Excess Returns Model

Excess Returns · Residual Income · Bank Valuation via Excess Returns

A valuation method for banks that prices the ability to generate returns above the cost of equity — an alternative to the Gordon model for fast-growing or transitioning banks.

ER_t = (ROTE_t − COE) × BV_{t−1}
Valuation

What it is

The Excess Returns Model (residual income model) values a bank as the sum of:

1. Book value (TBV) The starting point — tangible equity today.

2. Present value of excess returns (explicit phase) Over 10 years, we explicitly project how much value the bank creates above the cost of equity:

ER_t = (ROTEt − COE) × BV{t−1}

where:

Every year ROTE > COE, the bank "creates" value. Every year ROTE < COE, it destroys it.

3. Terminal value After year 10 (perpetuity): TV = (ROTE_terminal − COE) × BV_10 / (COE − g)

Resulting equity value = TBV + Σ PV(ER_t) + PV(TV)

This model is the banking equivalent of DCF — instead of FCF it discounts "economic profit" (return above the cost of capital).

Why track it

The Gordon model P/TBV ≈ (ROTE − g) / (COE − g) assumes ROTE is constant both today and in the future. For banks where:

  • ROTE changes significantly over time (rate transitions, restructuring, growing bank)
  • The explicit transition period matters (the bank is investing and ROTE temporarily declines)
  • The timing of reaching "steady state" is important

...the Excess Returns Model is more accurate.

Relationship to the Gordon model: If ROTE is constant in all years, the Excess Returns Model gives the same result as the Gordon model — they are equivalent approaches under the same assumptions.

Watch: how does the P/TBV implied by the model change across ROTE scenarios? Sensitivity to ROTE in years 1–3 is key — the market discounts distant returns more heavily.

Real-world example

Fast-growing bank (model example):

  • TBV today: €10bn, COE: 10%, g: 3%
  • Years 1–5: ROTE 8% (below COE — bank growing, investing in distribution)
  • Years 6–10: ROTE 13% (above COE — scaling, profitability rising)
  • Terminal ROTE: 14%

Gordon model would use current ROTE 8% → fair P/TBV < 1×. Excess Returns Model sees the transition phase and correctly prices P/TBV > 1× thanks to future ROTE of 14%.

Result: The model implicitly prices the "path to profitability" — value that the simple Gordon formula misses.

Watch out for

The model is sensitive to terminal ROTE and COE — just as DCF is to terminal growth and WACC. A small 1% change in ROTE_terminal at COE 10% and g 3% can shift the resulting equity value by 15–20%.

Do not project ROTE that has never been historically achieved without a compelling fundamental reason (new business mix, rates, regulation). The Excess Returns Model is a powerful tool — but garbage in, garbage out applies doubly here.