Kissell The Science of Algorithmic Trading and Portfolio Management

1. Auflage 2013
ISBN: 978-0-12-401693-4
Verlag: Elsevier Science & Techn.
Format: EPUB
Kopierschutz: 6 - ePub Watermark

E-Book, Englisch, 496 Seiten

The Science of Algorithmic Trading and Portfolio Management, with its emphasis on algorithmic trading processes and current trading models, sits apart from others of its kind. Robert Kissell, the first author to discuss algorithmic trading across the various asset classes, provides key insights into ways to develop, test, and build trading algorithms. Readers learn how to evaluate market impact models and assess performance across algorithms, traders, and brokers, and acquire the knowledge to implement electronic trading systems. This valuable book summarizes market structure, the formation of prices, and how different participants interact with one another, including bluffing, speculating, and gambling. Readers learn the underlying details and mathematics of customized trading algorithms, as well as advanced modeling techniques to improve profitability through algorithmic trading and appropriate risk management techniques. Portfolio management topics, including quant factors and black box models, are discussed, and an accompanying website includes examples, data sets supplementing exercises in the book, and large projects. - Prepares readers to evaluate market impact models and assess performance across algorithms, traders, and brokers. - Helps readers design systems to manage algorithmic risk and dark pool uncertainty. - Summarizes an algorithmic decision making framework to ensure consistency between investment objectives and trading objectives.

Robert Kissell, Ph.D., is President of Kissell Research Group, a global financial and economic consulting firm specializing in quantitative modeling, statistical analysis, and algorithmic trading. He is also a professor at Molloy College in the School of Business and an adjunct professor at the Gabelli School of Business at Fordham University. He has held several senior leadership positions with prominent bulge bracket investment banks including UBS Securities where he was Executive Director of Execution Strategies and Portfolio Analysis, and at JP Morgan where he was Executive Director and Head of Quantitative Trading Strategies. He was previously at Citigroup/Smith Barney where he was Vice President of Quantitative Research, and at Instinet where he was Director of Trading Research. He began his career as an Economic Consultant at R.J. Rudden Associates specializing in energy, pricing, risk, and optimization. Dr. Kissell has written several books and published dozens of journal articles on Algorithmic Trading, Risk, and Finance. He is a coauthor of the CFA Level III reading titled 'Trade Strategy and Execution,” CFA Institute 2019.”

Autoren/Hrsg.

Kissell, Robert

Inhaltsverzeichnis

1;Front Cover;1
2;The Science of Algorithmic Trading and Portfolio Management;4
3;Copyright Page;5
4;Contents;8
5;Preface;16
6;Acknowledgments;18
7;1 Algorithmic Trading;20
7.1;Introduction;20
7.1.1;Advantages;22
7.1.2;Disadvantages;23
7.2;Changing Trading Environment;24
7.3;Recent Growth in Algorithmic Trading;30
7.4;Investment Cycle;34
7.5;Classifications of Algorithms;35
7.6;Types of Algorithms;36
7.7;Algorithmic Trading Trends;39
7.8;Trading Venue Classification;40
7.8.1;Displayed Market;40
7.8.2;Dark Pool;40
7.8.3;Grey Pool;40
