The Bank That Was Growing Itself to Death
How a Regional Financial Institution Discovered That Its Best Year on Record Was Actually Its Worst Strategic Decision
By Pedro San Martín — Asher & Company
The numbers looked extraordinary. Loan portfolio up 34%. Net interest income at a five-year high. Market share in corporate lending climbing in three consecutive quarters. The CEO was preparing remarks for the annual shareholder meeting. The investment banking team was circling with acquisition targets. Everyone in the building was feeling good.
Everyone, that is, except the CFO.
She had been sitting on a problem for eight months. Not a problem anyone else could see in the standard financial reports — because the standard financial reports weren't built to see it. The bank — I'll call it AtlantaBanCo, a mid-size regional institution with operations spanning the southeastern United States, a correspondent banking corridor through Miami into LATAM, and a growing trade finance book tied to Central American and Colombian commercial flows — was, by every conventional metric, thriving.
But she had a number that didn't fit the narrative. A single output from their newly deployed Oracle EPM Cloud Profitability and Cost Management module.
Risk-Adjusted Return on Capital (RAROC) for the corporate lending book: 6.2%.
The bank's hurdle rate: 12%.
The CFO called Pedro San Martín at Asher & Company on a Friday afternoon. She didn't say much. She sent the file and three words: "Tell me I'm wrong."
She wasn't wrong.
The Battlefield Context
AtlantaBanCo had done what every ambitious regional bank does when rates are favorable and capital is available: it said yes. Yes to the USD 80M syndicated corporate loan at razor-thin spread. Yes to the mortgage portfolio acquisition that looked clean on NII but ugly on capital consumption. Yes to the anchor commercial client that brought ten subsidiaries — each with a full-service banking relationship and none of them individually profitable enough to justify the relationship management overhead they consumed.
The growth had been real. The revenue had been real. The capital deployment had been very real.
What had not been real — what had never been calculated — was the full cost-to-serve at the segment level. The activity costs. The regulatory capital charges. The credit risk-adjusted margin. The operational complexity premium that came with managing 47 distinct corporate relationships, each with bespoke covenant structures, each requiring dedicated credit officers, legal review cycles, and compliance reporting.
In short: AtlantaBanCo had been pricing its loans the way regional banks have priced loans for forty years — off a spread over SOFR, adjusted for perceived credit quality, blessed by a credit committee that looked at default probability but never, not once, asked: what does it actually cost us, fully loaded, to serve this customer?
Oracle PCM answered that question. The answer was brutal.
I. The Pressure Point
The pressure arrived from three directions simultaneously, which is usually how these crises work — not as a single dramatic event but as a convergence of signals that individually look manageable and collectively look like a wall.
The regulator. The OCC had included AtlantaBanCo in a thematic review of capital allocation practices at regional banks with significant LATAM correspondent flows. The question on the table: were risk weights accurately reflecting the economic reality of the cross-border portfolio? The implicit message: your capital adequacy looks fine on paper. We're not sure it looks fine in practice.
The rating agency. Moody's had placed AtlantaBanCo on review for possible downgrade — not because of credit quality concerns, but because the ratio of risk-weighted assets to Tier 1 capital had expanded 18 percentage points in 24 months. The bank had grown. But it had grown faster than its capital base could responsibly support. The rating agency's analyst wrote, in a note that the CFO read three times: "Revenue growth without commensurate return on risk-weighted assets is not a performance story. It is a risk story."
The board. A newly appointed independent director, a former regional bank CEO, asked a question at the Q3 board meeting that no one had a clean answer to: "Can someone show me the RAROC by business segment, fully loaded for cost-to-serve and capital charges?"
The silence in the boardroom lasted eleven seconds. The CFO timed it.
That was the moment AtlantaBanCo's relationship with Oracle PCM changed from a reporting tool to a survival tool.
