The headlines this week suggested New Zealand politics was locked in a student flat. Eminem, or something that sounded a lot like Eminem, was on the stereo.

Some guy called Willie was in a tizzy at his 21st, having failed to make the First XV.

Against this backdrop it is hardly a surprise that the prime minister's pre-budget speech on Wednesday attracted only modest coverage.

Disappointingly, Bill English didn't drop any of his trademark dope rhymes. He did, however, deploy a listicle, laying out 10 targets to replace the 10 set down five years ago.

Advertisement

They're laudable targets, on the whole, even if there remains no place for a target to reduce child poverty, which mysteriously continues to be just too hard to measure.

All of it, however, rests on Bill English's big idea, his signature policy approach: social investment.

As architect and cheerleader-in-chief of the social investment approach, Bill English avoids the criticisms of pragmatism and visionlessness his predecessor faced (arguablythe social investment approach, as furnished by his then finance minister, was the big idea of the John Key premiership), but is it really that big a deal? "It's the pretty straightforward idea," said English, extemporising off his prepared speech on Wednesday, "of intervening early to help the most at-risk people lead better lives and become more independent."

That hardly seems like a new frontier. Arguments for social spending have long been sold as a way to forestall greater expenditure down the track - directing resources to the top of the cliff. What is different now is the complexity of the data that informs those cliff-top interventions.

For example, the reservoir of government information reveals four factors that provide "key indicators of higher risk" of later becoming a burden on the state purse.

Those being, in Treasury's summation: "Having a CYF finding of abuse or neglect; being mostly supported by benefits since birth; having a parent with a prison or community sentence; having a mother with no formal qualifications".

Other government models incorporate dozens of other data points.

They're really going for it. A new Social Investment Agency is about to open, with responsibilities across every facet of social policy.

Amy Adams has been appointed to the new portfolio of minister responsible for social investment, with a $321 million boost in the 2017 budget.

It philosophically underpins the new, tellingly named Ministry of Vulnerable Children.

Planned reforms of the decile system in schools rely heavily on the social investment principles.

And the prime minister's drive comes from a moral core. He told the business audience (whom, he later reflected, "intuitively understand the idea of an investment"): "Too often, past governments have judged success only by what they spent, rather than what difference that spending made to people's lives" -a criticism, perhaps, of predecessors on both sides of the aisle - "yet changing lives is the whole point."

But as we charge down the data-bedazzled path of evidence-based social policy, there are more than a few reasons to take pause.

The social investment dawn follows the deluge of information that has become available - and crunchable - in recent years. The Dunedin longitudinal study, especially, has provided a wealth of evidence to recreate the Aristotlean "give me a child at seven" maxim in spreadsheet form.

Now, though, it's "give me a child at three". Last year a report based on the study and New Zealand government data, published in Nature: Human Behaviour, found that tests of brain health in children as young as three could identify those individuals likelier to in future "account for a disproportionate share of costly service use across a society's health care, criminal justice and social welfare system".

That's great, but also: gulp. Who's for a team screening of Minority Report? Are we OK with a 45 minute test on every three-year-old's brain health to pick out the ones that are likely to be liabilities? Consider, by the way, the way that study was presented in the mainstream press: "Future criminals revealed at age three" - a headline as wrong as it is unsurprising.

Beyond the potential of stigma, there are a bunch of other reasons to tread carefully on the road to the data Damascus. English says that vast amounts of social spending have historically been wasted, having had no impact on outcomes.

But can we be so sure about that? We can't measure everything after all. And while it clearly makes sense to target the most at-risk, might chucking everything into that basket mean neglecting a wider focus on welfare or opportunity or social mobility?

"Social investment" is in many ways just a warmer, fuzzier rebrand of what policy wonks have called the "forward liability investment model". Does a fixation on minimising future expenditure, while overlooking enhancements in welfare, create a distorted picture?

The Productivity Commission, for example, has warned against any approach that ignores the "wider costs and benefits of interventions", saying that a "slavish application of an investment approach based purely on costs and benefits to government might lead to perverse outcomes". Such as? "Some studies suggest that obesity might reduce future health costs as obese people die more quickly. A health system that sought only a reduction in future health costs might therefore do little, if anything, to discourage obesity."

It would be alarmist to suggest that the social investment approach might be keen on promoting obesity. But the example underscores the hazards of genuflecting before the data. In that context, it's unnerving to read Bill English's remarks about the triumph of evidence-based policy over human policy analysts.

"Public policy people have this view that everything they do is highly complex and very special," he said last year, adding: "But they're wrong. Policy is now a commodity - you can print world best practice off the internet. You don't need a department to know it, a 12-year-old can do it." Give me a policy wonk at 12? No thanks.

Are we always measuring the right things, measuring it correctly, asking the right questions of the data? Might we miss institutional biases? Is there a danger, in the phrase of self-described data nerd Keith Ng, of treating data as a magic lamp that gives us whatever we ask for? What about the implications of outsourcing? Everyone OK with the prospect of eager private contractor Serco as your local social investment banker?

That's a lot of questions. Sorry about that. But given the magnitude of the vision the government is promising it seems fair enough to ask.

One answer, perhaps, is to install an independent watchdog - a social investment commissioner, say - to keep an eye on the bold new approach.

This week Amy Adams heralded "the most remarkable transformation in how we deal with social services in New Zealand", and it has a lot to recommend it. But it should be scrutinised at least as hard as the numbers it's crunching.