(t) = [VL(t)/ VL(t-1)][PC(t) / VL(t)] ÷ [PC(t-1) / VL(t-1) ] News Price Discovery • Volatility of RC(t) = Vol [VL(t)/ VL(t-1)] + Vol [PC(t) / VL(t) ÷ PC(t-1) / VL(t-1)] Efficient Vol ©R. Schwartz Price Discovery Vol Equity Markets: Trading and Structure Slide 41 Information Complexities: Consequences Investor Expectations • Divergent • Can change independently at any time • Adaptive Simulation Results 55 54 53 Price Discovery • Random (multiple) equilibria • Path dependency • Accentuated intra-day volatility 52 Price 51 50 49 48 47 46 45 1 ©R. Schwartz 9 17 25 33 41 49 57 65 73 81 89 97 105 113 121 129 137 145 153 161 169 177 185 193 201 209 217 225 233 241 249 Equity Markets: Trading and Structure Events Slide 42 Implications • Volatility → Intra-day & longer run bubbles and crashes • Technical analysis →Validation of a price • Profitability of momentum and algorithmic trading →Path dependency • Importance of market structure → The quality of a network • Algorithmic trading ©R. Schwartz Equity Markets: Trading and Structure Slide 43 Algorithmic Trading ©R. Schwartz Equity Markets: Trading and Structure Slide 44 Question How Does Algo Trading Impact Price Discovery? ©R. Schwartz Equity Markets: Trading and Structure Slide 45 Remember Fragmentation Has a Temporal Dimension • An order can be fragmented → Slicing, dicing, & shredding → A breading ground for algo trading → Goal: make an elephant look like an ant → Warning: there can be negative consequences ©R. Schwartz Equity Markets: Trading and Structure Slide 46 From Elephants to Ants ©R. Schwartz Equity Markets: Trading and Structure Slide 47 If Algo Trading Helps Traders Individually, Does it Benefit Them Collectively? • Yes, in an electronic world, algos are essential • But some can lead to undesirable results • Lets look at one that might Slice & dice algos? VWAP algos? Momentum algos? Contrarian algos? An algo for an ant… ©R. Schwartz Equity Markets: Trading and Structure Slide 48 An Algo for an Ant That ant is a momentum player… ©R. Schwartz Equity Markets: Trading and Structure Slide 49 An Algo for an Ant ??? ©R. Schwartz Equity Markets: Trading and Structure Slide 50 The Ants Are Now in a “Circular Mill” ??? ©R. Schwartz Equity Markets: Trading and Structure Slide 51 The Wisdom of the Crowds James Surowiecki, 2005 • Who wants to be individually responsible for price discovery? •A consensus evaluation is likely to be more accurate than participants’ individual valuations • A “Wisdom of the Crowd” reality • But how wise is the crowd if everyone is keying off of what everyone else is doing? • There is no wisdom in a circular mill ©R. Schwartz Equity Markets: Trading and Structure Slide 52 Further Discussion ©R. Schwartz Equity Markets: Trading and Structure Slide 53 Question We have been focusing on trading, market structure, and price discovery: How well does anyone understand the markets? ©R. Schwartz Equity Markets: Trading and Structure Slide 54 … Both Sides of the Atlantic Have Experienced Major Change in • Technology • Regulation • Organizational Structure • Competitive Environment • Global Environment How Far Have We Come? ©R. Schwartz Equity Markets: Trading and Structure Slide 55 How Far Have We Come? • Commissions have shrunk • Trading volumes have soared • Implicit costs of trading have gone down (in the opinion of many) • A new, technologically savvy breed of traders are on the scene • Markets around the world are linked Have we arrived in the Promised Land? ©R. Schwartz Equity Markets: Trading and Structure Slide 56 Have We Arrived? Not quite Old issues are still with us • Small orders for large caps: no problem • Large orders for large caps: problems • All orders for mid & small caps: problems ©R. Schwartz Equity Markets: Trading and Structure Slide 57 More Old Issues • Transparency • Best Execution • Sub-Second Transaction Response Times • Elevated intra-day volatility • Spatial Consolidation/Fragmentation • Temporal Consolidation/Fragmentation – Slicing and dicing ©R. Schwartz Equity Markets: Trading and Structure Slide 58 Slicing & Dicing • Quantity discovery gets surprisingly little Fact Slicing and dicing are prevalent Average Trade Size at NYSE • 1988: 2,303 shares • June 2007: 297 shares Block Trading Volume at NYSE • 1988: 52 percent • June 2007: 20 percent ©R. Schwartz Equity Markets: Trading and Structure Slide 59 Decreasing Trade Size Average Daily Shares (Mils) Average Shares/Trades 1,600 5,000 1,400 4,000 1,200 1,000 3,000 800 2,000 600 400 1,000 200 NYSE Shares (right axis) NYSE Shares / Trade (left axis) ©R. Schwartz 6 Ja n0 5 Ja n0 4 Ja n0 3 Ja n0 2 Ja n0 1 Ja n0 0 Ja n0 Ja n9 Ja n9 9 0 8 0 Nasdaq Shares (right axis) NASDAQ Shares / Trade (left axis) Slide 60 Equity Markets: Trading and StructureSources: NYSE, Nasdaq, TowerGroup Why Does Market Structure Matter? • Why does market structure matter? Because price & quantity discovery are complex! • Why are these processes complex? Because shares do not have fundamental values! • Why don’t they? Because investors have divergent expectations! • Why do they? Because of the enormous complexity of information! ©R. Schwartz Equity Markets: Trading and Structure Slide 61 One More Statement It is time we recognize that we operate in a world where expectations are divergent, evaluations are adaptive, and analyze it as Such ©R. Schwartz Equity Markets: Trading and Structure Slide 62