Appendix for Solving Asset Pricing Models when the Price-Dividend Function is Analytic

Similar documents
332 Mathematical Induction Solutions for Chapter 14. for every positive integer n. Proof. We will prove this with mathematical induction.

Appendix - Normally Distributed Admissible Choices are Optimal

Multifactor Term Structure Models

Supplementary material for Non-conjugate Variational Message Passing for Multinomial and Binary Regression

Problem Set 6 Finance 1,

Consumption Based Asset Pricing

4.4 Doob s inequalities

ECE 586GT: Problem Set 2: Problems and Solutions Uniqueness of Nash equilibria, zero sum games, evolutionary dynamics

Fast Laplacian Solvers by Sparsification

arxiv: v1 [math.nt] 29 Oct 2015

Price and Quantity Competition Revisited. Abstract

Jean-Paul Murara, Västeras, 26-April Mälardalen University, Sweden. Pricing EO under 2-dim. B S PDE by. using the Crank-Nicolson Method

Foundations of Machine Learning II TP1: Entropy

Parallel Prefix addition

15-451/651: Design & Analysis of Algorithms January 22, 2019 Lecture #3: Amortized Analysis last changed: January 18, 2019

Understanding Annuities. Some Algebraic Terminology.

AMS Financial Derivatives I

Applications of Myerson s Lemma

3: Central Limit Theorem, Systematic Errors

Quiz on Deterministic part of course October 22, 2002

Tests for Two Correlations

Explaining Movements of the Labor Share in the Korean Economy: Factor Substitution, Markups and Bargaining Power

Maximum Likelihood Estimation of Isotonic Normal Means with Unknown Variances*

SOCIETY OF ACTUARIES FINANCIAL MATHEMATICS. EXAM FM SAMPLE SOLUTIONS Interest Theory

SUPPLEMENT TO BOOTSTRAPPING REALIZED VOLATILITY (Econometrica, Vol. 77, No. 1, January, 2009, )

Economic Design of Short-Run CSP-1 Plan Under Linear Inspection Cost

On the Moments of the Traces of Unitary and Orthogonal Random Matrices

COS 511: Theoretical Machine Learning. Lecturer: Rob Schapire Lecture #21 Scribe: Lawrence Diao April 23, 2013

Numerical Analysis ECIV 3306 Chapter 6

Linear Combinations of Random Variables and Sampling (100 points)

Dependent jump processes with coupled Lévy measures

Production and Supply Chain Management Logistics. Paolo Detti Department of Information Engeneering and Mathematical Sciences University of Siena

Lecture 7. We now use Brouwer s fixed point theorem to prove Nash s theorem.

EDC Introduction

c slope = -(1+i)/(1+π 2 ) MRS (between consumption in consecutive time periods) price ratio (across consecutive time periods)

On the discretization schemes for the CIR (and Bessel squared) processes

2.1 Rademacher Calculus... 3

TCOM501 Networking: Theory & Fundamentals Final Examination Professor Yannis A. Korilis April 26, 2002

Mode is the value which occurs most frequency. The mode may not exist, and even if it does, it may not be unique.

Finance 402: Problem Set 1 Solutions

Elements of Economic Analysis II Lecture VI: Industry Supply

arxiv: v2 [math.co] 6 Apr 2016

YORK UNIVERSITY Faculty of Science Department of Mathematics and Statistics MATH A Test #2 November 03, 2014

Basket options and implied correlations: a closed form approach

2) In the medium-run/long-run, a decrease in the budget deficit will produce:

4. Greek Letters, Value-at-Risk

Which of the following provides the most reasonable approximation to the least squares regression line? (a) y=50+10x (b) Y=50+x (d) Y=1+50x

Project Management Project Phases the S curve

Tests for Two Ordered Categorical Variables

arxiv: v3 [math.pr] 12 Dec 2017

Formation of Coalition Structures as a Non-Cooperative Game

Principles of Finance

Equilibrium in Prediction Markets with Buyers and Sellers

Option pricing and numéraires

Final Exam. 7. (10 points) Please state whether each of the following statements is true or false. No explanation needed.

