Description:
We document five facts about banks: (1) market and book leverage diverged during the 2008 crisis, (2) Tobin's Q predicts future profitability, (3) neither book nor market leverage appears constrained, (4) banks maintain a market-leverage target that is reached slowly, and (5) precrisis, leverage was predominantly adjusted by liquidating assets. After the crisis, the adjustment shifted towards retaining earnings. We present a Q-theory where notions of leverage differ because book accounting is slow to acknowledge loan losses. We estimate the model and show that it reproduces the facts. We examine counterfactuals where different accounting rules produce novel policy tradeoffs.