To achieve valid reform, we must know what works

Amidst our now decades-long productivity slowdown, I’m often asked what big ideas I have to usher in a new golden era of economic reform. In lieu of floating the dollar a second time, I typically don’t have much to add that isn’t obvious. With fewer low-hanging fruit, our mode of progress has settled into a groove of cautious incrementalism.

Illustration by Simon LetchCredit:

But I do have one idea I think could raise our productivity, economic growth, and living standards – all within the confines of this incremental system. And that is, simply, to figure out which policies actually work.

This may come as a shock but, in many areas, we have no idea. At least in economic policy, many of our pronouncements are based on little more than logic. Or studies carried out in other countries with very different institutions.

It’s often claimed that high company taxes are ultimately borne by workers in the form of lower employment and wages. That’s based on a very simple model of the economy. The only recent evidence we have is from company taxes imposed in American states and German municipalities.
That’s not to say they’re wrong. And I don’t advocate a mindless mantra of “evidence-based policy” at all costs, but wouldn’t it be nice to have a precise understanding of the effects Australia’s company tax has on investment, wages, employment, productivity, and so on? Because we don’t.

Economists get a bad rap, often deservedly. That common refrain – “I know it works in practice, but does it work in theory?” – may have been fair in a bygone era. But in the past 30 years, economics has undergone a profound transformation into an evidence-first social science.

Indeed, the Nobel Prize in economics last year went to Professors David Card, Joshua Angrist, and Guido Imbens for their critical contributions to this “credibility revolution”, while two years earlier it went to Professors Abhijit Banerjee, Esther Duflo, and Michael Kremer for their similarly critical contributions to the development of field experiments in economics.

This revolution has been enabled by massive advances in data. At least on this, Australia has made tremendous strides. Via the ATO’s ground-breaking ALife project, researchers can now analyse three decades of Australian personal tax records, which give us a sharp picture of exactly how people respond to new policy. We can see if they did what we said they’d do.

But while improved data has catalysed many high-quality studies, there is still a long way to go in shrinking the deficit in our understanding of Australia’s broad economic policy architecture. If the new government is serious about reversing Australia’s productivity slump, I can’t think of a better place to start.

What are the barriers? The first thing to keep in mind is not all evidence is created equal. How can we tell good evidence from bad? Well, it just so happens the entire academic discipline of economics is geared towards working out which claims are true and which are false – the “credibility” in “credibility revolution”.

So, in the first instance, we need more academics to work on Australian policy questions. Which is tough because the incentives are currently geared towards the US, which at least in economics is the centre of the academic world. (An attendee at a US talk I once gave asked: “Who gives a f$%# about Australia?” )

If we want academics to work on Australian policy questions, we need to adjust the incentives. There are many ways to do so, but we could start with root-and-branch reform of the way the Australian Research Council functions. And specific government departments could directly incentivise creation of the evidence of interest to them.

This would be aided by policymakers simply being more engaged with academics – inviting them to share their work, give seminars and meet with staff, provide guidance and training, or visit the department during a sabbatical. They could offer access to data in exchange for producing work of interest.

The government should also look to develop more of these capabilities in-house. Some of this is happening but we could and should do more. There should be a single shop, somewhere in government, solely devoted to using government data to evaluate the effects of policy. A massive public good.

The seeds of this were laid in now-minister Andrew Leigh’s proposal for an “evaluator-general”. I agree on the need for more field experiments in Australian policy. But there should be at least as much focus on the “quasi-experiments” embedded in countless policy reforms hiding in the data we already have.

Which leads us to the big barrier I haven’t mentioned. It’s not by accident there hasn’t been much momentum behind this idea. That’s because policymakers often don’t want to know the answers. If a policy doesn’t do what they said it would, or maybe even makes things worse, the political costs may be significant.

As with other reform ideas, this one is going to involve a bit of bravery among our elected (and unelected) officials. Our new government talks a big game both about wanting to improve accountability and wanting to turbocharge productivity. But does it have the guts to expose itself to the full scrutiny of an academic referee on the warpath?