2018 Spring Reports of the Auditor General of Canada to the Parliament of Canada Message from the Auditor General of Canada

2018 Spring Reports of the Auditor General of Canada to the Parliament of CanadaMessage from the Auditor General of Canada

Incomprehensible failures

An incomprehensible failure.

That’s how we described the Phoenix project in our audit report on building and implementing the Phoenix pay system.

That audit report lists the project’s missteps—so we know how the failure happened.

But the audit cannot explain why the failure happened.

Why did no one realize the project would fail? Why did no one stop and fundamentally reassess the project? The only explanation is that there were fundamental failures of project management and project oversight.

Why both failed is incomprehensible.

As I reflect on the incomprehensible failure that was Phoenix, I realize that two of our other audits point to another incomprehensible failure, although of a completely different type. I’m not talking about another project failure, but rather a systemic failure decades in the making. And this one has affected even more people than Phoenix.

It’s the incomprehensible failure of the federal government to influence better conditions for Indigenous people in Canada. Our recent audits are two more in a long line that bring to light the poor outcomes of Indigenous programs.

A pertinent fact about these two very different incomprehensible failures—Phoenix and Indigenous programs—is that both have been passed on from government to government.

Failure is not always bad. Failure happens and will continue to happen. It cannot and should not be eliminated. Failure is a way to learn and improve.

To use a sports analogy, every year one team succeeds and all the others fail, but they use that failure to try to get better. In sports, there is a term for the ultimate failure—that term is “own goal.” It’s used to describe what happens when a team scores a goal against itself.

Phoenix was a massive own goal.

It’s the own goals, the incomprehensible and avoidable failures that touch so many people and take so long and waste so much money to fix, that have to be stopped.

This leads me to echo a question posed by Donald Savoie in his book “Whatever Happened to the Music Teacher?” I have modified his question to make it about Phoenix.

How could Phoenix have failed so thoroughly in a system that has a management accountability framework; risk management policies; program evaluations; internal audit groups; departmental audit committees; accounting officers; departmental plans; departmental performance reports; pay-for-performance compensation; and audits by the Office of the Auditor General?

Our Phoenix audit didn’t answer that question because sometimes a traditional audit can’t get to the cause of a failure. That’s the purpose of this message—to explore the cause of incomprehensible failures in government.

But first, I want to say that I’m sure of one thing: Even if the government acts seriously on the recommendations in our Phoenix audit report, and implements all the lessons learned from the Goss Gilroy reportFootnote 1 on Phoenix, there will be another incomprehensible failure in the future.

Why am I so sure? Well, though I hope to be proven wrong, 40 years of our performance audits and the long history of incomprehensible failures of Indigenous programs tell me that the odds are stacked in favour of another incomprehensible failure in the future.

Can it be avoided?


But only with changes that go beyond the recommendations we made in our audit report.

The bottom line is that a change in the culture of the federal government will be the best hope to prevent incomprehensible failures in the future.

There has been a lot of discussion about lessons learned and blame, but little has been said about the culture that fosters incomprehensible failures.

Laying blame

Before I get into how government culture makes incomprehensible failures possible, I need to address the question of who to blame for the Phoenix failure.

I don’t think this is the most important question to ask, and I expect my answer will leave many unsatisfied.

And it’s a difficult question to answer because it’s as if the Phoenix project was set up to avoid responsibility—either by design or by accident.

There was no documented approval by anyone that the system should be launched.

Three different deputy ministers rotated through Public Services and Procurement Canada (PSPC) in the seven years before the launch of the system’s first wave in February 2016. The Deputy Minister who was in place when Phoenix was launched was only in the role for about a year, and he left shortly after the first wave was implemented.

PSPC tried to bring the whole Deputy Minister group into the launch decision at the last minute, but that group didn’t have any authority to make decisions.

The Treasury Board of Canada Secretariat realized late in the process that a problem was brewing, but it also didn’t have any project authority. So it was on the sidelines while the own goal was scored—reduced to having an external report prepared by Gartner,Footnote 2 which PSPC ignored.

So, as with many failures, there is plenty of blame to go around.

Our audit documents the project executives’ missteps, but that’s where the paper trail ends.

Ultimately, a Deputy Minister is accountable for what happens on his or her watch.

The Deputy Minister is the accounting officer of a department. In a 2007 report, the Standing Committee on Public Accounts said, “Accounting officers should ensure that financial administration in their department or agency meets the standards of compliance and prudence and probity.”

While the Deputy Minister was accountable for what happened on his watch, his staff didn’t give him the information he needed to understand the serious project risks, there was no independent project oversight, and he had been given responsibility for a complex department just as the Phoenix project was winding down.

These are not excuses—no Deputy Minister should avoid accountability because he or she didn’t know what was happening in his or her department. The former Deputy Minister should be held accountable for the decision to launch Phoenix, but he didn’t cause the problem.

The government that approved the Phoenix project in the first place should be accountable for the fact that the project didn’t have an appropriate oversight structure. Also, I don’t understand how anyone could believe that a project with a significant information technology (IT) component and a price tag of $300 million would be delivered on budget and would quickly produce savings of $70 million a year.

And then there’s the government in place when the decision to launch the system was made. It obviously didn’t ask enough questions to understand the serious risks attached to the project—the project was not in the mandate letter for the Minister of Public Services and Procurement, despite the controversy around the previous government’s decision to centralize compensation advisors in Miramichi.

I am not assigning political blame, but my view is that both governments had opportunities to prevent the incomprehensible failure that Phoenix became.

So my summary is that the project executives were to blame for the project failures, the Deputy Minister who was in place when the system was launched is accountable for the failure that happened on his watch, the former government is accountable for not having built an appropriate oversight mechanism into the project, and the current government is accountable to Parliament for the decision to launch the system. The current government is also responsible for fixing the problem, a fix that will have an incremental cost of more than $1 billion and will take years.

