Proxy voting records of Canada’s investment managers should serve as a wake-up call for pension funds and other investors who delegate proxy voting authority, according to the Shareholder Association for Research & Education (SHARE).

SHARE has released its third annual Key Proxy Vote Survey, which examines the proxy voting practices of Canadian pension investment management firms and proxy voting services.

“Despite heightened sensitivity to corporate governance issues, some investment managers still vote in favour of poor governance practices,” said Gil Yaron, SHARE’s director of law & policy, in a news release.

SHARE’s survey polled 91 investment management firms and four proxy voting services on how they voted proxies under their discretion in relation to 27 proposals from the 2003 proxy season. The proposals related to corporate governance issues such as board independence, dual-class share structures, excessive executive compensation, and corporate disclosure. SHARE says the survey results show wide discrepancies on how investment managers vote on specific issues.

Slightly less than half of the firms contacted for the survey agreed to participate. Of the 95 firms surveyed, 10 declined to participate and 47 did not respond to the survey.

“In light of a more engaged investor community, and of recent U.S. policy reforms on the disclosure of proxy voting records, it’s surprising that many investment managers still refuse to disclose this information,” said Yaron.”

“Investment managers should disclose their discretionary proxy voting practices, especially when they are voting on contentious corporate governance issues that could potentially impact investment performance,” said David McCaslin, senior VP of Greystone Managed Investments. “The survey distills the thousands of proxy votes down to a few critical issues and gives a picture of whether managers are voting in their clients’ best interests.”

The Survey report includes lists of firms that participated and did not participate, as well as individual firm scorecards.