On adaptation

Murray Mandryk's Wednesday column serves as a downright painful example of Monday morning quarterbacking - cherry-picking examples from seven decades of Saskatchewan governments to criticize "rash decisions" without recognizing the difference between reasonable experimentation and blatant cronyism. And under Mandryk's implicit standard for public-sector risk aversion (that if something could possibly prove to be anything less than an unqualified success, it's not worth doing), Saskatchewan's legislative assembly would be meeting around a donated table in a barn situated in the middle of the still-undeveloped prairie.

But Mandryk is far from the only voice suggesting that such a standard should apply to government decision-making. As much as I agree with Donald Savoie's overall criticism of the viewpoint that government should be run like a business, he too takes as a given that the public sector should be held to an impossibly high standard:

It doesn’t much matter in the private sector if you get it wrong 40 per cent of the time so long as you turn a handsome profit and increase market share. It doesn’t much matter in the public sector if you get it right 99 per cent of the time if the 1 per cent you get wrong becomes a heated issue in Question Period and the media.

But the consequence of accepting that view is that we're bound to be stuck with a sub-optimal set of choices. Private actors will effectively be free to go ahead even with projects which have a near-certainty of causing public harm (since a government can't be certain it will do good by intervening to stop them). And public projects with even a strong possibility of success and a minimal opportunity cost will be rejected out of hand.

So what alternative model can we use to make sure that government decision-making best serves the public interest - both in avoiding bad ideas, and encouraging the development of good ones?

I'll suggest that we should want our governments to apply substantially the same lens to regulatory and public policy decisions. In either case, we should consider the costs of both action and inaction - and recognize that stagnation in a system which fails to try new ideas is itself a detrimental outcome.

Fortunately, that lens is nicely described by Tim Harford in Adapt. Without getting into too much detail, I'll summarize his philosophy as applying an evolutionary model to public policy development: encouraging new ideas which have a plausible (if far from certain) chance of success, while ensuring that risks are taken only on a scale where their failure won't significantly affect the broader public interest.

As a corollary of that view of public policy, a single failed program doesn't necessarily signal a failed government. (In fact, the reasonableness of a government can be found in large part in its willingness and ability to evaluate its own decisions - including acknowledging the failure of pilot projects while recognizing how the lessons learned may lead to future success.)

Instead, the fundamental question is whether a government orients its activities toward a reasonable set of opportunities. And that should include both making choices at the cabinet table, and devolving survivable amounts of decision-making authority to lower levels.

Of course, the evolutionary model for public policy development also implies that we should look to preserve the positive results of successful adaptation. And it's here that Saskatchewan's right-wing governments have done the most damage: from giving away Crowns which could have held tens of billions of dollars of value for the public to taking a wrecking ball to a well-developed film industry without even a hint of reasonable explanation, the PCs and Sask Party have proven to be more interested in demolishing anything which might speak to the NDP's good decisions than in maintaining and improving ideas which have obvious benefits for the province.

In sum, we should expect our elected representatives to look for opportunities to advance the public interest. And while it's indeed essential for governments to make reasonable choices, we won't get those if we presume that doing nothing is the only safe option - or indeed a viable choice in general.

[Edit: fixed wording.]