II. The PACE Causal Model
III. The Data
Segment Revenue & Risk Profile
| Segment | Loan book USD M | NII USD M | Gross margin % | RAROC % | Hurdle rate % | Status |
|---|---|---|---|---|---|---|
| SME lending | 420 | 8.8 | 2.1% | 18% | 12% | VALUE CREATING |
| Corporate loans | 1,240 | 9.9 | 0.8% | 6% | 12% | BELOW HURDLE |
| Retail mortgages | 890 | 9.8 | 1.1% | 9% | 12% | BELOW HURDLE |
| Trade finance (LATAM) | 310 | 8.7 | 2.8% | 22% | 12% | VALUE CREATING |
| Total portfolio | 2,860 | 37.2 | 1.3% | 11% | 12% | BELOW HURDLE |
Oracle PCM Output: Fully Loaded Cost-to-Serve by Segment
| Segment | NII per USD M booked | Fully loaded cost-to-serve | Net economic margin | Capital charge | Economic P&L |
|---|---|---|---|---|---|
| SME lending | 21,000 | 8,400 | 12,600 | 6,200 | +6,400 |
| Corporate loans | 8,000 | 6,800 | 1,200 | 8,900 | −7,700 |
| Retail mortgages | 11,000 | 7,200 | 3,800 | 7,100 | −3,300 |
| Trade finance | 28,000 | 9,100 | 18,900 | 5,400 | +13,500 |
All figures in USD per USD 1M of outstanding balance
Activity Cost Breakdown: Corporate Lending (Oracle PCM Activity Tracing)
| Activity pool | Annual cost USD M | Cost per loan |
|---|---|---|
| Credit origination & underwriting | 4.2 | 210,000 |
| Covenant monitoring & reporting | 3.1 | 155,000 |
| Relationship management (RM time) | 5.8 | 290,000 |
| Legal & compliance review | 2.9 | 145,000 |
| Regulatory capital reporting | 1.7 | 85,000 |
| Total cost-to-serve per deal | 885,000 | |
| Average deal NII contribution | 640,000 | |
| Net economic loss per deal | −245,000 |
LATAM Corridor Exposure & Trade Finance Profitability
| Corridor | Volume USD M | Revenue USD M | RAROC % | Flag |
|---|---|---|---|---|
| Miami → Colombia | 94 | 2.6 | 24% | LOW |
| Miami → Central America | 67 | 1.9 | 19% | MEDIUM |
| Miami → Dominican Republic | 58 | 1.6 | 21% | LOW |
| Correspondent banking (Mexico) | 91 | 2.6 | 22% | LOW |
| Total LATAM trade | 310 | 8.7 | 22% |
IV. The Three Decisions
Option A — Reprice the Corporate Book: Raise Rates or Exit Relationships
Use Oracle PCM's customer-level economic P&L as the basis for a systematic repricing exercise. For each of the 14 economically unprofitable corporate relationships, present a repricing proposal that brings RAROC to the hurdle rate. Accept that some clients will leave.
PACE stage: Execute — immediate (0–6 months)
Expected benefits:
- Corporate RAROC improves from 6% to 13–15% on remaining book
- Risk-weighted asset density drops — improves capital adequacy ratios
- Eliminates the regulatory narrative risk with the OCC
- Stops the economic value destruction immediately
Risks:
- Volume loss of 20–35% of corporate book — NII drops short-term
- Cross-sell relationships become complicated to unwind
- Competitor banks absorb the exited clients — optics problem at board level
- Sales team morale and retention risk is real
Trade-off: The most economically correct option and the most organizationally painful one. The bank will look smaller on the outside and healthier on the inside. That requires board-level conviction, not just CFO conviction.
Option B — Reallocate Capital: Shrink Corporate, Grow SME and Trade Finance
Use natural run-off in the corporate book to redirect capital toward SME lending and LATAM trade finance. Implement a hard capital ceiling on new corporate originations. Set growth targets in SME and trade finance funded by freed capital.
PACE stage: Validate → Execute (6–18 months)
Expected benefits:
- Portfolio RAROC rises from 11% to 16–17% over 18 months without confrontational repricing
- Trade finance expansion deepens the LATAM corridor — strategic differentiator vs. larger competitors
- SME growth lowers concentration risk and improves capital efficiency
- Board gets a growth story, not a shrinkage story
Risks:
- Requires disciplined origination gates — sales teams must stop doing what they've been rewarded to do
- Capital reallocation is slow: corporate book has 3–5 year average tenors
- Trade finance requires operational capability investment in Colombia and Central America
- If run-off is slower than expected, capital remains trapped in below-hurdle assets
Trade-off: The most politically viable option internally. The risk is execution discipline — capital reallocation strategies fail when the sales culture doesn't change in parallel.
