One of the most striking features of the recent economic downturn was the number of countries that were affected. Virtually all advanced economies – and many developing countries – suffered large and simultaneous declines in output during the 2008/2009 recession. Figure 1 plots GDP growth rates of a select group of advanced economies and illustrates the common time-path of economic activity during this period. More broadly, the overall synchronisation of business cycles among advanced economies is high in absolute terms (a median correlation of 0.74 during the period 1995-2010), and indeed has grown substantially in recent decades (for the period 1980-1995 the median correlation was 0.29).
Figure 1. GDP growth rates of select OECD countries: 2002-2012
There has been increased attention to the sources of such business cycle synchronisation, particularly among policymakers attempting to limit the negative effects of international spillovers. A great deal of empirical work has documented the financial (see Peek and Rosengren 1997, 2000 and Kalemli-Ozcan et al. 2013) as well as policy channels (Gilchrist et al. 2014, for the case of monetary policy) of such spillovers. Comparatively less is known about the effects from ‘real’ (non-financial) channels such as international trade and investment. As discussed by di Giovanni and Levchenko (2010), there is an ongoing debate among economists as to whether these real linkages contribute to business cycle synchronisation, or whether the observed correlation is simply a by-product of some other causal mechanism. Disentangling the competing explanations is difficult, as economists generally do not have access to controlled experimental environments for the study of particular macroeconomic mechanisms.
New evidence on the transmission of shocks via input linkages
In a recent paper (Boehm et al. 2014), we provide new evidence for the role of trade and multinational firms in the cross-country transmission of shocks. Two novel contributions enable us to identify the specific mechanism at work.
- First, we link information on international ownership to firm-level data from the US Census Bureau and develop a new methodology for identifying intermediate input trade at the firm level.
- Second, the paper makes use of a ‘natural experiment’ of a large and exogenous shock – the 2011 Japanese Tōhoku earthquake/tsunami – to isolate causal spillovers from other potentially confounding factors.
The 2011 Tōhoku event was in fact a triple disaster beginning on 11 March, 2011 with an earthquake that measured 9.0 Mw – the fourth largest earthquake event ever recorded. Perhaps more damaging than the earthquake itself was the subsequent tsunami, which inundated entire towns and coastal fishing villages across Northeast Japan. Finally, the partial meltdown of the nuclear reactor near Fukushima had a devastating effect on that region, and the precautionary shut-down of all nuclear power plants in the country resulted in severe power outages that persisted for months.
In addition to the human tragedy caused by the cumulative effect of these disasters, there were severe consequences for Japanese industrial production. Manufacturing output fell by roughly 15 percentage points in the month of March (see Figure 2, left scale), only recovering to pre-shock levels five months later in August. As expected, US imports from Japan fell dramatically as well, though the large drop was recorded in April 2011, reflecting the several weeks of travel time for goods to reach the US. More surprising is the effect of this shock on the US economy, as measured by the deviations from trend in US manufacturing output during this time. US manufacturing output fell by about 1% in the month of April (see Figure 2, right scale), and remained significantly below prior levels for roughly six months following the Tōhoku event.
Figure 2. Manufacturing production in Japan and the US: Months surrounding the 2011 Tōhoku earthquake
Source: Boehm et al. (2014).
What were the mechanisms behind the transmission of this shock to the US economy? To answer this question, we turn to the detailed firm-level data at our disposal.
- In particular, we shed light on the way in which vertical production linkages served as a conduit for the transmission of this shock to the US.
The scope for shocks to the supply of imported inputs to affect US output depends critically on how easily the firm can substitute toward other inputs or find new suppliers. For example, a temporary disruption in oil production affecting a single small country is unlikely to seriously impact US firms since oil is a global commodity with many alternate sources of supply. Specialised firm-specific inputs may have fewer alternative sources, and hence disruptions to the supply of these inputs may exhibit greater spillover effects.
