Leases and financial instruments (IFRS 16 and 9)

# Leases and financial instruments (IFRS 16 and 9)

When Tesco adopted IFRS 16 in 2019, roughly £9 billion of lease liabilities materialised on a balance sheet that had, the day before, disclosed them only in a footnote. Nothing about the business changed, not a single store, not one supplier contract, not a pound of cash flow. Yet analysts recalibrated leverage models overnight, and the finance team spent months walking investors through why a company that looked identical was suddenly carrying billions more debt. This is the central problem of technical accounting judgement: the economics are constant, but the *reported* economics move, and the CFO owns the gap between the two.

IFRS 16 and IFRS 9 are the two standards most likely to make your headline metrics diverge from the story you've been telling the market. One rebuilt the liability side of the balance sheet; the other rewired how you recognise credit losses. Both convert what used to be quiet footnote disclosures into P&L volatility and covenant risk. Your job is not to master the accounting mechanics, your controllers do that, but to anticipate where the judgement calls land, how they cascade into ratios and covenants, and what you say before someone else frames it for you.

IFRS 16: the balance sheet grew, and so did the judgement

The mechanical story is well-rehearsed: operating leases came on-balance-sheet as a right-of-use (ROU) asset and a corresponding lease liability, and the straight-line operating lease expense split into depreciation (above EBIT) and interest (below it). The consequence every CFO must internalise is the *front-loading of expense*. Because interest is calculated on a declining liability balance, total expense is higher in the early years of a lease and lower later. For a company on a steady-state lease portfolio this washes out. For a company aggressively expanding its footprint, new retail, new logistics, new office space, reported profit is structurally depressed relative to the old world, even as EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.View full definition → rises because rent has vanished from operating costs.

That EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.View full definition → inflation is not cosmetic. It is where the real work begins.

The judgement calls that actually move the numbers

The standard hands you three levers, and each is a genuine estimate, not a mechanical output:

The discount rate. Where the rate implicit in the lease isn't readily determinable, which is nearly always for a lessee, you use your incremental borrowing rate (IBR). A lower IBR inflates both the ROU asset and the liability; a higher IBR shrinks them and shifts more expense into interest. Auditors scrutinise this because it is the single most sensitive input. The discipline: build an IBR framework that is defensible by lease term, asset class, and currency, and document it *before* the auditors ask. A retailer signing 15-year store leases and 3-year equipment leases at one blended rate is inviting a restatement conversation.

The lease term. IFRS 16 requires you to include renewal and termination options you are "reasonably certain" to exercise. This is where economics meet optimism. Extend the assumed term and you grow the liability, front-load more expense, and change your covenant profile. A CFO who lets operational managers assume every renewal option will be exercised is manufacturing balance-sheet leverage out of hope. Tie the term assumption to hard evidence, leasehold improvements, business-critical locations, exit costs, not to the ambitions in a five-year plan.

The lease-versus-service distinction. IFRS 16 only captures contracts that convey the right to control the use of an identified asset. Outsourcing, cloud, and "as-a-service" arrangements often *don't* qualify, which is precisely why the boundary is contested. A data-centre contract that specifies a particular set of servers may be a lease; a capacity-based cloud contract usually is not. The judgement here has migrated from accounting into procurement strategy: how you structure a contract now determines whether it hits your balance sheet.

Ratios, covenants, and the messaging you can't avoid

The ratio impact is predictable in direction and easy to underestimate in magnitude:

• Net debt and leverage ratios rise as lease liabilities join the debt stack.
• EBITDA rises (rent removed), so *EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.View full definition →-based* leverage may actually improve even as absolute debt climbs.
• EBIT and operating margin fall in growth phases due to front-loading.
• Operating cash flow improves, because the principal repayment element of lease payments moves to financing activities, flattering your cash conversion in a way that is entirely presentational.

The covenant question is where a CFO can get ambushed. Most loan agreements written before 2019 use "frozen GAAP" or explicitly define debt to exclude operating leases, meaning IFRS 16 doesn't technically breach anything. But newer facilities, and any renegotiation, force the issue. The action for Monday morning: audit every material facility for whether its financial covenants are calculated on a pre- or post-IFRS 16 basis, and confirm your banks and your models agree. A silent mismatch between how your treasury team calculates a covenant and how the lender's agent does is a discovered problem, and discovered problems cost you basis points and goodwill.

On investor messaging, the credible move is to lead with a clear bridge: pre-IFRS 16 and post-IFRS 16 metrics, side by side, consistently, every period. The market rewards continuity. When you change the denominator of your own leverage guidance without explanation, you invite the assumption that you're hiding deterioration. Say what changed, quantify it, and hold the presentation stable.

IFRS 9: provisioning for losses you haven't suffered yet

IFRS 9 did to the asset side what IFRS 16 did to the liability side, it made a footnote judgement into a P&L-moving estimate. The philosophical shift is the move from *incurred loss* to *expected credit loss* (ECL). Under the old regime, you provisioned when a loss event had occurred, a customer defaulted, a receivable soured. IFRS 9 requires you to provision for losses you expect *before* they happen, using forward-looking information.

