Recent years have witnessed a surge of international research in using micro-level data to answer empirical economic questions. This surge was possible because of the sharp increase in high-quality micro-level data across world, in particular in the US and Europe. The high-quality micro-level data are essential ingredients to understand economic behaviour and the impact of economic and social policy; however, having just the high-quality micro-level data is not sufficient to uncover the causal relationships. This is mainly because a large body of economic data are collected based on surveys and government registers rather than based on randomized experiments. Therefore, it is a central task in empirical economic research using non-experimental data to understand conditions under which the correlations or more generally associations obtained in statistical analysis can be interpreted as evidence for casual relationships. Studying these conditions is one of main econometric problems and generally viewed as an identification problem in econometrics. This research project aims to contribute to advances in understanding identification problems and developing estimation and inference methods using micro-level. In particular, the proposed research will: (1) develop point- and partial- identification results in microeconometric problems that occur naturally in empirical research in economics and social sciences; (2) develop semi- and nonparametric methods for estimation and inference in microeconometrics; (3) apply the state-of-the-art microeconometrics to important empirical problems in economics.