How thick is your armour? Measuring economic resilience to shocks in global production networks

Abstract

When economic shocks occur, whether at home or abroad, economic agents are expected to react to reduce the negative impact or amplify the positive effects. The ability of a country to contain economic losses can be defined as the resilience to economic shocks. Using the OECD’s annual Inter-Country Input–Output (ICIO) tables from 1995 to 2011, this paper investigates the relationship between changes in final demand and production structures for 61 economies. We found that, during economic downturns, countries that are able to prop up the economy through the domestic service sectors instead of domestic goods and foreign sectors are more resilient to negative shocks. Therefore, understanding the substitutability between goods and service sectors and between domestic and foreign sectors is crucial for gauging the potential risk to a country’s domestic economy from shocks abroad – whether economic, environmental, health-related or political.

KEYWORDS:

• Economic resiliency
• structural changes
• input–output
• global value chains

JEL Classifications:

• C14
• D57
• E12
• F47

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 See, for example, Miller and Blair (2009), Okuyama and Santos (2014), and Arto et al. (2015).

2 OECD (2014a) and OECD (2014b) discuss better social and economic policies to better withstand environmental, political, economic and social shocks.

3 The ICIO database provides split tables for China and Mexico to reflect a difference in production structures between processing exporters and domestic firms. See http://www.oecd.org/sti/ind/input-outputtables.htm.

4 The exchange rates are obtained from the OECD National Accounts and the UNSD (United Nations Statistical Division) National Accounts. The GDP deflators are constructed as GDP at current price divided by GDP at constant price, which are obtained from the National Accounts data of the United Nations (UNSNA).

5 We use the local polynomial regression method (Cleveland et al., 1992).

6 The charts for the remaining countries are available upon request.

7 This paper’s structural decomposition differs from the standard SDA in the following two points: (1) we conduct SDA for value-added decomposed with three final demand sources: domestic goods, domestic services, and foreign demands; and (2) given the total final demand of each country ($f_{ct}$), we apply the method of SDA to examine the contribution of changes in value-added ratios, the Leontief inverse, and the composition of final demand to the actual value-added. For the general discussion of SDA in an interregional IO model, refer to Rose and Chen (1991), Dietzenbacher and Los (1998), Akita (1994, 1999), Hitomi et al. (2000), and Miller and Blair (2009, chapter 13).

8 For this decomposition, we refer to Akita’s (1994, 1999) growth factor decomposition equation based on an interregional input-output model.

9 Figures B1–B3 of the online Appendix visualise the time-series behaviour of the three components (Value-added ratios, the Leontief inverse, and final demand share) for each country.

10 Brunei Darussalam is removed from the sample because there are some outliers which can influence the average of countries.