For banks this was seismic. But every corporate CFO with trade receivables, intercompany loans, contract assets, or debt investments now runs an ECL model. If your balance sheet holds financial assets measured at amortised cost or FVOCI, you provision for expected losses from day one.

The three-stage model and where judgement bites

IFRS 9 sorts financial assets into three stages:

• Stage 1: performing assets. You book 12-month expected losses.
• Stage 2: assets that have suffered a *significant increase in credit risk* (SICR) since origination. You book *lifetime* expected losses.
• Stage 3: credit-impaired assets. Lifetime losses, and interest recognised on the net carrying amount.

The migration from Stage 1 to Stage 2 is the cliff edge. When an asset "steps" from 12-month to lifetime ECL, the provision can jump dramatically in a single period. The definition of SICR is a matter of judgement, your judgement, and it determines the timing and size of that hit. This is why COVID-19 produced such divergent provisioning across companies with similar exposures: some tightened their SICR triggers and staged assets down early, taking large charges upfront; others held firm and absorbed losses later. Same economics, wildly different reported earnings paths.

Forward-looking information: the discretion regulators watch

ECL must incorporate reasonable and supportable forward-looking information, typically macroeconomic scenarios (base, upside, downside) with assigned probability weights. Here is the uncomfortable truth: this framework hands the CFO a lever over reported profit that is entirely legitimate and entirely discretionary. Shift your downside-scenario weighting from 20% to 30% and your provision rises; your earnings fall. Auditors and regulators know this, which is why the governance around scenario selection and weighting is now a board-level control, not a modelling detail.

The corporate CFO who thinks ECL is "a bank problem" is exposed. If you sell on credit, the *simplified approach* for trade receivables still requires a lifetime-ECL provision matrix built on historical loss rates adjusted for forward-looking conditions. When you enter a recession, that matrix should move, and if it doesn't, your auditor will ask why you're provisioning for a downturn using boom-era loss rates.

The messaging problem: volatility that isn't cash

The strategic communication challenge with IFRS 9 mirrors IFRS 16 but is sharper because the movements are less predictable. ECL charges are non-cash, forward-looking, and reversible, a provision taken in a downturn can unwind when the outlook improves, flattering a later period. Investors who don't understand this will read a large provision as a confession of bad lending or bad customers. The disciplined CFO separates, in every disclosure, the *underlying credit performance* from the *model-driven provision movement*, and explains the macro assumptions driving the latter. When Barclays and Lloyds took multi-billion-pound ECL charges in early 2020, the ones who communicated best were explicit: "This reflects our economic scenario weighting, not realised defaults." The market can price a model assumption. It cannot price a mystery.

Where the two standards converge on your desk

The reason to teach these together is that they interact, and the interaction is where sophisticated CFOs earn their seats.

They both widen the gap between reported profit and cash. IFRS 16 front-loads non-cash depreciation and interest; IFRS 9 front-loads non-cash provisions. Together they can depress reported earnings in a growth-and-caution environment while cash generation is perfectly healthy. If your incentive plans, your dividend policy, or your covenant headroom are anchored to reported profit, you are exposed to accounting-driven swings that have nothing to do with operational reality. Revisit whether your management incentives and internal targets should be defined on a pre-standard, cash-adjusted, or as-reported basis, and make that choice deliberately.

They both convert estimation into earnings management risk. The discount rate, the lease term, the SICR trigger, the scenario weights, these are all defensible ranges, and the point within the range you choose moves the P&L. The governance answer is the same for both: a documented, consistent methodology, applied through the cycle, changed only for evidenced reasons, and disclosed. Consistency is your defence against the accusation that you flex the estimates to hit a number. The moment you tighten SICR in a good quarter and loosen it in a bad one, you've traded credibility for a short-term print.

They both demand a proactive investor bridge. In both cases, the CFO's edge comes from framing the change before the analyst community does. Provide the reconciliations, hold the presentation constant, and separate accounting geography from economic substance. The companies that lost credibility in 2019-2020 weren't the ones with the biggest lease liabilities or provisions, they were the ones who let the market discover the impact and infer the worst.

Key Takeaways

• Audit every material debt facility for its covenant basis (pre- vs. post-IFRS 16) and reconcile it against your own models and your lenders' calculations. A silent mismatch is a discovered problem, and discovered problems are expensive.
• Build and document a defensible IBR and lease-term framework by asset class before the auditors ask. These two inputs move the balance sheet more than any other, and the lease-term assumption must be tethered to evidence, not to your growth ambitions.
• Treat your IFRS 9 scenario weighting as a board-level control, not a modelling detail. It is a legitimate but powerful lever over reported profit; apply it consistently through the cycle so no one can accuse you of flexing estimates to hit a number.
• In every disclosure, separate accounting geography from economic substance, provide consistent pre/post bridges for IFRS 16 and split underlying credit performance from model-driven ECL movements. Frame the change before the analysts do.
• Deliberately choose the basis for incentives, dividends, and internal targets given that both standards widen the gap between reported profit and cash. Don't let non-cash, estimate-driven volatility quietly distort decisions and rewards.