Funds Transfer Pricing in Banks: How FTP Destroys Client Centricity

Transfer pricing becomes toxic in banks when it stops being plumbing for allocating the cost of funds and becomes the basis for bonus pools. Teams then rationally defend their transfer priced margins against internal colleagues, the interlock process turns cross product deals into adversarial negotiations, and clients experience a bank whose parts do not cooperate. The author argues the mechanism should be replaced, not patched.

Every bank of any size runs on funds transfer pricing, and every bank of any size will tell you, if you catch someone honest in a quiet moment, that FTP is also quietly poisoning the organisation. Funds transfer pricing was supposed to be a plumbing mechanism, a way of fairly allocating the cost of funding, liquidity and interest rate risk between the business units that raise deposits and the business units that lend them out, so that nobody could claim credit for economics they didn’t actually create. That is a reasonable and even necessary idea, and it is why FTP sits at the centre of any serious treasury and balance sheet management function. The problem is what happens when a reasonable treasury mechanism becomes the load bearing wall for how every product team’s bonus pool is calculated, because at that point FTP stops being plumbing and starts being the thing the whole building is designed around.

1. The bonus pool problem

Once funds transfer pricing determines who gets paid, every product team has a rational incentive to optimise their own FTP adjusted number rather than the outcome the client actually experiences. This is not a story about bad people; it is a story about well designed incentives producing exactly the behaviour they were built to produce. A team whose bonus pool depends on the margin between their internal transfer price and their client price will, entirely predictably, defend that margin against every internal request that threatens it, even when the request comes from a colleague trying to serve the same client the bank claims to care about. The result is an organisation full of individually rational teams producing a collectively irrational client experience, which is one of the oldest and most well documented failure modes in economics, and banks somehow keep rediscovering it as if it were a surprise.

2. The interlock tax

The mechanism that is supposed to resolve these conflicts is the interlock process, and it is worth being honest about what interlock actually costs. Every time a deal crosses two product boundaries, someone has to convene a negotiation between teams whose compensation depends on the outcome of that negotiation, which means the negotiation is adversarial by design rather than collaborative by design. Multiply that by the number of cross product deals a bank does in a year and you get an enormous standing army of relationship managers, product specialists, and finance business partners whose actual job, whether or not it appears on an org chart, is fighting internal FTP battles rather than serving external clients. This is not a rounding error. In many institutions the fully loaded cost of interlock, measured in senior people’s time, in deal cycle time, and in deals that quietly die in committee rather than closing, rivals the cost of the front office it is meant to support.

3. The Hilux problem

The clearest illustration I know of is a genuinely small one, and that is exactly why it is worth telling. A business bank wants to lend against a farm, because the farm is a good credit and the relationship is valuable and the banker on the ground has done the work to understand both. The client, sensibly, also wants to insure the vehicle they use to run the farm, and the bank has an insurance team sitting one floor away that could easily write that policy. But the insurance team looks at a Hilux, prices it against their return on equity hurdle, and declines, because a single vehicle policy attached to a rural client does not clear the bar that the insurance business unit has been told to defend. The client walks out with a loan and no insurance, buys the policy from a competitor, and the bank has just demonstrated to its own customer that its various parts do not actually talk to each other in any way that benefits the customer. Nobody in that chain did anything wrong by the measure they were given. That is precisely the point.

4. Client centricity as a casualty

Every bank puts client centricity in its values statement, and most banks mean it when they write it, but values statements do not allocate bonus pools. Transfer pricing frameworks do. When the two are in tension, the FTP framework wins almost every time, because it is precise, measurable, and directly connected to compensation, while client centricity is aspirational, hard to measure, and connected to compensation only through whatever discretionary overlay a manager is willing to apply after the fact. A product team that yields margin to serve a client well is making a real sacrifice against a number someone else will judge them on at bonus time, and asking people to repeatedly sacrifice their own measured performance for an unmeasured collective good is not a sustainable design. It relies on individual heroism to compensate for a structural flaw, and organisations that rely on individual heroism at scale eventually run out of heroes.

