Are three buckets enough for asset impairment?

GLOBAL STANDARD SETTERS are working frantically towards a December deadline to put the finishing touches on some of the most contentious standards in financial reporting.

Impairment of financial instruments is among the ends to be tied off, and the standard is in sharp focus as the credit crisis highlighted glaring errors in the current model.

Financial services are particularly affected by the standard due to their complex financial dealings, but it also applies to other sectors.

Credit calamity

In 2008, banks were following a system of incurred loss reporting, meaning assets were marked down, or impaired, only once their value had demonstrably fallen.

Critics said this caused catastrophic shortcomings in financial early warning systems, meaning banks were unable to build up reserves for expected losses and were woefully unprepared when asset values suddenly went into freefall.

The International Accounting Standards Board was set to work remedying the issue and models for expected loss reporting that allow for forward-looking provisioning were developed.

Three-bucket solution

The current favourite is known as the three-bucket approach, designed to help banks make responsible provisions and decide when and how to impair assets and make corresponding provisions.

One pre-IFRS problem was earnings management, when banks would set aside provisions with little justification, only to release them in lean years to plump up earnings. Critics said this made it hard for investors to get a handle on banks’ true financial positions; from these concerns was born incurred loss reporting.

Post-crisis, the standard setters’ problem was therefore how to permit the judgement essential for expected loss provisioning without paving the way for a potential return to earnings management.

The three-bucket approach aims to break down assets according to impairments, keeping a tighter rein on provisioning and giving analysts a clearer picture of financial health.

Into bucket one goes ‘healthy’ assets, those for which banks expect a reasonable return and need only make minimal provisions. Bucket two is reserved for assets with some level of impairment, but which are not completely useless, while bucket three is for assets that are undeniably ‘bad’.

Throughout its life, the asset can move between buckets according to macro- and micro-economic triggers, hopefully allowing banks to make exactly the right provision at exactly the right time.

Bucket of mortgages

An example might be a bundle of mortgages. The bank grants the mortgages, and works out on the basis of historical data that it is likely to take an 80% return on them. It therefore makes provision for the 20% loss and the mortgage bundle sits in bucket one until a trigger makes re-evaluation necessary.

This trigger could be a macro-economic event such as falling oil prices, a contracting economy or rising unemployment. From this, the bank might deduce that a greater proportion of mortgage holders will struggle to pay and shift the asset bundle into bucket two, requiring higher provisions to be made.

For the mortgages to jump to bucket three, they must be demonstrably impaired, for example when the inhabitants of a town hit by unemployment begin defaulting on their mortgages. This is essentially an incurred loss model and would result in very high or 100% provisioning for the de-valued assets.

Unfinished business

Like all theoretical models, there is much uncertainty to be hammered out. What constitutes a bucket-moving trigger? When an asset is impaired, who decides whether the impairment is expected – therefore already provided for – or unexpected, meaning more cash should be set aside? How will auditors examine such a complicated model and will it really prevent earnings management if banks are determined to do it?

One expert said that banks must be disciplined when deciding how macro-economic changes affect buckets and provisioning. This would be the case from the start of the asset’s life, as account preparers could feasibly make excessive provisions for it in order to build up an earnings-management style cushion.

IFRS expert Andrew Spooner, partner at Deloitte, said there could be a way around this. When banks make a loan, they price it according to the perceived credit quality. The correct provisioning should be calculated from this, meaning an expensive loan to a risky borrower would trigger higher provisioning, whereas a ‘safe’ loan would require minimal cash reserves.

Provision questions

In the past, critics have been concerned that provisions were not built up fast enough. By the end of an asset’s life, 100% provisioning would be in place, but what if it failed in year two or year four? The bucket model attempts to address this by putting in place basic provisioning from day one and requiring higher cash reserves as soon as triggers – those bucket-moving cues – indicate that impairment risk is rising.

Currently, the IASB is suggesting that bucket one contain provision only for the first 12 months of an asset’s life. Critics say this could result in a similar problem as the old model, whereby a sudden impairment early in the asset’s life catches banks on the hop with too little provisioning in place.

Standard setters must make clear when provisioning should be increased but assets stay in the same bucket, as well as when triggers are so severe that they warrant a bucket hop.

They are also undecided on provisioning in bucket one after the first 12 months are up. Should cash reserves be ramped up to cover the lifetime of the loan, or should the asset move categories so that further provisioning can be made?

The model remains in its infancy but experts say the structure is pretty secure. Much of the outstanding work lies in the specifics: how should each bucket be defined, what are the triggers that make assets move buckets and which triggers justify higher provisioning without moving category?

Other questions swirl around the interdependencies between macro- and micro-economic factors. When does a big problem become a bank’s problem? How long does it take for climbing unemployment to filter through to mortgage holders, and what are the unforeseen events that might skew the picture?

The IASB’s September board meeting should cast some light on the issue, though standard setters could find it throws up as many questions as it answers.

Related reading

Fiona Westwood of Smith and Williamson.