Survey data can be viewed as incomplete or partially missing from a variety of perspectives and there are different ways of dealing with this kind of data in the prediction and the estimation of economic quantities. In this thesis, we present two selected research contexts in which the prediction or estimation of economic quantities is examined under incomplete survey data. These contexts are first the investigation of composite estimators in the German Microcensus (Chapters 3 and 4) and second extensions of multivariate Fay-Herriot (MFH) models (Chapters 5 and 6), which are applied to small area problems. Composite estimators are estimation methods that take into account the sample overlap in rotating panel surveys such as the German Microcensus in order to stabilise the estimation of the statistics of interest (e.g. employment statistics). Due to the partial sample overlaps, information from previous samples is only available for some of the respondents, so the data are partially missing. MFH models are model-based estimation methods that work with aggregated survey data in order to obtain more precise estimation results for small area problems compared to classical estimation methods. In these models, several variables of interest are modelled simultaneously. The survey estimates of these variables, which are used as input in the MFH models, are often partially missing. If the domains of interest are not explicitly accounted for in a sampling design, the sizes of the samples allocated to them can, by chance, be small. As a result, it can happen that either no estimates can be calculated at all or that the estimated values are not published by statistical offices because their variances are too large.

Income is one of the key indicators to measure regional differences, individual opportunities, and inequalities in society. In Germany, the regional distribution of income is a central concern, especially regarding persistent East-West, North-South, or urban-rural inequalities. Effective local policies and institutions require reliable data and indicators on regional inequality. However, its measurement faces severe data limitations: Inconsistencies in the existing microdata sources yield an inconclusive picture of regional inequality. While survey data provide a wide range of individual and household information but lack top incomes, tax data contain the most reliable income records but offer a limited range of socio-demographic variables essential for income analysis. In addition, information on the long-term evolution of the income distribution at the small-scale level is scarce. In this context, this thesis evaluates regional income inequality in Germany from various perspectives and embeds three self-contained studies in Chapters 3, 4, and 5, which present different data integration approaches. The first chapter motivates this thesis, while the second chapter provides a brief overview of the theoretical and empirical concepts as well as the datasets, highlighting the need to combine data from different sources. Chapter 3 tackles the issue of poor coverage of top incomes in surveys, also referred to as the ’missing rich’ problem, which leads to severe underestimation of income inequality. At the regional level this shortcoming is even more eminent due to small regional sample sizes. Based on reconciled tax and survey data, this chapter therefore proposes a new multiple imputation top income correction approach that, unlike previous research, focuses on the regional rather than the national level. The findings indicate that inequality between and within the regions is much larger than previously understood with the magnitude of the adjustment depending on the federal states’ level of inequality in the tail. To increase the potential of the tax data for income analysis and to overcome the lack of socio-demographic characteristics, Chapter 4 enriches the tax data with information on education and working time from survey data. For that purpose, a simulation study evaluates missing data methods and performant prediction models, finding that Multinomial Regression and Random Forest are the most suitable methods for the specific data fusion scenario. The results indicate that data fusion approaches broaden the scope for regional inequality analysis from cross-sectional enhanced tax data. Shifting from a cross-sectional to a longitudinal perspective on regional income inequality, Chapter 5 contributes to the currently relatively small body of literature dealing with the potential development of regional income disparities over time. Regionalized dynamic microsimulations provide a powerful tool for the study of long-term income developments. Therefore, this chapter extends the microsimulation model MikroSim with an income module that accounts for the individual, household, and regional context. On this basis, the potential dynamics in gender and migrant income gaps across the districts in Germany are simulated under scenarios of increased full-time employment rates and higher levels of tertiary education. The results show that the scenarios have regionally differing effects on inequality dynamics, highlighting the considerable potential of dynamic microsimulations for regional evidence-based policies. For the German case, the MikroSim model is well suited to analyze future regional developments and can be flexibly adapted for further specific research questions.

