Bank loan funds keep about four and a half percent of their assets in cash, against redemptions available every day.

Who holds a syndicated loan once it has been sold?

Institutions with very different obligations to their own investors. Federal Reserve staff describe the syndicated market as one where banks arrange and underwrite loans and then distribute them, mostly to collateralized loan obligations, loan funds and insurance companies.

Those three holders are not interchangeable, and the difference that matters is not credit appetite. It is what each one owes its own funders, and when. A collateralized loan obligation is term funded: its liabilities do not come due because a market fell. An insurance company holds against long-dated liabilities of its own. A loan mutual fund offers its investors their money back at the end of any business day.

That last obligation is the one worth understanding, because it is the only one in the list that can turn a price move into a forced sale within a week. The fund has not done anything unusual. Offering daily redemption against assets that take longer than a day to sell without moving the price is the ordinary business of an open-ended fund, and the Federal Reserve authors call it by its name: liquidity transformation.

None of which is a prediction of stress. It is a description of which holder in the chain has the shortest fuse, and therefore which one is worth watching for a borrower who wants to know how their paper is likely to behave in a bad month.

How large is that slice now, and which way is it moving?

Smaller, and shrinking. Between the fourth quarter of 2019 and the fourth quarter of 2025, the net assets of bank loan mutual funds fell by 19 billion dollars, or 21 percent, to 71 billion dollars.

The comparison that gives the number its shape is the neighbouring sector. Over the same six years, high-yield corporate mutual fund assets rose 13 percent to 263 billion dollars. Two fund sectors holding two below investment grade asset classes moved in opposite directions, and the loan sector is now roughly a quarter the size of the bond sector.

The flow data underneath is starker than the asset totals, because market moves partly offset it. On Morningstar Direct figures cited by the authors, from 2019 to 2025 bank loan funds saw net outflows of 23 billion dollars, equal to 27 percent of their assets. High-yield funds saw net outflows of 15 billion dollars, equal to 6 percent. Investors did not slightly prefer one to the other. They left one and largely stayed in the other.

Keep the scale honest. Seventy-one billion dollars is a modest share of a leveraged loan market that Federal Reserve staff put at roughly 1.4 trillion dollars at the end of 2025. This sector is not the loan market. It is the visible, daily-priced, daily-redeemable part of it, and that is precisely what makes its direction of travel worth reading.

How much cash do these funds hold against those redemptions?

About four and a half percent of net assets at the median bank loan fund, and it has been roughly stable there since 2024. The measure is a liquidity ratio: cash and cash equivalents, Treasury bills and short-term investment vehicles, divided by total net assets.

It used to be higher. The median bank loan fund's liquidity ratio climbed from 2020 to reach almost ten percent of net assets in 2021, then declined. The spread between the 5th and 95th percentiles has narrowed since 2021, so the sector has converged on the lower number rather than splitting into cautious and aggressive camps.

The other half of the picture moved the wrong way. The authors also track an illiquidity ratio: the share of net assets held in Level 3 assets, meaning positions valued using significant unobservable inputs because no observable market price exists. For the median bank loan fund that ratio rose above one percent in 2020, fell back, and since 2025 has been approaching the levels last seen in the pandemic spike.

Stable cash against rising hard-to-value holdings is the combination the authors flag, and they state the conclusion plainly: on balance, liquidity transformation risk in the median bank loan fund has increased. High-yield funds show the opposite. Their median liquidity ratio has held near four percent since 2020, their illiquidity ratio has been broadly flat, and their percentile range has narrowed, so on balance their liquidity transformation activity has declined.

Treat the level with the caution the authors do, and the direction as the finding. They say explicitly that they focus on time-series trends rather than point estimates, because funds classify assets differently. Some count instruments as cash equivalents that a stricter reading would not, which biases the liquidity ratio up. Fluidity between Level 2 and Level 3 classification biases the illiquidity ratio down. Level 3 also captures only the most extreme dimension of illiquidity, not a portfolio's full liquidity profile.