Dept of Mathematics and Statistics King Fahd University of Petroleum & Minerals

Elton, Gruber, Brown and Goetzmann. Modern Portfolio Theory and Investment Analysis, 7th Edition. Solutions to Text Problems: Chapter 4

Finite Math - Fall Section Future Value of an Annuity; Sinking Funds

COMPARISON OF THE ANALYTICAL AND NUMERICAL SOLUTION OF A ONE-DIMENSIONAL NON-STATIONARY COOLING PROBLEM. László Könözsy 1, Mátyás Benke 2

OPERATIONS RESEARCH. Game Theory

Games and Decisions. Part I: Basic Theorems. Contents. 1 Introduction. Jane Yuxin Wang. 1 Introduction 1. 2 Two-player Games 2

/ Computational Genomics. Normalization

A Constant-Factor Approximation Algorithm for Network Revenue Management

Financial mathematics

IND E 250 Final Exam Solutions June 8, Section A. Multiple choice and simple computation. [5 points each] (Version A)

Analysis of Variance and Design of Experiments-II

Comparison of Singular Spectrum Analysis and ARIMA

ACTA UNIVERSITATIS APULENSIS No 16/2008 RISK MANAGEMENT USING VAR SIMULATION WITH APPLICATIONS TO BUCHAREST STOCK EXCHANGE. Alin V.

Data Mining Linear and Logistic Regression

The first step in using market prices

Risk and Return: The Security Markets Line

SIMPLE FIXED-POINT ITERATION

A Study on the Series Expansion of Gaussian Quadratic Forms

An annuity is a series of payments made at equal intervals. There are many practical examples of financial transactions involving annuities, such as

A Comparison of Statistical Methods in Interrupted Time Series Analysis to Estimate an Intervention Effect

Taxation and Externalities. - Much recent discussion of policy towards externalities, e.g., global warming debate/kyoto

Elton, Gruber, Brown, and Goetzmann. Modern Portfolio Theory and Investment Analysis, 7th Edition. Solutions to Text Problems: Chapter 16

Raising Food Prices and Welfare Change: A Simple Calibration. Xiaohua Yu

arxiv: v2 [q-fin.pr] 12 Oct 2013

Random Variables. b 2.

Homework 9: due Monday, 27 October, 2008

Wages as Anti-Corruption Strategy: A Note

MgtOp 215 Chapter 13 Dr. Ahn

Interval Estimation for a Linear Function of. Variances of Nonnormal Distributions. that Utilize the Kurtosis

CS 286r: Matching and Market Design Lecture 2 Combinatorial Markets, Walrasian Equilibrium, Tâtonnement

occurrence of a larger storm than our culvert or bridge is barely capable of handling? (what is The main question is: What is the possibility of

Dept of Mathematics and Statistics King Fahd University of Petroleum & Minerals

Fall 2016 Social Sciences 7418 University of Wisconsin-Madison. Transactions and Portfolio Crowding Out

Spring 2010 Social Sciences 7418 University of Wisconsin-Madison. The Financial and Economic Crisis Interpreted in a CC-LM Model

II. Random Variables. Variable Types. Variables Map Outcomes to Numbers

Osaka University of Economics Working Paper Series No Hart Mas-Colell Implementation of the Discounted Shapley Value

Measures of Spread IQR and Deviation. For exam X, calculate the mean, median and mode. For exam Y, calculate the mean, median and mode.

Real Exchange Rate Fluctuations, Wage Stickiness and Markup Adjustments

arxiv:cond-mat/ v1 [cond-mat.other] 28 Nov 2004

PhysicsAndMathsTutor.com

Elton, Gruber, Brown, and Goetzmann. Modern Portfolio Theory and Investment Analysis, 7th Edition. Solutions to Text Problems: Chapter 9

A political-economic analysis of free-trade agreements: Comment

Final Examination MATH NOTE TO PRINTER

Centre for International Capital Markets

Functional Itô calculus and Applications. David-Antoine FOURNIE. Supervised by Professor Rama CONT

Transcription:

Appendx for Solvng Asset Prcng Models when the Prce-Dvdend Functon s Analytc Ovdu L. Caln Yu Chen Thomas F. Cosmano and Alex A. Hmonas January 3, 5 Ths appendx provdes proofs of some results stated n our paper. Frst, we prove the exstence and unqueness of the prce-dvdend functon Proposton to the ntegral equaton 6 n the vector space S Defnton. Second, we calculate the system of lnear equatons for the coeffcents of the power seres for the prce-dvdend functon. Fnally, we use Cauchy s ntegral formula to bound all the dervatves of the prce-expected dvdend functon 3 so that we can calculate the error terms 4, 5 and 6. Journal of Economc Lterature Classfcaton Numbers: G, G3, C63, D5 Key Words: Analytcty, Asset prcng. Department of Mathematcs, Eastern Mchgan Unversty. Department of Mathematcs, Idaho State Unversty, Pocatello, ID 839 Department of Fnance, Mendoza College of Busness, Unversty of Notre Dame, E-mal: thomas.f.cosmano.@nd.edu, Phone: 574 63-578, Fax: 574 63-555. Department of Mathematcs, Unversty of Notre Dame, Notre Dame, IN 46556.

