The Bank of England published a paper titled, "A simulation framework for sterling money market funds: estimating redemption capacity and evaluating liquidity requirements." The staff paper's summary says, "Money market funds (MMFs) aim to provide near-on-demand liquidity yet often hold assets that become hard to sell under stress, leaving them vulnerable to run-like redemptions. I build a simulation framework for sterling MMFs to estimate redemption capacity and failure probability across alternative redemption profiles and market-liquidity scenarios. Resilience of funds depends on both the timing of outflows and the effective liquidity of weekly liquid assets (WLA): front-loaded redemptions are most destabilising, and the benefit of asset sales shrinks as market depth thins. Removing the 30% WLA threshold effect – under which managers must consider measures to deter further redemptions – yields sizeable resilience gains by reducing cliff-edge behaviour. Under historically extreme shocks and without threshold effects, most resilience improvements come from holding WLA above the 30% regulatory minimum; in my simulations, gains concentrate around 40% WLA, with diminishing returns beyond."
Author Rishabh Kumar writes, "Money market funds (MMFs) aim to provide investors on-demand liquidity, often with same day or next-day access, while investing in instruments that can be difficult to liquidate quickly at par. That mismatch is a feature rather than a flaw: it is precisely what allows MMFs to earn a competitive yield. But it also makes them vulnerable to run dynamics especially during stress periods where investors might prefer cash. When redemptions rise abruptly, funds may be forced to generate cash by selling assets into stressed markets, potentially reinforcing price moves and tightening funding conditions. Because MMFs sit at the junction of short-term funding markets and cash management for non-financial corporations and financial institutions, disruptions can propagate beyond the sector and become a financial stability concern."
He continues, "Past episodes have illustrated these vulnerabilities in practice. During the March 2020 'dash for cash' (DFC) episode, funds with larger exposures to non-government assets, including U.S. Prime MMFs and EU Low Volatility Net Asset Value (LVNAV) funds, experienced substantial outflows. In the UK, the DFC episode triggered redemptions of roughly £25 billion (around 11% of sterling MMF assets), concentrated in EU-domiciled LVNAV funds (Financial Conduct Authority (FCA), 2023). Investors in need of liquidity made large MMF redemptions to raise cash quickly (Czech et al., 2021). The pattern echoes earlier stress episodes, including the Global Financial Crisis (GFC), when the Reserve Primary Fund 'broke the buck' after Lehman Brothers' default and the sector required extraordinary public backstops (McCabe, 2010; Bouveret, Martin, and McCabe, 2022)."
The paper tells us, "Regulatory reforms have tried to address these fragilities. The United States Securities and Exchange Commission's (SEC) 2010 MMF reforms and the European Union's (EU) 2017 Money Market Fund Regulation (MMFR) introduced liquidity management tools, minimum liquid-asset requirements, and constraints on portfolio composition (Bouveret, Martin, and McCabe, 2022). Yet recurring stress events suggest that liquidity mismatches remain a source of vulnerability for the sector. Moreover, the increase in interest rates from the 2020 pandemic-era lows introduced a separate tension: higher rates can reduce the market value of fixed-income holdings and interact with liquidity needs in stress. The Bank of England's Financial Stability in Focus highlights the potential for rate-driven valuation changes to intensify liquidity risks across non-bank financial institutions, including MMFs (Bank of England, 2023)."
It states, "A growing empirical literature documents MMF outflows and transmission channels during stress events. However, empirical work is necessarily backward-looking: it reflects realized outcomes under the institutional and policy conditions of a particular episode. In practice, those conditions may include use of emergency lending facilities and other central bank intervention that alter both investor behaviour and the liquidity of assets. This limits what can be inferred about MMF resilience under alternative market environments, different redemption paths, or counterfactual policy designs."
The piece explains, "This paper takes a complementary approach. I develop a forward-looking simulation framework for sterling-denominated MMFs that allows me to trace fund-level liquidity dynamics under a wide range of redemption shocks and market-liquidity assumptions. The framework delivers two outcomes that are directly relevant for policy. The first is redemption capacity: the maximum cumulative redemptions as a share of assets that a fund can meet over a specified horizon without breaching a requirement to hold a certain share of its assets in financial instruments that can be liquidated within one day. The second is failure probability: the share of simulated paths in which this daily liquidity constraint is breached under a given shock, redemption profile and time horizon. Using these metrics, I can compare in a unified way the resilience of MMFs to redemption stresses over different time horizons and profiles, depending on their initial share of WLA and a specified failure level of 10% of assets under management (AUM)."
It comments, "The model draws on behavioural mechanisms proposed in the recent literature (Bouveret, Martin, and McCabe, 2022; Dunne and Giuliana, 2022; Baes, Bouveret, and Schaanning, 2023) and explicitly varies the market environment. Specifically, I consider three liquidity settings: (i) a conservative benchmark in which secondary markets are effectively frozen and funds rely only on cash and maturing assets; (ii) a setting in which government securities can be sold to help meet redemptions; and (iii) an extension in which certificates of deposit (CDs) can also be monetised prior to maturity, under explicit limits intended to capture market depth. These market counterfactuals are chosen to reflect the different conditions seen in the past."
Kumar summarizes, "I have three main findings. First, in the frozen-market benchmark, redemption capacity rises monotonically with initial WLA and with the length of the stress horizon over which redemptions are spread, and it is strongly shaped by the timing of redemptions: front-loaded (left-skewed) paths are substantially more destabilising than historically observed unimodal paths. Second, allowing asset sales shifts the redemption-capacity frontier favourably, but the size of this improvement is highly sensitive to market depth. When secondary-market liquidity deteriorates, funds cannot reliably liquidate the assets counted as WLA at scale, so the effective buffer is smaller than the headline WLA measure and resilience is correspondingly weaker."
He adds, "Third, failure probabilities imply a non-linear trade-off in the design of liquidity buffers. Under historically extreme shocks, the largest reduction in failure risk is generated by increasing the current regulatory minimum to a higher buffer that lifts funds away from the region in which constraints tend to bind; in my simulations, that improvement is concentrated around a move to approximately 40% WLA. Beyond 40%, the marginal reduction in failure risk tends to diminish, especially under unimodal (historically observed) redemption paths. This pattern suggests that raising the minimum WLA above the current floor can materially improve resilience (especially for historically observed outflows), while also indicating where additional increases yield smaller incremental benefits."
Finally, the paper states, "These findings speak directly to ongoing policy debates, including the FCA review of sterling MMFs (FCA, 2023) and the SEC's 2023 reforms (SEC, 2023). The central issue is calibration: buffers must be large enough to reduce the likelihood of destabilising liquidity spirals, yet not so high that they materially impair intermediation. Crucially, the required buffer must also be credible in periods of stress."
