What is the difference between the standardized and IRB approaches to credit risk-weighted assets, and what's the trade-off?

Both compute credit RWAs (the denominator of capital ratios) but differ in who sets the risk parameters. Under the standardized approach, risk weights are prescribed by the regulator — based on exposure class and, often, external credit ratings (e.g., set percentages for sovereigns, banks, corporates, retail, mortgages) — simple and comparable but blunt. Under the internal ratings-based (IRB) approach, the bank uses its own models to estimate risk parameters: in Foundation IRB the bank estimates PD while LGD/EAD are supervisory-set; in Advanced IRB the bank estimates PD, LGD, and EAD itself, feeding regulatory formulas. IRB can produce lower, more risk-sensitive RWAs (capital better aligned to actual risk) — the incentive to use it — but it requires regulatory approval, robust data, validation and governance, and it created variability and 'RWA optimization' concerns. That's why Basel III's output floor now caps IRB RWAs at a percentage (72.5%) of the standardized result, limiting how much capital banks can save through internal models. The trade-off: risk sensitivity and capital efficiency vs complexity, model risk, and comparability.