7.8.4;Dark Pool Controversies;41
7.9;Types of Orders;42
7.10;Execution Options;42
7.11;The Trading Floor;44
7.11.1;Research Function;45
7.11.2;Sales Function;46
7.12;Algorithmic Trading Decisions;48
7.12.1;Macro-Level Strategies;48
7.12.1.1;Step 1—Choose Implementation Benchmark;49
7.12.1.2;Step 2—Select Optimal Execution Strategy;49
7.12.1.3;Step 3—Specify Adaptation Tactic;51
7.12.2;Micro-Level Decisions;52
7.12.2.1;Limit Order Models;53
7.12.2.2;Smart Order Routers;54
7.13;Algorithmic Analysis Tools;56
7.13.1;Pre-Trade Analysis;56
7.13.2;Intraday Analysis;56
7.13.3;Post-Trade Analysis;57
7.13.4;Rule-Based Trading;57
7.13.5;Quantitative Techniques;57
7.14;High Frequency Trading;58
7.14.1;Auto Market Making;58
7.14.2;Quantitative Trading/Statistical Arbitrage;60
7.14.3;Rebate/Liquidity Trading;60
7.15;Direct Market Access;62
7.15.1;Advantages;63
7.15.2;Disadvantages;63
8;2 Market Microstructure;66
8.1;Introduction;66
8.2;Market Microstructure Literature;68
8.3;The New Market Structure;70
8.4;Pricing Models;75
8.5;Order Priority;76
8.6;Equity Exchanges;76
8.7;New NYSE Trading Model;76
8.7.1;Designated Market Makers;77
8.7.2;Supplemental Liquidity Providers;78
8.7.3;Trading Floor Brokers;79
8.8;NASDAQ Select Market Maker Program;79
8.9;Empirical Evidence;80
8.9.1;Trading Volumes;80
8.9.1.1;Market Share;80
8.9.1.2;Large and Small Cap Trading;81
8.9.1.3;Do Stocks Trade Differently Across the Exchanges and Venues?;82
8.9.2;Volume Distribution Statistics;82
8.9.3;Day of Week Effect;84
8.9.4;Intraday Trading Profiles;86
8.9.4.1;Spreads;86
8.9.4.2;Volumes;87
8.9.4.3;Volatility;89
8.9.4.3.1;Intraday Trading Stability—Coefficient of Variation;91
8.9.5;Special Event Days;92
8.10;Flash Crash;95
8.10.1;Empirical Evidence from the Flash Crash;98
8.10.2;What Should Regulators do to SafeGuard Investors from Potential Future Flash Crashes?;102
8.10.3;Comparison with Previous Crashes;103
8.11;Conclusion;104
9;3 Algorithmic Transaction Cost Analysis;106
9.1;Introduction;106
9.1.1;What Are Transaction Costs?;107
9.1.2;What Is Best Execution?;107
9.1.3;What Is the Goal of Implementation?;108
9.2;Unbundled Transaction Cost Components;108
9.2.1;1. Commission;108
9.2.2;2. Fees;108
9.2.3;3. Taxes;108
9.2.4;4. Rebates;109
9.2.5;5. Spreads;109
9.2.6;6. Delay Cost;110
9.2.7;7. Price Appreciation;110
9.2.8;8. Market Impact;110
9.2.9;9. Timing Risk;111
9.2.10;10. Opportunity Cost;111
9.3;Transaction Cost Classification;111
9.4;Transaction Cost Categorization;113
9.5;Transaction Cost Analysis;113
9.5.1;Measuring/Forecasting;115
9.5.2;Cost versus Profit and Loss;116
9.6;Implementation Shortfall;116
9.6.1;Complete Execution;118
9.6.2;Opportunity Cost (Andre Perold);119
9.6.3;Expanded Implementation Shortfall (Wayne Wagner);120
9.6.3.1;Implementation Shortfall Formulation;123
9.6.3.1.1;Trading Cost/Arrival Cost;123
9.7;Evaluating Performance;124
9.7.1;Trading Price Performance;124
9.7.2;Benchmark Price Performance;125
9.7.3;VWAP Benchmark;125
9.7.4;Participation Weighted Price (PWP) Benchmark;127
9.7.5;Relative Performance Measure (RPM);128
9.7.6;Pre-Trade Benchmark;129
9.7.7;Index Adjusted Performance Metric;130
9.7.8;Z-Score Evaluation Metric;131
9.7.9;Market Cost Adjusted Z-Score;132
9.7.10;Adaptation Tactic;133
9.8;Comparing Algorithms;134
9.8.1;Non-Parametric Tests;135
9.8.1.1;Paired Samples;136
9.8.1.2;Sign Test;136
9.8.1.3;Wilcoxon Signed Rank Test;137
9.8.1.4;Independent Samples;139
9.8.1.4.1;Mann-Whitney U Test;139
