I'm strongly in favour of OpenID login over proprietary login as a way of controlling site access - it gives users choice and makes their lives a bit easier. In principle, this should, in turn, increase the number of logins - the stumbling block of login is being made more straightforward. Is there a valid reference in terms of traffic impact on the use of OpenID compared to proprietary login?
I know you are asking in particular about the OpenID implementation but maybe the following article published by MailChimp could help you: Social Login Buttons Aren't Worth It
They say the implementation of social login buttons didn't have the impact they were expecting and that the whole thing is actually not worth it, in their case because it is adding complexity to their login screen (because of additional buttons and options) and because there are other ways to decrease the number of login failures. (The impact of the social login implementation meant login failures dropped 3.48%).
I know a social network that got 20% decrease in login failures just by adding login with Facebook.
In my opinion the impact depends a great deal on the kind of service you are providing with your website. And specially you should think if your users would be willing to login with their private accounts (might not be the case, specially if your website is meant to be used on the working environment).