Since time immemorial, accountants have had an affinity with historical cost accounting. That might be an overstatement, but it certainly remains a key feature of accounting 101 and particularly so within bank financial reporting, where the majority of loans, deposits and long dated bonds (where held to maturity) are accounted for using amortised cost.

Banks typically hold long dated bonds for both liquidity purposes and yield and will often, at the outset, intend to hold the asset to maturity. The value of listed debt never stays in one place for long but what is known – bar any recoverability issues – is its value on redemption.

So why introduce unnecessary volatility to financial reporting by including gains or losses that are up until this point unrealised when you are looking to keep the asset to term? Effectively, it’s the financial reporting equivalent of being forced to sit through 90 minutes of commentary when you already know which team won and by how many points.

Enter the amortised cost approach: in its initial form, the fair value of the bond when acquired (less transaction costs), and thereafter, a predictable amortisation of the discount or premium reported over the remaining life of the bond. It was an approach that seemed to work well for long-term bonds. That is of course until you have issues with going concern, at which point, as was the case with Silicon Valley Bank (SVB), all bets are off.

As interest rates have risen, the value of outstanding long dated bonds has fallen. However, under an amortised cost approach you would be none the wiser. There is no obligation to write down assets despite unrealised losses lurking because, for example, if you acquired the bond at par or below, those losses will reverse if you hold the asset to maturity. Consequently, balance sheets remain unfettered and continue to appear sound.

While this certainly was a red flag for SVB, it was by no means alone in following this approach. Indeed, where the intention is to hold to maturity, the amortised cost approach is common across US banks.

Ultimately, what set SVB apart was its overweight exposure to the asset class and the absence of appropriate hedging to negate the interest rate risk. Hence, why use of amortised costs played such an important role in the downfall of the bank. It would be unfair to say the risks were not disclosed – but the details were in the notes, not as carrying values on the balance sheet.

However, this still begs the question – is there a point at which an asset’s value has decreased to such an extent that losses warrant recognition, irrespective of whether they are realised or not? The argument for is all the more compelling where recognition means the difference between reporting a profit or a significant loss or where the economics of holding the asset become less and less tenable.

Mark Rhys, Deputy President of ICAEW says: “With adequate liquidity, properly managed interest rate risk and no pressures from worried depositors, SVB’s accounting would have been less of an issue.”

Rhys says: “Rapid rate rises are really testing the market and banks therein and so it is not surprising to see this stretch to discussions around accounting policy.”

The applicable standard in the UK, IFRS 9, does at least include safeguards. For example, firms must apply a business model test when determining if they are eligible for the amortised cost approach. Consideration must be given to the historical and expected frequency of any bond disposals and the reasons why assets have been disposed of early, for example, due to credit deterioration. The standard also sets a high bar for any subsequent reclassification to amortised cost.

Rhys says arguments were waged in the aftermath of the 2007/08 financial crises that accounting policy was somewhat to blame. “Bad debts and losses were not recognised fast enough, despite red flags and apparent danger signs.”

More than a decade later and coming into economic troubles we had comforted ourselves that we could rest easily “reassured that those antiquated approaches to bad debts were long gone, and instead replaced by complex models to inform banks of their expected losses”.

While the failure of SVB is one driven by market risk rather than credit risk and the wider repercussions are now hopefully contained, “we are once again reminded that accounting policy decisions do matter and can have outsized implications, particularly where a bank is facing doubts as to its liquidity”.