What Brexit says about the credibility of experts and evidence

Economists and other experts overwhelmingly believed the United Kingdom had much to lose by voting to support Brexit and leaving the European Union. Why then, did a majority of UK voters choose to do so anyway? One reason may be that many citizens no longer trust experts, including economists. Some argue that a similar distrust is reflected in the US and other advanced countries.

Why have experts lost the public’s confidence? French economist Jean Pisani-Ferry suggests it’s because they have over-emphasized the diffuse benefits of an open economy and under-emphasized the more concentrated losses. University of Chicago professor Luigi Zingales blames a combination of monetary incentives that bias their perspective and an elitist worldview. Financial Times columnist Tim Harford argues that economists (and presumably experts in other fields) have played too passive a role in the public debate and may have to more assertively step into the public life to make their case heard.

Paradoxically, even as the public’s trust in their expertise is waning, many policy experts believe they are entering a “golden age of evidence-based policy.” In one sign of this growing interest, the Laura and John Arnold Foundation (which supports the Tax Policy Center) has substantially expanded its efforts to bring more evidence to bear on public policy. The Foundation emphasizes that programs should be rigorously assessed, evaluations should be transparent, and results should be made easily accessible so that they may be tested and potentially reproduced by independent experts.  It is all good advice.

The often-divided Congress has created a commission on evidence-based policymaking, proposed by now-House Speaker Paul Ryan (R-WI) and Senator Patty Murray (D-WA). My Urban Institute colleague Justin Milner describes its duties that include exploring how to make federal data sources more accessible to researchers, how to better integrate diverse government data sets, how to create a comprehensive data clearinghouse, and ways to protect the privacy and security of data.

An old anti-poverty demonstration project known as “Moving to Opportunity” is a good example of the value of evidence-based policy research. Its objective was to help low-income families move into middle-income neighborhoods. Unfortunately, University of Michigan economist Justin Wolfers has pointed out that early evaluations indicated that in several ways the program did not work as hoped.

But recent research by Stanford professor Raj Chetty and Harvard professor Nathaniel Hendren produced some startling—and very different—results. Using new data sources, they found that the adult income of children is strongly affected by where they grow up. A second study with Harvard professor Lawrence F. Katz concluded that children in the MTO program who moved to low-poverty neighborhoods before age 13 had substantially higher incomes as adults than those who did not move.

So how can this new evidence-based research reach its potential if there is so little trust in the experts doing the work? Complex studies can be hard to describe, and as a result the general public may see little value in research or, even worse, believe it is designed and conducted to support the interests of elites. More generally, the credibility of the studies may be open to question. They may also fear that research using government data (including tax data) could jeopardize their privacy.

Here are some suggestions about how to present new research and rebuild trust in those who produce it:

• At a minimum, research should be accompanied by readable summaries that help non-experts understand results without having to rely on the possibly incorrect descriptions (and spin) of others.
• Funding for studies should be transparent. Academic journals require revelation of funding sources. It should be standard for all policy analysis.
• When appropriate, research should be peer-reviewed. Peer review has problems: it can be slow, it does not ensure agreement among articles and it can omit findings important to policy experts that are less interesting to academics. But peer review is an important part of the scientific method and it must play a part in evidence-based policy.
• Federal agencies should enforce privacy rules and receive funding necessary to maintain them. Nothing will make the public lose confidence in the research process faster than a privacy breach.
• Finally, the federal government should build expertise among its own employees. But recruitment and retention of top performers is complicated by a number of factors, including relatively low salaries of federal employees with professional degrees.

We are doing a better job than ever of using evidence to analyze policy. Now, researchers have to convince the public of the value of this work.

This article first appeared at TaxVox.