Evidence of meeting #76 for Public Accounts in the 42nd Parliament, 1st session. (The original version is on Parliament’s site, as are the minutes.) The winning word was fraud.
A video is available from Parliament.
Liberal
Raj Saini Liberal Kitchener Centre, ON
Thank you.
Ms. Nycholat, you said something very interesting. It brought back my years in science when you talked about the coefficient of variance. You said something interesting. You used the numbers 0.04 and 0.06. I don't remember my statistics from 20 years ago, but you're talking about a p-value of 4% to 6%. Is that roughly right?
Vice-President, Procurement, Defence Construction Canada
Right, but I should qualify that I am not a statistician.
Liberal
Raj Saini Liberal Kitchener Centre, ON
No, no, that's okay. I'm not either, but I remember that. I'm getting excited because what I learned 20 years ago is coming back to me.
Liberal
Raj Saini Liberal Kitchener Centre, ON
You talk about the coefficient of variance, and I just want to get an idea, having also been in business. When five companies bid on a project, for example, and you're saying the coefficient of variance in the bid amount is between 0.04 and 0.06, you're dealing with companies that are probably mature and also are big companies that do this consistently. You're not going to invite new companies or disparate companies from other economic areas. You're dealing with a very small number of companies. Would you not expect a mature company that has been in business for such a long time to be somewhat in the same ballpark? I'm just wondering, because you mentioned that that coefficient would in some ways contribute to some amount of collusion. However, for me, having been in business for a certain amount of time, I think that after a certain period of time you would have an understanding of where the numbers should be, especially with the companies that you're dealing with that are large companies that do this on a continuous basis, and would be expected by their experience to have a ballpark that would be within that variance.
Vice-President, Procurement, Defence Construction Canada
Thank you very much.
It's a very good question and I totally agree with you. That's why we can't rely completely on artificial intelligence to draw conclusions. We have to have people who are going to review the data and be able to interpret it, people who know the industry. The people with whom I staff our procurement team, I select primarily from the construction industry. They are engineers, architects, or technologists who have worked in construction, because we need that knowledge to be able to interpret the data.
That's why we can't rely on just one statistical analysis. We have to do several of them so we can establish a trend. We cannot draw a conclusion from the coefficient of variance on its own. We have to do another one, and then another one, and another one, and we need to do further research into the file. It takes a lot of effort really to get down to being able to draw from it that, yes, there's more likely to be collusion than not.
Conservative
The Chair Conservative Kevin Sorenson
Thank you.
Thank you to all members, and to our Auditor General specifically, for appearing here on a good audit.
I like what Mr. Christopherson said. Sometimes we hear not so good reports. I want to commend you. It's obvious that you know your file and your numbers. Sometimes when we have crown corporations, I sit here wondering whether we really need the crown corporation, and why can't the private sector deliver this? I understand a bit better today what you do and how well you do it.
Thank you for what you do, and thank you for coming to our committee.
We are now adjourned.