INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted.
Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Aiso, if unauthorized copyright material had to be removed, a note will indicate the deletion., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy.
Higher quality 6" x 9” black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. ProQuest Information and Learning 300 North Zeeb Road, Ann Arbor, Mi 48106-1346 USA 800-521-0600 UMI ® ANALYSIS OF CREDIT RATING EQUITY INDEXES: VOLATILITY COMPARISONS AND OPTION CALIBRATION MANUEL SALES A DISSERTATION PRESENTED TO THE FACULTY OF PRINCETON UNIVERSITY IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY RECOMMENDED FOR ACCEPTANCE BY THE PROGRAM IN APPLIED AND COMPUTATIONAL MATHEMATICS JANUARY 2002 UMI Number: 3033036 Copyright 2001 by Sales, Manuel All rights reserved. UMI ® UMI Microform 3033036 Copyright 2002 by ProQuest Information and Learning Company.
All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P. Box 1346 Ann Arbor, MI 48106-1346 © Copyright by Manuel Sales, 2001.
All Rights Reserved _ Abstract In this thesis we analyze the properties credit segregated equity indexes. We form these indexes based on the Dow Jones Global Indexes and the Moody’s and Standard and Poor’s ratings databases. A methodology is implemented to test the hypothesis that two (or three) jointly Gaussian random variables with arbitrary correlation have the same variance. We employ this test to analyze the behavior of several of the Dow Jones indexes when segregated by credit rating.
We find that the fine structure of economce sectors is as important a determinant of the volatility patterns as the credit rating of the constituent companies. After discussing the minimum entropy calibration algorithm, and discussing a lower bound for the probability of feasibility of its Monte Carlo implementation, we analyze the problem of calibrating options to these indexes, based on the traded options on a few of the constituent companies. We find that the companies whose options we calibrate to need to account for over 90% of the index capitalization in order to attain any accuracy in the resulting index option prices. iii Acknowledgements First, I would like to thank my adviser, René Carmona.
He has been very helpful and has provided useful guidance, not only during the thesis writing months, but also throughout my graduate school experience. He has been an outstanding adviser and I am very grateful to him. Thanks are also due to Kian Esteghamat, who read this thesis in short notice yet provided useful insights on how to improve it. Ronnie Sircar and John Mulvey were also kind enough to sit on my committee, and provided many valuable suggetions.
I would also like to thank my friends for their support; in no particular order. Rob, Coral, Katerina, Kim, Toufic, Mike, Bing, Christine, Pavel and Tessy. My friends from Spain were also helpful; I didn’t get to see them as often as I would have wished but hopefully that can change now. Jorge, Raul, Nacho, German, Vicente, Pedro and all the rest.
I would also like to thank two of my friends for their valuable insights and help in the writing of this thesis. Jeffrey Schenker directed my attention to generating functions and how they could be useful to obtain a closed form solution to the feasi- bility problem of Section 5. Ernst Schaumburg was very helpful in too many ways to describe without embarrasing me. Just about every chapter has benefited from his comments and suggestions.
I also need to thank the Oravitz family for their support. Steve and Barbara, Steven, Bradley and Craig have all been there when I needed them; the last ten years of my life have been much more pleasant thanks to them. In a way, this thesis closes a chapter in my life which began when I was an exchange student in their home eleven years ago. Finally, I want to thank my family for all their help.
My brothers Joaquin, Nacho and Jaime were all very supportive and helpful. Our dog Kira was with us for almost the whole time that I was in America, and unfortunately won’t be there when I return, iv but I know she would have received me enthusiastically when I returned. My aunt Bibiana was not able to come to my undergraduate graduation and would have liked to come to this one. Unfortunately she passed away last year; she was very fond of me and I know she’s very proud and happy that I’m going back home.
I want to end this long list of acknowledgements by thanking the people who deserve it most. My parents have provided me with all the love and care that I could hope for, and have always encouraged me to do what was best for my education, even though it has meant that I have been living far away from home for ten years. I owe more than thanks, and I dedicate this thesis to them.20 00202 ee ee ee ee eee Acknowledgements .--0 22-205 eee eeeae List of Tables .00084 List of Figures. ee INTRODUCTION DOW JONES GLOBAL INDEXES AND COMPANY RATINGS o 2.1 Dow Jones Global Indexes .1 Companies included in the DJGIs.2 Calculation of the index value.OQẶ Q HQ HQ HH kg Và 17 2.217 Moody'* Ratings Database.2 Standard & Poor’s Ratings Database .3 Creating Indexes by Rating .3 Basic Properties of thelIndexe.0 0 2 ee ee ee ee 3.1 The Multivariate Gaussian Distribution.2 Likelhood Ratio Tests.2 Two dimensional comparisons .21 The Morgan-Pitman test.2 A Multidimensional test for variances.
Implementation of the two dimensional test.4 Three Dimensional Variance Ïlests.5 Implementing the three dimensionaÌ test. eee eee INDEX COMPARISONS 4.0 2 eee ee ee eee 4. ee eee ee ee ee ee 4.1 Comparing BBB, A, and AA+ index volatilities. ee ee eee 66 4.3 Q Q ee ee Technology Indexes.1 First volatility comparisons .2 Investment grade vs.
Non-investment grade indexes. Non-rated indexe.4 58+ 207 Drift rate comparisons.4 Basic Materials Indexes. 2 ee eee ee ee eee 4.1 High rating vs. Low rating vs.
