The seven million dollar typo

by David Broadland, November 19, 2010

The numbers the City presented for the cost of a new bridge and the cost to rehabilitate the current bridge were based on estimates done by Advicas Group. Those estimates were peer-reviewed by Stantec's Andrew Rushforth, but analysis suggests the numbers have been tweaked so the City’s high-end rehabilitation appears to be more expensive. We un-tweak the numbers.

City Hall says it will cost $77 million for a new bridge without rail on it and $80 million for what is now known as the “gold-plated” rehab.

I’ve been asking myself 3 question about these two numbers:

• Where did they come from?

• Can they be trusted?

• Do they represent the only options the City could have considered when deciding what to put to a referendum.

I think these are important questions and I’m going to try to provide answers.

The “peer-reviewed” Advicas cost estimates

First of all, the City’s numbers derive from the Class ‘C’ estimates done by Advicas Group Consultants in June of this year. 

The Advicas estimates were peer reviewed by Andrew Rushforth of Stantec Consulting Ltd. When something has been peer-reviewed, we tend to believe that it can be trusted. Right?

In the preface to Rushforth’s review he stated “the key finding will be to identify aspects that may not have been explored fully and to ensure that adequate funding to account for these is included in any estimate.”

But the road to hell is paved with good intentions, and when it came to the details of the cost estimates, I believe Rushforth, or his subcontractor, may have taken a shortcut.

In his study, Rushforth admitted “This review utilized data supplied by the project team and assumed that these quantities are a reasonable representation of the work that will be required.”

But Rushforth’s assumption seems to me to be ill-advised. The Advicas cost estimates for both the replacement bridge and the rehabilitation contain obvious errors of omission, calculation and substance.

Let’s start with the figures the Advicas estimates produced and then update them to what is currently being considered. I’ll then examine these estimates—perhaps more closely than Rushforth’s $50,000 study did—and keep a scorecard of the changes that ought to be made.

By the way, I should point out that MMM Group, the engineering company managing the bridge project, repackaged the Advicas numbers and added an amount for taxes and financing that Advicas did not include. So the numbers I start with may seem a few million dollars lower than what the City has been using. MMM provided no breakdown of the additional amounts they added, so for the purpose of comparing apples to apples, I am starting with the peer-reviewed Advicas estimates. I will put financing and taxes back at the end of this exercise.

For the replacement option, Advicas estimated a cost of $84,989,000.

For the rehabilitation option, which included a new pedestrian/cyclist bascule bridge, Advicas estimated $97,666,000.

When council took rail off the new bridge, that lowered its cost by $12 million. Using the Advicas figures, the new bridge then becomes $73 million. With rail no longer on the existing bridge, the Advicas estimate for the combination of refurbishment work that would need to be done to match the replacement option becomes $74.5 million.

So at the outset of this exercise, the two options are separated by only $1.5 million.

 

The $7 million typo

Now here’s where Andrew Rushforth’s assumptions about the “reasonable representation” of costs comes into play. We’ll start with something that Rushforth ought to have caught. It amounts to a $7 million dollar typo. Here’s how it goes.

Advicas estimated $6,197,095 for “Roads and civil works” for the new bridge.

Do you remember the pedestrian/cyclist bridge that was originally included as part of the refurbishment estimate? Remarkably, Advicas also estimated $6,197,095 for “Roads and civil works” for this bridge.

Is it really possible two completely different bridges, one designed for automobiles, a rail line, bicycle lanes and pedestrian walkways would have the same civil works as a much smaller one for pedestrians and bicyclists?

The individual breakdowns of costs that Advicas estimated for roads and civil works for both the new bridge and the pedestrian/cyclist bridge are exactly the same.

Does it make any sense that two completely different bridges would entail  exactly the same work for roads and civil works? No.

It seems to me someone copied and pasted the wrong information into one of these estimates. Rushforth’s $50,000 peer review falied to catch this simple mistake. One is wrong, but which one? For the answer to that, we need to go back to April 2009. 

Delcan did a “Class C” estimate for a new bridge back in 2009 that set the price for a new bridge at $63 million and the City insisted for many months that was a firm and reliable price. So let’s compare it with Advicas. Conveniently, Delcan detailed their estimate for roads and civil works almost exactly the way Advicas did. But Delcan estimated those costs to be $11.5 million.

In light of that, I would suggest the error in this case is in the amount assigned to the replacement bridge and that figure should be at least as high as Delcan’s $11.5 million. Didn’t the mayor say everything has increased since then?

So when the engineering, contingency and escalation costs are added to the difference, the cost of the new bridge goes up by $7.4 million. That’s the seven million dollar typo.

So here’s how the costs for the two bridges stands now: New bridge $80.4 million, refurbished bridge $74.5 million.

 

The missing paint

Let’s now look at something Advicas seems to have lost.  

Advicas allowed $3 million for painting the refurbished bridge, which, when engineering, contingency and escalation costs are included, compounds to $5 million.

And for the new bridge—which its architect, Sebastian Ricard, recently confirmed would be painted—how much does Advicas allow?

Zero dollars.

Think about that: Five million for painting the existing bridge. Nothing for painting the new bridge. If we’re comparing apples to apples, then shouldn’t both estimates include paint?

