The reliability of economic forecasts has been questioned again

Christine Lagarde, a former managing director of the International Monetary Fund who is now president of the European Central Bank, did not pull any punches in a speech she delivered at the annual World Economic Forum jamboree at Davos this month. She attacked economists as a tribal clique that does not reach out to experts outside their narrow world. She then said that part of this certitude may have to do with the models that economists depend on to make sense of the world.

Earlier, in August 2023, Lagarde, a lawyer by training, spoke on how policymakers need to negotiate what she described as an age of shifts and breaks. “We rely on past regularities to understand the distribution of shocks we face, how they will transmit through the economy, and how policies can best respond to them. But if we are in a new age, past regularities may no longer be a good guide for how the economy works." She called for clarity, flexibility and humility.

Two alternatives are worth considering—as complements to rather than replacements for the existing statistical models that economists use. The first is building alternative scenarios, or a set of possibilities that tell us what may happen in the future, rather than certainty about what will happen. The Anglo-Dutch energy giant Shell had pioneered this approach in the 1970s, another decade of dramatic shifts in the global economy as well as geopolitics.

To be fair, standard economic forecasts do offer probabilistic outlooks at various confidence levels. These are not often communicated to a wider audience. But sometimes they are, as in the “fan charts" that many, including the Reserve Bank of India, periodically publish. The challenge is to communicate such future possibilities to consumers, workers, investors and companies that build their expectations on point forecasts rather than a broad range. Scenarios can help policymakers focus not on a continuous range of possibilities, but on extreme events, such as a pandemic or war, that standard models are not built to anticipate.

The scenario approach has been sparingly used in India. For example, in its draft 12th Five-Year Plan released in 2012, the erstwhile Planning Commission had laid out three different economic growth trajectories for India in a way that was very similar to the Shell technique. This column had covered that shift back then (bit.ly/3S9mwqf). Perhaps the world will see more of this approach if uncertainty intensifies.

The second complementary approach is the greater use of case studies that help economic researchers access narratives of either success stories or failures in different countries and at various points of time. That would mean greater sensitivity to economic history as well as the idea that there are different ways of managing a modern economy, depending on ground realities, and thus a major shift in the way economics is taught in universities.

Economic models do have their use. They help economists focus on the few variables that matter more than a hundred others. Understanding any complex adaptive system is not easy—complex in the sense that there are many moving parts, adaptive in the sense that each part dynamically adjusts to changes in the others, and a system in the sense that the entire unit has a logic of its own that may seem at odds with the ways of each individual part. Economies are more like the human body or a climate system than a machine.

Economic forecasting has also improved with better data as well as more computing power. It also has a decent track record when there are no large shocks and the forecasting period is relatively short. Yet, the hard fact is that these forecasting models have performed inadequately in recent years, especially in the presence of exogenous shocks. Sometimes basic economic reasoning has worked better than complex models.

There is no need to throw the baby out with the bathwater, but definitely a case to depend more on alternative approaches—as well as subjective human judgement—as a complement to statistical models.

Lagarde quoted the Danish philosopher Soren Kierkegaard in the talk she gave in August 2023: “Life can only be understood backwards; but it must be lived forwards." And she added: “Since our policies operate with lags, we cannot wait for the parameters of this new environment to become entirely clear before we act. We have to form a view of the future and act in a forward-looking way. But we will only ever truly understand the effects of our decisions after the fact. So, we will have to establish new frameworks geared towards robust policy making under uncertainty".

And here is John Maynard Keynes writing in 1936, in the context of another economic crisis: “The classical theorists resemble Euclidean geometers in a non-Euclidean world who, discovering that in experience straight lines apparently parallel often meet, rebuke the lines for not keeping straight—as the only remedy for the unfortunate collisions which are occurring. Yet, in truth, there is no remedy except to throw over the axiom of parallels and to work a non-Euclidean geometry. Something similar is required today in economics."