Does a larger cash buffer actually stop the redemptions?

Not reliably, and the two stress episodes the authors examine point in opposite directions. In the week after the tariff announcement of 2 April 2025, bank loan funds that had entered the quarter with above-median liquidity saw average net outflows of 4.2 percent, against 2.7 percent for the below-median group. The higher-buffer funds lost more, though the difference is not statistically significant.

High-yield funds showed the same ordering in that week, at 2.1 percent of outflows for the above-median group against 1.1 percent for the below-median group, and there the difference is significant at the ten percent level. The authors read both results as consistent with the literature: riskier and less liquid funds tend to hold larger buffers deliberately, precisely because they expect to need them.

March 2020 reversed it, and only for loan funds. Bank loan funds with above-median liquidity going into the pandemic lost 6.7 percent of assets that month. The below-median group lost 14.2 percent. The gap of 7.6 percentage points is significant at the one percent level, and it is the single largest effect in the note. High-yield funds showed no meaningful difference between the two groups in the same month.

The authors' own conclusion is the careful one, and it is the useful one: precautionary liquidity management appears to mitigate redemption risk under normal market volatility, but on its own it does not explain fund behaviour during an extraordinary shock. The relationship between how much cash a fund holds and how much money leaves it is conditional on the nature of the stress, which is another way of saying that the buffer is not a fixed quantum of protection.

What should an owner or a sponsor do with this?

Ask where the paper is going to sit, and stop treating that as the arranger's business rather than yours. This is our read rather than the Fed's, which studies funds and says nothing about borrowers. But the composition of a holder base determines how a loan trades in a bad month, and that in turn shapes what a refinancing, an amendment or a consent request costs when one is needed.

The practical version is a small set of questions at the point of syndication. What proportion of the facility is expected to land with term-funded holders against daily-dealing funds. Whether the arranger expects to retain a position and for how long. Whether the paper is intended for a small group of relationship holders or a broad distribution. None of that is confidential, all of it is knowable while the deal is live, and almost none of it is asked once pricing has been agreed.

It also reframes what a secondary price means. A daily mark is a genuine advantage over a quarterly model, and the companion argument on this site is that a borrower should want the market that prices its debt honestly. The qualification is that a daily price set by a shrinking pool of daily-redeemable holders will move further on a given amount of selling than the same price set by a deeper one. Read a bad month accordingly, and do not confuse a liquidity move in the holder base with a verdict on the business.

And it sharpens the timing question on anything discretionary. If a facility is going to need an amendment, a consent or a repricing, the cost of getting it depends partly on who is holding the paper and what their own quarter has looked like. Approaching the holder base after a redemption wave and approaching it before one are two different negotiations, and only one of them is scheduled by the borrower.

What would change this picture?

The illiquidity ratio, which is the live series here rather than the fund size. It is the measure the authors flag as moving, and a sustained retreat from the pandemic-era levels it is now approaching would mean the median bank loan fund had rebuilt the margin between what it owes daily and what it can sell daily.

A turn in the flow direction would be the slower version of the same signal. Six years of outflows equal to 27 percent of assets is what has thinned this holder base, and the asset totals only partly show it because market moves offset flows. If the sector stopped shrinking, the argument here weakens on its own terms.

Hold the caveats where the authors put them. These are staff estimates rather than an official series, the liquidity ratio is biased upward and the illiquidity ratio downward by inconsistent classification between funds, the stress-episode samples cover 63 to 97 percent of sector assets depending on the episode, and two episodes are two observations rather than a pattern. The views in the note are the authors' own and are not those of the Federal Reserve System.

The durable finding survives all of that. One identifiable slice of the buyer base for syndicated loans offers daily redemption, has been shrinking for six years, holds a thin and no longer growing cash buffer, and is carrying more hard-to-value assets than it was. For a borrower, that is not a reason to avoid the market. It is a reason to know who ends up holding the loan, because that is the part of the transaction that outlives the negotiation.