The main hallmarks of the global financial crisis were too-big-to-fail institutions taking on too much risk with other people's money: excess leverage and default pressure resulting from contagion and counterparty risk. This paper looks at whether the Basel III agreement addresses these issues effectively. Basel III has some very useful elements, notably a (much too light "back-up") leverage ratio, a capital buffer, a proposal to deal with pro-cyclicality through dynamic provisioning based on expected losses and liquidity and stable funding ratios. However, the paper shows that Basel risk weighting and the use of internal bank models for determining them leads to systematic regulatory arbitrage that undermines its effectiveness. Empirical evidence about the determinants of the riskiness of a bank (measured in this study by the Distance-to-Default) shows that a simple leverage ratio vastly outperforms the Basel Tier 1 ratio. Furthermore, business model features (after controlling for macro factors) have a huge impact. Derivatives origination, prime broking, etc., carry vastly different risks to core deposit banking. Where such differences are present, it makes little sense to have a one-size-fits-all approach to capital rules. Capital rules make more sense when fundamentally different businesses are separated. JEL classification: G01, G15, G18, G20, G21, G24, G28 Keywords: Financial crisis, Basel III, derivatives, bank business models, distance-todefault, structural bank separation, banking reform, GSIFI banks