The federal government consistently “betrays” accepted best practice in modern government information technology procurement in several key areas, says a new study by Carleton University researchers.

Ottawa’s outdated, ineffective and costly approach to IT procurement is a historically overlooked but crucial driver of the government’s failing digital reforms, the authors argue in their study, “Breaking All the Rules: Information Technology Procurement in the Government of Canada.”

The Auditor General of Canada has warned since at least 2010, and most recently in a 2023 report, that the government remains precariously reliant on aging IT infrastructure, “undermining both internal operational efficiency and the reliability and quality of public-facing services,” says the Carleton researchers’ study.

At the same time, the government doesn’t have a credible or sufficiently resourced plan to adopt modern public digital infrastructures that are now well-developed in other jurisdictions – including cloud, digital identify and integrated, cross-government data and service platforms.

The government faced a gap, as of May 2022, in an estimated 7,000 digital roles, “with the policy profession and senior leaders in particular lacking the digital literacy required to effectively oversee policy and service design in a digital context,” the study says.

The study’s authors are: Amanda Clarke, associate professor of Public Policy and Administration; Sean Boots, former federal public servant/government IT policy expert and now public servant-in-residence at Carleton; and research assistants Chantal Brosseau and Anne Lajoie. Their study is currently under peer review for publication.

 “The federal government breaks almost all globally accepted best practice for modern public sector IT procurement, a reality which we argue helps explain why we have scandals like the ArriveCan debacle that’s still unfolding,” the study says.

“More importantly, we argue that unless we reform federal IT procurement so that it gets up to speed with widely accepted best practice in the field, any attempts to drive forward meaningful digital reform in the Government of Canada are bound to fail.”

The majority of IT contracts issued by the government have dollar values and contract terms that a strong body of evidence indicates will lead to project failures, the study notes.

The government’s supplier market consists of a small number of prominent IT vendors, where three vendors receiving more than $100 million in contracts annually, from 2017-2018 to 2021-2022, make up 23 percent of government IT contract spending. Those three vendors are IBM Canada, Bell Canada, and Microsoft Canada.

“The Government of Canada has a widely-acknowledged dearth of modern digital competency in house, and in certain departments, our analysis finds that private contractors outnumber in-house IT staff,” the study says.

 Based on the government’s own admission of its digital skillsets gaps, “we should expect that these in-house staff do not in many instances possess the skills required to sufficiently manage these contractors.”

 The government’s 2023 strategy, “Canada’s Digital Ambition,” acknowledged that “despite being one of the most connected countries in the world with over 94 percent of people have Internet at home, Canada has the lowest usage frequency of digital government services among a 2020 survey of 36 countries.”

 In another breach of global best practice, the study points out that federal government policies favour vendor-ownership of IP and data, and do not prioritize adoption of open source solutions, despite evidence showing open source generates more cost-effective, secure, publicly accountable and higher quality digital services.

Available contracting data indicate the government is failing to sufficiently capture ownership of data and intellectual property resulting from the IT products and services it buys, the study says. “This finding casts doubt on the ‘value for money’ at play in government IT contracts, and also suggests that the government is failing to sufficiently capture and steward potentially enlightening data about its operations and service users.”

Contract spending on IT consulting services increased more than 50 per cent in five years

The Carleton researchers’ study found that estimated government-wide contract spending on IT grew by 27 percent between 2017-2018 and 2021-2022, after correcting for inflation.

Spending on devices and equipment [hardware] has remained relatively stable. But spending on consulting services and software licensing have both increased by more than 50 percent during that period.

Federal departments spent $7.9 billion on IT consulting services, $4.3 billion on software licensing, $3.7 billion on devices and equipment, and $4.1 billion on “Other,” including telecommunications. These totals exclude IT spending from the Department of National Defence.

The majority of IT spending (53 percent) spending is allocated to contracts that break the global best practice of a $2 million-per-year threshold for likely project success. Among these “rule- breaking” contracts, the average contract value was $24 million, with a range of just over $2 million to $1.08 billion.

Contracts representing 57 percent of the dollar value on IT consulting services and software licensing also broke the global best practice of limiting contracts to three years. Among these “rule-breaking” contracts, the average contract length was 4.3 years, with a range of just over three years to 17.6 years in duration.

From 2017-2018 to 2021-2022, spending on IT consulting services has grown by 55 percent, after correcting for inflation, from a total of $1.17 billion to $1.82 billion per year (in constant 2019 dollars).

The total spent on IT consulting services over the five-year period was $7.72 billion (in constant 2019 dollars).

Throughout this period, there was a notable increase in departmental spending on IT consulting services from prominent management consulting companies. Over this period, one of these companies (Deloitte) became the government’s largest provider of IT consulting services, measured by dollar value, while two others (Accenture, and PricewaterhouseCoopers) are also among the 10 largest providers.

The data show the 10 largest IT vendors measured by contract dollar value represent 37 percent of the total estimated spending on IT contracts over 2017 to 2022. In total, the federal government spent $2.83 billion on IT consulting services contracts with these 10 firms over the period of analysis.

The study notes that best practice in modern IT procurement calls for a balance between in-house expertise and outside IT services provision.

For an initial point of comparison, the annual budget for the Canadian Digital Service, created to help build in-house IT expertise in the federal government, is $25.8 million annually. This compares with the $566 million spent on average annually on outsourced IT services from the 10 largest IT vendors, over the five years examined in the study.

