![Sample our Economics, Finance,Business & Industry journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5931/TOC-Subject-Sample-2021-Economics-Finance-Business-Industry.jpg"})
Abstract
This paper studies the effects of domestic and foreign demand impulses in euro area economies following the Great Recession of 2008–2009 and the Eurozone crisis of 2011–2012. Using a global Input–Output framework we apply a set of metrics to assess spillover effects of international trade in intermediates triggered by the dynamics of final demand. Our findings suggest that while cross-country trade spillovers have played a crucial role during the Great Recession, they have had a moderate impact when compared with the role of domestic sources of final demand during the Eurozone crisis. Hence, a strategy of coordinated fiscal austerity cannot be sustained by empirical evidence.
Acknowledgements
A preliminary version of this paper was presented and discussed within the framework of the Research Group PRIN 2009 Structural Change and Growth, of the Italian Ministry of Instruction, University and Research (MIUR). We thank all the group participants for their comments. We wish to thank two anonymous referees for their helpful and valuable comments and suggestions. The usual disclaimers apply.
Notes
38 Note that expression (EquationB1(B1) ) corresponds to the case analysed in the main text, as given by formula (Equation1
(1) 5).
39 This effect for Greece is approximately zero, and the cases where it deviates from zero is clearly negative. In fact, throughout the analysis in the main text, we have classified Greece into the group of countries with a negative ‘Total Inputs’ effect.
1. In fact, ‘the share of financing from surplus countries is larger than the share of exports to these countries’ (European Commission Citation2012, p. 111).
2. A global Input--Output framework requires to know precisely how much of commodity i in country r is bought by industry j in country s, and it is necessary to merge national Input--Output tables with merchandise and service international trade statistics.
3. Between 2008 and 2009, world imports fell by 9.99% at constant 2005 prices in US Dollars (Source: UNSD National Accounts Main Aggregates Database, December 2012 release).
4. Note that this amplified co-movement between trade flows, and income has seen an unprecedented increase during the last decades: the long-run elasticity of world trade to GDP between 1985 and 2000 has been estimated at 3.39 (Irwin Citation2002, p. 96).
5. The decline of real expenditure with its compositional effects have been considered of much greater importance than, for example, trade credit constraints (Levchenko et al. Citation2010).
6. Support for households’ purchasing power (reduction in VAT, direct taxes, social security contributions, as well as direct aid, such as income support for households and support for housing or property markets) accounts for about 40% of the total stimulus. Support for investment (infrastructures and public investment) and businesses (reduction of taxes and social security contributions, subsidies, export promotion, and so on) account for roughly 30% and 20% of the total stimulus, respectively. Labour-market actions (wage subsidies and active labour-market policies) account for about 10% of the total stimulus.
7. The estimation is made by means of an extended version of the ECB’s New Area-Wide Model (an open-economy DSGE model) with a richly specified fiscal sector.
8. It has even been concluded that ‘growth disappointments should be larger in economies that planned greater fiscal cutbacks’ (Blanchard and Leigh Citation2013, p. 3).
9. This is also the case of Auerbach and Gorodnichenko (Citation2012), who adopt an innovative method to identify fiscal shocks and apply it to a large set of OECD countries.
10. Thus, for an elasticity of 4, demand forces alone account for roughly 70% of the trade collapse. Crucially, the estimated elasticity of trade to GDP is high because the model allows for asymmetries in demand changes across sectors. Their analysis reflects in particular the role played by durable goods, which are both highly traded as a final demand component and tightly integrated into global supply chains.
11. However, due to the database they employ (GTAP 7 for year 2004), and the estimation method adopted for off-diagonal inter-regional trade matrices, results concerning year 2008--2009 might have to be treated with care.
12. The three components of final demand are: government consumption (cg), private consumption (cp), which includes the consumption expenditure of Non-profit institutions serving households (NPISH), and gross capital formation (gcf ), which includes gross fixed capital formation, changes in inventories and valuables. Appendix A below specifies the notation regarding the meaning and dimension of each vector and matrix used. All throughout the paper, vectors are indicated by lower-case bold characters (e.g. ), and are column vectors unless explicitly transposed (e.g.
), while matrices are indicated by upper case boldface characters (e.g.
), except for lower case characters with a hat (e.g.
