“No One Can Borrow” – Reassessing the “Junior Can’t Borrow” Equity-Premium Puzzle Resolution

“The equity premium puzzle” refers to the difficulty, shown in Mehra and Prescott’s seminal 1985 paper, of simulating a realistic equity premium in macro models. “Junior Can’t Borrow,” Constantinides, Donaldson, and Mehra’s (CDM) influential 2002 paper, seeks to resolve this puzzle. Their answer is to fully restrict borrowing by the young. This imposed reduction in bond supply raises the price of safe bonds and lowers their yield, producing a meaningful equity premium. But CDM’s model is highly stylized, an OLG endowment economy with just three periods and a specific configuration of age-specific risks. Our paper revisits CDM’s solution within a large-scale, stochastic, OLG economy. Our model has 80 periods, 81 state variables, and, otherwise, standard RBC modeling of production, endogenous factor returns, and TFP and depreciation macro shocks. For computational convenience, we posit hypothetical steeply rising soft-borrowing costs to shut down bond supply on an age-specific basis. As we show, eliminating borrowing by all, not just young generations, i.e. “No One Can Borrow”, not “Junior Can’t Borrow”, is required to “resolve” the equity premium puzzle. This holds despite our assumption of far larger macro shocks than observed. Overstating macro shocks is required to match the variability of the return on national wealth. Despite our larger-than-life macro shocks, our model’s agents can largely self insure by adjusting their saving and relying on good times to average out bad times. Self insurance explains why generations, regardless of age, are close to indifferent between holding or supplying bonds and equity. Once we omit our borrowing cost function, chosen for illustrative purposes not realism, the equity premium puzzle reemerges full force. Hence, unlike CDM, our paper reinforces, rather than resolves, the puzzle.