Issue 21: Plans within plans

Plus, policy adventurism, Alberta's open-source AI academy, and a dog mayor

WELCOME

The public service outlasts almost everything around it. It has been here longer than any single government, longer than most of the parties that form them, and longer than the careers of the people elected to lead.

I’ve been thinking about this status lately, as an asset the public service holds but does not leverage. There may be value in public services claiming greater responsibility for long-term stewardship of government, creating a stronger internal intellectual and operational infrastructure that allows them to better plan for the advice and execution that future policymakers will require of them. I want to know: can a public service, taking a long view of the decisions, consequences, and second-order effects of the current direction of government, design itself in such a way to anticipate the tasks it will be called upon to deliver decades in the future? Call it a benign Bene Gesserit or, the “constant public service”.

A client recently described her work to me in generations. She runs services for residents in acute need and has started planning with a twenty-year view: not merely because the problems she manages start early in people’s lives and compound, but because these problems get more expensive the longer they are left unaddressed. Her government pays either way, so taking the twenty-year view is how she intends to bend the cost curve.

That framing reminded me of the famous ‘seven generations’ principle in the Haudenosaunee Great Law of Peace. The Great Law establishes an obligation to those who come after; a duty owed to people who cannot advocate for themselves in the present. While not explicitly a planning horizon, it is a perspective that is intended to contextualize the decision-making of policymakers today.

Right now, the best existing example of this long view I can think of is the Parliamentary Budget Officer’s fiscal sustainability report, which projects federal, provincial, and territorial finances over seventy-five years. It forces present choices into a frame where their consequences become visible, whether or not the government of the day finds the results convenient. Arguably, an institution whose function includes a seventy-five-year view produces a different quality of advice than that which has a four-year view.

Thinking long-term would be a major component of the constant public service, but perhaps just as critical is valuing and cultivating institutional memory. Fearless advice requires wisdom, and not merely loyal but effective implementation requires the ability to learn from experience.

We know there are currently forces working against the development and maintenance of that experience in Canadian public services. Organizational churn moves people out of files before their judgment about those files has matured, machinery-of-government changes redistribute mandates, accountabilities, and records across new functions, and lessons learned are inconsistently captured and acted upon. Plus, thin absorptive capacity means that even where information exists, who has the time to access it? Constancy is an illusion.

This is more than just a cultural or capacity challenge, of course. Investment in people is often the first thing to get cut when budgets need to be tightened, or the last thing to receive investment when elected officials are setting mandates

The longevity of public services, then, is a latent asset rather than a standing advantage.

But let’s game out this thought experiment. Practically, what would building a constant public service include? On the fiscal side, this probably looks like a continued push for multi-year budgets, capital planning on a horizon that extends beyond the life of assets, and disciplined asset management plans and reserve policies. Administratively, it could include practices like outcomes-based procurement, routine evaluations of policy and regulatory performance, mandated service reviews, and more sophisticated case management systems. On the people side, it would include succession planning that identifies who holds critical tacit knowledge before they retire, formal mentorship and coaching that transfers judgment not just process, and recruitment agile enough to sustain a pipeline of critical talent. This isn’t public policy per se, but it builds the infrastructure for what makes good policy possible later.

Not all of this can be accomplished by senior public servants alone, particularly with respect to fiscal and administrative policy. And it is only on my most cynical days that I would advocate for the wholesale substitution of bureaucrats’ judgment for politicians’. Instead, the goal of the constant public service should be to extend the horizon over which politicians can meaningfully make choices. For example, a minister who inherits an asset plan can decide to defer a bridge repair and then be held to account for that decision if the bridge fails. A minister who inherits no such plan can only be surprised by the failure.

Indulge me in my Sir Humphrey-esque fantasies: What else would it take to build a public service that holds a horizon longer than the government it serves – without mistaking that horizon for a mandate?

This month’s issue reflects on the persistence of how we undermine effective policymaking, from the concept of “policy adventurism” to governments that have no written guidance for what good policy looks like. But there are good news stories, too: an incredible (and incredibly generous) AI learning tool from Alberta, and an example of building policy capacity in young people from Singapore.

One last item: I have started a subscriber referral program. More details will land in your inbox later this month, but if you’re interested in earning rewards for referring friends and colleagues, make sure to subscribe today.

THIS MONTH’S SYLLABUS

⚙️Think tank products worksheet, MERICS Think Tank Lab

I was musing on LinkedIn the other week about the tragedy of the policy report, thoughtfully researched and impeccably argued by a think tank or advocacy organization only to immediately disappear on release. Happily, there is help: This simple table from the MERICS Think Tank Lab provides a list of dozens of possible products your organization could produce, depending upon the audience, time/effort required, your goals, and other criteria. Instead of writing a report, how about organizing a site visit, filming a short-form video, or running a morning briefing? A handy reference to print out and hang above your desk for inspiration.

