The average American worker would have to work 331 years to match last year’s pay of the typical chief executive in corporate America, according to pay comparisons released Tuesday by the AFL-CIO.

The union group said the average CEO’s pay last year was $11.7 million, or 331 times the average U.S. worker’s annual pay of $35,293. That’s down from a ratio of 354 in 2012, when average CEO pay was $12.3 million, according to the AFL-CIO.

While the CEO-worker pay gap is down from recent years, AFL-CIO President Richard Trumka called the extreme difference between top executives’ compensation and rank-and-file workers’ pay a “short-sighted” policy that hurts companies in the long run because low-paid employees can’t afford to buy their products.

“America’s CEOs … are cannibalizing their own consumer base. It’s wrong. It’s unfair, and it’s bad economics,” Trumka said in a news release.

Companies and some experts counter that such CEO pay levels are needed to attract and retain top talent. CEO pay soared much faster than corporate profits or most employees’ pay in recent decades, and became a potent political issue during the Great Recession. In 1980, the average CEO’s pay was 42 times the average worker’s salary.

But executive pay consultant Todd Sirras said top executives have been getting smaller pay raises lately, partly because investors are weighing in through so-called “say-on-pay” votes at annual shareholder meetings. The nonbinding votes are required under the Dodd-Frank financial reform law enacted in 2010.

“It’s had a real impact,” said Sirras, managing director at Sembler Brossy, a Los Angeles-based executive compensation consulting firm.

Another rule in the Dodd-Frank law that is still pending will require companies to disclose the ratio of their CEOs’ pay to their employees’ typical pay. The AFL-CIO hopes the rule, expected to require such disclosures starting in 2016, will pressure companies to raise employee pay and lower top executives’ pay.

However, Sirras said he expects the pay ratio rule to have little impact because — unlike the say-on-pay vote results — it will be difficult to meaningfully compare one company’s pay ratio with others.

“It will be interesting, but it won’t help anybody make decisions,” he said. “It will make headlines.”