Panel data analysis has become an important tool in applied econometrics and the respective statistical techniques are well described in several recent textbooks. However, for an analyst using these methods there remains the task of choosing a reasonable model for the behavior of the panel data. Of special importance is the choice between so-called fixed and random coefficient models. This choice can have a crucial effect on the interpretation of the analyzed phenomenon, which is demonstrated by an application on research and development spillover. (author's abstract) / Series: Forschungsberichte / Institut für Statistik

This thesis studies the value of branding in the smartphone market. Measuring brand value with data available at product level potentially entails computational and econometric challenges due to data constraints. These issues motivate the three studies of the thesis. Chapter 2 studies the smartphone market to understand how operating system platform providers can grow one of the most important intangible assets, i.e., brand value, by leveraging the indirect network between two user groups in a two-sided platform. The main finding is that iPhone achieved the greatest brand value growth by opening its platform to the participation of third-party developers, thereby indirectly connecting the consumers and the developers via its app store effectively. Without the open app store, I find that iPhone would have lost its brand value by becoming a two-sided platform. Hence these findings provide an important lesson that open platform strategy is vital to the success of building platform brands. Chapter 3 solves a computational challenge in structural estimation of aggregate demand. I develop a computationally efficient MCMC algorithm for the GMM estimation framework developed by Berry, Levinsohn and Pakes (1995) and Gowrisankaran and Rysman (forthcoming). I combine the MCMC method with the classical approach by transforming the GMM into a Laplace type estimation framework, therefore avoiding the need to formulate a likelihood model. The proposed algorithm solves the two fixed point problems, i.e., the market share inversion and the dynamic programming, incrementally with MCMC iteration. Hence the proposed approach achieves computational efficiency without compromising the advantages of the conventional GMM approach. Chapter 4 reviews recently developed econometric methods to control for endogeneity bias when the random slope coefficient is correlated with treatment variables. I examine how standard instrumental variables and control function approaches can solve the slope endogeneity problem under two general frameworks commonly used in the literature.

This thesis studies the value of branding in the smartphone market. Measuring brand value with data available at product level potentially entails computational and econometric challenges due to data constraints. These issues motivate the three studies of the thesis. Chapter 2 studies the smartphone market to understand how operating system platform providers can grow one of the most important intangible assets, i.e., brand value, by leveraging the indirect network between two user groups in a two-sided platform. The main finding is that iPhone achieved the greatest brand value growth by opening its platform to the participation of third-party developers, thereby indirectly connecting the consumers and the developers via its app store effectively. Without the open app store, I find that iPhone would have lost its brand value by becoming a two-sided platform. Hence these findings provide an important lesson that open platform strategy is vital to the success of building platform brands. Chapter 3 solves a computational challenge in structural estimation of aggregate demand. I develop a computationally efficient MCMC algorithm for the GMM estimation framework developed by Berry, Levinsohn and Pakes (1995) and Gowrisankaran and Rysman (forthcoming). I combine the MCMC method with the classical approach by transforming the GMM into a Laplace type estimation framework, therefore avoiding the need to formulate a likelihood model. The proposed algorithm solves the two fixed point problems, i.e., the market share inversion and the dynamic programming, incrementally with MCMC iteration. Hence the proposed approach achieves computational efficiency without compromising the advantages of the conventional GMM approach. Chapter 4 reviews recently developed econometric methods to control for endogeneity bias when the random slope coefficient is correlated with treatment variables. I examine how standard instrumental variables and control function approaches can solve the slope endogeneity problem under two general frameworks commonly used in the literature.

In this dissertation, I consider linear, binary response correlated random coefficient (CRC) panel data models and a truncated CRC panel data model which are frequently used in economic analysis. I focus on the nonparametric identification and estimation of panel data models under unobserved heterogeneity which is captured by random coefficients and when these random coefficients are correlated with regressors. For the analysis of linear CRC models, I give the identification conditions for the average slopes of a linear CRC model with a general nonparametric correlation between regressors and random coefficients. I construct a sqrt(n) consistent estimator for the average slopes via varying coefficient regression. The identification of binary response panel data models with unobserved heterogeneity is difficult. I base identification conditions and estimation on the framework of the model with a special regressor, which is a major approach proposed by Lewbel (1998, 2000) to solve the heterogeneity and endogeneity problem in the binary response models. With the help of the additional information on the special regressor, I can transfer a binary response CRC model to a linear moment relation. I also construct a semiparametric estimator for the average slopes and derive the sqrt(n)-normality result. For the truncated CRC panel data model, I obtain the identification and estimation results based on the special regressor method which is used in Khan and Lewbel (2007). I construct a sqrt(n) consistent estimator for the population mean of the random coefficient. I also derive the asymptotic distribution of my estimator. Simulations are given to show the finite sample advantage of my estimators. Further, I use a linear CRC panel data model to reexamine the return from job training. The results show that my estimation method really makes a difference, and the estimated return of training by my method is 7 times as much as the one estimated without considering the correlation between the covariates and random coefficients. It shows that on average the rate of return of job training is 3.16% per 60 hours training.

