Three essays in applied microeconomics: incentives, structural shift, and the unintended distributional consequences of economic change

Abstract

In the first part of the thesis, I use American administrative data from 1990 until 2018 to examine the effect of state merit aid scholarship programs on pre-collegiate education attainment, measured by high school dropout rates. Because these programs were adopted by different states at different times, I use a dual identification strategy that combines the Synthetic Control Method (SCM) and panel matching. Both methods construct counterfactual comparisons from valid donor states and address the limitations of traditional Two-Way Fixed Effects Difference-in-Differences (TWFE DID) models in staggered-adoption settings. I then use TWFE DID estimates, and alternative dropout-rate measures as robustness checks. The SCM results suggest that merit-aid scholarship programs increase high school dropout rates by 1.3 to 5.9 percentage points across regional specifications. Panel matching provides a second main estimate in the same direction, and the TWFE DID robustness estimates are directionally consistent with these findings.

In the second part of the thesis, I systematically document the Chinese labor share decline and delineate various channels in explaining such trends. Moreover, from the aggregate industrial level, I demonstrate that an increase in the concentration ratio contributes to the decrease in the labor share in China. Though the rise of Chinese superstar firms drove the labor share down, the Chinese new entrant firms grew at a faster pace and took over more market share which contributed to the labor share drop too.

In the third part of the thesis, I examine the Chinese Industrial Yearbook Data from 1997 to 2013 with the Chinese Patent Reference Dataset from 1985 to 2018 to identify the causal inference between Chinese "superstar firms" and their declining labor share. In addition, I adopt the Melitz and Polanec (2015) method to identify and decompose the sources of labor share decline. The empirical results suggest when superstar firms in China double their patent inventions, it is estimated to decline their labor share by 1.189 percentage points. Moreover, when a Chinese firm invests in its technology advancement, the jump effect, measured by the impact of having the first patent, is estimated to contribute to a 1.9 percentage points decline in its labor share. Lastly, using the preferred dynamic GMM specification, I show that technology progression is expected to lead to an average labor share decline of 14.1 percentage points.

Date of Award	15 Nov 2026
Original language	English
Awarding Institution	University of Nottingham
Supervisor	Jingong Huang (Supervisor) & Paolo Epifani (Supervisor)