When WeWork’s founder and CEO Adam Neumann stepped down last year after a failed IPO, he received a lot of press over his $1.7 billion exit package and his dual-class shares. Wells Fargo & Co. split the chairman and CEO roles after its fraudulent account opening scandal.

Governance always comes into sharp focus when there’s a systemwide breakdown. But a review by Stanford University law professor David Larcker and researcher Brian Tayan suggests that many of the assumptions people make about what constitutes good governance are rarely evidence based.

For example, the research is mixed on whether a dual-class share structure harms a company’s performance. The argument in favor of dual-class shares is that giving founders and executives more voting power insulates them from the short-term thinking of the public markets. A 2018 study, for example, found that performance of companies with dual and single-stock classes was fairly similar over time. Other studies have come up with mixed results.

During the 2019 proxy season, investors’ most popular governance-related demand was for an independent chairman to lead the board, not the CEO. A 2013 survey of 48 companies found independent chairmen had no impact on corporate performance, risk or executive pay practices. But there’s some modest evidence that having a lead independent director does improve performance, the Stanford University researchers found.

“In our view, while corporate governance is really important, there’s a lot of loose thinking and bad empirical support for a lot of these ideas,” Larcker said in an article about the paper. “We’re just trying to get the story straight. Here are these ideas — let’s take a look at them and see if they make sense. Are they supported by research, or just conjecture?”

The researchers conclude that there’s a lot of “loosey-goosey governance,” a tendency to generalize and advocate for one-size-fits-all solutions, as well as a lack of clear definitions. “You see remedies being proposed that may not actually be beneficial,” Tayan said.

The research is even mixed on the subject of gender and racial board diversity and its impact on performance. Multiple studies have been done over the years, with some finding that board diversity improves performance and some not. One study of Norwegian companies after an imposed gender quota for boards in 2003 found that quotas hurt governance quality and firm value. And on that subject, it appears that support for board diversity is on the decline inside boardrooms. A recent survey by PwC found only 38% of directors said in 2019 that gender diversity on a board is very important, down from 46% in 2018. Directors seem to be fatigued of the issue, with 63% saying investors devote too much attention to gender diversity, up from 35% who thought so the year prior.