Publications

An update of "Why Has CEO Pay Increased So Much?" after the crisis; the theory holds well.

Like rational agents, consumers learn, but like myopic agents, consumers respond to recent events more than events that occurred just a few months ago.

We propose an account of the movement of GDP volatility in the past 50 years, including the great moderation and its undoing.

Ten puzzling features of the macro-finance data are natural outcomes of a model where investors have time-varying perceptions of the risk of economic disaster.

Dynamic model of CEO pay featuring manipulation and private saving.  Proposes a reform of CEO compensation to address problems that led to the financial crisis.

Brief, non-technical summary of the framework for achieving tractable incentive contracts developed in "Tractability in Incentive Contracting" and used in "The Effect of Risk on the CEO Market and Dynamic CEO Compensation."

Develops a framework that delivers tractable (i.e. closed-form) optimal contracts, with few restrictions on the utility function, or noise distribution.

Risk aversion causes distortions in talent allocation--risky firms hire less talented CEOs.  Tractable market equilibrium allowing calibration of costs of corporate governance.

Take a production economy which has good business cycle properties:  to fix its asset pricing properties (e.g. to have a high and volatile returns to equities), add disaster risk in the right places, i.e. "disasterize" this economy.

We build cities "from the bottom up" by clustering areas obtained from high-resolution data, and find that a beautiful Zipf's law for population and for areas, for cities about 12,000 inhabitants in the USA and 5,000 inhabitants in Great Britain.

To estimate a Pareto exponent with an OLS regression, it's best to specify it with log(Rank-1/2)=a-b log(Size), and use the standard error s.e. (b)=b (2/n)^0.5, as OLS standard errors strongly biased.