Credit risk of derivative securities includes the risk of underlying company and the risk of seller's nonfulfilment of contracts. Take bonds for example, we regard Treasury bills as default-free bonds, and corporate bonds as risky bonds. When the liability of property of derivative securities underlying company is less than 1, we regard the company is of bankruptcy. And then the seller of derivative securities will break the contract. The essay extends two period risky bonds pricing valuation of Jarrow and Turnbull(1995) to multiperiod situation, and derive arbitrage-free condition. Furthermore, we derive formulae of risky bonds prices by assuming the logarithm of the odds ratio of an underlying company's bankruptcy probability satisfies an AR(1) or MA(1) processes. Empirical data of Rebar, Chinarebar, Ceon are studied, time series models are established for logarithm of odds ratios. In most cases, we find that the log odds ratios can be well fitted by AR(1) models.

Les travaux de la thèse portent sur les théorèmes limites pour des modèles stochastiques à forte dépendance. Dans la première partie, nous considérons des modèles AR(1) à coefficient aléatoire. Nous identifions trois régimes asymptotiques différents pour le schéma d’agrégation conjointe temporelle-contemporaine lorsque les processus AR sont indépendants et lorsque les AR possède des innovations communes. Ensuite, on discute de l’estimation non paramétrique de la fonction de répartition du coefficient autorégressif à partir d’un panel de séries AR(1) à coefficient aléatoire. Nous prouvons la convergence faible du processus empirique basé sur des estimations des coefficients autorégressifs non observables vers un pont brownien généralisé. Ce résultat est ensuite appliqué pour valider différents outils d’inférence statistique à partir des données du panel AR(1). Dans la deuxième partie de la thèse, nous nous concentrons sur les modèles spatiaux en dimension 2. Nous considérons des champs aléatoires construits à partir des polynômes Appell et de champs aléatoires linéaires. Pour ce modèle non linéaire, nous étudions la limite de ses sommes partielles normalisées prises sur des rectangles et prouvons l’existence d’une transition d’échelle. Enfin, nous abordons la même question pour le modèle de germes-grains aléatoire. Nous mettons en évidence l’existence de deux points de transition dans les limites de ces modèles. / The thesis is devoted to limit theorems for stochastic models with long-range dependence. We first consider a random-coefficient AR(1) process, which can have long memory provided the distribution of autoregressive coefficient concentrates near the unit root. We identify three different limit regimes in the scheme of joint temporal-contemporaneous aggregation for independent copies of random-coefficient AR(1) process and for its copies driven by common innovations. Next, we discuss nonparametric estimation of the distribution of the autoregressive coefficient given multiple random-coefficient AR(1) series. We prove the weak convergence of the empirical process based on estimates of unobservable autoregressive coefficients to a generalized Brownian bridge and apply this result to draw statistical inference from panel AR(1) data. In the second part of the thesis we focus on spatial models in dimension 2. We define a nonlinear random field as the Appell polynomial of a linear random field with long-range dependence. For the nonlinear random field, we investigate the limit of its normalized partial sums over rectangles and prove the existence of scaling transition. Finally, we study such like scaling of the random grain model and obtain two-change points in its limits.

In the doctoral dissertation, we consider problems of testing and estimating changed segment with unknown starting position and duration of epidemic state in the autoregressive first-order model. The proposed tests are based on partial sums of model residuals and model-parameter partial-estimator polygonal line processes. We derive asymptotic results for these processes in Holder spaces. The behavior of test statistics under the null hypothesis of no change and alternative is provided. Empirical power analysis has shown that tests are more powerful when absolute values of model parameter are quite large or autoregressive process changes from a stationary state to a nonstationary one. We prove the consistency of the least square changed-segment estimators and provide their convergence rates. / Disertacijoje nagrinėjamas pirmos eilės autoregresinio modelio pasikeitusio segmento testavimo ir vertinimo uždavinys. Aprašomo modelio epideminio pasikeitimo pradžia ir ilgis nėra žinomi. Pasiūlyti kriterijai pasikeitusio segmento testavimui, kurie pagrįsti modelio paklaidų įvertinių dalinių sumų ir modelio parametro dalinių įvertinių laužčių procesais. Šiems procesams gautos ribinės teoremos Hiolderio erdvėse. Nurodomas testų statistikų ribinis elgesys esant teisingai nulinei ir alternatyviajai hipotezėms. Iš empirinio kriterijų galios tyrimo rezultatų matyti, kad pasiūlytų testų galia didžiausia aptinkant pasikeitimus iš stacionarios būklės į nestacionarią arba esant artimoms vienetui modelio parametro reikšmėms. Taip pat įrodoma, kad mažiausių kvadratų metodu gauti pasikeitusio segmento pradžios ir ilgio įverčiai bei autoregresinio modelio su pasikeitusiu segmentu parametrų įverčiai yra suderintieji bei pateikiamas jų konvergavimo greitis.

