61/2 Decades of Global Trade and Income
'New Normal' or 'Back to Normal' after GTC and GFC?
Author: Sanjay Kalra

Contributor Notes

Author’s E-Mail Address: skalra@imf.org

Global merchandise trade expanded rapidly over the last 6½ decades and its relationship with global income has seen ebbs and flows. This paper examines the shifts in this relationship using time series data over 1950-2014 and situates it in the current and longer term context. The conjunctural context comes from, among other things, the “great trade collapse” (GTC) and the global financial crisis (GFC) in 2009, and developments since then. The longer term context comes from the relative role of “globalization” and “technology” shocks in accounting for the short and long run variance of global exports and income. The paper estimates trade and income elasticities using ADL models taking account of structural breaks, and impulse response functions from structural VARs. The estimated SVAR model provides a lens to ask whether global trade and income are in a “new normal’ or only “back to (an old) normal” after the GTC and GFC.

Abstract

Global merchandise trade expanded rapidly over the last 6½ decades and its relationship with global income has seen ebbs and flows. This paper examines the shifts in this relationship using time series data over 1950-2014 and situates it in the current and longer term context. The conjunctural context comes from, among other things, the “great trade collapse” (GTC) and the global financial crisis (GFC) in 2009, and developments since then. The longer term context comes from the relative role of “globalization” and “technology” shocks in accounting for the short and long run variance of global exports and income. The paper estimates trade and income elasticities using ADL models taking account of structural breaks, and impulse response functions from structural VARs. The estimated SVAR model provides a lens to ask whether global trade and income are in a “new normal’ or only “back to (an old) normal” after the GTC and GFC.

I. Motivation

The co-movement of global trade and income has regained prominence in the context of the “great trade collapse” (GTC) in 2009 following the global financial crisis (GFC) in 2008Q4 and the sluggish recovery since then. In 2009, nominal and real global exports fell by 26 and 13 percent, respectively. Export values and volumes recovered quickly in 2010 to surpass the pre-crisis level. However, the average growth rate during 2011-14 has been only 5¼ and 3¼ percent, respectively. During the same period, global growth averaged a little over 2 percent. From a longer term perspective, global exports expanded rapidly during the past 6½ decades (1951-2014). Nominal and real merchandise exports grew at annual average rates of nearly 9½ percent and 5½ percent, respectively (Figure 1). Over the same period, global real GDP (income) grew at an average annual rate of 3½ percent (Table 1).1

Figure 1.
Figure 1.

Global Exports, 1950-2014

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

Table 1.

Global Exports and Income: Growth Rates

(Average annual percent change, 1951-2014)

article image
Sources: WTO (2015) and author’s calculations.

The long run relative growth rates have generated a “rule of thumb” that the rate of growth of real exports is double the growth rate of real income, i.e., the elasticity of global exports to income—a measure that is often taken to summarily represent the relationship between the two variables—is 2. Notwithstanding this “rule of thumb”, there is variation in the literature on the elasticity of global trade to income. A part of the variation across the studies comes from differences in data sources, sample period, specification of the estimated equation, and whether the reference is to short term or long term elasticity.

Against the backdrop of the slowdown in the growth of global exports, especially compared to the 2000s and prior to the GFC, the relationship between global trade and income has generated significant comment.2 The commentary had varied from concern over whether the trade slowdown reflects lost “mojo” (Davies, 2013) to the argument that there is no a priori reason to expect that the relationship should remain invariant (and with a high elasticity) in all time spans (Krugman, 2013).

This paper revisits this discussion in a tractable, dynamic framework with extended time series data. In particular, the extended time span helps to put the current discussion in historical perspective and ask how large the GTC and GFC were relative to previous shocks. The longer time span also allows a look at the episode from a “forecast” perspective: what would have been expected during 2009-14 from various time points if the past had continued into the future? For example, what would the forecast for global exports and income have been in 1990, 2000 and 2008? In addition, such an analysis is useful to ask which of the effects mentioned in the literature is significant, over which time periods, and with what relative importance. The remainder of the paper is organized as follows. Section II sets the stage, with a brief selected literature review, for the estimation framework of Section III. The empirical results are also presented in Section III. Section IV interprets the GTC and GFC with an estimated SVAR model.