Data fusions are becoming increasingly relevant in official statistics. The aim of a data fusion is to combine two or more data sources using statistical methods in order to be able to analyse different characteristics that were not jointly observed in one data source. Record linkage of official data sources using unique identifiers is often not possible due to methodological and legal restrictions. Appropriate data fusion methods are therefore of central importance in order to use the diverse data sources of official statistics more effectively and to be able to jointly analyse different characteristics. However, the literature lacks comprehensive evaluations of which fusion approaches provide promising results for which data constellations. Therefore, the central aim of this thesis is to evaluate a concrete plethora of possible fusion algorithms, which includes classical imputation approaches as well as statistical and machine learning methods, in selected data constellations. To specify and identify these data contexts, data and imputation-related scenario types of a data fusion are introduced: Explicit scenarios, implicit scenarios and imputation scenarios. From these three scenario types, fusion scenarios that are particularly relevant for official statistics are selected as the basis for the simulations and evaluations. The explicit scenarios are the fulfilment or violation of the Conditional Independence Assumption (CIA) and varying sample sizes of the data to be matched. Both aspects are likely to have a direct, that is, explicit, effect on the performance of different fusion methods. The summed sample size of the data sources to be fused and the scale level of the variable to be imputed are considered as implicit scenarios. Both aspects suggest or exclude the applicability of certain fusion methods due to the nature of the data. The univariate or simultaneous, multivariate imputation solution and the imputation of artificially generated or previously observed values in the case of metric characteristics serve as imputation scenarios. With regard to the concrete plethora of possible fusion algorithms, three classical imputation approaches are considered: Distance Hot Deck (DHD), the Regression Model (RM) and Predictive Mean Matching (PMM). With Decision Trees (DT) and Random Forest (RF), two prominent tree-based methods from the field of statistical learning are discussed in the context of data fusion. However, such prediction methods aim to predict individual values as accurately as possible, which can clash with the primary objective of data fusion, namely the reproduction of joint distributions. In addition, DT and RF only comprise univariate imputation solutions and, in the case of metric variables, artificially generated values are imputed instead of real observed values. Therefore, Predictive Value Matching (PVM) is introduced as a new, statistical learning-based nearest neighbour method, which could overcome the distributional disadvantages of DT and RF, offers a univariate and multivariate imputation solution and, in addition, imputes real and previously observed values for metric characteristics. All prediction methods can form the basis of the new PVM approach. In this thesis, PVM based on Decision Trees (PVM-DT) and Random Forest (PVM-RF) is considered. The underlying fusion methods are investigated in comprehensive simulations and evaluations. The evaluation of the various data fusion techniques focusses on the selected fusion scenarios. The basis for this is formed by two concrete and current use cases of data fusion in official statistics, the fusion of EU-SILC and the Household Budget Survey on the one hand and of the Tax Statistics and the Microcensus on the other. Both use cases show significant differences with regard to different fusion scenarios and thus serve the purpose of covering a variety of data constellations. Simulation designs are developed from both use cases, whereby the explicit scenarios in particular are incorporated into the simulations. The results show that PVM-RF in particular is a promising and universal fusion approach under compliance with the CIA. This is because PVM-RF provides satisfactory results for both categorical and metric variables to be imputed and also offers a univariate and multivariate imputation solution, regardless of the scale level. PMM also represents an adequate fusion method, but only in relation to metric characteristics. The results also imply that the application of statistical learning methods is both an opportunity and a risk. In the case of CIA violation, potential correlation-related exaggeration effects of DT and RF, and in some cases also of RM, can be useful. In contrast, the other methods induce poor results if the CIA is violated. However, if the CIA is fulfilled, there is a risk that the prediction methods RM, DT and RF will overestimate correlations. The size ratios of the studies to be fused in turn have a rather minor influence on the performance of fusion methods. This is an important indication that the larger dataset does not necessarily have to serve as a donor study, as was previously the case. The results of the simulations and evaluations provide concrete implications as to which data fusion methods should be used and considered under the selected data and imputation constellations. Science in general and official statistics in particular benefit from these implications. This is because they provide important indications for future data fusion projects in order to assess which specific data fusion method could provide adequate results along the data constellations analysed in this thesis. Furthermore, with PVM this thesis offers a promising methodological innovation for future data fusions and for imputation problems in general.

There is a wide range of methodologies for policy evaluation and socio-economic impact assessment. A fundamental distinction can be made between micro and macro approaches. In contrast to micro models, which focus on the micro-unit, macro models are used to analyze aggregate variables. The ability of microsimulation models to capture interactions occurring at the micro-level makes them particularly suitable for modeling complex real-world phenomena. The inclusion of a behavioral component into microsimulation models provides a framework for assessing the behavioral effects of policy changes. The labor market is a primary area of interest for both economists and policy makers. The projection of labor-related variables is particularly important for assessing economic and social development needs, as it provides insight into the potential trajectory of these variables and can be used to design effective policy responses. As a result, the analysis of labor market behavior is a primary area of application for behavioral microsimulation models. Behavioral microsimulation models allow for the study of second-round effects, including changes in hours worked and participation rates resulting from policy reforms. It is important to note, however, that most microsimulation models do not consider the demand side of the labor market. The combination of micro and macro models offers a possible solution as it constitutes a promising way to integrate the strengths of both models. Of particular relevance is the combination of microsimulation models with general equilibrium models, especially computable general equilibrium (CGE) models. CGE models are classified as structural macroeconomic models, which are defined by their basis in economic theory. Another important category of macroeconomic models are time series models. This thesis examines the potential for linking micro and macro models. The different types of microsimulation models are presented, with special emphasis on discrete-time dynamic microsimulation models. The concept of behavioral microsimulation is introduced to demonstrate the integration of a behavioral element into microsimulation models. For this reason, the concept of utility is introduced and the random utility approach is described in detail. In addition, a brief overview of macro models is given with a focus on general equilibrium models and time series models. Various approaches for linking micro and macro models, which can either be categorized as sequential approaches or integrated approaches, are presented. Furthermore, the concept of link variables is introduced, which play a central role in combining both models. The focus is on the most complex sequential approach, i.e., the bi-directional linking of behavioral microsimulation models with general equilibrium macro models.