Non-rated indexes. Low rating indexes.3 High rating vs -- BBBindexe. Non-rated indexe.46 Drit Rate Comparisons. Q Q Q Q SH HH ee ee viii CALIBRATION OF INDEX OPTIONS 119 5.
200 eee eee ee .2 Calibration by Maximum Entrop.1 General theory: PDE approach.2 Solution by Monte Carlo simulations.3 An illustrative example.3 Probability of Feasibility .3 Probability of Feasibility .4 Application to Index Options.1 Calibrating options on correlated stocks .2 High correlation comparisons -- .43 - - - {so Pricing options on an index. ee ee ee ee 159 CONCLUSION AND FUTURE RESEARCH 161 PROOFS OF MAIN RESULTS 164 COMPUTATION OF THE INDEXES 183 B. ee ee es 185 B.2 Standard & Poor’s Ratings.2-2 ee ee ee eee .- 191 INDEX FORMATION 197 SOME ADDITIONAL PROGRAMS 205 List of Tables 2.1 Dow Jones Global Indexes Industry Groups.2 Moody’s Long-Term Credit Rating Classes.3 Standard & Poor’s Long-Term Credit Rating Classes.1 3 dimensional test: 80%, 85%, 90% and 95% quantiles of 2 and —2log A.1 p-values for tests for equality of drifts, high and low rating basic ma- terials indexes. eee ee ee eee eee 116 9.1 Pricing different options with 2,000 scenarios: equal weighting for each scenario and entropy calibrated prices.2 Probabilities of feasibility.
Top row: number of benchmark assets. Leftmost column: Number of simulated scenarios.3 February 1, 2000: sixteen benchmark options for Cisco Systems. Cisco’s stock price was $109.4 February 1, 2000: Calibrated and observed prices for 25 Microsoft options. Microsoft’s stock price was $98.5 February 1, 2000: eight benchmark options for Oracle Corporation.
Oracle was trading at $50.6 Index options: Cisco is 50% on February 1, 2000. Fourth column are prices calibrated only to Cisco options.7 Performance of partial calibration algorithm. All quantities are per- centages.8 Performance of partial calibration algorithm. All quantities are per- centages.
QOQ Q Q Q Q Q Q ce eee 185 B.2 Moody’s ratings and numerical equivalents.3 Standard & Poor`s ratings and numerical equivalents. Q Q Q Q Q Q Q Q v2 196 List of Figures 2.1 S&P500 Index and log returns .2 Two Dow Jones Global Indexes. ee eee ee 26 2.3 Medium and Large Capitalization High-rating Telecommunications companies index. Top graph: index time series (left axis) and number of companies in the index (right axis).
Bottom graph: log returns series.4 Medium and Large Capitalization Low-rating Telecommunications com- panies index. Top graph: index time series (left axis) and number of companies in the index (right axis). Bottom graph: log returns series.1 Power of variance test: probability of not rejecting Hp when comparing a series of 30% variance with others of variances 80%-5% and correla- tion 5%.2 Quantile-quantile plot of two vectors of test statistics generated with different covariance matrices.3 Histogram of 95% quantile level of 10,000 test statistics generated with 192 different covariance matrices.4 Quantile-quantile plot of 12 distribution and 60 day window test statis- tic distribution.5 Quantile-quantile plot of x2 distribution and 1000 day window test statistic distribution.00 ee ee ee ee eee 51 41 Volatility estimates of the Financial AA+ index: GARCH a(t) and 30 day sample volatility.2 Volatility estimates of the Financial AA+ index: GARCH o(t) and 60 day sample volatility.3 Number of companies in four Financial sector indexes (from top to bottom): Rated companies index, High rating, Low rating and Non- Investment grade.4 Medium and Large Capitalization Financial companies rated AA or higher index. Top graph: index time series (left axis) and number of companies in the index (right axis).
Bottom graph: log returns series.5 Two Financial companies indexes. Top graph: companies rated A. Bottom graph: companies rated BBB.6 60 day sample volatilities (annualized) of Financial indexes: BBB, A and AA+.7 p-values for 60-day sample variance test: BBB, A and AA+ financial tre ‹ TA .8 p-values for 60-day sample variance test: A and AA+ financial indexes.9 Finacial Services Industry Groups indexes (left axes) and number of companies (right axes). ee ee ee eee 68 4.10 Finacial Services Industry Groups’ annualized volatilities.11 Percentage of capitalization of each Finacial Services Industry Group in AA + (top graph) and BBB (bottom graph) indexes.12 Composition of A index capitalization percentage.13 Several drift rate comparions involving Real Estate companies, some of which are close to being signifcant at the 95% levelL .14 Medium and Large Capitalization Investment grade Technology com- panies index.
Top graph: index time series (left axis) and number of companies in the index (right axis). Bottom graph: log returns series.15 Medium and Large Capitalization Non-investment grade Technology companies index. Top graph: index time series (left axis) and number of companies in the index (right axis). Bottom graph: log returns series.16 Medium and Large Capitalization Non-rated Technology companies index.
Top graph: index time series (left axis) and number of compa- nies in the index (right axis). Bottom graph: log returns series.17 Technology indexes 60 day sample volatilities (annualized): Non—investment grade, Non-rated and investment grade .18 p-values for variance test of three technology indexes.19 Technology indexes variance test. Top graph: annualized volatilities of investment and non~investment grade indexes. Bottom graph: p-— values for variance test and 5% conldence level .20 Medium and Large Capitalization Non-ratedTechnology companies in- dex.