Option C — Build a RAROC-Linked Pricing Engine Inside Oracle PCM
Invest 18–24 months building a forward-looking pricing engine within Oracle PCM that makes RAROC the primary input to loan pricing. Every new origination runs through the model. Relationship managers see the economic P&L before the deal goes to credit committee. Incentive structures are restructured around risk-adjusted returns, not volume.
PACE stage: Formulate → Validate → Execute (18–24 months)
Expected benefits:
- Permanent structural fix — prevents the problem from recurring
- Creates a genuine competitive capability: most regional banks cannot price this way
- Aligns the entire origination process with shareholder value creation
- Positions AtlantaBanCo as a sophisticated risk manager — relevant to both the OCC and Moody's
Risks:
- Eighteen to twenty-four months leaves the existing book unaddressed
- Requires significant change management — RM culture, incentive redesign, credit committee process
- Legacy core banking system requires substantial ETL work to feed PCM cleanly
- Economic value destruction continues while the engine is being built
Trade-off: The only option that doesn't just solve the current problem — it eliminates the conditions that created it. The CFO who chooses Option C is betting on transformation, not repair.
V. Lessons Learned
1. Revenue growth and value creation are not the same thing. In banking, they can be opposites. AtlantaBanCo's best revenue year was also its worst year for economic value creation. Every dollar of NII growth in the corporate book came with a negative economic margin after capital charges and cost-to-serve. The bank was not generating returns. It was generating activity — and mistaking the activity for returns.
2. The credit committee is not a profitability committee. In most regional banks, the credit committee decides whether to make a loan. It does not decide whether the loan is worth making. That distinction — between creditworthiness and economic value creation — is precisely what Oracle PCM closes. A borrower can be creditworthy and unprofitable. Most banks don't find out until it's too late.
3. Relationship banking is a cost structure problem masquerading as a revenue strategy. The corporate clients that generated the most NII also generated the most cost-to-serve. Until Oracle PCM traced those activities to specific relationships, AtlantaBanCo had no mechanism to know that its deepest relationships were its most economically damaging ones.
4. RAROC without activity-based costing is just math without a foundation. Many banks calculate RAROC. Few calculate it correctly, because they use allocated overhead rather than traced activity costs. AtlantaBanCo's corporate RAROC looked like 9% on allocated costs. It looked like 6% on traced costs. That three-percentage-point gap — applied across a USD 1.24 billion portfolio — represented approximately USD 37 million in annual economic value destruction.
5. The LATAM trade finance corridor was the hidden crown jewel. While the organization was obsessing over the corporate book, the highest-returning segment in the bank was operating quietly and underinvested in the Miami-to-LATAM corridor. RAROC of 22%. Capital intensity below the portfolio average. Oracle PCM made this visible for the first time. Your best business may not be where your most senior people are spending their time.
6. Incentive misalignment is a costing problem in disguise. AtlantaBanCo's relationship managers were paid on volume and NII. They were never paid on RAROC. You cannot hold people accountable for a metric you cannot calculate. Oracle PCM made the metric calculable. The incentive redesign became possible only because the measurement became possible.
7. When growth destroys value, urgency is not optional. The CFO had the data for eight months before the board finally saw it. Eight months of continued origination in a below-hurdle corporate book. Oracle PCM's value is not the reports it produces. It is the decisions it accelerates.
Pedro San Martín is Managing Partner at Asher & Company, a specialized advisory firm in Enterprise Performance Management, profitability architecture, and strategic finance across the Americas and Europe. He works with CFOs and finance leaders to connect EPM technology — including Oracle EPCM — to decisions that move the needle on enterprise value.
asher.company | Battlefield Lessons — real decisions, real consequences, no textbook answers.