9.8.1.4.2;Median Test;141
9.8.2;Distribution Analysis;142
9.8.2.1;Chi-Square Goodness of Fit;142
9.8.2.1.1;Kolmogorov-Smirnov Goodness of Fit;143
9.9;Experimental Design;144
9.9.1;Proper Statistical Tests;145
9.9.2;Small Sample Size;145
9.9.3;Data Ties;145
9.9.4;Proper Categorization;146
9.9.5;Balanced Data Sets;146
9.10;Final Note on Post-Trade Analysis;146
10;4 Market Impact Models;148
10.1;Introduction;148
10.2;Definition;148
10.2.1;Example 1: Temporary Market Impact;149
10.2.2;Example 2: Permanent Market Impact;149
10.3;Graphical Illustrations of Market Impact;150
10.3.1;Illustration 1—Price Trajectory;150
10.3.2;Illustration 2—Supply-Demand Equilibrium;151
10.3.3;Illustration 3—Temporary Impact Decay Function;154
10.3.4;Example—Temporary Decay Formulation;156
10.3.5;Illustration 4—Various Market Impact Price Trajectories;157
10.4;Developing a Market Impact Model;158
10.4.1;Essential Properties of a Market Impact Model;159
10.5;Derivation of Models;161
10.5.1;Almgren & Chriss—Market Impact Model;161
10.5.2;Random Walk with Price Drift—Discrete Time Periods;162
10.5.3;Random Walk with Market Impact (No price drift);163
10.6;I-Star Market Impact Model;165
10.7;Model Formulation;166
10.7.1;I-Star: Instantaneous Impact Equation;166
10.7.2;Market Impact Equation;167
10.7.2.1;Derivation of the Model;167
10.7.2.2;Cost Allocation Method;168
10.7.2.3;I* Formulation;170
10.7.2.4;Comparison of Approaches;172
10.7.3;Underlying Data Set;173
10.7.3.1;Imbalance/Order Size;173
10.8;Parameter Estimation Techniques;176
10.8.1;Technique 1: Two-Step Process;176
10.8.1.1;Step 1: Estimate Temporary Impact Parameter;176
10.8.1.2;Step 2: Estimate ai Parameters;177
10.8.2;Technique 2: Guesstimate Technique;179
10.8.2.1;Technique 3: Non-Linear Optimization;179
10.8.3;Model Verification;179
10.8.3.1;Model Verification 1: Graphical Illustration;180
10.8.3.2;Model Verification 2: Regression Analysis;180
10.8.3.3;Model Verification 3: Z-Score Analysis;180
10.8.3.4;Model Verification 4: Error Analysis;181
11;5 Estimating I-Star Model Parameters;182
11.1;Introduction;182
11.2;Scientific Method;183
11.2.1;Step 1: Ask a Question;183
11.2.2;Step 2: Research the Problem;183
11.2.3;Step 3: Construct the Hypothesis;183
11.2.4;Step 4: Test the Hypothesis;183
11.2.5;Step 5: Analyze the Data;184
11.2.6;Step 6: Conclusion and Communication;184
11.3;Solution Technique;185
11.3.1;The Question;185
11.3.2;Research the Problem;185
11.3.3;Construct the Hypothesis;190
11.3.4;Test the Hypothesis;192
11.3.5;Data Definitions;194
11.3.6;Universe of Stocks;195
11.3.7;Analysis Period;195
11.3.8;Time Period;195
11.3.9;Number of Data Points;195
11.3.10;Imbalance;195
11.3.11;Side;196
11.3.12;Volume;196
11.3.13;Turnover;196
11.3.14;VWAP;197
11.3.15;First Price;197
11.3.16;Average Daily Volume;197
11.3.17;Annualized Volatility;197
11.3.18;Size;198
11.3.19;POV Rate;198
11.3.20;Cost;198
11.3.21;Estimating Model Parameters;198
11.3.21.1;Sensitivity Analysis;200
11.3.21.2;Cost Curves;205
11.3.21.3;Statistical Analysis;206
11.3.21.4;Error Analysis;206
11.3.21.5;Stock Specific Error Analysis;208
12;6 Price Volatility;212
12.1;Introduction;212
12.2;Definitions;213
12.2.1;Price Returns/Price Change;213
12.2.2;Average Return;213
12.2.3;Volatility;215
12.2.4;Covariance;215
12.2.5;Correlation;216
12.2.6;Dispersion;216
12.2.7;Value-at-Risk;216
12.2.8;Implied Volatility;217
12.2.9;Beta;217
12.3;Market Observations—Empirical Findings;218