5. Keep the plumbing. Abandon the management system.

The temptation at this point is to conclude that funds transfer pricing itself should be torn out of the bank, and that temptation should be resisted, because it hands a treasury specialist an easy way to dismiss the whole argument on technical grounds. FTP genuinely is the right mechanism for allocating funding cost, liquidity risk and interest rate risk across a balance sheet, and no bank of any size can run without something doing that job well. The distinction that actually matters, the one worth defending under real scrutiny, is between FTP as a treasury and risk mechanism and FTP as a management system. Used to price liquidity and funding risk between the treasury function and the business units that consume it, FTP is doing exactly what it was designed to do. Used to decide whether a customer relationship is valuable, to construct rigid product silos, and to determine individual people’s bonuses, the same mechanism becomes something else entirely, something the original treasury logic was never built to carry.

The alternative, then, is not to remove FTP but to strip it back to that original job and refuse to let it drift any further. A bank should be organising its management and compensation decisions around three questions instead, asked at the level of the client and the balance sheet as a whole rather than at the level of any single product. First, is capital being deployed efficiently, meaning is the bank’s scarce capital going to the clients and the exposures that generate the best risk adjusted return for the institution overall, not the best return for whichever product team happens to be holding the capital at that moment. Second, is the client actually being served, meaning does the farmer walk out with both the loan and the insurance, because the relationship as a whole is profitable and durable, even if one product line within it looks unremarkable in isolation. Third, is the bank hitting its cost targets, meaning is the enormous overhead of interlock, of duelling finance business partners, of committees convened to referee internal pricing disputes, being stripped out and redirected toward people who actually talk to clients. None of these three questions require FTP to answer, and none of them should be decided by it. They require a bank willing to measure itself, and pay itself, on capital efficiency, client outcomes, and cost discipline at the group level, keeping funds transfer pricing exactly where it belongs, inside treasury, and nowhere near the org chart or the bonus letter.

6. The Tesla lesson

It is worth stepping outside banking entirely to see how strange this obsession with product level profitability really looks from the outside. Tesla was barely profitable, and for long stretches not profitable at all, for something like eighty percent of its existence as a public company, and yet it became worth more than a trillion dollars, because the market was not pricing the margin on any individual car sold in any individual quarter. It was pricing a vision, executed with relentless efficiency, of what the company was building toward and how completely it intended to serve the people who bought into that vision. SpaceX tells the same story in an even more extreme form, having spent years blowing up rockets on landing pads while simultaneously becoming the most valuable private company on the planet, because nobody serious was asking whether the Falcon 9 program was hitting its internal transfer price against the Dragon capsule division. The value came from a coherent, obsessively client focused vision pursued with real operational discipline, not from a finance function slicing the company into product silos and demanding each one defend its own margin in isolation.

That is the comparison a bank running heavy FTP driven management should sit with, uncomfortably, because it is not really a comparison about industries, it is a comparison about what an organisation chooses to optimise. A bank that lets its insurance team decline a Hilux policy to defend a return on equity hurdle is optimising for the health of a product silo’s quarterly number. A bank that thinks like Tesla or SpaceX would ask instead whether serving that farmer completely, efficiently, and at real capital discipline is building the kind of institution clients want to concentrate their business with for decades. The trillion dollar lesson is not that profitability doesn’t matter, both companies eventually had to and did prove out the economics. The lesson is that vision and client centricity, backed by efficient execution, create value that a spreadsheet of internal prices between business units can never see coming and will actively prevent an organisation from building.

