Financial instruments II: credit losses and hedge accounting

The global financial crisis of 2008 exposed a weakness in the old incurred loss models: banks could not recognize credit losses until a loss event had occurred, so allowances were too little and too late. Both standard setters replaced those models with forward-looking expected loss models, but they could not agree on a single one. IFRS 9 recognizes twelve months of expected losses at first and lifetime losses only after a significant increase in credit risk. The FASB's current expected credit loss (CECL) model, in ASC 326, recognizes lifetime expected losses from the day an asset is recognized. For a food manufacturer the effect is smaller than for a bank, but the same models apply to trade receivables, loans and contract assets. The second half of the lesson turns to hedge accounting, where the frameworks again differ in important details.

Expected credit losses: the IFRS 9 general approach

CECL requires an entity to estimate expected credit losses over the contractual life of a financial asset measured at amortized cost, adjusted for expected prepayments, from initial recognition. The estimate reflects historical loss experience, current conditions and reasonable and supportable forecasts, reverting to historical experience for periods beyond the forecast horizon. There are no stages: the allowance is always lifetime. For comparison, IFRS 9's general approach starts with twelve months of expected losses (stage 1) and moves to lifetime losses only after a significant increase in credit risk (stage 2) or when the asset is credit-impaired (stage 3), which produces a cliff effect when an asset changes stage. Available-for-sale debt securities are not in CECL's scope; their credit losses are recognized through an allowance under a separate impairment model in ASC 326-30.

Measuring expected losses requires three inputs: the probability of default (PD) over the relevant horizon, the loss given default (LGD), the share of the exposure that would not be recovered, and the exposure at default (EAD). The estimate is unbiased and probability-weighted, considers a range of possible outcomes and reflects the time value of money. Forward-looking information, such as forecasts of unemployment, sector insolvencies or commodity prices, must be incorporated. Banks build complex models for these inputs; a manufacturer like Tamarind Bay uses simpler approaches but faces the same principles.

# Credit losses at December 31, 2026 (invented). IFRS 9 expected credit losses vs ASC 326 CECL.
# 1. Supplier loan of 25.0 (lesson 7), amortized cost.
ead, lgd = 25.0, 0.45
pd12, pd_life = 0.012, 0.045          # at origination
print("Supplier loan")
print(f"  IFRS 9 stage 1 (12-month ECL): {ead * pd12 * lgd:.3f}")
print(f"  ASC 326 CECL (lifetime from day one): {ead * pd_life * lgd:.3f}")
pd_life2 = 0.110                      # after the supplier lost a major customer
print(f"  After significant increase in credit risk -> IFRS 9 stage 2 (lifetime): {ead * pd_life2 * lgd:.3f}")
print(f"  CECL updated lifetime estimate: {ead * pd_life2 * lgd:.3f}")
# 2. Trade receivables: provision matrix with a forward-looking overlay.
buckets = [("Current", 168.0, 0.004), ("1-30 days", 41.5, 0.012), ("31-60 days", 12.2, 0.045),
           ("61-90 days", 4.6, 0.140), ("Over 90 days", 3.1, 0.480)]
overlay = 1.10                        # forecast: higher grocery-sector insolvencies
print("\nTrade receivables (lifetime losses under both frameworks)")
print(f"{'Age':<14}{'Balance':>9}{'Hist. rate':>12}{'Adjusted':>10}{'Allowance':>11}")
tot = tot0 = 0
for b, bal, r in buckets:
    a = bal * r * overlay; tot += a; tot0 += bal * r
    print(f"{b:<14}{bal:>9.1f}{r:>12.1%}{r * overlay:>10.2%}{a:>11.2f}")
print(f"{'Total':<14}{sum(x[1] for x in buckets):>9.1f}{'':>22}{tot:>11.2f}")
print(f"Allowance without the forward-looking overlay: {tot0:.2f}")
print("ASU 2025-05 expedient: assume current conditions persist for current receivables ->", f"{tot0:.2f}")

