We study how Small Business Administration (SBA) employees respond to salient defaults. Using novel data to identify employees transferring across SBA offices, we find that defaults on SBA loans in their previous workplace reduce SBA loans in their current workplace. The effect is independent of local risk conditions and the informational content of the non-local defaults, consistent with a mechanical updating of risk perceptions among local SBA employees and inconsistent with rational learning. The local SBA loan market becomes geographically clustered and concentrated among fewer borrowers and lenders, especially those who have prior relationships with the SBA, suggesting higher barriers for participation.

We build a risk management model that incorporates variable capital utilization and procyclical collateral capacity. The former means that capital utilization determines production, which affects capital depreciation and risk exposure, linking capital utilization to firms' risk management decisions. The latter means that the ability to borrow and hedge increases with expected profitability. Using a new dataset on hedging and capital utilization of oil and gas producers, we employ novel identification strategies and find that hedging is positively correlated with corporate liquidity and expected profitability, whereas utilization is negatively correlated with liquidity. These results support the key predictions of our theory.

We establish that macroprudential policies limiting capital flows can curb risks arising from corporate foreign currency borrowing in emerging markets. Using detailed firm-level data from India, we show that propensity to issue foreign currency debt for the same firm is higher when the difference in short-term interest rates between India and the US is higher, i.e., when the dollar `carry trade' is more profitable; this behavior is driven by the period after the global financial crisis. The positive relationship between issuance and the `carry trade' breaks down once regulators institute more stringent interest-rate caps on foreign currency borrowing. Riskier borrowers such as importers and those with higher interest costs cut issuance most. Firm equity exposure to foreign exchange risk rose after issuance in favorable funding conditions and emerged as a source of external sector vulnerability during the `taper tantrum’ of 2013. Macroprudential policy action limiting capital flows is able to nullify this effect, such as during the market stress due to the COVID-19 pandemic.

Since ineffective debt resolution perpetuates zombie lending, bankruptcy reform has emerged as a solution. We show, however, that lender-based frictions can limit reform impact. Exploiting a unique empirical setting and novel supervisory data from India, we document that a new bankruptcy law had muted effects on lenders recognizing zombie borrowers as non-performing. A subsequent unexpected regulation, targeting perverse lender incentives to continue concealing zombies, increased zombie recognition particularly for undercapitalized and government-owned banks, highlighting the role of bank capital and political frictions in sustaining zombie lending. Resolving zombie loans allowed lenders to reallocate credit to healthier borrowers who increased investment.

I study the relative importance of lending and deposit-taking for bank value. Comparing outcomes for winning banks to runner-up bidders in failed bank auctions, I find winners experience a 1.5% abnormal return and this increase is mainly due to deposits, not loans. After acquisition, the winning bank cuts lending to the failed bank’s borrowers and closes branches but it retains almost all acquired deposits. These deposits are not channeled into lending elsewhere. Rather, the acquirer is able to lower deposit rates, reflecting increased market power. Multiple results are independent of the failed bank, suggesting the findings have broader relevance.

Using changes in the composition of the US House Financial Services Committee as a shock to a region’s political importance, I provide evidence that financial institutions alter lending patterns depending on whether a county is represented by a member of the committee. The effects are asymmetric – on gaining a member, counties see no immediate change but on losing a member, there is a decline in home mortgage loans originated. This asymmetry is consistent with models that emphasize reputation building in the market for political favors. Effects are greater where the politician receives less direct contributions suggesting that these indirect contributions might be substitutes for direct giving. In the presence of limits on campaign contributions, these results emphasize alternate channels that firms may employ to influence politicians.