The Five Types of Liquidity Risk a Bank Actually Has to Manage

The types of liquidity risk are not one risk but five, and treating them as a single number is how you understate your real exposure. This post breaks liquidity risk into the five categories that matter in practice, shows how each one behaves, and explains how they feed each other under stress so a small problem turns into a crisis.

The video below is a walkthrough of the same material, so watch it first if that is how you learn, then read on.

Why liquidity risk is not one risk

If you treat liquidity risk as a single number, you will manage it badly. It splits into distinct categories, each with its own drivers, its own early warning signs, and its own place in the regulatory toolkit. Funding, market, contingent, intraday and structural risk behave differently and fail differently.

The reason this matters is not academic. The categories are correlated. When one goes wrong it tends to pull the others with it. A market freeze raises haircuts, which triggers collateral calls, which drain your intraday buffer, which spooks short term lenders, which becomes a funding run. Manage them in isolation and you will understate your real exposure, because you will miss the feedback loops that turn a bad day into a bad month.

Funding liquidity risk: borrowing short, lending long

This is the classic one, and it comes straight from what banks do for a living. They fund long dated assets (a 25 year mortgage, say) with short dated liabilities (overnight deposits, commercial paper, repo). That maturity mismatch is the business model, not a mistake.

The risk is that your short term funding needs to roll, and one day it does not roll. Deposits leave. Wholesale lenders decline to renew. The asset side, meanwhile, matures on its own slow schedule and cannot be accelerated. So the cash you owe today comes due faster than the cash owed to you arrives.

Confidence is the trigger. Funding does not drain smoothly. It disappears in steps as counterparties reassess you, and the reassessment is often driven by rumour rather than fundamentals. A firm that looked solvent on Friday can be gone by the following Wednesday, not because the assets went bad but because the funding stopped.

Market liquidity risk: when you cannot sell what you planned to sell

Your liquidity plan almost certainly assumes you can convert some assets into cash quickly. Market liquidity risk is the risk that when you actually try, the market is not there at the price you assumed.

This shows up in two ways. First, the price gaps down because everyone is selling the same thing at once, so your fire sale crystallises a loss. Second, the haircut widens, so the same bond raises less cash in repo than it did last week. Both mean your assumed buffer is smaller than it looked on paper.

This is exactly why HQLA quality is defined so carefully. Level 1 assets (top quality government bonds and central bank reserves) hold up in stress and attract minimal haircuts. Level 2 splits further: Level 2A (for example certain high grade covered bonds and public sector debt) takes a heavier haircut, and Level 2B (assets such as some corporate bonds and equities) takes a heavier one still. Under the standard Basel caps, illustratively, total Level 2 is limited to 40% of the buffer and Level 2B to 15%. If you fill your buffer with assets that are liquid on a calm day but not on a stressed one, you have market liquidity risk hiding inside what you called your buffer.

The judgement call sits in the assumptions. A bond can be liquid for months and then not liquid for the two weeks you happen to need it. Historic trading volumes tell you about normal conditions, not about the moment everyone reaches for the exit together.

Contingent liquidity risk: the off balance sheet demands

The first two categories are on the balance sheet, where you can see them. Contingent liquidity risk sits off it, which is what makes it dangerous. These are obligations that only become cash demands when a condition is met, and the condition tends to be met precisely when everything else is going wrong.

The usual suspects:

• Undrawn committed credit lines. Corporates draw their facilities in a stress because their own funding has dried up. You promised the cash. Now they want it, on the same day your own funding is under pressure.
• Guarantees and standby facilities. Same pattern. They convert to real outflows when the underlying party is in trouble.
• Derivative collateral calls. Two things bite here. Variation margin flows when mark to market moves against you, and in a sharp market move that call can be large and it is due now. Initial margin can also rise as volatility increases, so you post more even before any loss crystallises. Both are demands for cash or high quality collateral at the worst moment.

The common thread is correlation. These demands are not random. They cluster in the same window as your funding stress and your market stress. That is why you cannot treat them as low probability tail events priced independently. They arrive together.

Intraday liquidity risk: timing within the day

Here is the one that most end of day models miss entirely. Intraday liquidity risk is about timing within the day, not your position at the close. You can end the day flat and still have failed at 11am.

Payment systems settle continuously. You send and receive payments throughout the day, and the timing does not net neatly. If large outflows land before your matching inflows arrive, you need enough liquidity in the system to bridge the gap. Run out mid morning and you cannot make a payment. That payment was someone else's expected inflow.

That is the cascade. In a real time gross settlement system, one bank holding back or failing to pay starves the next bank of liquidity it was counting on, which delays its payments, and so on. A short lived shortfall in one place propagates across the system. The shortfall need not last long to do damage.

Managing this means knowing your intraday profile: when your big obligations settle, how much collateral you have pledged at the central bank, how much throughput depends on inflows you do not fully control. It is an operational discipline as much as a treasury one.

Structural risk and how the categories interact under stress

Structural liquidity risk, also called maturity transformation risk, is the slow burn version. It is not about a single stress event. It is the shape of the balance sheet itself creating an ongoing vulnerability. If you are consistently funding long assets with unstable short liabilities, you carry structural risk every single day, stress or no stress.

This is what NSFR is built to surface. It compares your Available Stable Funding (ASF), the funding you actually have weighted by how stable it is, against your Required Stable Funding (RSF), the stable funding your assets require, over a one year horizon. A weak ratio tells you the shape is fragile before any event forces the issue. The ILAAP is where you argue whether your structure is sustainable given your specific business and appetite.

Now put the categories together, because the interaction is the whole point:

1. Markets freeze, so market liquidity falls and haircuts widen.
2. Wider haircuts and adverse price moves trigger contingent collateral calls.
3. Meeting those calls drains your intraday buffer at the worst moment.
4. Counterparties notice the strain and pull short term funding.
5. Underneath it all, weak structural shape means you had little slack to begin with.

None of these is severe on its own. Chained together they are a crisis. That is why managing each risk in a separate silo understates the true exposure. The correlation between them is the exposure.

Where each risk meets the regulatory toolkit

The concepts map onto things you already report, which is the easiest way to make them stick.

• Funding risk lives inside the LCR. The LCR is principally a funding measure: it stresses 30 days of outflows against a buffer of HQLA, with run off rates doing most of the work. Market liquidity enters only indirectly, through the HQLA haircuts and eligibility classification rather than as a direct measure. The mechanics come down to net stressed outflows measured against a buffer of HQLA. The Basel Committee's LCR standard sets out the framework in full.
• Structural risk is the NSFR over the one year horizon, comparing ASF against RSF, and the narrative sits in the ILAAP, where a firm sets out the stresses it believes it should survive.
• Intraday risk is captured through intraday monitoring metrics (the BCBS 248 indicators) and firm specific intraday reporting, and it also features in PRA110 style cash flow reporting. PRA110 itself is primarily a cash flow mismatch return, not the intraday monitoring tool.
• Contingent risk shows up across the LCR outflow assumptions and in your stress scenarios, where committed facilities and collateral calls get their own outflow treatment.

If you can point to where each category is captured, you can also spot where your own framework is thin.

The practical takeaway

Pick one recent stress scenario and trace a single shock through all five categories rather than testing each in isolation. Follow the chain:

1. A market freeze widens haircuts.
2. Wider haircuts trigger collateral calls.
3. Those calls drain your intraday buffer.
4. The strain pulls your short term funding.
5. Check whether your structural slack absorbs any of it.

If your model treats these as separate lines that never move together, it is telling you a calmer story than reality will. When you are ready to look at the numbers, laying them out by bucket and currency in Python shows how these risks actually move together.