Overview
This series explains the Federal Reserve paper A Static Capital Buffer is Hard To Beat through its full policy problem. In a calibrated multi-shock banking model, limited liability and deposit insurance can make socially inefficient risky lending privately attractive, while a convex payoff creates a safe-to-risky regime switch. Productivity, investment, and volatility shocks can all change the relevant incentive, so credit-to-GDP has a mild, statistically tenuous link to lower U.S. real GDP two years later without identifying the needed capital response. The paper finds that a modest static buffer can stay close to an information-rich no-failure benchmark when the threshold is hard to observe, but risk weighting may help under reliable classification. The result is conditional, preliminary, and model-bound.
Learning outcomes
- Limited liability and deposit insurance can make socially inefficient risky lending privately attractive.
- Convex bank payoffs make small capital changes capable of switching lending from safe to risky.
- Productivity, investment, and volatility shocks can all raise risk-taking pressure through different mechanisms.
- In the calibrated model, optimizing simple indicator rules largely removes their cyclical movement, leaving a modest static buffer close to the no-failure benchmark.
- Risk weighting offers potential gains only when project risk can be classified and enforced reliably.
Keywords
Preview every chapter · Unlock full text, audio, and video from chapter 4
PlansFAQ
Why can a static capital buffer beat a cyclical rule?
It can be more robust when aggregate indicators do not identify the shock changing banks' relative incentives. In the paper's calibrated model, optimized cyclical rules largely become static, while a fixed margin limits regime crossings without claiming precise information.
Why do banks choose risky loans in the model?
They may choose them because limited liability preserves the upside while deposit insurance shields depositors from default losses. The risky project has a lower expected return, but the bank's residual payoff is convex.
Does a high credit-to-GDP ratio show that capital should rise?
No, an above-trend credit-to-GDP ratio does not by itself identify that capital should rise. In the paper's U.S. 1980:Q1–2024:Q4 data it has a mild, statistically tenuous link to lower real GDP two years later; in the model, volatility can change risk-taking with little movement in the ratio.
When can risk-weighted capital beat a broad buffer?
It can beat a broad buffer when supervisors can reliably classify and enforce the risk of individual projects. The paper's 0.22 percent consumption-equivalent gain assumes perfect classification, so it is a stylized thought experiment rather than a general real-world estimate.
What does the no-failure capital rule actually require?
It requires the lowest capital level that, given the current shocks and model state, keeps safe lending more attractive than risky lending. It is a full-information benchmark: the paper does not present it as a directly observable formula or universal regulatory prescription.