Economic orthodoxy dictated the reaction of the Union government to the prolonged slowdown in the Indian economy. In September, finance minister Nirmala Sitharaman cut corporate tax rates by ordinance and, while urging the passage of the record giveaway into law in the ongoing Parliament session, she described it as a powerful stimulus to investment, growth and employment.
Few questioned the underlying logic — unsurprising since in the dominant paradigm of economics, it is almost axiomatic that success is about getting the incentives right and reserving maximum rewards for the most productive elements in society. It is a beguilingly simple chain of reasoning. Tax cuts bolster corporate savings, which in turn promote investment.
When the highest honour in the economics discipline — the Nobel Prize — has just gone to a trio of practitioners who have made experimentalism the key test of validity, the persistence of this mythology should occasion surprise. It does not take the much-celebrated process of “randomised controlled trials” (RCTs) to conclude that “supply side” economics has caused growing deficits, cutbacks in welfare spending, and huge accretions to corporate treasuries. That is an observation applicable to every instance where “supply side” has become the dominant policy paradigm.
The consequence of reserving disproportionate gains for higher income strata is an unequal pattern of growth, driven mostly by the consumption of the rich and the “wealth effect”, or the spending prompted by increasing asset values. Speculative bubbles in asset markets are a constant feature of this pattern of growth. And that is a fact revealed by the most cursory observation.
In many ways, the 2017 Nobel award to Richard Thaler for his contributions to “behavioural economics”, and this year’s prize for Abhijit Banerjee, Esther Duflo and Michael Kremer for their contributions to the use of RCTs in the study of development, focus attention on the crisis within the discipline. Like all experimental procedures, RCTs are an abstraction from the larger world. They purport, however, to arrive at inferences that could be applied across a wide range of conditions.
Through randomised trials in diverse situations that share common developmental challenges, RCTs seek a pattern that would apply to the wider reality. RCTs, in the narration of its proponents, provide “evidence based” answers to complex questions. And this is a far more valuable way of spending budgetary resources and donor dollars, than a plodding quest through the thickets of politics and society for the elusive key to development.
In the idiom of the theory of knowledge, RCTs are a way of acquiring verifiable methods of intervention in the cause of development. But like all such processes contingent on verifiable results, it is vulnerable to the “black swan” event: One instance among many that produces a contrary result. How then does one arrive at general conclusions that could apply across situations posing diverse developmental challenges?
This is equivalent to the choice that the theory of scientific discovery faced, between “verification” and “falsification” as the touchstone of authenticity. The two are asymmetric in their relationship to experimental observation. A single observation of a black swan is sufficient to refute the generalisation that “all swans are white.” In contrast, an infinite number of observations of white swans would not support the definitive verification of the same proposition, since that black swan could be lurking just around the corner.
This compels the adoption of a criterion other than “verifiability” as the pathway to an “evidence based” approach. The philosopher Karl Popper required that a scientific proposition be “falsifiable” in principle, that is, it should specify the set of conditions under which it could be disproven.
By this criterion, a minimal requirement for admitting the validity of the RCT procedure would be that it should specify when remedies applied in one context are likely to fail. Involving parent-teacher associations in rural schools may promote better education outcomes in five out of 10 situations, but learning would be incomplete without identifying the reasons for the five failures. That is where the RCTs process bumps up against the realities of market dynamics, distributional relations and antagonisms, and the fallibility of political institutions.
If reproducibility is the first challenge that RCTs fail to surmount, “scalability”, or their application across wider expanses and populations, is the next. In a milieu dominated by “supply side” orthodoxy, the growth metric becomes a singular focus. The proposition that incentives for the rich would promote growth may hold up in some measure, but the related assertions that growth boosts budgetary resources and contributes to distributive justice, have been demonstrably proven false. This is no “randomised” observation, but one supported across a wide range.
This is what should in the theory of scientific advance, lead to a “paradigm shift” or a radical break with existing practices. Theory has taken some hesitant steps in that direction, though without great impact on practice. Restoring its relevance requires not experimentation, but the reintroduction of politics as its indispensable component.
Sukumar Muralidharan teaches at the school of journalism, OP Jindal Global University, Sonipat