APPENDIX We use the followng notaton n the proofs: Cn = n = φ = φ n n =,,,..., Φx = π x e s K 3 = K e x K + σ γ φ α+φ α, K 4 = K 3 [ + Φ φk σ for ds for all x R, θ = x + σ + φ α γ, ], K 5 = K 4 e φ K σ + φk θ and K 6 = π K e σ γ + σ [ γ+k ]. Q s defned n 8 whch satsfes the ntegral equaton 9. A seres of lemmas are now proved whch are used n the proof of Proposton. s x a e Lemma : Let a > and ϕx = x+a Φ a σ. σ ds for all x R. Then ϕx Proof: Ths follows from the fact that ϕ has a unque global maxmum at x =. Lemma 3: For any real numbers A and k wth k, we have e σ s A +k s ds e k σ +k A [ + Φkσ]. Proof: Note that σ s A + ks = σ s A kσ + k σ + ka. e σ s A +k s ds = = e k σ e ka e k σ +k A e σ s A ks ds + e = e k σ +k A A+kσ e σ s A +ks ds σ s A+kσ ds + e ka e σ s A+kσ ds + = e k σ +k A + A+kσ e s σ ds + e s σ ds A kσ A kσ e s σ ds By Lemma, we acheve e σ s A +k s ds e k σ +k A [ + Φkσ]. e σ s A kσ ds e σ s A kσ ds

Lemma 4: For any f S, the functon T f defned by T fx = K 3e φk x fse σ [s φx+θ] ds les n the vector space S. Thus, T defnes a lnear transformaton from the vector space S to tself. Proof: By the defnton of S, we can fnd postve constants M and k such that fx Me k x. Usng Lemma 3, we deduce T fx K 3e φk x MK 3e φk x fs e σ [s φx+θ] ds e σ [s φx+θ] +k s ds MK 3 [ + Φkσ]e φk x+ k σ +k φ x + θ MK 3 e k σ +k θ [ + Φkσ]e φ k+ K x. The contnuty of T f can be easly checked by applyng the standard argument n real analyss; for example, see Theorem 56 n Kaplan 956. Now construct a sequence of functons { Q n n =,,,... } by settng Q = and Q n+ x = K + K 3e φk x and set Q = n= [Q n+ Q n ]. Recall that Cn = n = φ = φ n Cn < φ, and c lm n Cn = φ. Lemma 5: Q n+ x Q n x K K4 n e σ K n =,,,.... Q n se σ [s φx+θ] ds satsfes a Cn + = φ Cn +, b n = C + K θ n = C e Cn K x for

Proof: These nequaltes can be easly proven by nducton on n. Lemma 6: The seres Q = n= [Q n+ Q n ] s unformly convergent on any bounded closed nterval. Proof: Fx a bounded closed nterval [a, b]. For any x [a, b], by Lemma 5 we see [Q n+ x Q n x] K n= n= K4 n e σ K n = C + K θ n = C e Cn K a + b. The rato test together wth Weerstrass M-test see Theorem 3 Kaplan 956 mples the unform convergence of the seres Q = n= [Q n+ Q n ] on [a, b], provded lm K 4e σ K Cn + K θ Cn e [Cn+ Cn] K a + b = K 4 e σ n Lemma 7: The functon Q les n the vector space S. K φ + φk θ = K 5 <. Proof: The contnuty of Q follows from Theorem 3 n Kaplan 956 and Lemma 6. By Lemma 5, we also have Qx K n= K n= K4 n e σ K n = C + K θ n = C e Cn K x K4 n e σ K n = C + K θ n = C e φk x. 6. Lemma 8: For any x R, we have lm n [Qs Q ns]e σ [s φx+θ] ds = Proof: Ths s an mmedate consequence of Theorem 3 n Kaplan 956 and Lemma 3

Proof of Proposton : Equvalently, we are gong to show that the equaton 9 has a unque soluton Q n the vector space S. Exstence Applyng the lmt as n approaches to the equaton shows that the functon Q = n= [Q n+ Q n ] = lm n Q n s a soluton to the equaton 9 n the vector space S. Unqueness Suppose that Q S s another soluton to the equaton 9. Then Q and Q satsfy the functonal equaton Q Q = T Q Q. By the defnton of S, we can also fnd postve constants M and k such that Qx Qx Me k x. Usng the results n the proof of Lemma 4, we see Qx Qx T Q Q x MK 3 e k σ +k θ [ + Φkσ]e φ k+c K x for all x R. Applyng ths computaton n tmes gves rse to Qx Qx { n } MK3 n e φ k+c K σ + φ k+c K θ [ + Φσ φ k + σc K ] e φ n k+cn K x. Note that lm K 3e φ n k+cn K σ + φ n k+cn K θ [ + Φσ φ n k + σcn K ] n = K 3 e φ K σ + φk θ [ + Φ ] φk σ = K 5 <. We can fnd a postve nteger N and a postve real number δ < so that for any n N, K 3 e φ n k+cn K σ + φ n k+cn K θ [ + Φσ φ n k + σcn K ] δ. Then for all n N, we have Qx Qx { N } MK3 N e φ k+c K σ + φ k+c K θ [ + Φσ φ k + σc K ] δ n N e φ n k+cn K x. 4