Disertacijoje nagrinėjamas pirmos eilės autoregresinio modelio pasikeitusio segmento testavimo ir vertinimo uždavinys. Aprašomo modelio epideminio pasikeitimo pradžia ir ilgis nėra žinomi. Pasiūlyti kriterijai pasikeitusio segmento testavimui, kurie pagrįsti modelio paklaidų įvertinių dalinių sumų ir modelio parametro dalinių įvertinių laužčių procesais. Šiems procesams gautos ribinės teoremos Hiolderio erdvėse. Nurodomas testų statistikų ribinis elgesys esant teisingai nulinei ir alternatyviajai hipotezėms. Iš empirinio kriterijų galios tyrimo rezultatų matyti, kad pasiūlytų testų galia didžiausia aptinkant pasikeitimus iš stacionarios būklės į nestacionarią arba esant artimoms vienetui modelio parametro reikšmėms. Taip pat įrodoma, kad mažiausių kvadratų metodu gauti pasikeitusio segmento pradžios ir ilgio įverčiai bei autoregresinio modelio su pasikeitusiu segmentu parametrų įverčiai yra suderintieji bei pateikiamas jų konvergavimo greitis. / In the doctoral dissertation, we consider problems of testing and estimating changed segment with unknown starting position and duration of epidemic state in the autoregressive first-order model. The proposed tests are based on partial sums of model residuals and model-parameter partial-estimator polygonal line processes. We derive asymptotic results for these processes in Holder spaces. The behavior of test statistics under the null hypothesis of no change and alternative is provided. Empirical power analysis has shown that tests are more powerful when absolute values of model parameter are quite large or autoregressive process changes from a stationary state to a nonstationary one. We prove the consistency of the least square changed-segment estimators and provide their convergence rates.

Binomial AR(1) process is a model for integer-valued time series with a fi- nite range and discrete time. It has the binomial marginal distribution and the AR(1)-like autocorrelation structure. This thesis deals with deriving some ba- sic properties of this process, methods of parameter estimation and goodness of fit testing. Three methods of parameter estimation are presented: Yule-Walker, the conditional least squares and the maximum likelihood method together with proofs of their asymptotical properties. Next, the goodness of fit testing is pre- sented. At first, two known methods based on the marginal distribution and the autocorrelation function are summarized. Then our own method is added, based on the probability generating function. Several simulations are provided to show the properties of all tests. The application of this model is illustrated on a real dataset. 1

My final thesis firstly addresses basic knowledge of the theory of stochastic processes. This is firstly due to the author's effort to make the thesis more comprehensible, and also due to the need for introduction of key concepts. The autoregressive model AR(1) is defined in the thesis through basic linear time series models, and in this model, the estimation of model parameter by the method of least squares is introduced. For this estimation, the theoretical findings of the thesis are extended through the classical limit theory. Furthermore, the models with their parameter dependent on number of observations are introduced and models of NIAR (1) are defined. Classical limit theory for least squares estimation is then enriched by the limit theory in these models. The category of more general models is introduced and using the acquired knowledge, the features for the model AR (1) are derived. This thesis deals with this issue in models of NIAR (1) and its area of interest is also the bootstrap. The theoretical part of the thesis is supplemented by a practical part represented by numerical studies.

ABSTRACT PORTFOLIO SELECTION AND RETURN PERFORMANCE: An Application of the Black-Litterman Method in the Istanbul Stock Exchange Bozdemir, Mehmet Burak M.Sc, Department of Financial Mathematics Supervisor : Assist. Prof. Dr. Seza Dani