7. Metric punishment

There is a deeper failure hiding underneath all of this, and it shows up most clearly at the executive committee table rather than in the interlock room. When every product line has its own carefully negotiated transfer price and its own tidy scorecard, the exco loses the ability to genuinely challenge each other, because every challenge can be deflected with a metric that says everything is fine. If a colleague raises a concern about how a business is actually performing for clients, the answer arrives instantly and confidently in the form of a dashboard, and the dashboard says the numbers are green, so the conversation ends there rather than starting there. That is metric punishment, the practice of using a measurement as a shield against a debate the organisation actually needs to have, and any exco that has quietly stopped being able to challenge itself because the metrics always say ok has already failed as a leadership team, whether or not anyone in the room has noticed yet.

The genuinely dangerous part is that the metrics are not simply wrong when this happens, they are often precisely calibrated and precisely misleading at the same time. A product can be showing green across every measure the organisation has agreed to track, and be a genuinely terrible product for clients, because the metrics were built to capture FTP adjusted margin and cost efficiency rather than whether a farmer walked out with the insurance he actually needed. Equally, and this is the part people find harder to sit with, a product can be flashing red on every internal measure, hitting none of its transfer pricing targets, missing its return on equity hurdle every quarter, and still be quietly one of the most valuable things the bank does, because it is the product that keeps the whole client relationship together even though no single metric was ever designed to see that. The label the metric gives a product, ok or terrible, is very often the exact opposite of the truth, and an exco that trusts the label instead of interrogating the label has handed its own judgement over to a spreadsheet.

Capitec is the clean local proof of this, and it is worth dwelling on because it undercuts the excuse every FTP driven management defender eventually reaches for, which is that none of this works without obsessive internal measurement. Capitec runs the lowest cost to income ratio of any major bank in the market, and it does so while famously paying the least institutional attention to cost to income ratio as a thing to be managed for its own sake. The efficiency is not the product of a finance team relentlessly interlocking product lines against a CTI target, it is the byproduct of an organisation built around a simple, coherent, obsessively client focused model that never had to be reconciled against a dozen competing product level metrics in the first place. The bank that watches its ratio least ends up with the best ratio, and the bank that watches its ratios most tends to produce exactly the kind of internal warfare this whole piece has been describing. That is not a coincidence, it is the same lesson as Tesla and SpaceX, applied at home, and it is the most direct evidence available that vision and client centricity outperform metric management even on the metric management’s own terms.

8. The killer SLA myth

I once worked inside an organisation where I personally wrote more than two hundred SLA metrics governing the relationship between technology and the business units it served. Every single one of those metrics was green, quarter after quarter, and the clients were miserable the entire time, which should have been the end of the SLA program right there, except it wasn’t, because the organisation had a different theory for why things still felt broken. Each year the business believed there was a killer SLA still waiting to be found, one more contract, one more clause, one more precisely worded metric that would finally close the gap between what the dashboard said and what the client actually experienced. They hired lawyers to draft these agreements properly, as though the problem had been a lack of legal precision all along rather than a lack of anyone actually being accountable for whether the client was happy. Each year, nothing changed, except that the organisation needed more people to track the growing pile of metrics and SLAs, which is its own quiet confirmation that the theory was wrong, because a mechanism that requires ever more headcount to sustain the appearance of health is not a mechanism that was ever going to produce health.

The SLA program and FTP driven management are the same disease wearing different clothes. Both start from the same reasonable sounding premise, that a relationship between two parts of a business can be made to work by writing the right contract between them, and both end up in the same place, an organisation that has hired an army of people to manage the paperwork of internal accountability instead of building a culture where people are simply accountable to the client. Two hundred green metrics did not produce one satisfied client, because the metrics were never actually measuring the thing that mattered, they were measuring whether technology had done what technology promised on paper, which is a completely different question from whether the business unit’s client got what they needed. No SLA was ever going to be the killer SLA, because the entire premise, that the fix was a better contract rather than a better organisation, was broken from the start. Funds transfer pricing deserves the same verdict. Keep it as the treasury mechanism it was built to be. Never again let it decide who gets paid.