Human Capital & Careers

An 18th Century Solution to Human Capital Reporting Standards

“Bayes’ Theorem” on probability offers a bridge between two schools of thought about whether proposed standards are a good idea.
John BoudreauAugust 27, 2012

Ever hear of Thomas Bayes? He was a “nonconformist” minister in the 1700s who developed a theory of probability that became a pillar of modern statistics. “Bayes’ Theorem” held, in part, that the value of information is greater when it is more likely to change a decision, when being wrong has large consequences, and when the decision maker is capable of acting on it and willing to do so. Failing any one of those conditions, information has little value.

These musings of an 18th century “nonconformist” also offer insights to the present-day debate over financial-reporting standards for human capital.

In Draft American National Standard of Investor Metrics for HR, the Society for Human Resource Management (SHRM) states that companies should report or “indicate their reasons for not doing so” in six areas: (1) spending on human capital, (2) ability to retain talent, (3) leadership depth, (4) leadership quality, (5) employee engagement, and (6) human-capital discussion and analysis. Yet members of the Human Resources Policy Assn. (HRPA) produced a letter and accompanying web page opposing the standards. The letter said the proposed standards add to an already-excessive reporting burden, few investors request such information or find it material to decisions, the standards cannot be compared across companies, and their costs outweigh their benefits.

Both the SHRM and HRPA include smart, well-meaning, and dedicated professionals. Can their conflict be resolved?

Back to Bayes’ Theorem. Suppose you own an ice cream stand, and you must decide each day whether to open it or not. If it rains you should be closed, and if it doesn’t you should open. You also might want to decide whether to pay for a location-specific weather forecast, and if so, how much you should pay for it. There is a way to calculate that value, but the decision about whether to open boils down to this: if you have no weather forecast, the best decision is to open, so you will open every day and take your lumps when it rains. A weather forecast is valuable only when it correctly forecasts rain, and if you always close the stand when it predicts rain. (By the way, if you think it’s far-fetched to pay for a weather forecast, consider your next big family reunion or wedding. There are fee-based services, such as, that sell pinpoint location weather forecasts for such events.)

It turns out that the accounting profession takes a rather Bayesian perspective in its general guidance about reporting standards. The FASB Statement of Accounting Concepts (No. 8, September 2010, page 1, Section OB2) says, “The objective of general purpose financial reporting is to provide financial information about the reporting entity that is useful to existing and potential investors, lenders, and other creditors in making decisions about providing resources to the entity.”

Applying the Bayes and Financial Accounting Standards Board logic to human-capital reporting standards, we would ask questions like: What will investors do without the new data? How often would the data suggest changing their decisions? How often would the data be correct? What is the value of changing the decision? Are investors likely to change their decisions based on the new information? While debate about the standards has been robust, it has not focused much on such questions.

A High Standard
Are there human-capital measures that offer insights about company performance beyond existing financial measures, and where would we find them? In his book Intangibles: Management, Measurement and Reporting (Brookings Institution Press, 2001), Baruch Lev argues that predictions of organizational performance are less accurate when they’re based only on the tangibles included in traditional financial statements, while factoring in intangibles like human and social capital offers enhanced predictive power. There is also a largely untapped repository of published scholarly research on elements of human capital, and their relationship to organizational performance, that might reveal human-capital measures that predict organizational performance beyond existing financial measures. A test of whether HR measurement standards will predict future performance is their consistency with this research.

Yet even if human-capital reporting standards satisfy the question posed at the beginning of the previous paragraph, Bayes’ Theorem raises a second, perhaps even tougher demand: decision makers must act on the HR information. Just as in the ice cream-stand example, the forecast is worthless unless you close when the forecast calls for rain. The most valid human-capital reporting fails if it falls on deaf ears. A key element of any discussion of the proposed standards should be how to educate and motivate investors and decision makers to use them. That question is a focus of the “evidence-based management” movement, which considers how to improve the connection between management research and management practice including why decision makers typically ignore such research.

Thomas Bayes never anticipated human-capital reporting standards, but much of today’s debate turns on his ideas. Bayes’s principles have improved measurement for centuries, and the insights of this 18th-century minister should be more prominent in the current discourse about HR reporting standards.

John Boudreau is a professor at the University of Southern California’s Marshall School of Business and research director at the school’s Center for Effective Organizations.