This study uses a random coefficient estimation procedure to test the hypothesis that much of the volatility in the U.S./Canada real exchange rate over the time period 1971 through 1999 is due to the relative price of non-traded goods to traded goods. The model specification used in this study provides estimates of the sensitivity of movements in the U.S./Canada real exchange rate to movements in both the relative price of traded goods and the relative price of non-traded goods to traded goods in each of the two countries. I test for purchasing power parity in each of the two components of the model and address the question of volatility through the examination of the time profile of the respective coefficient estimates. The empirical results support the conclusion that the average value of the coefficient on the relative price of non-traded goods to traded goods component is smaller than that on the relative price of traded goods component. However, purchasing power parity in both components can not be rejected when the period of study is limited to 1971 through 1994. Furthermore, examination of the time profile of the random coefficients on the relative price of non-traded goods to traded goods component suggests that it is much more volatile and, therefore, quite significant in capturing the volatility in U.S./Canada real exchange rate movements. With regard to purchasing power parity in both the traded goods component and the non-traded goods to traded goods component, these results are consistent with the implications of the theory of purchasing power parity. However, they are not entirely consistent with the evidence presented in recent literature. Specifically, evidence presented in recent studies can not support perfect purchasing power parity in either traded goods or non-traded goods and leads to the conclusion that non-traded goods are much less significant, if at all, in the determination of the U.S./Canada real exchange rate. This inconsistency with recent literature is most likely a result of the fact that the random coefficient modeling technique used in this study allows the coefficients to vary over time and, thereby, enables the volatility of both components to be captured in the model. Therefore, given the apparent significance of the relative price of non-traded goods to traded goods, the volatility of this component can logically be expected to significantly contribute to the volatility in the U.S./Canada real exchange rate. / Master of Arts

No description available.

No description available.

This work investigates mainly panel data models in which cross-sections can be considered independent. In the first part, we summarize results in the field of pool models and one-way error component models with fixed and random effects. We focus especially on the ways of estimating unknown parameters and on effects significance tests. We also briefly describe two-way error component model issues. In the second part, estimators of first order autoregressive panel data model parameters are derived, for both fixed and random parameters case. The work proves unbiasedness, consistency and asymptotic normality of selected estimators. Using these features, hypothesis tests about corresponding parameters are derived. Application of models is illustrated using real data and simulated data examples. Powered by TCPDF (www.tcpdf.org)