For example, Public Services and Procurement Canada, Canada Border Services Agency, and Immigration, Refugees and Citizenship Canada all have possible ranges of in-house to contractor ratios that would mean they rely on more outsourced IT workers than in-house IT staff, the study says.

“Such a ratio would be unimaginable in other core public sector job categories, such as policy analysis, program evaluation, or communications, and is particularly striking given the central role that the IT function plays in delivering key public services in the digital age.”

Contracts still being signed for federal IT projects likely to fail

The study’s authors say their findings come at a pressing moment in federal public administration and speak to ongoing, high-stakes risks that jeopardize the delivery of federal public services.

The federal government is currently responsible for more than 8,700 applications which demand maintenance and renewal, often resting on aging IT infrastructures that will need to be addressed – in many cases, through procurement initiatives – in the coming years, according to the Treasury Board of Canada Secretariat.

High-value, long-term IT contracts that the evidence suggests are bound to fail continue to be signed, the study says – most recently in May 2022, with the federal government signing a $193-million contract with Deloitte to support Employment and Social Development Canada’s benefits delivery modernization.

“This project’s outcomes will affect the millions of Canadians receiving major federal social security benefits; its failure could leave individuals without essential supports, and could in turn significantly breach public confidence in the state.”

Agile pilot projects, innovation labs, and comparatively small investments in IT talent recruitment and training are “swimming upstream” against an institutionalized culture of IT procurement that betrays accepted best practice, the study says.

“Federal IT procurement is thus at odds with responsible public money stewardship, singles Canada out amongst its peers as a digital government laggard, and ultimately, threatens the quality of programs affecting the public’s welfare,” the study says.

The study makes several policy recommendations including:

• Implement formal spending controls, as adopted in the U.K., and/or a mandated commitment to modular contracting, as advocated in U.S. policy documents. This could help the federal government end its propensity to establish large, long-term contracts that invite project failure and promote vendor lock-in.
• Actively promote a more competitive market for IT service contracts.

Similar to the UK’s Digital Marketplace21, a key goal of this effort should be to engage small-scale, specialized digital vendors that don’t normally engage in government procurements and that are outside of the National Capital Region.

For firms participating in these new procurement approaches, barriers to entry should be dramatically lower than current government procurement requirements, which impose significant administrative burdens that favour incumbents and large, well-resourced firms, the study says.

• Improve in-house public service technology capacity in the federal government, an initiative that depends on a number of urgent policy changes.

These changes include establishing market-competitive pay scales for software developers and cybersecurity experts (in a separate classification distinct from IT support and system administration roles).

Federal departments should be allowed to classify and hire technology staff that report directly to program and business teams (outside of departmental CIO and IT divisions).

The government needs to create “dual-stream” career progression models for technology staff that enable compensation at the highest pay scales without management responsibilities (e.g.. “individual contributor” progression models, used in most modern technology companies), the study says.

The government should provide pathways to meet bilingualism requirements for technology staff in order to increase the available talent pool.

Ottawa also needs to make permanent exemptions for digital specialists in the “return to office” rules applied to the federal public service, to allow the government to recruit these experts from anywhere in Canada, and to align with the expectations of technologists that expect to be able to work in distributed teams from a location of their choosing.

• Eliminate provisions preventing procurement of open source software in the Policy on Title to Intellectual Property Arising Under Crown Procurement Contracts (2015).

This policy represents a clear recipe for ongoing lock-in to the vendors producing custom software for the government, reducing departments’ ability to share and reuse resulting software and likely leading to frequent cases where the government pays for the same or comparable software multiple times over, the study says.

The government also should begin a government-wide effort to exclusively procure and publish open source software. “In doing so, it would follow the steps that peer countries have taken to reduce vendor lock-in, increase software reuse, and improve stewardship of taxpayer dollars.”

Constant scrutiny required to drive changes

As the situation now exists, archaic corporate policies deny public servants access to widely accepted modern digital work tools, the study says. Policies promoting modern digital ways of working, such as the Digital Standards2, “are optional and widely ignored in the daily business of government operations.”

“Risk aversion, excessive oversight, reporting burdens and entrenched organizational silos render it incredibly difficult, and in some instances, impossible, for federal public servants to work across disciplines, to iterate and learn from service users, and to keep pace with now broadly accepted best practice in modern service design.”

The failings of government IT projects that result from a dearth of in-house competency and an over-reliance on outside contractors can reinforce the public sector technology talent gap, the study says. Outside technology talent becomes less inclined to pursue government technology jobs, given their reputation for lags and failures.

Ultimately, the scale of changes required for better federal government IT procurement outcomes – and better public service delivery, as a result – is large enough that it depends on political support and leadership, the study says.

Public servant-led reform initiatives over the past several years, and decades of Auditor General scrutiny, haven’t been able to disrupt the patterns of vendor dependency that are visible in the continued growth of poorly managed IT consulting contracts.

For long-established IT managers and leadership in the public service, there is an expectation that they can “wait out” transformation efforts, given the short turnover of senior leadership nominally leading these efforts.

Concludes the study: “The vendor-dependency status quo is sufficiently established that public, media and political scrutiny is likely the only way to see it change.”

See also: "Why federal government IT projects are doomed to failure."

R$