Others will disagree with my views of blame, accountability, and responsibility. Of course, I believe that accountability is important, but assigning blame—while it may satisfy some—won’t fix anything.

Most of the people who created the problem have already moved on. The names and faces have changed and will change, and governments have changed and will change, but incomprehensible failures will continue.

A standard lessons-learned exercise won’t prevent future incomprehensible failures. Phoenix is a defining moment—a wake-up call—that goes well beyond lessons learned. It needs to lead to a deeper understanding and correction of the pervasive cultural problems at play.

An obedient culture that puts itself at risk of failure

Organizational culture is often talked about, but it is difficult to define or measure, so it needs to be described.

What follows is my description of the culture in the federal government—a culture that has evolved with each passing decade.

I want to be clear that the current government did not create this culture—it inherited it—but it now has an opportunity to shape it for the better.

In a Westminster democracy, there needs to be a healthy tension between the political perspective and the public service perspective. If this tension is in balance, government programs will work for people.

Governments will always have the right to do something other than what the public service advises—a government is accountable to the voters. But if the tension between the two perspectives is not in balance, the risk of failure increases.

Any elected government, regardless of political party, values quick action intended to resolve societal problems. It believes that quick action will produce political benefit within an election cycle.

The public service needs to champion a longer-term perspective to make sure that government programs are sustainable—even if it won’t bring short-term political benefit. Both views are important.

Our audit on replacing Montréal’s Champlain Bridge shows that when the two perspectives are not balanced, decisions are delayed and only a crisis triggers action. A government is forced to react to a crisis, even if it didn’t create it. Delays in making decisions bring a large price tag.

Ministers and public servants have different views about program implementation. To a Minister, the announcement of a new program is action—implementation is detail. To a Deputy Minister, a program announcement is just the beginning. A Deputy Minister knows that successful implementation will take a lot of diligent work, perhaps over years.

The political perspective also means that ministers tend not to be excited by back-office administration projects like payroll projects—they cost money that could have been spent on more visible initiatives, and they don’t bring political benefit. Voters only hear about them when they go wrong.

Ministers expect back-office projects to be strictly managed, so project budgets and timelines are set in stone early—and they are usually impossible to meet. If an IT project is delivered on time and on budget, it will almost certainly not do what it was supposed to do—that was what happened with Phoenix.

Ministers also want to harvest savings from back-office projects too quickly, because they are the only visible outcome of the project. Often, however, transformative projects need transition funding to work as intended, but this is often overlooked. If a government reduces budgets too quickly, the result can be a failed transition and higher, rather than lower, costs in the long term.

The ministerial focus on the short term explains why the Indigenous file has been so intractable. A long-term view has to dominate that file, but because it usually only brings political problems in the short term, government tries to stay in the safe space of administering payments instead of being an active partner with Indigenous people to improve outcomes. The measure of success has become the amount of money spent, rather than improved outcomes for Indigenous people.

In the current culture, the two perspectives are out of balance, with the political perspective being dominant. This is largely because of instant digital communication, which means that politicians are more concerned with message and image management.

Ministers now expect public servants to implement programs without making a media-worthy mistake. No politician wants to be blamed for a mistake—real or perceived—that makes the news cycle.

Because of the dominant political perspective, there has been an erosion of Deputy Minister influence. Two signs of this are the short tenure of deputy ministers and the increased influence of ministerial staff.

In the last few years, there have been four different deputy ministers in Public Services and Procurement Canada and in National Defence. How could a Deputy Minister understand the issues in those two large departments in such a short tenure?

A large cadre of ministerial political staff give policy advice to the same ministers that deputy ministers are responsible for advising, so it’s harder for a Deputy Minister to be heard. This means that it’s easier for a Deputy Minister to just implement the will of the Minister without question rather than provide fearless advice on the pitfalls that could arise and how to avoid them. This is how deputy ministers keep the trust of their ministers, and keep whatever influence they have.

The result is an obedient public service that tries to eliminate risk and mistakes, which of course is not possible, so it has to try to avoid responsibility for those mistakes.

In this culture, for a public servant, it is often better to do nothing than to do something that doesn’t work out. If, however, action can’t be avoided, people search for plausible deniability—a way to deny responsibility for a mistake.

Policies are applied as cover to avoid blame. There is a reverence for checking boxes: If all the policies and procedures were followed—if all the boxes were checked—then the flaw must be in the system.

And of course, every public problem leads to more rules and boxes to check. Compliance with all government rules has become impossible because of their sheer volume.

I have provided a long description of the current culture. My short description is that the culture has created an obedient public service that fears mistakes and risk. Its ability to convey hard truths has eroded, as has the willingness of senior levels—including ministers—to hear hard truths. This culture causes the incomprehensible failures it is trying to avoid.

The culture has to change

My description of the culture in the federal government is bleak—but I hope it will start the process of change.

This culture is the answer to the two questions that I asked earlier: Why did Phoenix fail? And, in the government’s highly controlled environment, how was it possible for Phoenix to happen?

As I said earlier, the bottom line is that the culture has to change.

I don’t have a set of instructions to fix a broken government culture. But I know that the first step has to be to describe the current culture, which I have attempted to do, although I may not have captured everything. The second step is to admit that the culture problem is real and that it urgently needs to be fixed.

How to fix it will be up to the government and the public service.

The silver lining is that while there is a culture problem, the recent public service survey shows that the average public servant wants the culture to change, and wants to work in a culture that focuses on results for people.

The government has a choice. It can either perpetuate the current culture and its problems—including the incomprehensible failures—or it can change that culture and reap the benefits of programs that work for people.