), indicating diagonal matrices with the vector elements on the main diagonal. Moreover,
is used to represent sum vectors of different dimensions according to the corresponding subindex (e.g.
,
,
).
13. In order to aggregate, two basic summation rules that we will use throughout are: and
. For an accurate specification of vectors
,
,
,
see Appendix A.
14. We thank an anonymous referee for calling our attention to the need to justify our procedure of introducing a multi-sectoral accounting framework and then reporting aggregate country-level results. The key point to be grasped is that country-level aggregation is performed only after disaggregated sectoral computations have been carried out. There is a crucial difference between aggregating first and then inverting a matrix, with respect to inverting a matrix first and then aggregating the results, due to the fact that matrix inversion is a non-linear operation. Hence, compositional effects still play a crucial role, as each component is specified in its full sectoral dimension, and it is only when results at the most detailed level are obtained that a bottom-up aggregation is performed. This notwithstanding, exploring the industry-level differences composing a country-aggregate spillover figure is a very relevant question for further research, especially within a global Input--Output accounting framework.
15. Operator ∘ indicates the element-wise scalar product between two matrices of the same dimension (the Hadamard product), i.e. .
16. From equations (Equation2(2) ) and (Equation4
(4) ), matrix
can also be expressed as
. Defining
as the diagonal matrix of direct labour requirements per unit of monetary gross output of each industry in every region, gives
.
17. The WIOD Project has been funded by the EC as part of the Seventh Framework Programme, and it has been developed and deployed by a Consortium of European institutions from the Netherlands, Spain, Austria, Germany, Belgium, France and Greece. See http://www.wiod.org/ for details. The database can be accessed for free.
18. The fixed product sales structure assumption has been used in the WIOD Project to obtain a square Input--Output system from a set of International Supply and Use Tables. See Timmer (Citation2012) for details.
19. The 41 regions included are: each of the EU27 countries, the US, Canada, Mexico, Brazil, China, India, Japan, South Korea, Australia, Taiwan, Turkey, Indonesia, Russia, and an aggregate RoW region covering the Rest of the World.
20. In order to simplify exposition and comments we have concentrated on a subset of 11 amongst the 17 Euro Area countries, noting however that this subset accounts for 98% of the GDP of the Eurozone.
21. All throughout tables and figures we identify these 11 economies by their ISO3 code: Germany (DEU), the Netherlands (NLD), Austria (AUT), Belgium (BEL), Finland (FIN), France (FRA), Italy (ITA), Spain (ESP), Greece (GRC), Portugal (PRT) and Ireland (IRL).
22. The reading key for a representative row of Table is as follows: for a given row, we have that in each of the years analysed (1995, 2002, 2007, 2009). Hence, for the case of Germany (DEU), in 1995, 80.8% of its income (i.e. gross value added) was generated by German final demand, 7% by the rest of the EZ, 2.9% by the rest of the EU27 (other than the EZ) and 9.3% by final demand coming from the rest of the world, noting that 80.8 + 7.0 + 2.9 + 9.3 = 100.
23. This asymmetry might suggest that the presence of a common market is of greater importance than the institution of a common currency in explaining the productive integration among European economies.
24. These results are in line with recent findings on Global Value Chains (GVC) using the WIOD database: ‘[a]veraged across products, Belgium, Ireland and the Netherlands had the most fragmented GVCs in 2008, followed by Germany, the Czech Republic, and Hungary, where fragmentation increased at a high pace since 1995. We also find that in 1995, European value chains were mainly fragmented across other EU countries. Afterwards, however, there has been a strong trend towards increased participation of non-European countries’ (Los et al. Citation2013)
25. The reading key for a representative row of Table runs as follows. If we consider the case of Germany (DEU), we note that between 2008 and 2009, demand-induced GDP reductions originating in a drop of German final demand amounted to 0.43 (percentage) points (), while the comprehensive figure for the income reduction due to decreasing final demand (irrespective of its source of origin) was 3.79 points (
). German domestic final demand fell by 2.31 points (
), while world final demand, excluding Germany, fell by 2.21 points (
). Spillover figure
indicates that 82% of the fall in German final demand (of 2.31 points) was borne by income reductions in foreign countries, while the elasticity value of
indicates that the response of own income to changes in foreign demand has been more than proportional.
