INTERVIEW WINTER 2025 // Canadian Government Executive / 23 servants that need to be hired to deliver it. Every new program or regulation that goes to Treasury Board has a cost estimate, including the number of people needed. If a government wants to reduce immigration levels, over time you’ll need fewer people processing applications. Different governments can make different choices on what they want the size of government, not the public service, but government to be. And that will have consequences. That’s how it works. However, there are ways to make things more efficient. IT investments can let you do more with fewer people, and artificial intelligence could have a huge impact. There’s some discomfort about how government should use AI, but if we can strike the right balance, we can gain significant productivity. The fundamental choice for governments, and this has been true in all spending reviews going back to (Prime Minister Brian) Mulroney when there were attempts to say we’ll just kind of constrain hiring. Ultimately there was no sustainability for that in terms of changing the deficit track. What you have to do if you want to reduce the size of government is do a deep program review that looks at what are the things we want to stop doing. And what are the things that we no longer think are core to what the federal government is doing. As was true during the program review in the in the Chretien/Martin (PM and finance minister) period. Q: McLAUGHLIN: Fundamentally, you’re saying that if you really want to permanently reduce size, you have to ask existential questions about the role of the federal government—whether it should be in a certain area. Otherwise, you might get only short-term savings. Is that a fair summary? FLACK: Absolutely. That’s based on 40 years of spending reviews at both the federal and provincial levels. Yes, you can find shortterm savings, say, by cutting Parks Canada’s capital maintenance budget or deferring bridge maintenance—both have happened— but 10 years later, you’ve got a major repair backlog. Another example: the Science and Technology Museum in Ottawa had to close when the roof was literally in danger of collapsing. Cutting capital budgets might show savings on this year’s balance sheet, but the long-term cost is much higher. The Canadian experience with these reviews has been you do not get sustainable spending cuts if you just do this sort of marginal thing of a little bit here, and then we’ll limit this over there, and will ask our departments to absorb their operating costs. You have to do a deeper review to say what we’re going to stop doing or delivering. Q: TURNBULL: Sometimes you hear about whether there’s an “optimal size” for the public service, which strikes me as odd. It really depends on what government is trying to do, right? FLACK: I agree. I’m a bit of a fiscal hawk, so the real constraint is the market constraint on your ability to carry debt. We have a recent history in Canada of adjusting government size when markets lost confidence. But once you choose your spending levels and program mix, there’s a cost to delivering that, including human capital. If I had another wish beyond the Australian one, it would be to bring operational considerations to the front of the budgetary decision-making process. I’m a fan of how many provinces have the Treasury Board process before budget approval, so design choices are baked in early. Example: If a government says there’s a gap in training for the trades and it wants to invest $100 million, you could do a transfer to the provinces, which already have training programs. That’s a low-cost administrative overhead. But if you say the federal government is going to build its own training programs nationwide, that creates an increased public service head count to deliver them. I’d like to bring that discussion up front in government, rather than after we’ve made the decision. Q: TURNBULL: Absolutely. Otherwise, it’s confusing for everyone. People end up thinking government isn’t transparent about costs, which undermines trust and leads to questions about competence. FLACK: Exactly. In engineering, there’s a trade-off between complexity and efficiency, and the same is true for public policy. If you want to micro-target a program to a specific group, that can reduce costs to certain beneficiaries, but the overhead is high because you need a complex system to deliver it. If, on the other hand, you have a more simplified program, generally, you’re able to automate that and do that with much lower administrative overhead costs. That’s the trade-off. Q: McLAUGHLIN: So, is this an issue of political discipline? . What is needed are the politicians making choices that are informed by operational considerations. Or, does the public service not have the right skills and operational chops? What’s your experience? FLACK: It’s not just political actors. Historically, there’s been a challenge between policy developers and operators who weren’t always at the table together. At ESDC, we’d worked hard to integrate them, so when COVID hit, we were making policy decisions in nearly real time. If you’d walked into the room, you wouldn’t necessarily know who a policy person was and who was an operational person. It was fully integrated. Q: McLAUGHLIN: But how do we institutionalize that? FLACK: One way is for deputy ministers, even those from a policy background, to spend a significant amount of their time on major IT transformations and operational issues. That started to happen, but it needs to accelerate. Structurally, as I mentioned, I’d consider doing the Treasury Board approval before the budget. That informs cabinet about the actual costs and design before the government announces details. Otherwise, you sometimes see the details announced in the budget, and the department hasn’t even seen them, leading to suboptimal implementation. TURNBULL: I think that’s a great place to leave it. Thank you, Graham, for an excellent conversation. FLACK: I really enjoyed it—always great talking to both of you. Hope we can do it again soon.
RkJQdWJsaXNoZXIy NDI0Mzg=