Government by trial and error: a discussion with Tim Harford
Tim Harford, Financial Times "undercover economist" and presenter of Radio 4's More or Less, author of Adapt: Why Success Always Starts With Failure
Speaker:
Tim Harford, Financial Times "undercover economist" and presenter of Radio 4's More or Less, author of Adapt: Why Success Always Starts With Failure (2011)
Discussant:
Professor John Van Reenen, Director, Centre for Economic Performance, LSE
Harford's book argues that because the world is complex and unpredictable, most interesting problems are inevitably solved through a process of experimenting and adapting - whether the experiments are formal controlled trials or more informal explorations.
These themes are also explored in Institute's new report, System Stewardship: the future of policy making.
Event Report
Tim Harford's book, started out as an investigation of how economics could solve the most important issues facing Government, such as climate change and terrorism. But he changed focus - adapted - after he realised the most interesting problems are solved through a process of trial and error.
The marvel of the market...
The economy works through trial and error - whether in creating a cheap toaster - near impossible to reproduce at home (though someone called Thomas Thwaites tried) or supplying the 10 billion goods available in London. Firms' fortunes wax and wane - within 2 years of the publication of the 1982 bestseller In Search of Excellence one third were in financial distress (some later recovered). They were excellent - for a while. 10% of US businesses go bankrupt each year. Those individual failures are a source of growth within the wider economy, a selection mechanism that rids the economy of bad ideas and retains the good ones. Economies with more failures have higher growth rates.
But the market can't solve everything...
Many problems are not amenable to market solutions. But governments are much less willing to allow trial and error - which means that selection and improvement mechanism is missing. Government should be more willing to experiment and set-up randomised trials or other forms of evaluation where possible. This was standard practice in medicine - and saved lives.
The Cochrane Collaboration was established to run systematic reviews of the effects of health interventions - amassing all the data available from good studies and less good ones, to reach conclusions and point to better practice - and was particularly valuable in undermining the "God complex "of assuming you know the answers (which applies to government as well as doctors)! But the more recent Campbell Collaboration, trying to do the same for social policy, was having less impact.
Learning to take more risks would need a culture shift to be more open to the prospect of failure. At present, the adversarial nature of politics (and the media) hinders such an approach. 50 things might work but the one failure becomes a major problem for the government. In contrast, businesses can have many failures, as long as these are small, and produce big successes. It was better to be Silicon Valley - failing fast but survivably, than Wall Street - too big to fail.
John Van Reenen was concerned about the lack of evidence based policy making. EMAs had been subject to rigorous studies, but had been scrapped. Meanwhile, the latest NHS reforms have virtually no evidence base. There was a huge pool of academics that are keen to carry out those evaluations. Data was now much more available than ever before - and a time of austerity - when funds were short - was a great time to be looking for what worked.
Government needs to set up a system to evaluate policy from the start, and create institutions that are independent enough to look at evidence and make informed decisions (e.g. NICE). The US Healthcare plan had built in resilience to politics to allow successful trials to be rolled without reauthorisation by Congress.
In discussion, doubts were raised about whether Ministers, civil servants or decentralised decision-maker had the right incentives or skills to make the best use of data and concerns about what role was left for politicians if decisions were handed over to the technocrats. But Tim Harford stressed that politicians would still be there to make the big political choices on redistribution, on spending priorities, but on the basis of what worked not poor data and absent evidence.
Event report by Edward Marshall
More information
- Topic
- Policy making
- Keywords
- Economy
- Publisher
- Institute for Government