Supplier loan
  IFRS 9 stage 1 (12-month ECL): 0.135
  ASC 326 CECL (lifetime from day one): 0.506
  After significant increase in credit risk -> IFRS 9 stage 2 (lifetime): 1.238
  CECL updated lifetime estimate: 1.238

Trade receivables (lifetime losses under both frameworks)
Age             Balance  Hist. rate  Adjusted  Allowance
Current           168.0        0.4%     0.44%       0.74
1-30 days          41.5        1.2%     1.32%       0.55
31-60 days         12.2        4.5%     4.95%       0.60
61-90 days          4.6       14.0%    15.40%       0.71
Over 90 days        3.1       48.0%    52.80%       1.64
Total             229.4                             4.24
Allowance without the forward-looking overlay: 3.85
ASU 2025-05 expedient: assume current conditions persist for current receivables -> 3.85

For the 25.0 million supplier loan, CECL requires the lifetime loss of 0.506 million from day one, while IFRS 9 would initially require only the 12-month loss of 0.135 million. In November 2026 the supplier lost a major customer, and the lifetime PD rose from 4.5 to 11.0 percent. Under CECL, Tamarind Bay simply updates its lifetime estimate to 1.238 million. Under IFRS 9 a Canadian competitor would have to decide whether credit risk had increased significantly, which it almost certainly had; the loan would move to stage 2 and the allowance would jump to the same lifetime figure. The difference between the frameworks is greatest for new, performing assets, where CECL front-loads the allowance.

Trade receivables and the provision matrix

For trade receivables, CECL also requires lifetime losses, and a provision matrix is a common method. The program applies a 10 percent forward-looking overlay to Tamarind Bay's historical loss rates because economists forecast higher insolvencies among regional grocers in 2027, raising the allowance from 3.85 to 4.24 million. ASU 2025-05, issued in July 2025 and effective for annual periods beginning after December 15, 2025, offers relief: all entities may elect a practical expedient that assumes current conditions at the balance sheet date do not change for the remaining life of current accounts receivable and current contract assets arising under ASC 606, removing the need for reasonable and supportable forecasts for these short-term balances. Entities other than public business entities may also make a policy election to consider collection activity after the balance sheet date. If Tamarind Bay elects the expedient in 2026, its allowance would be the 3.85 million based on current conditions, although it must still consider current conditions, such as known customer difficulties, at the balance sheet date.

For comparison, Canadian private enterprises using ASPE still apply an incurred loss model, recognizing impairment only when there are indications such as significant financial difficulty of the debtor, so their allowances are typically smaller than those under either CECL or IFRS 9.

Hedge accounting: why it exists

Hedge accounting is optional. It exists because the normal measurement rules can create accounting mismatches: a derivative used to hedge a risk is measured at fair value through profit or loss, while the hedged item, such as a forecast purchase, is not recognized at all until it occurs. Without hedge accounting, profit would swing with the derivative even though the economic position is hedged. Hedge accounting changes the timing of recognition so that gains and losses on the hedging instrument and the hedged item affect profit in the same period. In exchange, the entity must designate the relationship at inception, document the risk management objective and strategy, and show that the hedge is expected to be effective.

Tamarind Bay's cash flow hedge of natural gas purchases

Tamarind Bay's Frozen plants use large amounts of natural gas for cooking and steam. To protect its margins, it hedges about 70 percent of forecast gas purchases for the next twelve months with pay-fixed commodity swaps and designates them as cash flow hedges of highly probable forecast purchases. ASC 815 requires the hedge to be highly effective, generally interpreted as offset within 80 to 125 percent, both at inception and on an ongoing basis. After ASU 2017-12, an entity may perform the initial quantitative test and then use qualitative assessments if facts support it, and, for a highly effective cash flow hedge, it records the entire change in the fair value of the hedging instrument in OCI, without separately measuring ineffectiveness.