26. As discussed in section 3.2.3, structural decompositions are not unique. We have already provided a detailed explanation of the rationale for adopting the specification contained in expression (Equation1(1) 5) above, while a sensitivity analysis for all possible six decompositions is performed in Appendix B, showing that with the adoption of the present formulation the conclusions obtained in the analysis below are not compromised. The reading key for a representative row of Table runs as follows. If we consider Germany (DEU), between 2008 and 2009, the change in labour inputs per unit of gross industry output (direct labour coefficients) contributed to an increase of 1.77 million employment units or 4.42 points (column (i) in absolute value and percentage points, respectively), the effect of changing total (direct and indirect) intermediate input coefficients was responsible for a negative effect of 428,000 employment units or 1.06 points (column (ii)), while domestic final demand reductions implied a decrease in employment of 1.35 million units or 3.37 points (column (iii)). The combination of these three effects is reported under column ‘Total’, which, for the case of Germany, was slightly negative (a net reduction of 5000 employment units), meaning that labour hoarding practices compensated to a great extent the fall in final demand and intermediate input coefficients. Percentage points are computed with respect to the 2008 employment level reported under column ‘Level 2008’.
27. The reading key for a representative row of Table runs as follows. The row for Germany (DEU) describes the percentage distribution of demand-induced changes in GDP by source of origin (for a total of - 100.00 in column (9), the negative sign implying that aggregate income change has been negative). Thus, we read that domestic final demand (column (4)) accounted for 11.26 points of demand-induced GDP reductions. Columns (1)--(3) disaggregate further the domestic component, having that private and government consumption (cp in column (1) and cg in column (2)) go in an opposite direction with respect to the negative effect exerted by gross investments (gcf in column (3)), and noting that 8.05 + 15.46 − 34.76 = − 11.2 (i.e. column (4)). The remaining 88.74% of income reductions is explained by changes in foreign (final) demand (as detailed in columns (5)--(8)): 29.83 points due to the rest of the EZ, 21.92 points due to the rest of EU27 (other than the EZ), 10.46 points due to the USA, and 26.55 points due to the rest of the world.
28. We thus considered growth rates to be uniform across industries and destination country for a given component u of final demand from activating country s. Appendix C reports the set of growth rates utilised for this exercise.
29. We thank an anonymous referee for calling our attention on the interplay between final demand and technical coefficients in estimates of isolated spillover effects.
30. The reading key for Table follows the same logic as previously specified for the case of Table .
31. The reading key for Table follows the same logic as previously specified for the case of Table .
32. The figure for Ireland is particularly striking (--463%), noting that it has been almost entirely offset exclusively by foreign demand coming from outside the EU, USA and BRIC (+449% under column (11) of the Table).
33. In fact, the case of Ireland emerges with its own peculiarities, being the only country that was able to take advantage of growth in the US, BRIC and the RoW to the point of (almost entirely) offsetting the negative demand impulses coming from other Eurozone economies.
34. Moreover, note that the effects coming from the remaining six Euro Area economies (not explicitly analysed), as well as from the rest of EU countries -- columns (7) and (8), respectively -- are clearly of a smaller order of magnitude (with the exception of the presumable influence of the UK on Belgium and Ireland), as compared to those coming from PIIGS.
35. Note, in fact, that even the greatest world exporter, China, has seen a reversing trend in its income dynamics: ‘[d]omestic final demand for non-tradables has become the main source of growth’ (Timmer et al. Citation2012, p. 2, italics added).
36. The key point no. (4) above already suggests this conclusion. Furthermore, as an extension to this line of argument, a preliminary analysis of cross-income elasticities has been carried out in order to identify those economies that could both benefit the most from a coordinated fiscal stimulus and, at the same time, induce the highest amplifying effects on others. Unfortunately, due to space constraints, the analysis could not be included in the final version of the paper. However, for the interested reader, these additional results are available upon request.
37. Of course, it would be better that a demand-led policy should be undertaken at the EU level, together with structural policies as well as more effective crisis-management tools to counteract huge macroeconomic shocks (such as the sovereign debt crisis). But the size of the EU budget rules out this solution and the latest agreement concerning the 2014--2020 budget shows that EU institutions are moving toward the opposite direction (even contradicting many official proposals such as the blueprint for a deep and genuine EMU prepared by the EU Commission in 2012).