Pillar 1 vs Pillar 2 capital: what an ICAAP has to add

What each number is#

Pillar 1 capital is the minimum a bank must hold under standardised or approved internal calculation rules for credit, market and operational risk — a requirement computed the same way, from the same formulas, for every institution in scope.

Pillar 2 capital is whatever additional capital a bank concludes it needs once it has assessed its own risk profile properly, together with whatever a supervisor concludes it needs after reviewing that assessment. The first is a calculation the bank performs. The second is a judgement the bank has to defend.

The price of a common formula#

The distinction exists because standardisation buys comparability at the cost of accuracy. A formula that must apply to a diversified universal bank, a monoline consumer lender and a state-owned development institution cannot be calibrated correctly for any of them.

It is deliberately built to be roughly right across a population, which means it is systematically wrong in the specific — sometimes conservative, sometimes generous, and rarely in a direction the bank can predict without doing the work. Pillar 2 is the mechanism that reintroduces the specific.

What Pillar 2 has to close#

Three gaps do most of the work.

1. The first is risks the formulas do not capitalise at all: concentration, interest-rate risk in the banking book, sovereign and country exposure, business and strategic risk, model risk, conduct and legal exposure, pension obligations and, increasingly, climate transition and physical risk.
2. The second is risks the formulas capitalise inadequately for a particular institution — a standardised operational-risk charge driven by gross income tells a bank almost nothing about its own control failures, and a credit charge calibrated on a broad population may sit well below the loss experience of a portfolio concentrated in a single sector.
3. The third is the forward view: Pillar 1 is a point-in-time requirement computed on today’s exposures, while an ICAAP has to hold capital adequate for the plan the board approved and for the stress that plan might meet.

A portfolio the formula cannot see#

Concentration is the clearest illustration and the most commonly mishandled. Standardised credit risk weights are calibrated on the assumption of a well-diversified portfolio; they do not increase because a bank’s twenty largest exposures represent a large share of its capital, because half its book sits in one sector, or because its collateral is overwhelmingly one asset class in one city.

Nothing in the Pillar 1 number changes. Everything about the bank’s actual loss distribution does.

An ICAAP that acknowledges concentration in narrative and then assigns it no capital has identified the gap and declined to close it, which reviewers read as the weaker of the two possible answers.

Where capital and funding meet#

Interest-rate risk in the banking book works the same way from a different direction. It generates no Pillar 1 charge, yet a bank funding long fixed-rate assets with short repricing liabilities carries an exposure that can move economic value and net interest income substantially.

The ICAAP is where that exposure is measured, capitalised or explicitly managed within limits, and where the interaction with the funding structure becomes visible — which is also the point at which the capital and liquidity assessments start to describe the same balance sheet rather than two separate ones.

Judged from outside#

A capital add-on is additional capital a supervisor requires an institution to hold above its Pillar 1 minimum, set after reviewing the bank’s own assessment and the evidence behind it.

Add-ons typically arise from one of three findings:

• a material risk the bank has not capitalised at all
• a quantification the reviewer considers insufficiently conservative or insufficiently evidenced
• a governance weakness — inadequate challenge, unvalidated models, a framework that does not influence decisions — that makes the whole assessment less reliable

The third category is worth noting, because it means capital can be required for a process failure rather than a portfolio one, and it is the category a bank most easily prevents.

Adding the same risk twice#

The most common technical error in a first ICAAP runs in the opposite direction: double counting. Teams quantify each material risk independently, sum the results, add the total to the Pillar 1 requirement, and arrive at an internal capital number that no reasonable reading supports.

The overlap is real and it has three sources:

• a risk already partly capitalised under Pillar 1 then capitalised again in full under Pillar 2
• the same loss counted twice through two lenses, as when a sector downturn is captured in both concentration risk and a credit stress scenario
• the absence of any diversification recognition, since the assumption that every material risk crystallises simultaneously and at full severity is more conservative than the evidence supports

The remedy is a documented aggregation approach that states which overlaps were removed, which diversification benefit was recognised and why, and what would happen to the conclusion under a more conservative aggregation.

One page, or four chapters?#

Presentation decides how much of this work survives review.

The most readable ICAAPs show an explicit bridge: the Pillar 1 requirement as the starting point, then each addition with its rationale and derivation, then each overlap removed with its justification, then any diversification benefit with its method, arriving at the internal capital requirement — and finally the comparison of that requirement against available capital resources, in the baseline and under each stress scenario.

A reviewer who can follow that bridge in one page has been given the argument. A reviewer who has to assemble it from four chapters has been given a document.

A floor, not an answer#

The relationship between the two pillars is therefore not additive in the way the arithmetic suggests. Pillar 1 sets a floor computed on a population; Pillar 2 asks whether that floor is the right number for this institution, and requires the bank to say so in its own words, with its own evidence, and to act on the answer.

A bank whose internal assessment lands close to its Pillar 1 requirement has not necessarily done poor work — but it should be able to explain why the standardised calibration happens to fit its portfolio, and reviewers will ask.