Snce lm n δ n N = and lm n e φ n k+cn K x = e φ K x, we obtan Qx = Qx for all x R. Ths completes the proof of Proposton Solvng for the coeffcents n the analytc Prce-Dvdend Functon The ntegral equaton 6 may be wrtten as P x = K e K x + K e K x σ γ Ix. where Ix = s e γs σ P x + φx + sds. To solve the ntegral equaton we post that P x s analytc n whch case t has the functonal form. Consequently we may wrte P x + φx + s as P x + φx + s = e K x +φx+s x + φx + s x k. Usng the Bnomal theorem 6 we may rewrte ths equaton as P x + φx + s = e K x +φx+s Substtutng ths result nto Ix we obtan Ix = e K x +φx x + φx x k s. x + φx x k γ where γ = e γ+k s s σ s ds. The evaluaton of the ntegral γ entals the use of a change of varable followed by the evaluaton of all the moments of the normal dstrbuton. γ = K 3 a where K 7 = π e σ [ γ+k ] and a j= σ [ γ + K ] j σ j + j /Γ[ j+ ]. We can now wrte Ix as K 7 e K x +φx n k the ntegral equaton becomes e K x P x = K + K 6 e K x +φx 5 x + φx x k a = j x + φx x k a so that x x k.

Next we fnd the undetermned coeffcents b, =,... n. In order to do ths, we collect all the functons of x on the left hand sde of ths equaton and use Taylor s theorem to take an n th order Taylor expanson around x. Ths produces a system of n + lnear equatons n the varables b, =,... n. Defne the followng functons qx = K x + φx, r,k x = x x + φx k, and w,k x = expqxr,k x n l= l! wl,k x x x l. Substtutng for w,k x nto the ntegral equaton gves us K + K 6 l= l! wl,k x x x l a = x x k. Fnally, equate the coeffcents on the left and rght hand sde of ths equaton to yeld b = K + K 6 w,k x a and b l = K 6 l! wl,k x a, where l =,..., n. These Equatons are an n + system of lnear equatons n the b s. Analytc error of the Taylor polynomal approxmaton seres The complex functon Qz = e K z P z s analytc and t s expressble as a Taylor Qz = Q k z z z k for any complex number z, k! where z = x. We may use Cauchy s ntegral formula to estmate the error of the Taylor polynomal approxmaton Qz n Q k z z z k! k near z = z. Wrte C r for the crcle of radus r > centered at z = z n the complex plane. Cauchy s ntegral formula see Corollary 5.9 n Conway 973 gves Q k z = k! π C r Qz dz for k =,,,.... z z k+ Each pont z = x + y on C r satsfes x r x x + r and r y r. Set A = K n= K4 n e σ K n = C + K θ n = C. 6

Usng the result n the proof of Lemma 7, we obtan Qs Ae φk s for all s R, and Qz K + K 3 e φkx+y = K + K 3e φk φ x+ y σ K + AK 3e φk φ x + y σ Qs e [s φx θ φy] σ ds Qse σ [s φx θ] ds e σ [s φx θ] + φk s ds K + AK 3 e φk x + φ K x + φ y σ + φ K σ + φk θ [ + Φ = K + AK 3 e φk x + θ + φ y σ + φ K σ [ + Φ φk σ K + AK 3 e φk r+ x + θ + φ r σ + φ K σ Defne B r = K + AK 3 e φk r+ x + θ + φ r σ + φ K σ ] [ + Φ φk σ [ + Φ φk σ ] φk σ ]. ]. Cauchy s ntegral formula for Q k z together wth the above estmate to Qz gves rse to Q k z k! Qz dz π z z k+ = k! π Qz + re θ π re θ dθ re θ k+ k! π C r π Qz + re θ r k dθ B rk! r k. Ths bound allows us to estmate the analytc error 4. FOOTNOTES 6 The bnomal theorem states that we may wrte x +φx+s x k = k φx x k s. x + 7