Macroeconomic uncertainty characterize agents beliefs about future borrowing and asset prices, especially under constrained debt conditions. The project aims to provide an empirical evaluation
of agents beliefs in a model where macroeconomic uncertainty is acknowledged to affect prices and quantities in debt market under alternative statistical and theoretical forms, as in Hansen and Sargent (2019).
This setup implies a structure of agents beliefs which, deviating from the rational expectation paradigm, embodies the idea that borrowers and policy makers are affected by multiple sources of uncertainty and behave according to both risk and ambiguity attitudes. The aim is to be able to empirically distinguish among uncertainty within a model (risk); uncertainty across a set of available known models (ambiguity); and uncertainty about each model (model misspecification).
The aim is to quantify the impact of different types of investors' beliefs about models for discounted expected utilities, evaluate their behaviour in a model where we introduce an Euler equation (consumption/saving decisions) with occasionally binding borrowing positions. Then, we will inspect whether and how uncertainty attitudes (risk/ambiguity) fluctuate over time determining debt and asset prices cyclical facts. This will be done by both using aggregate time series and cross-sectional data at the risky asset level.
The analysis is, therefore, designed to offer an historical quantification of how macroeconomic uncertainty is priced in debt markets, when agents form beliefs about the future value of their wealth and the state of the economy. Debt accumulation is constrained by an occasionally-binding collateral constraint. A standard recursive utility model is employed and adjusted to embody three key additional ingredients: i) a comprehensive setup for different uncertainty sources; ii) state- dependency in uncertainty attitudes; iii) limited borrowing capacity.
Despite the growing body of theoretical research explaining the role of beliefs formation for leverage pro-cyclicality and asset prices, the empirical evidence is still struggling to keep the pace and contribute with a sound and extensive framework. Therefore, the calibration itself of macroeconomic models lacks of empirical guidance on the value of key deep parameters.
The purpose of this study is, therefore, to help filling this gap by empirically testing a well-recognized recursive preference model with two key ingredients. First, the project aims to provide a flexible and extensive framework for estimating uncertainty premia on risky assets. Uncertainty may manifest under differently-identifiable beliefs attitudes, distinct in nature (risk versus ambiguity and model misspecification), over states of the economy and across different sets of risky assets. Given that uncertainty in beliefs affects the way agents discount future outcomes, its representation in asset pricing models generally consists in an additional pricing kernel factor depending on the future value function. Different specifications of uncertainty attitudes affect the way the value function defines this risk pricing augmentation. Secondly, debt market conditions will be modelled through an occasionally binding collateral constraint, entering the estimated Euler equation through the shadow price of debt. This is a key addition to standard asset pricing models, especially in the context of analyses where borrowers' beliefs about the future valuation of collateral matter.
The aim of the project is to lay down a comprehensive empirical assessment on the conditions at which different uncertainty attitudes over the cycle can amplify leverage fluctuations. This would either help clarifying where the empirical evidence stands on the matter or advance on the calibration of macroeconomic models of the leverage cycle.