Using survey data on secured funding arrangements provided by broker–dealers for their clients—a class of contracts that includes bilateral repo—we document that financing rates, collateral haircuts, lending maturities, and position limits move strongly together over time and across asset classes. Liquidity of the underlying securities, as opposed to their volatility or credit risk, is the main driver of this behavior, with dealer balance-sheet constraints also playing a role in the funding of less-liquid security types. A simple model of dealer–client interaction rationalizes these findings. Instrumenting with changes in market conventions, we find that funding conditions had little effect on cash securities markets between 2011 and 2019, but the tightening of terms during the market stress of early 2020 likely impaired liquidity and reduced asset returns to some degree.

In a theoretical context where innovators borrow loans or settle for state-owned enterprise (SOE) sponsorship for their projects, we examine the effects of banking prudential regulations and their interaction with corporate leverage on the patenting partnership choice in China using a unique matched patent-firm-bank loan dataset for 15,623 observations in the 2013–17 period. We use a unique instrumental variable (IV) strategy to identify idiosyncratic bank prudential reform shocks associated with the post-2012 Basel III regulation and find prudential metrics (corporate leverage) of the financiers (firms) to positively (negatively) influence SOE patenting partnership choice, though prudential regulation mitigates the latter. Prudential reforms therefore come at a cost of further SOE dominance. However, conditional on an innovation project being SOE sponsored, we find positive spillover effect from the SOE’s employment mandate to loan productivity. Our results are robust across different IV strategies, alternative measures, sub-sample and mechanism analyses.