The thesis comprises four chapters. Chapter 1 serves as an overall introduction to the thesis. Chapters 2, 3, and 4 are the three main chapters that explore enterprise performance from three different perspectives. The first half of Chapter 2 primarily investigates the endogenous determinants of firm size using a panel vector autoregression (PVAR) model. The second half introduces a Markov chain-based method to describe a firm’s transition between different size categories over time, eliminating the long time period data requirement of conventional autoregressive models. Transition trend and transition entropy are defined and calculated based on the obtained transfer matrices. Chapter 3 evaluates the impact of external shocks on firm performance, considering the establishment of science and technology industrial parks (STIPs) in China. The policy impact is estimated using a quasi-experimental analysis, calculating the average treatment effect by comparing the performance of treated and control firms. Advanced matching approaches, including propensity score matching (PSM) and coarsened exact matching (CEM), are employed to obtain more precise results. Chapter 4 aims to advance the counterfactual inference framework used in Chapter 3. By integrating the latest machine learning approach, generative adversarial networks (GAN), the impact of external shocks on firm performance can be accurately estimated, addressing most of the problems in traditional matching estimators. Additionally, three performance tests are conducted to verify the accuracy of the proposed GAN-ATT estimator.
Date of Award | 28 Jun 2023 |
---|
Original language | English |
---|
Awarding Institution | |
---|
Supervisor | Kerry Papps (Supervisor) & Judith Delaney (Supervisor) |
---|
- Causal Inference
- Enterprise Dynamics
- Policy Evaluation
- GAN-ATT Estimator
Studies on the Dynamics and the causal Inference of Enterprises in China: (Alternative Format Thesis)
You, B. (Author). 28 Jun 2023
Student thesis: Doctoral Thesis › PhD