John Lewis and Selien De Schryder.
Did the dramatic fall in global trade exports in 2008/9 stem from a temporary shock, or did it herald a permanently lower level of trade? The answer has important implications for the UK’s growth prospects. If it was a permanent hit, then the UK’s weak export growth after the crisis is more explicable, but external growth sources have become less fruitful. We find that advanced economies as a whole exhibited a high degree of bouncebackability, since exports recovered to levels consistent with the pre-crisis relationship between output and exports. But this masks considerable variability at the country level, with UK exports in particular recovering by less than expected.
In the wake of Lehman Brothers’ collapse, world trade fell faster and farther than at any time on record. And as Baldwin (2009) and others have noted, this fall appeared far larger than could be explained even by the sharp falls in GDP at the time. This prompted a substantial debate about how to interpret the great trade collapse (GTC): some argued that this represented a permanent hit to world trade, either because of hidden protectionism, or because the process of outsourcing and lengthening supply chains had finally run its course. But others argued this was merely a temporary shock reflecting financial constraints, the split of GDP between investment and other components, inventory adjustment or heightened uncertainty.
These two sets of explanations imply quite different things for the extent of bouncebackability in world trade. Bouncebackability is a term, if not a word, which is familiar to many sports fans and is generally defined as the ability to recover quickly from a setback. If the GTC represented a permanent hit to the relationship between trade and GDP, we wouldn’t expect to see trade bouncing back to levels implied by the pre-crisis relationship, but if the GTC was merely down to temporary shocks, we would expect to see trade recovering to levels implied by a pre-crisis model. In a recent paper, we attempt to test these two competing hypotheses empirically, and find that the evidence points to the latter, temporary shocks, explanation.
An earlier post on this blog by Kate Stratford argued that world trade has been weak relative to world GDP because since 2012 global GDP growth has been more concentrated in emerging markets which suck in fewer imports per dollar of GDP growth than advanced economies. In response, Paul Krugman argued that import intensities are not structural parameters, and that the underlying cause was a slowing in the growth of world supply chains. Whilst the focus of this piece is on a different time period, it does have relevance for that debate. Since we find that world trade bounced back to a level consistent with the pre-crisis world GDP: world trade relationship, that goes against the idea that globalisation and lengthening of supply chains had plateaued prior to the crisis. If they had plateaued, we would expect a model using pre-crisis coefficients to predict a bounce back in world trade greater than actually occurred.
The first step is to estimate a model of the determinants of exports for 19 advanced economies, using pre-crisis data. We use the common correlated effects estimator of Pesaran (2006), an approach which explicitly allows for the fact that common factors, i.e. those which are outside those expressed in the model’s variables, may have influenced exports across all countries. Alongside the more traditional determinants of exports – trade-weighted rest of world GDP and the unit labour cost based real exchange rate – we also include a variable which captures the effects of the shifting composition of global trade. The intuition here is that if the makeup of world trade moves away from the type of goods that the UK exports, that would hurt UK exporters, even if the total volume of world trade remained unchanged.
When we split up the real exchange rate into the nominal exchange rate and a measure of relative unit labour costs, we find a different response of exports to each of these elements. A real exchange rate depreciation stemming from a fall in the nominal exchange rate has a smaller effect on exports in both the short and long run than an equivalent-sized one stemming from a fall in unit labour costs. This is consistent with the argument that the competitiveness boost from a nominal depreciation is partially offset by a rise in import input costs. Or that firms “look through” nominal exchange rate changes, perhaps reflecting Obstfeld and Rogoff’s result that nominal exchange rates are more volatile than economic fundamentals.
The next step is to use the estimates to construct a forecast of what each country’s exports “should have been” based on the outturns of the exogenous variables and the common correlated effects. We then aggregate these up to yield a forecast for the path of advanced economy goods exports as a whole. This is shown below in figure 1, alongside the actual outturn of exports.
Figure 1: Actual vs forecast exports of advanced countries as a whole
Overall, the forecast (yellow diamonds) fits the data (dashed line) fairly well. But whilst it captures the GTC fairly well, most of the fall in the forecast profile is accounted for by the common correlated effects (dark blue bars) – i.e. factors outside of the variables included in our model. And so, the fall in exports in 2008/9 is far sharper than the weaker world GDP could account for (the yellow bars).
But the key result of interest is what happens in the subsequent recovery- actual exports recovered to a very similar level to what a pre-crisis model would have predicted. The contribution of common correlated effects dies away to virtually zero. In short, exports bounced back to exactly the level we would have expected given the outturn of GDP and the pre-crisis relationships between GDP and trade, which is evidence in favour of the “temporary shock” explanation.
So having apparently ruled out the hypothesis that there was a structural break in the determinants of exports around the time of the GTC, it is interesting to delve in more depth into the performance of individual countries. Comparing each countries actual path of exports, with the forecast based on the pre-crisis model is a way of assessing whether each country over or under performed relative to its peers. Figure 2, below, plots the level of exports at the end of our post-crisis sample, relative to the forecast benchmark.
Figure 2: Actual minus forecast exports for individual advanced countries
When ones adds up the forecasts on either a weighted or unweighted basis, there appears to be no systematic over- or underperformance. But there is a pretty wide degree of variation in performance across countries. The best performers are the “core Europe” economies of Germany, Austria and the Netherlands, those at the bottom of the rankings are Japan, Finland and Denmark. And the UK has performed relatively poorly – with exports around 8% weaker than would have been expected. Even allowing for the fact that nominal exchange rates boost exports less than an improvement in relative unit labour costs, and for the weak external growth environment, UK goods exports should have been stronger in the wake of sterling’s 2007/8 depreciation.
So in aggregate, advanced economy exports did exhibit commendable bouncebackability, consistent with the view that the GTC was just a temporary shock rather than a structural one. But for the UK, and some other economies, the export performance was not quite so resilient.
John Lewis works in the Bank’s Research Hub, Selien De Schryder works at Ghent University.
If you want to get in touch, please email us at bankunderground@bankofengland.co.uk. You are also welcome to leave a comment below. Comments are moderated and will not appear until they have been approved.
Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.