12.4;Forecasting Stock Volatility;221
12.4.1;Volatility Models;221
12.4.1.1;Price Returns;222
12.4.1.1.1;Data Sample;222
12.4.1.2;Historical Moving Average (HMA);223
12.4.1.3;Exponential Weighted Moving Average (EWMA);224
12.4.1.4;Arch Volatility Model;224
12.4.1.5;GARCH Volatility Model;225
12.5;HMA-VIX Adjustment Model;225
12.5.1;Determining Parameters via Maximum Likelihood Estimation;227
12.5.1.1;Likelihood Function;227
12.5.1.2;Estimation Results;228
12.6;Measuring Model Performance;228
12.6.1;Root Mean Square Error (RMSE);229
12.6.2;Root Mean Z-Score Squared Error (RMZSE);229
12.6.3;Outlier Analysis;230
12.6.4;Results;230
12.6.5;Problems Resulting from Relying on Historical Market Data for Covariance Calculations;233
12.6.5.1;False Relationships;233
12.6.5.2;Degrees of Freedom;238
12.7;Factor Models;240
12.7.1;Matrix Notation;242
12.7.2;Constructing Factor Independence;243
12.7.3;Estimating Covariance Using a Factor Model;244
12.8;Types of Factor Models;246
12.8.1;Multi-Index Models;247
12.8.2;Macroeconomic Factor Models;247
12.8.2.1;Cross-Sectional Multi-Factor Models;248
12.8.3;Index Model;246
12.8.3.1;Single Index Model;246
12.8.4;Statistical Factor Models;250
13;7 Advanced Algorithmic Forecasting Techniques;254
13.1;Introduction;254
13.2;Trading Cost Equations;255
13.2.1;Model Inputs;255
13.3;Trading Strategy;256
13.3.1;Percentage of Volume;256
13.3.2;Trading Rate;257
13.3.3;Trade Schedule;257
13.3.4;Comparison of POV rate to Trade Rate;258
13.4;Trading Time;258
13.5;Trading Risk Components;259
13.6;Trading Cost Models—Reformulated;260
13.6.1;Market Impact Expression;260
13.6.1.1;I-Star;260
13.6.1.2;Market Impact for a Single Stock Order;260
13.6.1.3;Market Impact for a Basket of Stock;262
13.7;Timing Risk Equation;262
13.7.1;Timing Risk for a Basket of Stock;267
13.8;Comparison of Market Impact Estimates;267
13.9;Volume Forecasting Techniques;270
13.9.1;Daily Volumes;270
13.9.1.1;Definitions;270
13.9.1.2;Daily Forecasting Analysis—Methodology;271
13.9.1.3;Variable Notation;271
13.9.1.4;ARMA Daily Forecasting Model;271
13.9.1.5;Analysis Goal;272
13.9.1.6;Forecast Improvements;276
13.9.1.7;Daily Volume Forecasting Model;276
13.10;Forecasting Monthly Volumes;277
13.11;Forecasting Covariance;282
13.12;Efficient Trading Frontier;284
13.12.1;Single Stock Trade Cost Objective Function;286
13.12.2;Portfolio Trade Cost Objective Function;286
14;8 Algorithmic Decision Making Framework;288
14.1;Introduction;288
14.2;Equations;289
14.3;Algorithmic Decision Making Framework;291
14.3.1;1) Select Benchmark Price;291
14.3.1.1;Arrival Price Benchmark;291
14.3.1.2;Historical Price Benchmark;292
14.3.1.3;Future Price Benchmark;294
14.3.2;Comparison of Benchmark Prices;295
14.3.3;2) Specify Trading Goal;295
14.3.3.1;1. Minimize Cost;296
14.3.3.2;2. Minimize Cost with Risk Constraint;298
14.3.3.3;3. Minimize Risk with Cost Constraint;299
14.3.3.4;4. Balance Trade-off between Cost and Risk;299
14.3.3.5;5. Price Improvement;300
14.3.3.6;Further Insight;302
14.3.4;3) Specify Adaptation Tactic;303
14.3.4.1;Projected Cost;304
14.3.4.2;Target Cost Tactic;307
14.3.4.3;Aggressive-in-the-Money;308
14.3.4.4;Passive-in-the-Money;310
14.3.5;Comparison across Adaptation Tactics;312
14.3.6;Modified Adaptation Tactics;313
14.3.7;How Often Should We Re-Optimize Our Tactics?;313
15;9 Portfolio Algorithms;316
15.1;Introduction;316
15.2;Trader’s Dilemma;317
15.2.1;Variables;318
15.3;Transaction Cost Equations;319
