Stress-Testing the APTA & TVC: When Stochastic Cashflow Models Cross into "Wishful Thinking"

Stochastic cashflow versus deterministic realities: auditing the APTA and TVC under COBS 19

Under the FCA’s defined benefit transfer regime, the Appropriate Pension Transfer Analysis (APTA) and Transfer Value Comparator (TVC) form the technical core of the suitability assessment.

However, in both Section 166 past-business reviews and Financial Ombudsman Service (FOS) adjudications, cashflow modelling remains one of the most heavily scrutinised areas. Under COBS 19.1.2CR, firms may deploy stochastic cashflow modelling to support advice, provided that projected outcomes at the 50th percentile are no less conservative than if the analysis had been prepared under the prescribed deterministic assumptions in COBS 19 Annex 4A and Annex 4C.

When compliance auditors look under the hood of financial planning software, they frequently find that optimistic stochastic models have been used to paper over structural cracks in transfer sustainability. And that’s not good enough.

The File Review Failure: Masking Depletion Behind Averages

The fatal flaw in retirement cashflow modelling occurs when software projections produce cosmetically reassuring charts that ignore real-world friction.

The 50th-percentile real return illusion

Models that rely on median stochastic returns assume that average performance protects the client, ignoring early market shocks. Under decumulation, an investor does not retire into an average; they retire into a specific, unforgiving sequence of annual returns.

The “omitted fee drag” deficit

Models projecting 5% or 6% real returns while stripping out platform administration, discretionary fund manager (DFM) charges, ongoing adviser retainers, and underlying fund transaction fees. Total advice friction typically amounts to 1.50% to 2.20% per year. Omitting or minimising these deductions artificially inflates capital longevity by 5 to 15 years.

Mismatched inflation modelling

Defined benefit schemes provide statutory indexation (often RPI or CPI capped at 2.5% or 5%). Cashflow models that assume a flat 2.0% inflation rate across a 30-year retirement understate the compounding power of the guaranteed DB escalator against escalating late-life health and living costs.

Auditing the Cashflow Suite: Deterministic Hurdle vs Stochastic Reality

A robust file audit contrasts the prescribed TVC hurdle against the stochastic decumulation engine.

What Quality Assurance Looks For

Gross versus net cashflow runs

Under COBS 19 Annex 4A (3R), all charges that may be incurred as a result of the transfer — ongoing adviser fees, platform charges, portfolio management, and discretionary dealing fees — must be deducted directly from the cashflow model. If cashflows are run gross and charges are relegated to a footnote disclosure, the model is misleading under COBS 19.1.2ED and breaches the fair, clear, and not misleading rule.

Testing sequence-of-returns vulnerability

Does the model apply an early-years market shock? Applying a 15%–25% equity drawdown in Years 1–3 of retirement reveals whether a transferred pot survives, or whether fixed withdrawals crystallise permanent capital losses.

Defensible longevity horizons

Under COBS 19 Annex 4A (7), the analysis must plan for a reasonable period beyond average life expectancy. Stopping a cashflow run at age 82 or 85 because capital exhausts at that point is an automatic fail. Reviewers expect models to stress-test sustainability through to age 95 or 100.

Summary

Stochastic modelling software is a diagnostic stress-testing tool, not a sales aid. When cashflow models rely on aggressive return assumptions or downplay fee drag to bypass the TVC hurdle, they leave the advisory firm with uninsurable liabilities at the complaint stage.