Financial Stability Policies and Bank Lending

The aim of this research project is to answer a question that has been widely debated among policymakers and academic economists: Should increased systemic risk in the financial sector be addressed with conventional monetary policy that "leans against the wind" (i.e. sets a higher interest rate than would be warranted based on inflation or real economic considerations alone), or should it be countered with more targeted, so-called "prudential" measures (such as bank-specific capital requirements)? Existing studies on the effects of different financial stability policies are predominantly theoretical and yield conflicting conclusions. Economic trade-offs are one reason for the lack of consensus in the literature. On one hand, generalized interest rate increases cannot be as easily circumvented as targeted prudential measures because they affect everyone in a given monetary area. In contrast, prudential measures primarily focus on specific financial institutions deemed unsound, while leaving healthy ones unaffected. This allows room for regulatory arbitrage, reducing the effectiveness of targeted prudential measures in containing systemic risk. On the other hand, leaning against the wind could result in significant economic collateral damage as a higher interest rate usually depresses aggregate demand. Targeted prudential interventions are designed in a way to minimize precisely these negative consequences. Compelling empirical studies on the comparative effects and effectiveness of different financial stability policies face several challenges. Since the two described types of measures are often seen as alternatives, it is hardly possible to compare both while keeping the environment and time constant. Additionally, valid control groups are almost never available. Financial sectors subjected to financial stability policies (treatment group) exhibit increased systemic risk. Hence, they fundamentally differ from financial sectors that remain without economic policy interventions because they show no build-up of potentially dangerous risks (control group). The present research project aims to overcome these challenges by exploiting a research design rooted in economic history. It relies on regional monetary policy differences within the United States. Between 1914 and 1935, the twelve (still existing) Federal Reserve districts were able to pursue different monetary policies. In May 1920, four of these districts decided to increase their interest rates to counter a significant expansion of bank lending. Four other districts resorted to a prudential measure. The remaining four districts did not take any special action. Enabling a local comparison of detailed bank-level data in small bands around district boundaries, this historical episode thus provides unique conditions to identify the causal effects of different financial stability policies, while keeping environment and time fixed.

This paper shows that a little-known policy experiment by the U.S. Federal Reserve in 1920 may offer valuable lessons for how central banks can protect the financial system today. The main finding is that a "progressive discount rate" (PDR) - a central bank policy that raised borrowing costs only for banks that took on excessive debt - slowed credit growth, helped impacted banks survive longer and reduced economic distress, compared to a simpler across-the-board interest rate hike. The study compares two approaches that different Federal Reserve districts adopted during the post-World War I American credit boom. Some districts raised interest rates uniformly ("leaning against the wind," or LAW). Others introduced the PDR, which kept normal borrowing costs low but penalized banks that borrowed heavily from the central bank. By analyzing bank records and economic data from that time, the paper finds three main results. First, the PDR caused banks to cut back on lending more than LAW did. Banks in PDR districts reduced lending, because they feared that unexpected shifts in their funding needs could suddenly push them into the higher penalty rates. This "expectations channel" made banks more cautious and encouraged them to build up stronger liquidity buffers, which also improved their resilience to shocks further down the road. Second, banks under the PDR experienced significantly lower failure rates than those under LAW. This means the PDR not only slowed risky lending but also made banks sturdier during the ensuing economic depression of the early 1920s. Third, counties whose banks were exposed to the PDR saw less economic hardship and higher land values in the years after the policy. By stabilizing banks, the policy thus also protected local economies. The broader implication is that central banks policies that specifically target the riskiest institutions - and shape their expectations about future borrowing costs - can be effective in boosting financial stability, while avoiding unnecessary harm to healthier parts of the economy. Although the experiment took place a century ago, the findings echo today's debates about how to prevent credit booms from ending in crises. They suggest that tools which adjust costs for central bank borrowing in proportion to risk-taking could complement current approaches, such as uniform rate hikes or quantitative credit regulations. In practice, this could mean more flexible ways to manage financial risks - potentially shielding savings and local economic activity from the fallout of banking crises.