MISUNDERSTANDINGS
As might be expected of a novel piece of research, some aspects of BB68 were misunderstood. One misunderstanding involved confusing our two principal results, which we described as "usefulness" (on which dimension accounting income received a positive rating) and "timeliness" (a negative rating). Smith Bamber, Christensen, and Gaver (2000) show that the subsequent literature essentially ignored the timeliness result, and instead conveyed the erroneous impression that earnings announcements release substantial new information . They attribute the error to researchers over-generalizing the results from the unrepresentative sample of small firms in Beaver (1968).Ball and Shivkumar (2008) attribute it also to researchers comparing event-day volume or price volatility with that of an average day, thereby depicting an apparent "spike," whereas there are only four such event days each year and their volume and volatility constitute a small fraction of volume or volatility aggregated over the whole year. We suspect that another reason for ignoring our negative rating on timeliness was the understandable desire of researchers to show that “accounting matters.” The belief that accounting income definitely matters was in fact our starting point, when we reasoned it must matter to have survived as an important institutional fact; but the evidence we reported indicates the reason it matters cannot involve timeliness.
Another misunderstanding concerned the implications of our results for the extant literature. We note above that although our research design did not directly test any of the alternative reporting schemes proposed in that literature, we were able to test the literature’s proposition that the existing reporting scheme produced meaningless numbers. Nevertheless, researchers of the “old school” generally dismissed price-based research because it did not test the alternative reporting schemes, ignoring our evidence that users did not act as if accounting income is totally devoid of meaning. Thus, in this review of price-based Chambers (1974, 50) argued:
It just does not tackle the kind of question which has concerned those whose exploratory or research processes are criticized in the opening paragraphs of the paper. For that reason it offers no mode of testing the propositions of those who have concerned themselves with the betterment of accounting theory or practice.
The counter-factual in the form of any proposed alternative to the status quo is by definition unobservable, not only to us, but also to the proponents of such schemes. But that does not mean price-based research cannot be used to test propositions mad by those proponents. We believed we reported evidence contradicting the proposition, asserted regularly in the literature since canning (1929), that accounting income is devoid of meaning because it is not founded on a homogeneous set of principles, rules, and methods. Furthermore, there were wider implications, since this novel result brought into question the whole way fo thinking on which that proposition and the proposed alternative accounting schemes were based.
A related misconception is that there is little or no overlap between “positive” empirical research and “normative” a priori research. A distinction between them, which dates at least as far back as Hume, was introduced to the economics literature in 1891 by John Neville Keynes, father of john Maynard Keynes, as follows:
A positive science may be defined as a body of systematized knowledge concerning what is; a normative or regulative science as a body of systematized knowledge relating to criteria of what ought to be, and concerned therefore with the ideal as distinguished form the actual. (emphasis in the original)
Keynes’ distinction was rekindled by friedman (1953), and it dovetailed well with the emphasis at Chicago on testing theories against the data, which is what we set out to do.
While our research was conducted in the spirit of positive economics, as noted above we viewed it as having distinct normative implications. Recall that the prior literature contained propositions about “what is” (such as the thesis that earnings are meaningless) as well as “what ought to be.” In rejecting the “what is” proposition, we believed we brought into question the “what ought to be” propositions emanating from the same way of thinking. Our departure from the mode of research in the prevailing literature was more to do with empirical testing versus a priori theorizing, and not that we adopted a positive versus a normative perspective.
A related concern sometimes is raised, that “positive” accounting research in the decades after our paper was published has done little to assist standard-setters in developing financial reporting norms. This concern is grounded to some degree in the misconception that there is little or no overlap between “positive” empirical research and “normative” a priori research. BB68 was the genesis of a now deep and deepening body of research on how users utilize accounting information, including investors, analysts, lenders, rating agencies, litigants, and competitors, and its use in compensation, supply, and other contractual contexts. There was no systematic evidence on any of these uses prior to 1968. In elemental economic terms, it is difficult to see how an industry can develop optimal supply properties without knowledge of demand. How can one design accounting standards without a systematic body of knowledge on how financial statements actually are used? In the sense, the modern “positivist” literature allows arm-chair theorizing to be replaced with theorizing based on systematic observation.