Diese Dissertation umfasst drei Aufsätze zu verschiedenen Themen aus dem Bereich der Mikroökonometrie. Das erste Kapitel ist eine gemeinsame Arbeit mit Christoph Breunig und umfasst semi/nichtparametrische Regressionsmodelle, in denen die abhängige Variable einen nicht-klassischen Messfehler aufweist. Es werden Bedingungen erarbeitet, unter denen die Regressionsfunktion bis auf eine Normalisierung identifiziert werden kann. Zur Schätzung wird ein neuer Schätzer entwickelt, bei dem eine Rang-basierte Kriteriumsfunktion über einen sieve-Raum optimiert wird und dessen Konvergenzrate hergeleitet. Das zweite Kapitel beschäftigt sich mit der Schätzung von bedingten Dichtefunktionen von zufälligen Koeffizienten in linearen Regressionsmodellen. Es wird ein zweistufiges Schätzverfahren entwickelt, in dem zunächst eine Approximation der bedingten Dichte Koeffizienten hergeleitet wird. In einem weiteren Schritt können diese Funktionen mit generischen Methoden des maschinellen Lernens geschätzt werden. Des Weiteren wird auch die Konvergenzrate des Schätzers in der L2-Norm hergeleitet sowie dessen punktweise, asymptotische Normalität. Im dritten Kapitel wird ein neuer und einfach umsetzbarer Ansatz zur Schätzung semi(nicht)parametrischer diskreter Entscheidungsmodelle, unter Berücksichtigung von Restriktionen auf die funktionalen Parameter des Modells, vorgestellt. Die untersuchten Modelle weisen funktionale Parameter auf, die bestimmte funktionale Formen aufweisen. Zentraler Teil der Arbeit ist die Entwicklung eines GLS-Schätzers über einen geeigneten sieve-Raum, der aus I- und B-Spline Basisfunktionen unter geeigneten Restriktionen basiert. Es wird gezeigt, dass sich die Berücksichtigung der Restriktionen auf die funktionale Form positiv auf die Konvergenzrate des Schätzers in einer schwachen Norm auswirkt und so notwendige Bedingungen für die asymptotische Normalität semiparametrischer Schätzer einfacher erreichen lässt. / This dissertation comprises three individual papers on various topics in microeconometrics. In the first chapter, which is joint work with Christoph Breunig, we study a semi-/nonparametric regression model with a general form of nonclassical measurement error in the outcome variable. We provide conditions under which the regression function is identifiable under appropriate normalizations. We propose a novel sieve rank estimator for the regression function and establish its rate of convergence. The second chapter deals with the estimation of conditional random coefficient models. Here I propose a two-stage sieve estimation procedure. First, a closed-form sieve approximation of the conditional RC density is derived. Second, sieve coefficients are estimated with generic machine learning procedures and under appropriate sample splitting rules. I derive the $L_2$-convergence rate of the conditional RC-density estimator and also provide a result on pointwise asymptotic normality. The third chapter presents a novel and simple approach to estimating a class of semi(non)parametric discrete choice models imposing shape constraints on the infinite-dimensional and unknown link function parameter. I study multiple-index discrete choice models where the link function is known to be bounded between zero and one and is (partly) monotonic. In the paper I present an easy to implement and computationally efficient sieve GLS estimation approach using a sieve space of constrained I- and B-spline basis functions. The estimator is shown to be consistent and that imposing shape constraints speeds up the convergence rate of the estimator in a weak Fisher-like norm. The asymptotic normality of relevant smooth functionals of model parameters is derived and I illustrate that necessary assumptions are milder if shape constraints are imposed.

Les processus autorégressifs à bifurcation (BAR) ont été au centre de nombreux travaux de recherche ces dernières années. Ces processus, qui sont l'adaptation à un arbre binaire des processus autorégressifs, sont en effet d'intérêt en biologie puisque la structure de l'arbre binaire permet une analogie aisée avec la division cellulaire. L'objectif de cette thèse est l'estimation les paramètres de variantes de ces processus autorégressifs à bifurcation, à savoir les processus BAR à valeurs entières et les processus BAR à coefficients aléatoires. Dans un premier temps, nous nous intéressons aux processus BAR à valeurs entières. Nous établissons, via une approche martingale, la convergence presque sûre des estimateurs des moindres carrés pondérés considérés, ainsi qu'une vitesse de convergence de ces estimateurs, une loi forte quadratique et leur comportement asymptotiquement normal. Dans un second temps, on étudie les processus BAR à coefficients aléatoires. Cette étude permet d'étendre le concept de processus autorégressifs à bifurcation en généralisant le côté aléatoire de l'évolution. Nous établissons les mêmes résultats asymptotiques que pour la première étude. Enfin, nous concluons cette thèse par une autre approche des processus BAR à coefficients aléatoires où l'on ne pondère plus nos estimateurs des moindres carrés en tirant parti du théorème de Rademacher-Menchov. / Bifurcating autoregressive (BAR) processes have been widely investigated this past few years. Those processes, which are an adjustment of autoregressive processes to a binary tree structure, are indeed of interest concerning biology since the binary tree structure allows an easy analogy with cell division. The aim of this thesis is to estimate the parameters of some variations of those BAR processes, namely the integer-valued BAR processes and the random coefficients BAR processes. First, we will have a look to integer-valued BAR processes. We establish, via a martingale approach, the almost sure convergence of the weighted least squares estimators of interest, together with a rate of convergence, a quadratic strong law and their asymptotic normality. Secondly, we study the random coefficients BAR processes. The study allows to extend the principle of bifurcating autoregressive processes by enlarging the randomness of the evolution. We establish the same asymptotic results as for the first study. Finally, we conclude this thesis with an other approach of random coefficient BAR processes where we do not weight our least squares estimators any more by making good use of the Rademacher-Menchov theorem.