Let’s assume—a dangerous thing to do, I admit—that painting the new bridge will cost half as much as painting the old bridge. That means we should add $2.5 million to the Advicas estimate for the new bridge.

So with paint accounted for, the numbers are now: New bridge $82.9 million. Refurbished bridge $74.5 million.

 

The ignored peer review

I now want to talk about something that was ignored.

Rushforth’s $50,000 peer review didn’t make many recommendations, but it did recommend the contingency allowance for the new bridge be raised from 15 percent to 20 percent. But MMM Group and the City ignored their consultant’s recommendation. 

Claiming that an estimate has been peer reviewed and then ignoring the reviewer’s recommendation seems to me to be hypocritical. And it also seems imprudent.

When Delcan produced their $63 million Class C estimate for a new bridge, they included a contingency of 30 percent. If  MMM and the City had listened to the peer reviewer, they would have increased the estimate for the new bridge, compounded by the escalation factor, by $3.3 million. So let’s do that.

So here are the new numbers: New bridge $86.2 million. Refurbished bridge $74.5 million.

 

The inflated engineering costs

What I want to look at next is engineering costs and whether they have been applied consistently in the estimates.

In Delcan’s Class C estimates for refurbishment and replacement, they based engineering costs on a percentage of predicted construction costs.

Advicas used that formula for engineering costs for the replacement bridge, but not for any of the work associated with refurbishment.

Instead, for the refurbished bridge, the engineering cost was calculated on the SUM of construction costs plus contingency. This had the affect of unnecessarily inflating the cost of refurbishment by $2.5 million. Since we’re trying to compare apples with apples, let’s adjust the estimate for refurbishment downward by that amount.

The score is now: New bridge $86.2 million. Refurbished bridge $71.9 million.

 

Financing and taxes

So let’s now put back the amounts we took off at the top for “financing and taxes” that MMM Group had inequitably added to the Advicas estimates. Let’s make the adjustment equitable, and add 4.5 percent to each.

So here’s the final score in the price game: New bridge $90 million. Gold-plated refurbished bridge $75 million.

 

What this all means

So, can the numbers the City has presented be trusted?

I believe those numbers have been fudged to show a replacement bridge would cost less than a rehabilitated bridge.

And what about the numbers I’ve shown with all the adjustments? Do those numbers reflect any objective truth about what a new bridge will cost or what a rehabilitated bridge should cost? I have my doubts.

I think the design of a new bridge is in its infancy, far too early for cost certainty. I believe the numbers presented by Advicas have more to do with political agendas and personal aspiration than engineering.

In the days following the successful counter-petition, pro-replacement councillors were faced with either forging ahead with replacement or backing down and losing face. And whether they knew it or not, the price they thought was firm wasn’t. But they chose to forge ahead anyway.

I want to show you an excerpt from a document from January 2010, recently obtained from the City through an FOI request. It’s part of a memo sent to the City’s bridge project manager, Mike Lai, from Joost Meyboom, MMM Group’s chief consultant on the project.

At the time, it was clear the counter-petition challenging the City’s decision to replace the bridge was going to be successful and, perhaps worried that a referendum debate would expose the flimsiness of previous cost estimates, Lai posed a series of questions to Meyboom. 

The replacement option was then estimated at $63 million, and one of the questions Lai asked Meyboom was, “Do we have a firmer budget estimate for the project? Are there scope items that we may need to scale back on or eliminate to maintain the budget?”

Meyboom responded, saying, in part, “The current estimate is based on limited engineering and a preliminary geotechnical investigation.” (He recently admitted than no additioal design work has been done on the bridge since the time he made this statement)

The excerpt shows the reductions in scope Mr Meyboom suggested to keep costs from rising.

Meyboom memo

Several of the scope reductions that Meyboom suggested were possible—and his rationalizations of those points—call into question the arguments the City has made for replacing the bridge. Had Meyboom’s suggestions been implemented, the City would have appeared to have lost control of the project. Refurbishment would have looked like a no-brainer.

So the pro-replacement councillors and staff needed a strategy that would inflate the cost of refurbishment, and councillor Madoff’’s apples to apples doctrine accomplished that nicely. She insisted a refurbished bridge must have the same seismic performance as a new bridge and the same amenities as a new bridge. Somewhere along the line, the idea that it should also have the same life expectancy came into play. Effectively, the pro-replacement councillors and staff wanted to compare two new bridges, one of which just happened to look like the old bridge.

Once new costs for replacement and refurbishment were obtained, they appear to have been pushed, prodded and tweaked to produce the desired result: that refurbishment would be more expensive than replacement.

So here we are, on the eve of the referendum, and I would hazard the City and their consultants have no firm idea of what a new bridge will cost.

And what about a common-sense refurbishment? One that has not been captured by the politcal agendas of the pro-replacement councillors or the personal aspirations of City Hall staff?

Delcan’s 2009 Class C estimate of $23.6 million took many months to develop and prices have changed little since then. There’s good reason to believe it would still hold—or at least could be made to hold if we had a city council with the will to keep costs down.

If the referendum produces a “No,” the next step must be to revisit the common sense, least-cost option developed by Mr Meyboom two years ago and expanded upon by those who followed him at Delcan.

David Broadland is the publisher of Focus Magazine.