The author of this contribution to the discussion group on this site bears the sole responsibility for both the substance and the style of the contents. The purpose of the discussion group is to elicit comments and to promote debate on specific topics. As such, the views expressed on any of the issues raised are not to be attributed to the IMF.
Comments on IMF Working Paper WP/01/209: ?The Treatment of Nonperforming Loans in Macroeconomic Statistics? by Adriaan M. Bloem and Cornelis N. Gorter.
Several issues for discussion are flagged at the end of the paper. I have therefore framed my comments along these lines.
1. What should be the definition of a bad or nonperforming loan? Should there be different definitions for different circumstances? Are deposits and trade credit to be treated in the same way as loans if they become impaired?
For loans the definition should follow the practice advocated by banking supervisors in the context of the Basle principles. As part III points out there is not yet an agreed international standard on this, so one needs to be developed in that context. Once agreed, it should be encouraged/required from all countries. The IMF proposals seem to be a good start.
Deposits are different, they are placed with financial institutions, and are largely protected by deposit protection schemes.
2. Which of the following options is most appropriate in the context of national accounts and related macroeconomic statistics concerning the valuation of loans: (i) continue the present 1993 SNA conventions. (ii) change the loan valuation rules to reflect provisions for nonperformance, (iii) continue the current approach but show provisions for nonperforming loans as memoranda items.
Option (ii) is the conceptually correct one to follow in the long run. However this should only be in the context of wider changes towards market/fair values for all financial assets and liabilities on the balance sheet ? perhaps part of the next major update to SNA and other manuals. The IMF discussion paper on accruals accounting for interest flows sets the scene for this, and allowance for non-performing loans can and should be incorporated into this framework.
In the short term, option (iii) would provide valuable additional information, probably best in the context of banking statistics, rather than the national accounts.
3. Should the international statistics manuals contain more criteria on when exactly a loan is non-recoverable and should ve completely written off?
I think commercial accounting conventions have to be followed here. However, government accounts do not always follow such conventions, and some firmer guidance may be needed in the Government Finance Statistics manual.
4. Should the international statistics manuals be changed to allow price fluctuations in loans that are expressed in domestic currency? How should this be implemented in practice? are there better ways to improve the informative value of thenational accounts in tracking impaired loans?
Yes, but within a wider market/fair value context (see Q2 above)
5. Should the national accounts cease to record interest accrual on impaired loans? Are financial institutions to be treated different in this respect from other holders of credit instruments?
No. But total interest recorded for a portfolio of loans would be reduced under my preferred option below. Interest flows will continue to require balancing between and across all sectors of the economy.
6. Should the international statistics manuals define an income concept that includes expected or actual losses on financial claims? Should a difference be made for that purpose between 'normal' losses and 'catastrophic' losses?
Yes and Yes. Normal, expected losses are factored into decisions on interest rates charged by financial institutions. The expected amounts that will not be collected due to nonperforming loans should be subtracted in order that pricing and FISIM is not over-estimated. Provisions made by financial institutions for such regular losses (i.e. not those specific to a particular loan) would be one means of estimating this amount. The analogy would be with other similar flows in the current SNA. For example, accrued taxation (which should not be larger than amounts expected to be collected) and regular expected losses from retail turnover.
Catastrophic losses are different and should continue to be treated as write offs and other changes in the balance sheet.
Philip Turnbull, Statistics Advisor
Pacific Financial Technical Assistance Centre
NOTE: These comments are not written from the point of view of PFTAC, or my current statistical post.