💻“It’s time to name what’s ailing Canadian governance: policy adventurism” by Charles Lammam, Policy Economics

Do you feel the heat? The 🔥Policycraft Hot Take of the Month🔥 is back! Charles Lammam has named a moment he sees in Canadian public policy: “policy adventurism”. What he calls “the habit of solving for visibility instead of competence” is what some media commentators have cynically dubbed, “the announcement is the policy”. Charles is a thoughtful wonk and economist so he describes what he’s seeing with more substance: a government committing money, political capital, and scarce bureaucratic attention to something new while a proven fix for the same problem goes untried because it is politically harder. He builds his case with examples from the Vancouver condo purchase to the steady proliferation of federal agencies layered on top of institutions that already hold similar mandates. What I appreciate most is that he takes seriously a government resource that is frequently forgotten: the finite hours public servants have to design and run programs well.

⚙️Alberta AI Academy and Government of Alberta · GitHub

An incredible resource from the Government of Alberta, both championed and led by a Deputy Minister, and deeply in-tune with the state of AI knowledge among public servants: a small group of non-users, a small group of sophisticated power-users, and a lot of people just trying to keep their skills relevant. The Alberta AI Academy runs in three graduated levels (from prompting and verification fundamentals through reusable agents and on to building enterprise-grade applications) and more than two thousand Alberta public servants have taken its courses since launch in September 2025. What’s especially special about this resource is that Deputy Minister of Technology and Innovation Janak Alford teaches much of it himself, on camera, which says a lot about how seriously Alberta is treating AI up-skilling. Also notable is that the entire curriculum is open access with no login required for individuals, as well as published alongside the government’s code on GitHub under an MIT licence, which means any government or agency in the country can lift it wholesale. I recommend you take them up on it!

🎓“Guiding the provision of quality policy advice: the 5D model” by Christopher Walker and Sally Washington, Policy Design and Practice

This article suggests a model for the production of quality policy advice that is both practical and aligned with theory - and then the authors tested it against real-world governments. Walker and Washington set out five domains of inquiry (Demand, Discovery, Design, Decision, and Delivery) and shared them with senior officials from Australia and New Zealand to see how it compares to existing practice. Their findings should worry us: fewer than half of the officials they spoke to said their organization had any formal guidance for developing policy advice. I’m afraid to ask if the results in Canada would be any different.

💻Policy co-creation helps to ‘build muscles’ of policymakers to engage with citizens by Si Ying Thian, GovInsider

If you’re not concerned about the state of young people, you may be living under a rock. This year’s World Happiness Report placed Canada 71st of 147 countries when only those under 25 are counted. A July survey from Abacus Data found that, among Canadians aged 18 to 29, only 43 percent say democracy is always preferable to any other kind of government, and just 31 percent believe real change can come through voting and normal politics.

Perhaps we should be taking a cue from Singapore. Its Youth Panels, run out of the National Youth Council, put 140 young people into sustained deliberation with government agencies on a series of policy issues, producing written recommendations that can be tabled in Parliament. What’s particularly innovative is that these panels are more about policy co-creation than consultation; before deliberating, members went through workshops on context setting, policymaking, and design innovation, so the government was building the citizens’ capacity as deliberately as it was building its own. Singapore is attempting to address not just youth policy concerns, but the depth and sophistication of their civic engagement as well.

⚙️The Prosocial AI Index: What Governments Need to Know Before Deploying AI by Cornelia C. Walther, CIGI

This is something I’ve been thinking and writing about recently: what are not merely the useful uses for AI in policymaking, but the appropriate uses for AI? The answer is a mix of technical, legal, and democratic accountability concerns. Cornelia Walther, writing for CIGI, has proposed a tool called the Prosocial AI Index that should help policymakers determine whether a given system is fit for public use before the public carries any risk. The Index is a sixteen-cell matrix: four questions about the system (is it Tailored to this setting, Trained on the right data, Tested before and after release, Targeted at an authorized purpose) compared against four questions about what it is for (Purpose, People, Profit, Planet). Walther grounds the whole thing in cases that have gone badly, and pitches it as a very doable ninety-minute exercise.

🎙️GovLove: A podcast about local government, Engaging Local Government Leaders

Over the past year, I’ve worked very closely with a number of municipalities and have fallen in love with their public servants: endlessly creative with how they connect to their communities and manage tight budgets, always ready to roll up their sleeves, and good-humoured (and occasionally gossipy) in a way that makes the work fun. It’s one of the reasons why I bang on about how important it is to include municipal government in the public policy discussion: they’re highly constrained but frequently agile, and are doing the policy work that is closest to people and their real lives. This podcast, GovLove, celebrates everything municipal government and includes interviews with public servants and other experts on a dazzlingly wide range of topics. While focused on the US, there are a number of episodes with wider relevance: on cutting-edge techniques like participatory mapping, on Gen Z in the local government workforce, and on making your municipality “smarter” with technology. There’s even an interview with a dog mayor!