The big flaw in the SEC’s CEO pay-ratio rule by S. Kumar @FortuneMagazine August 6, 2015, 2:38 PM EDT E-mail Tweet Facebook Linkedin Share icons Five years after it became law, the Securities and Exchange Commission has finally adopted a rule requiring public companies to disclose their CEO’s pay as a ratio of the median compensation of average employees. The rule is well intentioned. CEO pay in 2014 was an eye-popping 373 times that of an average worker, according to data compiled by the AFL-CIO, and a sharp rise from 331 times in 2013. This imbalance contributes to America’s growing wealth gap and accompanying social and political inequities. Requiring companies, especially large public corporations, to disclose how richly their CEOs are paid would provide valuable information for shareholders and possibly help the larger national debate about economic fairness. The only problem is that the SEC’s formula for the CEO pay ratio is flawed and unlikely to produce the results it wants. By using the median compensation of average employees in the denominator of the ratio, the SEC is inadvertently skewing the calculation. For example, Company A has four employees who are paid $15,000 a year, one employee who is paid $50,000 a year, and four employees who are paid $80,000 a year. The median (the midpoint) here is $50,000, even though half the number of average employees are paid far below that wage. That would reduce the CEO pay ratio and paint an inaccurate picture of what may really be going on at the company. In addition, the SEC rule provides some leeway to companies in calculating the median, including using statistical sampling (a slippery slope open to subjectivity) which could possibly enable them to game the system. Take the scenario where the median compensation of an average employee was too low, thereby producing a high ratio. In that case, a company could simply hire one more person on the higher end of the income distribution to increase the median, or fire someone on the lower end. In the above example, that means the median would rise to $65,000 (the average of the two middle numbers, $50,000 and $80,000), further skewing the CEO pay ratio. That defeats the central purpose of the SEC’s rule, which is to create greater transparency in how lopsided CEO pay is relative to that of the average worker. A better approach would be for the SEC to require companies to use the weighted average pay of average employees to compute the ratio. A weighted average would take into account not only how much various workers got paid but how many workers got paid at different levels. That, in turn, would weight the denominator for the ratio towards the largest segment of the worker pool, which is as it should be. Another benefit of the weighted average method is that it would produce a consistent result across companies (the weighted average concept is less sensitive to the addition or subtraction of a single employee as a median is) and therefore enable the public to compare the CEO pay ratio amongst different companies – knowing that they are comparing apples to apples. The SEC rule is a step in the right direction but needs to be fixed in order to be effective. S. Kumar is a tech and business commentator. He has worked in technology, media, and telecom investment banking.