Research

Abstract. The bank lending channel is heterogeneous across firms. Using matched bank-firm credit data, we develop a framework that estimates the effects of firm-demand and bank-supply credit shocks, allowing for interactions between bank and firm unobserved factors. Bank shocks can have heterogeneous effects across firm types, where firm type is unobserved for the econometrician. We decompose credit growth dynamics into time-varying firm and bank-firm type interaction effects. We uncover significant heterogeneity in the bank lending channel: i) The effect of bank shocks varies considerably across the identified firm types, ii) During the Great Recession, more exposed banks severely contracted their loan supply to some firms but shielded others, iii) We uncover a bankfirm matching channel: the transmission of shocks crucially depends on the bank-firm network, with credit growth dropping by up to 20% in a counterfactual randomly matched network, iii) Accounting for interactions helps to more precisely estimate the impact of bank shocks on firm real outcomes.

[ESRB Working Paper No. 68]

Abstract. I propose and estimate a model in which bank and firm net worth influence macroeconomic fluctuations. Firm net worth affects borrowing capacity. Bank net worth is necessary for intermediation: banks lend to firms and expand the flow of funds using their diversified portfolio as collateral subject to aggregate risk. Aggregate firm and bank net worth determine activity. The net worth distribution is only relevant in banking crises, in those, shocks to bank net worth have an effect beyond standard financial frictions models. A novel bank accelerator helps explain output, bank net worth, lending, and spreads during the 1990, 2001, and 2008 recessions.

Coming soon.

Coming soon.

Abstract. We study the optimality of macroprudential policies in an environment where banks provide liquidity to firms. Informational frictions between banks can cause interbank market freezes, prompting firms to accumulate their own liquid assets. Liquidity hoarding by firms in turn reduces the demand for bank loans and bank profitability, makes interbank market freezes even more likely, and may ultimately trigger a self-fulfilling bad equilibrium. Such “liquidity panics” provide an additional rationale for liquidity requirements on banks, which alleviate frictions in the banking sector and, paradoxically, can increase aggregate investment. Instead, policies encouraging bank lending can have the opposite effect.

[Previous versions of this paper circulated with titles "Equity allocation and risk-taking in the intermediation chain" and "Demand for safety, risky loans: A model of securitization"]

Abstract. Safe asset demand increases loan risk. This arises in a competitive model in which securitization vehicles create safe assets by pooling loan payoffs purchased from loan originators. Equity investors allocate their wealth between originators, who need skin-in-the-game due to moral hazard, and vehicles, who need loss-absorption capacity against aggregate risk. An increase in demand for safety fosters safe asset creation through a securitization boom: originators sell more of their loan payoffs to vehicles, equity is reallocated from originators to vehicles, and the two effects contribute to an increase in loan risk. The model is consistent with a broad set of facts in the run-up to the Global Financial Crisis.

Abstract. We study optimal government support following a rare disaster that creates heterogeneous firm liquidity needs. Firms’ increase in debt reduces their output due to moral hazard. Banks are subject to a minimum capital requirement that limits deposit insurance costs upon bad aggregate shocks. Without government support, firms’ moral hazard and banks’ funding frictions reinforce each other amplifying output losses. Optimal support is implemented with firm-specific transfers combined with the provision of aggregate risk insurance through a capital requirement relaxation and a public preferred equity stake in banks. Our results shed light on suboptimality features in the actual policy responses to Covid-19 lockdowns.

Abstract. We analyze government interventions to support firms facing liquidity needs during a lockdown in a competitive model of financial intermediation. Banks and firms have legacy balance sheets at the lockdown date. Firms' liquidity needs can be financed by banks that are subject to risk-weighted capital requirements and funded with insured deposits. An increase in firms' overall claims to external investors aggravates moral hazard problems and reduces expected output. The government can support firms directly through transfers or indirectly through guarantees to new bank loans or reductions in the capital requirement. As a result of the diversification of idiosyncratic firm risks conducted by banks, a reduction in the capital requirement only creates costs for the government following negative aggregate shocks that lead to banks' failure. A pecking order on the government policies that maximize output as a function of the government's budget is derived. For low budget, a reduction in capital requirements is optimal and is fully transmitted to firms through increases in banks' leverage. For medium budget, the capital requirement reduction becomes slack and needs be combined with transfers to firms or loan guarantees. For high budget, transfers are strictly necessary.