Zhesheng Qiu
Interests: Macroeconomics, Bounded Rationality, Money and Banking

Job Market Paper

Finite Depth of Reasoning and Monetary Policy

This paper examines the role of households' expectations for monetary policy. First, I uncover a unique pattern which I call “Asymmetric Reasoning” in growth expectations from Michigan Survey of Consumers. Although households do observe output growth from one year ago, they do not use this information to forecast one year ahead growth, but use interest rate related information to do so. This cannot be explained by information frictions, limited attention or adaptive learning. Second, I interpret this pattern as finite depth of reasoning, and formalize it by a level-k equilibrium on top of a small scale New Keynesian model with incomplete markets. I establish a Recursive Level-k Equilibrium to handle endogenous state variables under level-k reasoning, and obtain a state space representation. In the model, households are less aware of the nature of output dynamics because it requires higher level of rationality to understand. The model is estimated using both macroeconomic data, and expectations data. I also calibrate the level of market incompleteness to obtain a marginal propensity to consume (MPC) of the unconstrained households (7%) close to that in state-of-art heterogeneous agent models. Third, I find that level-k amplifies the output response to monetary shocks by more than 50% compared to the rational expectations benchmark. The role of ``Asymmetric Reasoning" can be viewed as a missing self-stabilizing general equilibrium effect in households' expectations. Households over-react to monetary policy because they are over-confident on output growth when output level is high. The MPC is crucial for these results.


Search-Based Sticky Prices (with José Víctor Ríos Rull)

We pose a directed search style shopping friction on top of an otherwise standard New Keynesian structure, and nest it as a special case. Firms are not free to increase prices not only because it reduces the quantity demand, but also because it induces more competition on the supply side, and hence reduce the likelihood that a produced good can be sold. When the good market is tighter in aggregate, there is less room for a firm to reduce its own market tightness by posting a lower price, so that all firms have higher desired price markup. As a result, our model can produce procyclical price markup and labor productivity conditional on monetary shocks, which is consistent with our empirical findings in SVAR. In our model, inflation is partly pulled by demand, in contrast to the standard model in which it is completely pushed by cost. Our model performs equally well in other aspects, compared with a medium scale New Keynesian DSGE model, and provide new insights on how demand creates its own supply.

Rehypothecation and Intermediary Leverage

This paper provides a theory of endogenous leverage through rehypothecation in collateralized intermediation. Overcollateralization with a rehypothecation option arises as an optimal contract between broker dealers and their clients due to clients' private information on collateral quality. Knowing less about the quality of the collateral, broker dealers are able to repledge the collateral with lower margins to obtain cheap cash flows from intermediating the collateral. The cheap cash flow increases broker dealers' risk-shifting incentive, and induces potential risks of losing the collateral, to the clients with high collateral quality. A compensation for these risks arises to make the high quality clients break even, but becomes an information rent for the lower quality ones. Hence, clients' private information on collateral quality enables the broker dealers with positive net worth to obtain higher leverage through cheap cash flow, but also makes the cheap cash flow more expensive. As a result, the private information on collateral quality first raises broker dealers' leverage, and then reduces it. In the over-the-counter market, this induces too much leverage, and too little leverage, respectively. The results shed light on why rehypothecation flourishes before the 2008 financial crisis, and then collapses in U.S.

Housing and Saving with Finance Imperfection (with Yanbin Chen and Fangxing Li)

Annals of Economics and Finance 14, no. 1 (2013): 207-248.

Emprical facts show that controlling for net wealth, more housing wealth leads to higher marginal propensity to consume. This indicates that housing wealth is illiquid. In theory, it is the refinancing and repayment constraint, instead of the downpayment constraint, that makes housing wealth illiquid. In quantitative work, the emergence of housing market does not reduce Chinese households' non-housing saving rates, and increases the aggregate households' saving rates dramatically.

Teaching Experience

Teacher Assistant:
Statistics for Economists - Spring 2014
Economics for Business - Fall 2013
Introductory Macroeconomics - Fall 2012, Spring 2013


Jośe-Víctor Ríos-Rull (advisor)
Department of Economics
University of Pennsylvania
507 McNeil Building,
3718 Locust Walk
Philadelphia, PA, 19104, USA

Dirk Krueger
Department of Economics
University of Pennsylvania
511 McNeil Building,
3718 Locust Walk
Philadelphia, PA, 19104, USA

Harold Cole
Department of Economics
University of Pennsylvania
436 McNeil Building,
3718 Locust Walk
Philadelphia, PA, 19104, USA


I am on the job market and will be available for interviews during the AEA meeting in Philadelphia on January 5-7, 2018.