To address this issue, the International Accounting Standards Board (IASB) has proposed an alternative (see attached document). It's a portfolio approach that requires the lender to continuously project total expected losses. The expected losses are then amortized over the life of the portfolio and netted against interest income. For example if you project a 20% total principal loss on the portfolio over the next 5 years, you would be deducting 4% of the initial portfolio face value from the interest income going forward. It's a constant dollar amount taken out of interest income every year. That means if loans default at some constant rate, interest income will drop off, while the provision will stay constant, creating a possibility of net interest loss.
Here is a comparison of the current method with the "expected loss model":
IASB: Interest revenue that is recognised will reflect the allocation of expected credit losses over the life of the instrument. This is a better reflection of the ‘economic’ interest that the lender expects to earn from an asset over its life than today’s approach. Hence, it avoids inappropriate front-loading of interest revenue.
This approach becomes more problematic when the expectations for credit losses suddenly change. That may mean that the reserve has been "under-accrued", and IASB would say that you have to take that difference into P&L immediately. And this concept makes it the trickiest portion of the proposal.
IASB: Using the proposed impairment method, credit loss expectations are updated each period. Any changes to initial expectations of credit losses will be recognised immediately in P&L. This change could be an increase in expected losses, or a reduction (reversal) of past expected losses (including the initial expected loss estimate).
If one uses CDS spreads for example to imply credit loss expectations, this method amounts to a form of mark to market. It effectively means that rather than holding "banking book" loans at par (current methodology), banks would be required to take a mark to market hit amortized over the expected life of the portfolio. And that could be bad news for banks that are thinly capitalized - these reserve requirements may make them insolvent.
Expect a massive industry (and political) backlash against this accounting methodology going into effect. In addition, if IASB adopts this proposal, it may impact the convergence of the US GAAP and the IAS standards, which has been the ultimate industry goal in recent years.