Reading a liquidity position with pandas in twenty lines

Spreadsheets are fine until the model grows. When the same liquidity calculation runs every day across many entities, a few lines of Python pay for themselves quickly. Here is a small worked example that mirrors the LCR logic.

Start with a cash ladder. Each row is a bucket of outflow with a category that drives its run off rate.

import pandas as pd

ledger = pd.DataFrame({
    "category": ["retail", "retail", "wholesale", "facility"],
    "amount":   [500_000, 250_000, 400_000, 300_000],
})

run_off = {"retail": 0.05, "wholesale": 1.0, "facility": 0.10}
ledger["run_off"] = ledger["category"].map(run_off)
ledger["outflow"] = ledger["amount"] * ledger["run_off"]

Now the coverage number. HQLA sits on one side, stressed outflows on the other.

hqla = 900_000
net_outflows = ledger["outflow"].sum()
lcr = hqla / net_outflows
print(f"LCR: {lcr:.0%}")

From here the path is obvious. Pull the ledger from your warehouse instead of hard coding it, load the run off rates from a config table the risk team owns, and write the result to a dashboard. The calculation stays readable, and the assumptions stay auditable.

That is the real win. Not speed, but a model a reviewer can actually follow.