LITERATURE AND THEORETICAL BACKGROUND
The Literature on Business Failures
When attempting to understand the characteristics of failures, researchers are forced to rely on incomplete data which has increased the misunderstanding of the distinction between business failures and other forms of business closure. Duncan & Handler suggest that a failure should be defined as "a firm that has gone out of business with losses to its creditors" (1994:7). However, many other closures might be incorrectly classified as failures. These include the transfer of firm ownership, the fulfillment of the owners intentions, retirement without succession, the pursuit of other opportunities, or even a change in the firms name to reflect broadened market offerings. Williams (1993) found that Dunn & Bradstreet had incorrectly classified 50.3% of the firms which were considered as having failed, with 57.6% of these firms still at the same address and the same phone number. The myth that four out of five businesses fail in the first five years, was challenged by Sexton & Upton (1986) more than a decade ago. The research by both Williams and by Duncan & Handler, cited above, suggest failure rates far below the 80% which is so generally accepted.
The problem with understanding the issue of closures versus failures is compounded because it is so difficult to obtain data on firms which have closed/failed. Thus, researchers have utilized various data sources in the area of business closures/failures. This includes the elimination of the firms from tax records (Hall, 1994), state employment (ES202) files (Birley, 1984), telephone directory white pages, membership in local Chambers of Commerce, ES202 files, and actual counting of businesses (Kalleberg, Marsden, Aldrich & Cassell, 1990), and the examination of how many firms remained in an existing sample after a period of time (Williams, 1993). Each of these approaches potentially restricts our understanding the characteristics which distinguish business failures from other types of business closure.
Even when failure rates can be accurately determined, unexplained variability may remain. Haswell & Holmes (1989) suggest industry, geographic location, and organizational form as possible sources of differing failure rates. Duncan & Handler (1994) show different failure rates associated with the size of the firm at startup.
Finally, in addition to great variability in the definition and measure of firm failures, we have even less insight into the reasons that firms discontinue operations. For example, Peterson, Kozmetsky & Ridgway (1983), used the perceptions of existing firms to explain why firms fail. Bruce (1996) suggests that small firms fail due to poor management, based on the perceptions of bankers and accountants; Laitinen (1992), developed a model to predict failure among newly founded firms based entirely on existing firms, with failures defined as cash insolvency. The problem with these approaches is that failed firms may be inherently different from existing firms. Thus, a prediction model based on existing firms may or may not be a reasonable indicator of the likelihood of firm failures. This, combined with the probability that firm closure cannot be equated with firm failure, makes it clear that our understanding of the phenomenon is incomplete.
1997 Babson College All Rights Reserved
Last Updated 03/19/98