Utilisation: Why Borrow Rates Move

The utilisation rate is the share of a lending pool that has been lent out rather than sitting idle. It sets what borrowers pay and what depositors earn at the same moment, and it changes with every deposit, borrow and repayment. A pool sitting at 40% utilisation rate is mostly idle cash; a pool at 95% is almost fully committed.

Most people watch the advertised borrow APR and never look at the number underneath it. The APR is an output. Utilisation is the input, it is published before you sign anything, and it tells you what happens to your position when the pool fills up.

• One ratio, two prices. The same number sets the borrow cost and the deposit yield, which is why they cannot be considered separately.
• The curve is not a straight line. Almost every pool has a kink where the rate climbs sharply, and that kink is the real risk boundary.
• Withdrawals need idle cash. At high utilisation your balance is intact but not immediately redeemable.
• Utilisation is reflexive. A rate spike attracts deposits and repels borrowers, and the adjustment is not instant.

What the utilisation rate measures

Utilisation is borrowed funds divided by lendable funds. Nothing more complicated than that, and it answers a single question: how much of the pool’s money is currently out the door.

The formula and its one adjustment

In most pools, utilisation is total borrows divided by total supplied minus reserves. The reserve adjustment matters: a slice of interest is skimmed to a protocol treasury, and that slice is not lendable. Ignore it and your calculated ratio drifts above the published figure on pools with a meaningful reserve factor.

The complement is equally useful. Available cash equals total supplied minus total borrows, adjusted for reserves. That number is the only thing standing between depositors and a withdrawal queue, and it is usually displayed one line below the yield.

Why idle capital is not wasted capital

Cash held in a pool earns nothing, which looks like inefficiency. It is really an insurance buffer. Every depositor has a claim redeemable at par on demand, and those claims exceed the borrows by design. The buffer is what makes the claim credible.

When the buffer shrinks, the pool is not broken. It has simply converted redeemable deposits into illiquid loans. The loans are still performing; the exit is what has tightened.

How the utilisation rate moves the borrow rate

The borrow rate is a function of utilisation, defined in advance by a parameter set, and the shape of that function does most of the work.

The kink

Nearly every lending market uses a two-part curve. Below a target utilisation — commonly somewhere between 70% and 90%, depending on the venue and the asset — the rate rises slowly. Above that target, the slope steepens by several multiples. The target is the kink.

The reason is a queueing problem dressed as a price. Below the kink, the pool wants borrowers, so credit is cheap. Above it, the pool needs deposits more than it needs borrowers, so the rate is raised hard enough to force repayments and pull in new supply. On volatile collateral the kink usually sits lower, because the pool needs more headroom for liquidations.

What the two slopes feel like

A gentle first slope means a small move in utilisation produces a small move in your cost. A steep second slope means the last few percent of utilisation can multiply the rate. This is the part borrowers underestimate: the danger is not the rate you sign at, it is the rate you get pushed to.

Utilisation is not a measure of how popular a pool is. It is a measure of how little spare cash is left in it.

A worked example

Take a stablecoin pool with 10,000,000 supplied, no reserve adjustment, a base rate of 0%, a kink at 80%, a first slope worth 4 percentage points at the kink, and a second slope worth 60 points across the final 20%.

With 8,200,000 borrowed, utilisation is 82%. The rate is 4% plus 60% multiplied by 2/20, which is 10%. The deposit yield, after a 10% reserve factor, is 10% × 0.82 × 0.9, or about 7.4%.

A single borrower then draws another 900,000. Borrows reach 9,100,000 and utilisation reaches 91%. The rate becomes 4% plus 60% multiplied by 11/20, or 37%. Deposit yield climbs to roughly 30.4%. The borrower’s cost has nearly quadrupled on a 9-point move in one ratio.

Now a depositor adds 500,000. Supply is 10,500,000 against 9,100,000 borrowed, so utilisation falls to 86.7% and the rate drops to about 24%. Deposit yield falls to 18.7%. Adding capital cut the yield by more than a third, which is the reflexivity in one line.

Where the utilisation rate misleads you

Three situations make the ratio say something other than what it appears to say.

Full utilisation is a freeze

At 100%, available cash is zero. Withdrawals do not fail because the protocol is insolvent; they wait because every unit of the asset is lent to someone who has not repaid. Withdrawal can only resume as borrowers repay or new deposits arrive. Some markets queue exits by time, some by share, and some simply block them.

Recursive borrowing inflates the number

A depositor can supply collateral, borrow against it, re-supply the borrowed asset and borrow again. Each loop adds to both supply and borrows, and the ratio it produces describes the same underlying capital several times over. A pool at 88% utilisation may be occupied by a handful of looping positions whose unwind removes most of the supply at once. High utilisation plus concentrated looping is a different risk from high utilisation plus many independent borrowers.

Deposit yield is not the borrow rate

The gap between them is utilisation multiplied by one minus the reserve factor. In a quiet pool that gap is large in percentage terms and small in absolute ones. Reading a 30% deposit yield as evidence of 30% borrower demand is a common error; the actual borrower cost might be 37% and the pool may still be emptying.