This paper introduces a new internal propagation mechanism into a baseline multi-sector RBC model by focusing on the cross-sectional interdependencies across the population that are  generated by the discrepancy between individuals' consumption behavior and their production activities. In the model, there are two types of agents (L and H-skill) who differ in terms of the occupations that they are specialized in and the composition of their consumption baskets. In particular, both types of agents have non-homothetic preferences which induce a strong desire to consume the good that the other type produces. The resulting mechanism, in response to a positive investment shock, works as follows: as H-skill agents are employed in a more capital-intensive sector, their productivities, and so their incomes, initially increase relatively more compared to the L-skill agents. But they spend their additional income mostly on service goods in which L-skill agents are specialized in, so the demand for L-skill labor subsequently increases as well. On the other hand, as their income rises, L-skill agents start consuming relatively more capital intensive goods, which in turn generates additional demand for H-skill labor that is not related to the original technology shock. Consequently, this mechanism feeds into itself and a circular interaction between the two types of agents emerges as a result of the imposed preference structure and the division of labor. I show, through a series of quantitative experiments, that this interaction serves as a very strong internal propagation mechanism that can significantly increase the amplification of exogenous shocks- a long quest in the RBC literature. I also discuss the way the same mechanism gives rise to an endogenous variation in the capital content of the goods consumed over the business cycle and how this endogenous variation generates more persistence in the model. 

In this paper, I study the effects of an exogenous investment shock that is correlated with the skill distribution in the population. In the model, agents, who have heterogeneous productivities, are sorted into different occupations based on their comparative advantages over the business cycle. I also assume that each  occupation is associated with a different consumption bundle. This latter assumption implies that as agents switch between low and high-skill jobs their consumption behavior changes in a discrete fashion as well, which in turn gives rise to non-linear income expansion paths at the aggregate level without assuming non-homothetic preferences at the individual level. In this environment, I show that after a positive investment shock, the fraction of agents working in the goods sector increases on impact and those who start working in the goods sector replace their service intensive consumption baskets with goods intensive ones, and therefore the demand for goods increases more relative to the services along the expansion, which in return results in even more agents being hired in the goods sector. I show quantitatively that this endogenous mechanism alone can generate sizable amplification and persistence, even without labor-leisure choice, in contrast to the standard models. The model also suggests an alternative resolution to what is known as the productivity puzzle in the literature. This is because the one-to-one link between output and productivity does not exist in the model, in contrast to the standard models. Instead, the productivity is determined by the aggregate skill distribution and how agents are sorted between two sectors over the business cycle. 

I study the model in Morris and Shin (2002) under strategic information transmission constraints between the public authority and private agents in order to investigate how the precision of the public signal affect the quality of the communication. In the model, a central bank (CB) communicates with two private agents. All three players have different interests over the state of the economy and receive informative but noisy signals when the state is realized. I find that CB only cares about the average bias of the private agents, and CB can truthfully reveal her signal only if the average bias is below a certain threshold. More importantly, as the CB's signal precision increases, the possibility of truthful communication increases in her signal precision which in return enhances her ability to communicate with the private agents.

Research In Progress
Boundaries of Economic Growth: Ownership Inequalities and Finance
On The Mathematical Structure of Pareto Optimal Exchange (Edgeworth Box) Economies
The Existence of Correlated Equilibria under Perfect Competition