We do seem to be talking more about economics – what it should do and look like. But there is still a whiff of revolution about calls for the discipline to be more evidence-based and, well, scientific. This article, by Philip Aldrick in the Times yesterday, argues for more careful scientific approaches, and this is worth noting. Of course, in the natural sciences, this would be taken as read. Drugs need to be extensively trialled before they are sold and used to treat disease in humans. But for some reason, in the social sciences, theory and ideology have the ability to shape policy just as much as evidence.
Aldrick’s piece cites two studies launched by Nesta, a UK Innovation think tank, roughly seven years ago. The first was a retrospective review of the effectiveness of business clusters; do small businesses do better when they are closely located and can share location and labour advantages? The second was a randomised controlled trial on whether tax relief for small creative companies worked. The results of the studies were not their most important findings however.
For the sake of finishing a story, the first study proved relatively inconclusive, and could not find any clear correlation between clusters and growth. The second study found that tax and financial incentives were helpful in the short term for small creative businesses, but after 12 months any advantage had faded.
So, what was the main impact? The reason these two studies are remarkable are for their illustration of research methods. While retrospective reviews – generally the majority of most empirical work in the social sciences – can only look for correlation, randomised controlled trials (RCTs) can go deeper, further, and can identify causative factors. In other words, we can target specific factors and identify why things happen. This is important because it means we can be more scientific about what works, how it works, and why. And this means we can begin to base policy on evidence rather than theory. RCTs also offer a way of measuring the extent of policy impacts. By having a test group and a control group, we can gauge the extent to which a policy really makes a difference. And that means we can evaluate whether a policy is financially and economically viable. So, RCTs offer a way of seeing not only what works, but how much.
Why is this news? Similar to other recent posts on here, there is increasing discussion of economics and how the discipline can be improved in the mainstream media. Aldrick’s argument is that economics – both the research and the formulation of policy – can and should be more scientific in its approach. And to this end he calls for more RCTs and longer term studies testing causation before policy is enacted. The government has launched the Business Basics Fund with Nesta to carry out trials investigating, among other things, productivity. UK productivity lags behind that of other countries, attributed generally to poor management practices. But how can management practices be altered to improve productivity?
Questions like this lend themselves readily to RCTs where different techniques can be trialled in comparison with a control group. Nevertheless, there are questions of macroeconomics that are not suitable for trials. We cannot test interest rates or tariffs, for example, against control groups. And this remains a problem for the larger questions tackled in macroeconomics, where theory remains a significant influencer of policy.
Calls for greater use of careful empirical data in shaping economic, legal and social science policy is not new though. Economic sociology, economic sociology of law, and sociolegal approaches have long stressed the need for analysis and understanding to be based firmly in the real world, on real data, and about real people. Increasing access to big data and AI could enhance this. As Aldrick states, “Economics is a social science. Why not make it more scientific?”
Clare's research blog 