15.3.1;Market Impact;320
15.3.2;Price Appreciation;320
15.3.3;Timing Risk;321
15.3.4;One-Sided Optimization Problem;321
15.4;Optimization Formulation;321
15.4.1;Constraint Description;322
15.4.1.1;Objective Function Difficulty;324
15.4.1.2;Optimization Objective Function Simplification;324
15.5;Portfolio Optimization Techniques;325
15.5.1;Quadratic Programming Approach;325
15.5.2;Trade Schedule Exponential;327
15.5.3;Residual Schedule Exponential;328
15.5.4;Trading Rate Parameter;329
15.5.4.1;Market Impact Expression;329
15.5.4.2;Timing Risk Expression;330
15.5.5;Comparison of Optimization Techniques;331
15.6;Portfolio Adaptation Tactics;335
15.6.1;Description of AIM and PIM for Portfolio Trading;336
15.6.2;How Often Should We Re-Optimize?;338
15.7;Managing Portfolio Risk;339
15.7.1;Residual Risk Curve;339
15.7.2;Minimum Trading Risk Quantity;341
15.7.3;Maximum Trading Opportunity;342
15.7.4;When to Use These Values?;343
15.7.5;Program-Block Decomposition;344
15.8;Appendix;347
16;10 Portfolio Construction;350
16.1;Introduction;350
16.2;Portfolio Optimization and Constraints;351
16.3;Transaction Costs in Portfolio Optimization;354
16.4;Portfolio Management Process;358
16.4.1;Example: Efficient Trading Frontier w/ and w/o Short Positions;359
16.4.2;Example: Maximizing Investor Utility;359
16.5;Trading Decision Process;360
16.6;Unifying the Investment and Trading Theories;362
16.7;Cost-Adjusted Frontier;367
16.8;Determining the Appropriate Level of Risk Aversion;369
16.9;Best Execution Frontier;370
16.10;Portfolio Construction with Transaction Costs;371
16.10.1;Quest for best execution frontier;373
16.10.1.1;Return;374
16.10.1.2;Risk;374
16.11;Conclusion;378
17;11 Quantitative Portfolio Management Techniques;380
17.1;Introduction;380
17.2;Are the Existing Models Useful Enough for Portfolio Construction?;382
17.2.1;Current State of Vendor Market Impact Models;383
17.3;Pre-Trade of Pre-Trades;386
17.3.1;Estimation Process;387
17.3.2;Applications;391
17.3.2.1;Example 1;391
17.3.2.2;Example 2;392
17.3.2.3;Example 3;392
17.3.2.4;Example 4;392
17.4;How Expensive Is It to Trade?;393
17.4.1;Acquisition and Liquidation Costs;396
17.4.2;Portfolio Management—Screening Techniques;399
17.5;MI Factor Scores;403
17.5.1;Derivation of the MI Factor Score for Shares;403
17.5.2;Current State of MI Factor Scores;405
17.5.3;MI Factor Score Analysis;405
17.6;Alpha Capture Program;407
17.6.1;Example 5;408
17.6.2;Example 6;409
17.6.3;Alpha Capture Curves;412
18;12 Cost Index & Multi-Asset Trading Costs;414
18.1;Introduction;414
18.2;Cost Index;415
18.2.1;Cost Basis;416
18.2.2;Cost Strategy;417
18.2.3;Normalization Process;419
18.2.3.1;Customized Indexes;421
18.3;Real-Time Cost Index;422
18.3.1;Back-Testing;427
18.3.2;Market Impact Simulation;429
18.3.3;Simulation Scenario;431
18.4;Multi-Asset Class Investing;434
18.4.1;Investing in Beta Exposure and Other Factors;434
18.4.2;Beta Investment Allocation;438
18.5;Multi-Asset Trading Costs;439
18.5.1;Global Equity Markets;440
18.5.2;Multi-Asset Classes;441
19;13 High Frequency Trading and Black Box Models;448
19.1;Introduction;448
19.2;Data and Research;450
19.3;Strategies;451
19.3.1;Statistical Arbitrage;451
19.3.2;Triangular Arbitrage;455
19.3.3;Liquidity Trading;458
19.3.4;Market-Neutral Arbitrage;459
19.3.5;Index and Exchange Traded Fund Arbitrage;461
19.3.6;Merger Arbitrage;462
19.4;Evaluation;465
19.5;Summary;469
20;References;472
21;Index;484