The term structure of risk premia with heterogeneous preferences and beliefs

The purpose of this study is to offer a general equilibrium model of economy capable of explaining the changes in the term structure of equity risk premia as a result of the global financial crisis (GFC) of 2008-09, when the global economy experienced an unprecedented shock to the aggregate consumption. The data on the realised risk premia on dividend futures presented in Binsbergen and Koijen (2016) suggested the possibility of a change in slope of the term structure around the time of the GFC in three(Europe,theUK, and Japan)out of four economies considered: the term structure went from upward to downward-sloping. My own analysis of prices of dividend swaps on S&P500, Euro Stoxx 50, Nikkei 225, and FTSE 100 from 2005 to 2016 presented in Chapter 2 shows that the term structure went from upward to downward-sloping in Europe and the UK and from upward-sloping to flat in the US and Japan. In Chapter 1, I consider a continuous-time "Lucas-tree" economy with a long-run risk in the form a two-state Markov chain regime switching with the representative agent endowed with Epstein-Zin-Weil preferences. I find that the sign of the slope of the term structure is determined by the parameters of preferences and does not change its sign in response to consumption shocks. Chapter 2 is dedicated to an economy with constant growth and volatility of consumption populated by two agents having CRRA preferences with heterogeneity in both their preferences and beliefs. I find that in this setting, the sign of the slope may changes in response to significant consumption shocks and derive the conditions, under which it happens. I also analyse the effect of various heterogeneities, and calibrate the model. In Chapter 3 I solve the same economy but with two Epstein-Zin-Weil agents with differences both in preferences and beliefs and demonstrate the effects, as well as benefits to calibration, of adding heterogeneity in the elasticity of intertemporal substitution. As an extension, I also consider the effect of adding a long-run risk in the form of a 2 two-state Markov chain regime switching considered in Chapter 1.Open Acces