Narrative reporting: how alike are like-for-likes?

Brian A Rutherford considers the development of narrative reporting in the operating and financial review

Demand from analysts for so-called narrative accounting information has grown rapidly in recent years and the largest corporations have been reasonably willing to respond to that demand: indeed, some time between 1996 and 2000, the balance tipped in favour of words and now more than half the space in company annual reports is devoted to prose (Spice Up the Story: A Survey of Narrative Reporting in Annual Reports, Arthur Andersen, 2000). Given the purposes for which accounting information is used, however, it is hardly surprising that much of this �narrative� revolves around quantified disclosures. Most of the innovative narrative information currently being published comes in the Operating and Financial Review (OFR) and one of the most popular forms of quantified disclosure is �like-for-like� comparison of sales trends through time.

Although there is a UK ASB pronouncement on the content of the OFR, it remains only advisory and there are no financial reporting standards in the area. Both analysts and preparers are becoming concerned about the lack of comparability of like-for-likes, although, as usual, preparers view the prospect of an accounting standard with a degree of ambivalence - Safeway was recently quoted in the press as being concerned about the lack of regulation of like-for-likes but added that it was not calling for a new standard (Accountancy Age, 21 November 2002). The ASB itself apparently thinks that like-for-likes are straightforward - Allan Cook, its technical director, was quoted as saying so in the same article. How badly a standard is needed depends in part on the variety of practice currently possible, and actually employed, and how well users can understand what individual preparers have done. This article looks at practice among retailers in the FTSE 100 in providing information on sales trends on a like-for-like basis. Retailing is a relatively straightforward activity, so that, if this sector is difficult to understand, it is likely that the position in other sectors will be worse.

Current disclosures

There are nine retailers in the FTSE 100 (see Panel 1) and they all provide information on like-for-like sales in their OFR, or equivalent. Eight use the phrase �like-for-like�, all without definition, rather implying that it is a term of art, the meaning of which will be familiar to readers. Ironically, using a pseudo-technical term in this way obviates the need to define it and may thus actually convey less information than a narrative which avoids giving the statistic a name. That was certainly the case here because the company not using the term, Tesco, simply explained that, of its total growth in sales, �6.2% came from existing stores and 2.9% from net new stores�. Since these two numbers account for the overall increase in sales, the sentence comes close to defining the basis for the like-for-like comparison. Another company, Morrison, also provided a breakdown of overall sales growth between the like-for-like figure and other sources and, here again, we can work back to a partial definition of the like-for-like calculation. In this case like-for-like sales exclude �new space� (not necessarily the same as new stores) and also adjust for a difference in the number of weeks between the reporting and comparative periods. No company explicitly defined like-for-like sales (one, Boots, gave a glossary in its annual report but it did not include this term). No other company came close to providing even the level of definition available in the narratives of Tesco and Morrison. Dixons did, however, attribute part of the change in like-for-like sales to temporary closures for refits, indicating that temporary closures did not result in exclusion from the like-for-like base. This was the only other instance of commentary helping to build up a definition of the calculation.

The problem of defining like-for-likes received more press attention than usual recently as a consequence of Boots� decision to change the basis of its calculation for the half year to 30 September 2002. This was heralded in a statement, posted on the company�s website, which went to some lengths to provide both the formulae for the previous and revised definitions and a worked example. However, the effect of the switch was to express the absolute change in like-for-like sales as a percentage of like-for-like sales in the base period rather than, as previously, of all sales in the base period. What counted as like-for-like sales remained unchanged - and defined only as sales in like-for-like stores.

The possibilities

Panel 2 lists some of the ways in which sales figures for retailers could be adjusted to promote comparability through time. Figures adjusted in any of the ways listed could plausibly be labelled �like-for-like�. Although some of the bases listed cannot be applied by all retailers (for example, sales volume is only appropriate where sales can be reduced to a common physical measure), most are available in all cases. Like-for-like data published by one company could be based on any combination of the bases listed - from a single adjustment to the whole list - and could include other factors or, indeed, only factors not included in the list. Obviously, another company�s like-for-likes could be based on a completely different set of adjustments.

Furthermore, any of the bases mentioned in Panel 2 could be applied using a variety of different approaches. Panel 3 illustrates this point using one of the bases mentioned in Panel 2, and the one that appears to be used most frequently in practice, namely excluding stores that have not contributed effectively to sales in both periods. But what counts as contributing in both periods? Obviously a store that wasn�t open in the comparative period should be excluded, but what about one that was open for part of the comparative period and the whole of the current period? What about stores that have been refurbished over an extensive part of one or other of the periods? And if, for example, a newly opened store is located in the same town as one that has closed, should both be excluded or is one a substitute for the other? The newly opened store is likely to be larger than its predecessor and, thus, not strictly comparable, but then will a store that is extended at some time in the two periods be excluded from the like-for-like calculation?

And, if stores which are opened or closed part way through a period are excluded from the calculation, as it can quite properly be argued they should be, suppose managements accelerate or delay the opening or closure to influence the application of the treatment?

Now in a �typical� year it may well be that none of the variations suggested in Panel 3 would make much difference to the final calculation. But this is often true of the differences in accounting policy that exercise the ASB. What matters is what happens in an atypical year - for example, one where a substantial refurbishment programme is underway. And the range of bases available in Panel 2 is so great that it seems likely that preparers� choices would make a significant difference even in a typical year.

What analysts need

The OFR combines narrative reporting with financial and other quantified disclosures outside the framework of the main financial statements. Where there are numbers, there is scope for different ways of doing the calculation, with the meaning of the result depending to some degree on the method used, especially if users are comparing one company with another. The small survey of practice in a relatively uniform and straightforward sector described earlier suggests that companies are using different methods and failing to disclose the methods used. The range of alternatives available suggests that practice across the corporate sector generally may exhibit a substantial degree of variability, damaging comparability even in �normal� years.

The nature of what is being measured would make it difficult to establish standardised ways of preparing like-for-likes - especially at this early stage in their development, when very little is known about analysts� preferences. Nonetheless, the current situation is unsatisfactory for users and potentially dangerous for preparers and for confidence in financial reporting generally. It would seem sensible for preparers to disclose an outline of the adjustments they have made in calculating like-for-likes, both to provide users with information about what has been done in individual cases and as a first step to exploring practice across the sector, to see how varied it is and whether particular methods need to be made mandatory - or outlawed. If preparers take this step voluntarily, that would be very useful. If they do not, even this preliminary step will not be easy to mandate in the short term, since publishing an OFR at all, and thus everything in it, is currently only a matter of authoritative guidance. The new, statutorily-backed OFR, to be introduced as part of the Government�s reform of company financial reporting (Cm 5553, Modernising Company Law, 2002), will make the regulation of like-for-likes a great deal easier - though how long it will be before the new OFR arrives remains to be seen.

Brian A Rutherford is professor of accounting at Canterbury Business School, University of Kent. The author�s study of the OFR, Half The Story: Progress and Prospects for the Operating and Financial Review, is published by ACCA as Research Report no. 80.