# Cash flow hedge of forecast natural gas purchases with pay-fixed swaps (ASC 815; invented).
# After ASU 2017-12, a highly effective hedge records the whole change in the swap's fair value
# in OCI; amounts are reclassified to earnings when the gas affects cost of sales.
periods = [  # cumulative changes in fair value, millions of dollars
    ("Q1", 1.10, -1.05), ("Q2", 2.40, -2.28), ("Q3", 0.90, -0.97), ("Q4", 3.20, -2.95)]
print(f"{'Quarter':<8}{'Swap':>8}{'Hedged item':>13}{'Offset':>9}{'Highly effective?':>19}{'To OCI':>9}")
for q, swap, item in periods:
    ratio = -swap / item
    he = 0.80 <= ratio <= 1.25
    print(f"{q:<8}{swap:>8.2f}{item:>13.2f}{ratio:>9.0%}{'yes' if he else 'NO':>19}{swap if he else 0:>9.2f}")
print("Dollar-offset within 80-125% is a common way to show the hedge is highly effective.")
print("No ineffectiveness is measured separately; the full change is deferred in OCI.")

Quarter     Swap  Hedged item   Offset  Highly effective?   To OCI
Q1          1.10        -1.05     105%                yes     1.10
Q2          2.40        -2.28     105%                yes     2.40
Q3          0.90        -0.97      93%                yes     0.90
Q4          3.20        -2.95     108%                yes     3.20
Dollar-offset within 80-125% is a common way to show the hedge is highly effective.
No ineffectiveness is measured separately; the full change is deferred in OCI.

Because the hedge is highly effective in every quarter, the whole change in the swap's fair value is deferred in OCI, and nothing is recognized in earnings until the hedged gas purchases affect cost of sales, at which point the accumulated amount is reclassified into the same line item. IFRS 9, by contrast, would recognize the excess of the hedging instrument's cumulative change over the hedged item's cumulative change as ineffectiveness in profit or loss, about 0.25 million by the fourth quarter on these figures, and would require the deferred amount to be included in the initial cost of the purchased inventory as a basis adjustment. The US approach produces less volatility in earnings for highly effective hedges; the IFRS approach shows imperfections in the hedge as they occur.

Hedge accounting judgments

Several judgments recur. The first is whether a forecast transaction is highly probable: a purchase that the company expects but has not committed to must be supported by budgets, past patterns and the company's ability to carry it out, and hedging 100 percent of forecast volumes invites challenge. If a forecast transaction is no longer expected to occur, the accumulated amount in OCI must be reclassified to profit immediately, and a pattern of forecasts that fail to occur can call into question the entity's ability to forecast at all. The second is the hedge ratio and sources of ineffectiveness, such as differences in timing, location or quality between the derivative's underlying and the hedged item, or the credit risk of the counterparty. The third is the hedged risk component: IFRS 9 permits hedging a separately identifiable and reliably measurable component of a non-financial item, such as the benchmark gas price embedded in a delivered price, and ASU 2017-12 introduced a similar contractually specified component approach in US GAAP.

The IASB published a request for information on the post-implementation review of IFRS 9's hedge accounting requirements in September 2026, so the IFRS model may change. The FASB continues to make targeted improvements to ASC 815; companies should check for recent updates when designing new hedge programs.

Disclosures

Credit risk disclosures explain the entity's credit risk management practices, the inputs, assumptions and techniques used to measure expected losses, how forward-looking information is incorporated, changes in the allowance during the period, the credit quality of financial assets and concentrations of risk. Hedge accounting disclosures cover the risk management strategy, how hedging affects the amount, timing and uncertainty of future cash flows, and the effects on the financial statements, including amounts in the hedge reserve and, under IFRS 9, ineffectiveness recognized. Tamarind Bay's largest customer accounts for 14 percent of revenue, a concentration that must be disclosed and that also matters for the going concern assessment in lesson 12.