In preparation for his FASRI Roundtable discussion, Stephen Penman has helpfully shared his CEASA White Paper on Financial Statement Presention, which you can download here.

The paper discusses disaggregation in some depth, providing guidelines for how line items should be disaggregated on both the income statement and the balance sheet.  I need to read more carefully, but Stephen’s proposals seem more similar to what the FASB and IASB are proposing than they are different:  disaggregation on the balance sheet should separate items with different risk profiles, while disaggregation on the income statement should allow better prediction of future cash flows.

More novel, to my eyes, is Stephen’s advocacy of a ‘forward looking earnings’ number.  Stephen starts with the problem that comprehensive income includes many highly transitory items, which can be a problem in using earnings to predict the future.  From the paper:

We suggest the following solution. Report per-share (comprehensive) Earnings to Common Shareholders but also report a second per-share number that bears on the future. This is not a forecast, but rather current Earnings to Common Shareholders stripped of items reported in the disaggregated income statement that do not bear on the future. Discarded items would include one-time items and transitory gains and losses which, though recurring in the future, do not predict next year’s gain or loss. One way to think about it is as follows: What number would a (diligent) analyst forecast as forward earnings? What income statement line items are omitted from the analyst’s forecasting spreadsheet? Clearly that analyst is not forecasting the one-time items of the previous period that will not repeat, nor components of earnings that are not predictable. The second earnings-per-share number might be called Forward-Looking Earnings (to Common) with the understanding that it is to align with “forward earnings” that analysts forecast. In this way, financial reporting and analysts’ forecasts would align such that analysts’ forecast errors could be determined by comparing their forecast to next year’s realization of Forward-Looking Earnings (and not earnings-per-share that includes gains and losses, write-downs, etc. that they could not be expected to forecast). Not only would a summary number be produced with forward-looking information but a discipline would be imposed on the analyst community, so that investors who use analysts’ forecasts would know just what is being included in their “forward earnings” and in a “forward” P/E ratio.

With its focus on forecasting, our proposed income statement template is designed to identify items to be stripped out in determining Forward-Looking Earnings. The components retained could be presented in a second column in the income statement that totals to Forward-Looking Earnings, or as a supplementary schedule. We are aware that a division between forward-looking and other income invites opportunistic reporting. A requirement that management justify classifications in the accounting policy footnote mitigates, as does rigorous auditing. The FASB-IASB might engage the analyst community on the matter, in the hope that a strong consensus would emerge to which management would then be committed.

Well, maybe not that novel.  This sounds to me somewhat similar to the notion of ‘headline earnings’ proposed a good 15 years ago by the UK Society of Investment Professionals:

UKSIP Headline Earnings has been used in the calculation of the price/earnings ratios of the United Kingdom companies in the Financial Times since 1994 and has also been used by computer bureaus and other intermediaries in the United Kingdom. At the same time Headline Earnings has been adopted in South Africa not only for use by but also as a standard definition for the announcement of company results.

The aim, as set out in Paragraph (vii) of the Summary of SIP 1 is “… to define an earnings figure calculated on a standard basis, which can be used as an unambiguous reference point …”. It should be noted that SIP 1 in defining Headline Earnings in accordance with this aim delineates clearly between such a reference point and an earnings figure calculated as a measure of the company’s maintainable earning capacity, which in the view of SIP 1 is a figure which cannot be calculated on a standard basis, but will vary from company to company and in other ways.

I will be interested to hear Stephen’s thoughts on the similarity of these earnings notions, and on how this idea could be made workable — and also, whether this earnings number would need to be reported as part of GAAP, or could be something outside of accounting standards.