Many companies have been noisily publicizing their cushy parental leave policies in recent years, telling the world they are bestowing workers with an increasingly generous length of time to take off and bond with their newborns.

Deloitte now gives new moms and dads 16 weeks of paid leave. Etsy hands out a full six months. Netflix said in 2015 that its workers could take off a child's first year.

But while the headlines may be making a splash, the more generous policies aren't making much of a dent in the overall numbers.

The average number of weeks employers are giving workers actually fell slightly over the past decade, even as a greater percentage of companies offered 12 weeks or more in annual leave, according to new data from a nationally representative sample of more than 900 companies.

"Even though those have been the focus of media stories, they don't show the whole picture," said Ellen Galinsky, senior research adviser for the Society for Human Resource Management, which publishes the survey.

In 2005, Galinsky said, employers with at least 50 workers allowed an average maximum of 15.2 weeks in maternity leave, compared with 14.5 weeks in 2016. That average is slightly higher than the length of time found in 2014 and 2012 but lower than the one in 2008 and 2005. The data were first reported by Bloomberg News.

The report also shows that more companies are paying women who go on parental leave — 58 percent of companies now offer some pay during leave. The number has been roughly flat since 2012 but is up from 46 percent in 2005.

However, the percentage of employers offering workers 100 percent of their regular pay during leave has actually dropped, according to Galinsky's data, declining from 17 percent in 2005 to just 10 percent last year.

Galinsky and senior researcher James T. Bond noted that the drop in weeks is probably best explained by some differences in the sample. The study is not longitudinal — in other words, the same 900 companies weren't examined over the 11-year period — but is a nationally representative sample of the U.S. economy each year the study is done, weighted by employer size.

Still, Galinsky thinks what is clear is that the ample perks giving companies in certain industries good P.R. haven't yet translated into better benefits for everyone.

Jena McGregor writes for the Washington Post.