Science and subterfuge in economics

NEW DELHI — Mainstream economics has a tendency to decide on some “established” conclusions, and then hold to them, notwithstanding all evidence to the contrary. This is bad enough, but what may be worse for a discipline that lays claim to being a science is the lack of insistence on the replicability of empirical results. This is both standard and essential in most natural sciences; in economics, by contrast, there is mostly indifference and occasionally even fierce resistance to it. In some cases, the data that must be used to replicate conclusions are denied to other researchers.

The reason is often deeply political, because the results that are promoted and disseminated accord with visions of the economy that support particular ideological positions and associated policy stances. For example, empirical work that supports fiscal austerity or market deregulation is cited extensively and becomes the basis for advancing those particular policy outcomes. Very rarely is such work subject to the scrutiny, for example, challenging its assumptions and questioning its statistical procedures, that would be the norm for research in the natural sciences.

Consider the claim made by Stephen Moore and Arthur B. Laffer that the Trump tax cuts in the US would not only pay for themselves, but also actually bring down the government deficit while generating more private investment. Their claim was completely wrong, but somehow economic reality seems to have had little impact on those who continue to believe the assertion of the Laffer Curve that lower tax rates will generate higher tax revenues.

Now, a new paper by Servaas Storm effectively demolishes another famous trope of neo-liberal economics: the argument that labour market “rigidities” depress output and employment. One of the empirical investigations most often cited for this argument is a paper by Timothy Besley and Robin Burgess using manufacturing data across Indian states for the period 1958-92. Besley and Burgess claimed to show that pro-worker regulations in some states resulted in lower output, employment, investment and productivity, and even increased urban poverty, relative to states that did not adopt such regulations.

This conclusion came to underpin the conventional wisdom that labour-market regulation is harmful for industrial expansion, and that the way to increase production and employment in manufacturing is to promote more labour-market “flexibility” by repealing laws that protect workers. This wisdom prevailed not only in India; it influenced policies accordingly across a wide range of developing countries. Although various economists raised serious concerns about the methodology Besley and Burgess adopted, their criticisms never gained much traction among policymakers.

But Storm’s critique is more fundamental, because his study reports a failure to replicate Besley and Burgess’ findings and demonstrates that their conclusion concerning the impact of labour regulation on manufacturing performance is statistically non-robust. He finds that the results are not just inconsistent with the authors’ own theoretical assumptions, but are also internally contradictory and empirically implausible. Storm comes to the devastating conclusion that “the paper is a professional embarrassment... it almost perfectly illustrates how a combination of scientific pretension and a deep desire for respectability can lead to a gratuitous empiricism in which priors trump evidence”.

So how did Besley and Burgess get away with it, and why have such results not been more comprehensively trashed in the literature and in policy circles? After all, this article was published in a top-tier double-blind peer-reviewed economics journal. It was used to justify a wave of labour-market deregulation across the world, actively harming workers. The deep complicity of the economics profession, and of the mainstream academic journals that confer “respectability” on such research, needs to be called out for this.

It is no secret that mainstream economics has operated in the service of power. John Kenneth Galbraith noted in 1973 that establishment economics had become the “invaluable ally of those whose exercise of power depends on an acquiescent public”. If anything, economists’ embrace of that role has grown stronger since then. But it has also made the subject less relevant and reduced its legitimacy and credibility. Economists are no longer seen by much of the public to be asking the right questions or seeking to answer them with integrity.

To recover credibility, economics needs to become more open to criticism of assumptions, methods and results. The inconvenient truths spoken by dissenting voices cannot be ignored indefinitely. Sooner or later, reality bites.

Jayati Ghosh is professor of Economics at Jawaharlal Nehru University in New Delhi, executive secretary of International Development Economics Associates and a member of the Independent Commission for the Reform of International Corporate Taxation. Copyright: Project Syndicate, 2019. www.project-syndicate.org