Applying import-adjustmed demand methodology to trade analysis during the COVID-19 crisis: what do we learn?

In this paper, we estimated the standard (macro-economic) import equation over the period 1995-2021Q2, using an import intensity-adjusted measure of aggregate demand (IAD) calculated from input-output tables at country level, and compared the results with regressions using GDP. Initially introduced by Bussière (2013), this "synthetic" concept of IAD was perfected, inter alia, by the IMF (2016) and by us (2017), with a view to explaining the "missing" trade flows unpredicted by GDP-based import models during the trade collapse of 2009 and subsequent recovery from it. At the time, it appeared that the integration of IAD helped predict over three-quarters of the changes in global imports, a better performance than if using GDP (two-thirds) or any other measure of aggregate demand. We had found much value to this method, as a complement to existing analytical tools, enabling to measure the relative importance of each component of demand in the variations of country/global imports, over entire economy cycles (a phase of trade expansion, a sudden collapse and a recovery). Moreover, by weighting each aggregate demand component by its direct and indirect traded inputs, import-adjusted integrated a supply-side dimension to such macro-economic modelling. By extending our estimates to cover global trade during the (on-going) Covid-19 pandemic (1995-2021 Q2), we found the IAD-based model to continue performing well, predicting 79% of changes in global imports during the period 1995-2021Q2 (10 percentage points more than when using GDP). We also found that, on average, 97% of the difference in global import growth between the pre-pandemic (2012-2019) and the pandemic period (2020), was attributable to IAD. Most of the variations in imports can be explained by changes in the growth of investment and exports, the two-most trade-intensive elements of demand, by 29% and 45%. The variations of consumption also accounted for a significant share of global import variations during this period (25%).