31 Août: Soutenance de Thèse de Johanna Schauer

• Thomas CHANEY, Chercheur TSE, Université Toulouse Capitole.
• Franck PORTIER, Chercheur TSE, Université Toulouse Capitole.
• Marty MESTIERI, Université de Northwestern
• Sang Yoon LEE, Université de Mannheim

This thesis focuses on the development of heterogeneous agent DSGE models to investigate questions of inequality, education and financial development. It is composed of three chapters. 

In the first chapter, Martí Mestieri, Robert Townsend and I use household-level data from Mexico, to document patterns between schooling, entrepreneurial decisions and household characteristics, e.g., assets, talent of members, and age of the household head. Motivated by our findings, we develop a heterogeneous-agent, overlapping-generations, dynasty model. Households jointly decide over their life-cycle on (i) kids’ human capital investments (schooling) and (ii) parents’ entry, exit and investment into alternative entrepreneurial modes (subsistence and modern). With financial constraints, all these are co-determined. We show that a calibrated version of our model can account for the broad correlation patterns uncovered in data within and across generations, e.g., a non-monotonic relationship between educational choices and assets across occupational modes, growth in profits and employment for modern firms only, and dynastic persistence across generations in education and wealth. Endogenous human capital acquisition is a key driver of inequality and intergenerational persistence. Eliminating this channel from the model would decrease the top 10% income share by 56%. Eliminating collateral constraints would increase average household expenditure by 7.1% and benefit the middle class, reducing top and bottom expenditure shares. It would also reduce by 29% the correlation between household assets and kids’ schooling levels. 

In the second chapter, Sonali Jain-Chandra and I look at recent developments in India, which show impressive growth and progress in eradicating extreme poverty but rising inequality. We investigate the contribution of high skill premia and inequality in education levels to this development. We set up a model of heterogeneous agents, who endogenously choose their skill levels and occupations while being financially constrained. Moreover, we account for the large variance in the quality of education and the duality of the economy. We calibrate the model to India’s economy today making use of detailed survey data. Evaluating different policies we find that targeted unconditional cash transfers can lower inequality and increase output and educational attainment at relatively low costs if designed carefully. 

In the final chapter, I investigate the role of payroll-based lending in the creation of entrepreneurship and its effect on inequality. Payroll-based credit has expanded rapidly in recent years, especially in emerging markets. I develop a heterogeneous agents DSGE model, that captures the specific characteristics of these countries and loans. I calibrate the model to the South African market using micro data from the National Income Dynamics household survey and macro moments. I find that the increase in payroll-based lending from 2008 to 2014 leads to a decrease in entrepreneurship and productivity and an increase in inequality. Restricting access to payroll-based loans for households with low income levels can improve output and inequality in the long-run, while a substitution by developmental loans targeting the poorest can only be effective if they are large enough.