Imagine a car that crashed in 2008. Basel II was driving, the brakes were weak $($capital$)$, the airbags didn’t deploy $($liquidity$)$, and the GPS kept saying “Recalculating” $($risk models$)$. The global economy was the passenger — and it flew through the windshield.

So, Basel III came along as a total financial rehab program, and in December 2017, it went into reform-overdrive.

Let’s step through these changes like we’re debugging a really expensive, really global software patch.


🔧 Why More Reforms? Didn’t Basel III Already Fix Things?

Despite Basel III’s initial patchwork, the post-2009 stress tests revealed something scary — banks still had creative ways to game the system using internal models. The goals of the 2017 update were to:

  • Make capital requirements more robust and risk-sensitive
  • Limit banks’ ability to get “creative” with internal models
  • Add leverage ratio buffers for those “too-muscular-to-fail” banks $($G-SIBs$)$
  • Introduce the output floor — which ensures no bank can dig its way under standardized capital requirements

🤔 If we’re tightening the rules, where do we start? With the foundation: credit risk.


🏦 Standardized Approach for Credit Risk: Risk-Sensitive Reality Check

Basel III upgraded the Standardized Approach $($SA$)$ from “paint-by-numbers” to “realistic shading.”

  • Now, risk weights are granular. Residential mortgages are no longer treated equally — if your house has a bad loan-to-value ratio, expect a higher capital requirement.
  • It also reduces overreliance on credit ratings — because apparently, letting agencies call junk “AAA” didn’t work well last time.
  • This SA revamp lays the foundation for the output floor.

🤔 But what if banks say, “We’ll just use our fancy internal models instead”?
Let’s see what Basel did to that playbook.


🧠 Internal Ratings-Based $($IRB$)$ Approaches: From Freedom to Floor

Before, banks could use advanced internal models $($A-IRB$)$ and estimate things like:

  • Probability of Default $($PD$)$
  • Loss Given Default $($LGD$)$
  • Exposure at Default $($EAD$)$

But turns out — giving each bank its own formula for capital was like letting students grade their own exams. So Basel said:

  • No more A-IRB for big players $($large corporates, financial institutions$)$ — now it’s F-IRB or bust
  • Introduced input floors for PD, LGD, and EAD
  • Made banks document their model logic — no more “trust me, bro”

🤔 Wait… so even derivatives weren’t immune to model manipulation?


🔄 CVA Risk Framework: When Derivatives Bite Back

Credit Valuation Adjustment $($CVA$)$ is the risk that your counterparty in a derivative trade goes poof before settling.

During the financial crisis, this was a big source of losses — and Basel didn’t forget:

  • Now, market risk components of CVA must be modeled
  • Say goodbye to IRB-based CVA — it’s Standardized or Basic Approach only
  • The framework is now aligned with market risk sensitivities — so your hedge must actually hedge

🤔 We’ve tackled model abuse and counterparty risk. But what about that wild beast — operational risk?


💥 Operational Risk Framework: Goodbye Complexity, Hello Common Sense

Operational risk — a.k.a. “when someone at the bank messes up or misbehaves” — wasn’t covered well before. Basel III used to allow four different methods $($AMA and three SA options$)$ — and none worked.

So now?

  • One simplified SA for all banks
  • Capital requirement is based on:
    • Income $($more money = more risk of mistakes$)$
    • Historical losses $($because past misconduct predicts future disasters$)$

🤔 Great! Now how do we stop banks from pretending they’re smaller or less risky than they are?


🏋️ Leverage Ratio Framework: Stop Hiding That Bulk

Basel III also revised the leverage ratio — think of it like forcing bodybuilders to step on a scale.

  • G-SIBs must maintain a leverage buffer, met using Tier 1 capital
  • Adjustments to exposures now include derivatives and off-balance sheet items — no more hiding weights in the gym locker

🤔 Okay, but what if a bank still finds a clever way to use internal models to lower its capital needs?


⛔ Output Floor: No Digging Below This Line

LO 64.c:

The output floor says: “You can’t go lower than 72.5% of what the Standardized Approach says you owe in capital.” No more model trickery to squeeze capital down to risky levels.

Specifically, risk-weighted assets $($RWAs$)$ must be the higher of:

  1. The bank’s internal + standardized RWA total
  2. 72.5% of RWAs calculated using only the standardized approach

Standardized Approaches Used in the Floor:

  • Credit Risk: Use the new Basel III SA
  • Counterparty Credit Risk: SA-CCR
  • CVA Risk: SA-CVA or BA-CVA
  • Securitization: SEC-ERBA, SEC-SA, or 1,250% risk weight $($ouch$)$
  • Market and Operational Risk: Both use SA

🤯 In short: The output floor is like minimum wage — no matter how good you are at accounting magic, you must meet it.


🏁 Final Thought: Basel III’s 2017 Reforms — The Financial Firewall 2.0

Basel III’s 2017 reforms tightened the screws on every corner of the banking world:

  • More accurate risk weights
  • Restrictions on internal model usage
  • Better safeguards for CVA and operational risks
  • Clear-cut leverage and output limits

These changes weren’t about punishing banks. They were about ensuring that next time — if the sky falls — we won’t be frantically Googling “how to bail out a bank without a taxpayer revolt.”