With this in mind, we focus on US manufacturing firms’ exposure to Japanese inputs. The US manufacturing affiliates of Japanese multinationals are – almost universally – highly dependent on inputs from Japan. These inputs comprise almost one-quarter of the cost share in these firms’ US production, and the large majority (86%) is purchased intra-firm. Moreover, these firms are large; taken as a whole, they represent over 60% of the total US imports of intermediate inputs from Japan.
The US manufacturing affiliates of Japanese multinationals suffered large drops in US output in the months following the shock, roughly one-for-one with the fall in imported inputs (see Figure 3).1
- This behaviour indicates that there is virtually no scope for substitution to other inputs for these firms – such that the shock passed directly to their US operations following the shock in Japan.
Estimates for this substitutability are much lower than those typically used by economists to evaluate the cross-country transmission of shocks via international trade.2
Figure 3. Relative imported inputs and output (proxy) of Japanese affiliates in the US
Source: Boehm et al. (2014).
- Our results also suggest that these firms reduced their expenditures on non-Japanese inputs in line with the fall in Japanese inputs. This implies that other firms in the US economy – upstream or downstream from the Japanese affiliates – would have suffered from the shock even without direct input exposure to Japan.
Such network effects can propagate the shock throughout an economy, and can serve to both amplify and extend its impact (see Carvalho 2014). A related paper (Carvalho et al. 2014) documents that such propagation of the shock to upstream and downstream firms via these network linkages did indeed occur in Japan following the Tōhoku event.
Our findings suggest that global supply chains are sufficiently rigid to play an important role in the cross-country transmission of shocks. We believe there are three key lessons for policymakers resulting from this research. First, the responsiveness of trade patterns to exchange rate changes may be muted in the short-run. Second, trade linkages and increased international input sourcing – which in part follow from foreign direct investment by multinational firms – will not always lead to greater diversification of country-specific supply chain risk. Finally, in light of low input substitutability, policies affecting the production decisions of firms – particularly those of multinational firms – should include sufficient notice in order to prevent short-run disruptions.
Disclaimer: Any opinions and conclusions expressed herein are those of the authors and do not necessarily represent the views of the US Census Bureau. All results have been reviewed to ensure that no confidential information is disclosed.
Boehm, C E, A Flaaen, and N P Nayar (2014), “Input Linkages and the Transmission of Shocks: Firm-Level Evidence from the 2011 Tōhoku Earthquake”, Working paper.
Carvalho, V M (2014), “From Micro to Macro via Production Networks", Journal of Economic Perspectives, 28(4): 23-48.
Carvalho, V M, M Nirei, and Y U Sato (2014), “Supply Chain Disruptions: Evidence from the Great East Japan Earthquake", Working paper.
di Giovanni, J and A Levchenko (2010), “Putting the Parts Together: Trade, Vertical Linkages, and Business Cycle Comovement”, American Economic Journal: Macroeconomics 1(1): 95-124.
Gilchrist, S, V Z Yue, and E Zakrajsek (2014), “The Response of Sovereign Bond Yields to U.S. Monetary Policy”, Mimeo.
Kalemli-Ozcan, S, E Papaioannou and F Perri (2013), “Global Banks and Crisis Transmission”, Journal of International Economics 89(2): 495-510.
Peek, J and E S Rosengren (1997), “The International Transmission of Financial Shocks: The Case of Japan”, American Economic Review 87(4):495-505.
Peek, J and E S Rosengren (2000), “Collateral Damage: Effects of the Japanese Bank Crisis on Real Activity in the United States”, American Economic Review 90(1):30-45.
1 These estimates are based on a treatment effects regression using a control group, properly adjusted for industry and size differences, that represents the counterfactual time-path of imported inputs (and production) corresponding to firms not exposed to the Tōhoku shock.
2 The estimates for non-Japanese firms are slightly higher than those for the Japanese affiliates, but still much lower than traditional estimates used by economists.