Investors, analysts, and other stakeholders rely on a slew of materials when measuring how businesses perform. Some of those materials are issued directly by company executives, and they typically include a particularly important set of documents: management earnings forecasts. Unsurprisingly, management forecasts do not always perfectly predict actual earnings. However, according to HKUST’s Yue Zheng and colleagues, incorrect forecasts can often be useful in assessing a company’s future earnings prospects.

“Traditionally,” the authors say, “forecasting errors are viewed as backward-looking measures, but not as sources of forward-looking information.” They point out stakeholders typically view managers’ forecasting errors (MFEs) purely as blemishes on management’s credibility. Yet such errors may be either intentional or unintentional, depending on managers’ incentives and forecasting ability. “Therefore,” the authors note, “whether MFEs contain predictive information is ultimately an empirical question.”

To answer this question, the researchers conducted a rigorous series of analyses of 14,449 annual management forecasts from 1995 to 2016. Their results showed that inaccurate forecasts can actually be “informative about earnings beyond the forecast period.” In other words, when a forecast contains earnings projections that are not realized in the current reporting period, there is a chance that those earnings will eventually be realized in subsequent periods. Due to this forward-leaning nature, forecasting errors can contain a significant amount of valuable information about a company’s future earnings.

Some forecasting errors appear to be more useful than others. “Distinctions should be made among [companies that] make forecasting errors,” the authors advise. Take, for instance, companies that issue forecasts relatively early in the reporting year. If their forecasts overstate the company’s earnings, the study shows that those earnings are likely to materialize in subsequent periods.

Another distinction identified in the study is whether forecasting errors are overly pessimistic or overly optimistic. By isolating both types of errors and analyzing them separately, the researchers found that overly optimistic forecasts stood a better chance of being borne out over time. That is to say, earnings not realized in the current forecast may come to fruition in future periods, particularly if the errors reflect incorrect expectations about the timing of a particular economic event. (As an example of this type of timing error, the authors note a case in which a company does not reap the benefits of an advertising campaign as quickly as expected.) Overly pessimistic forecasts, meanwhile, do not share the same fate. The predictive nature of forecasting errors “is concentrated in optimistic … errors,” the researchers find. In contrast, “pessimistic forecasts are often the result of a form of expectation management,” designed to cast the company in a positive light.

The authors highlight several other distinctions to consider, thereby providing a blueprint for identifying the forecasting errors that are most likely to yield valuable¾that is, predictive¾information about a company’s future earnings. They uncover convincing evidence that forecasting errors can be interpreted not as indicators of mismanagement but as useful forward-looking tools. The practical implications of their research are significant, especially for stakeholders who want to refine their own expectations of future company earnings. By incorporating these novel findings, they are likely to make more accurate projections.