Darbe nagrinėjamas AR(1) procesų $X_{i,t} = a_i X_{i,t-1} + \vep_t, \ i=1,\cdots, N$ su atsitiktiniais n.v.p. koeficientais $a_i \in (-1,1)$ ir n.v.p. bendrais triukšmais $\{\vep_t\}$ agregavimas, kai triukšmai priklauso $\alpha-$stabilaus dėsnio normaliajai traukos sričiai, $(0< \alpha \le 2)$. Nagrinėjamas atvejis, kai atsitiktinių koeficientų tikimybinis tankis auga į begalybę taškuose $a = 1 $ ir $a=-1$. Gautos sąlygos, kurioms esant egzistuoja ribinis agreguotas procesas $\bar X_t = \lim_{N \to \infty} N^{-1}\sum_{i=1}^N X_{i,t} $, išnagrinėta kada jis turi ilgalaikę atmintį. Taip pat parodyta, kad atitinkamai normuotos $\bar X_t$ dalinės sumos konverguoja į trupmeninį $\alpha-$stabilų judesį. Ir esant tam tikroms sąlygoms ribinis agreguotas procesas $\{\bar X_t\}$ turi LRD(SAV) (angl. long-range dependence (sample Allen variance)) sąvybę, bei skirstinių ilgąją atmintį. Šis darbas išplečia kai kuriuos P. Zaffaroni rezultatus nuo baigtinės dispersijos atvejo iki begalinės dispersijos atvejo. / Aggregation of random coefficient AR(1) processes $X_{i,t} = a_i X_{i,t-1} + \vep_t, \ i=1,\cdots, N$ with i.i.d. coefficients $a_i \in (-1,1)$ and common i.i.d. innovations $\{\vep_t\}$ belonging to the domain of attraction of $\alpha-$stable law $(0< \alpha \le 2)$ is discussed. Particular attention is given to the case of slope coefficient having probability density growing regularly to infinity at points $a = 1 $ and $a=-1$. Conditions are obtained under which the limit aggregate $\bar X_t = \lim_{N \to \infty} N^{-1}\sum_{i=1}^N X_{i,t} $ exists and exhibits long memory, in certain sense. In particularly, I show that suitably normalized partial sums of the $\bar X_t$'s tend to fractional $\alpha-$stable motion, and that $\{\bar X_t\}$ satisfies the long-range dependence (sample Allen variance) property and distributional long memory. The present paper also extends some results of P. Zaffaroni from finite variance case to infinite variance case.

The purpose of this master’s thesis is to investigate to what extent ESG assets and traditional benchmarks affect one another. Since sustainable investment is a growing segment of the financial market, investors need to be informed about how it may affect their portfolios, and by extension if it can be used for portfolio diversification. By using an AR(1)-GARCH(p,q) model and a Diebold-Yilmaz spillover approach, we can measure the spillover effects between ESG indices and other benchmark indices for both return and volatility. We find that country-level ESG indices are more integrated with other country-level ESG indices than other assets, and that country-level ESG indices transmit more to the MSCI world ESG index, MSCI world equity index, Crude oil, Gold, and our currency index EUR/USD. These findings hold true for both return and volatility spillover. Thus, our policy implications are that including country-level ESG assets in the portfolio can decrease portfolio risk and help minimize the contagious effects of shocks on the portfolio.

<p>In recent years, two estimators have been proposed to correct the bias exhibited by the leastsquares (LS) estimator of the lagged dependent variable (LDV) coefficient in dynamic regression models when the sample is finite. They have been termed as &lsquo / mean-unbiased&rsquo / and &lsquo / medianunbiased&rsquo / estimators. Relative to other similar procedures in the literature, the two locationbased estimators have the advantage that they offer an exact and uniform methodology for LS estimation of the LDV coefficient in a first order autoregressive model with or without exogenous regressors i.e. ARX(1).</p> <p><br /> However, no attempt has been made to accurately establish and/or compare the statistical properties among these estimators, or relative to those of the LS estimator when the LDV coefficient is restricted to realistic values. Neither has there been an attempt to&nbsp / compare their performance in terms of their mean squared error (MSE) when various forms of the exogenous regressors are considered. Furthermore, only implicit confidence intervals have been given for the &lsquo / medianunbiased&rsquo / estimator. Explicit confidence bounds that are directly usable for inference are not available for either estimator. In this study a new estimator of the LDV coefficient is proposed / the &lsquo / most-probably-unbiased&rsquo / estimator. Its performance properties vis-a-vis the existing estimators are determined and compared when the parameter space of the LDV coefficient is restricted. In addition, the following new results are established: (1) an explicit computable form for the density of the LS estimator is derived for the first time and an efficient method for its numerical evaluation is proposed / (2) the exact bias, mean, median and mode of the distribution of the LS estimator are determined in three specifications of the ARX(1) model / (3) the exact variance and MSE of LS estimator is determined / (4) the standard error associated with the determination of same quantities when simulation rather than numerical integration method is used are established and the methods are compared in terms of computational time and effort / (5) an exact method of evaluating the density of the three estimators is described / (6) their exact bias, mean, variance and MSE are determined and analysed / and finally, (7) a method of obtaining the explicit exact confidence intervals from the distribution functions of the estimators is proposed.</p> <p><br /> The discussion and results show that the estimators are still biased in the usual sense: &lsquo / in expectation&rsquo / . However the bias is substantially reduced compared to that of the LS estimator. The findings are important in the specification of time-series regression models, point and interval estimation, decision theory, and simulation.</p>