II. Selected Literature Survey

Irwin (2002) estimates the trade elasticity for three sample periods: pre-World War I (1870-1913), interwar period (1920-38) and post-World War II (1950-2000).3 It also provides estimates of the elasticities over subsamples, identified by structural breaks in the data. The paper concludes that trade grew slightly more rapidly than income in the late nineteenth century, with little structural change in the trade-income relationship. However, during the interwar and post-war periods, there were structural breaks. These breaks—for the post-war period—are identified in 1974 and 1985. With these structural breaks, the estimated elasticities for the three subsamples imply that since the mid-1980s trade was more responsive to income than in any of the other periods (short and long-term elasticities of 1.55 and 3.39, respectively), although the results cannot directly determine the reasons for the increased sensitivity of trade to income. As backdrop, the trade policy regime differed across periods, from the bilateral treaty network in the late nineteenth century to interwar protectionism to post-war GATT-WTO liberalization. The commodity composition of trade had also shifted from primary commodities to manufactured goods.4 While providing these clues, the paper does not empirically estimate the contribution of these factors.

Several studies revive the determinants of trade growth noted by Irwin (2002). Most recently, Constantinescu et. al. (2016) notes turbulence in global trade in 2015 and suggests that China’s rebalancing and transition to a new growth path are already contributing to trade volatility and would continue to shape developments in the foreseeable future. Other studies have also noted the role of China in shaping global trade developments, especially in emerging markets through input demand and other channels. Developments in China may, of course, not be the only factor shaping global trade. From a conjunctural standpoint, there is now an extensive literature which examines factors that may have contributed to the oversized decline in global trade in 2009. These studies, which include Altomonte et. al. (2012), Baldwin (2009), Bems et. al. (2010, 2011, 2013), Buono (2013), Levchenko et. al. (2010), Bussiere et. al. (2013), Ferrantino and Taglioni (2014), and Abiad et. al. (2014), attribute the collapse variously to a sharp contraction in aggregate demand (income), concentrated on trade-intensive components, aggravated by inventory effects and trade financing constraints. Constantinescu et. al. (2015) and Bems et. al. (2010, 2011, 2013), for example, examine the role of the commodity composition of trade and vertical specialization while Freund (2009) and several studies in Hoekem (2015) point to the role of cyclical factors.5 Other studies in Hoekem (2015) attribute the slowdown to structural factors, including the composition of trade, the end of the integration of central/eastern Europe and China into the global trading system, and limits of vertical integration in global value chains (GVC).6 The policy implications of the role of cyclical versus structural factors in the trade slowdown are, of course, quite different.

At the same time, Irwin (2002) acknowledges—as is the case with many of the studies referred to above—the possible endogeneity of exports consistent with the notion of trade as an “engine of growth”. Indeed, Irwin and Terviö (2002) shows that more trade has led to higher income throughout the 20th century, with the exception of the interwar period. From an estimation standpoint, incorporating this bi-directional relationship requires the use of a simultaneous equation model. For purposes of this paper, the strong presumption from these studies is that the relationship of trade and income is bi-directional and that it has varied over different time spans.

III. Global Trade and Income over 6½ Decades

Consistent with the discussion of Section II, we examine the relationship between global exports and income over 1950-2014 using several empirical specifications. In what follows, we first estimate single equations models and then multivariate equation models of global exports and income. The data for the empirical estimation (Figure 2) is taken from WTO (2015) which shows the exponential increase of global exports over 6½ decades as global income rose tenfold.

Figure 2.
Figure 2.

Global Exports and Income, 1950-2014

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

A. Income as Determinant of Exports

Specification and estimation

We start with a single equation autoregessive distributed lag (ADL) specification. The estimated equation is:

(1)xt=β0+β1yt+β2yt1+β3xt1+ɛxt

where x and y are global exports and income, respectively, and ɛxN[0,σx2] is the error term.7 Both the variables are in logs, real, and indexed at 2005=100, as shown in Figure 1. In this specification, the short-term elasticity of exports to income (θS) is β1 and the long-term elasticity (θL) is β1+β21β3. The results of this benchmark equation for the full sample periods are shown in Table 2. θS and θL are estimated at 2.1 and 1.8, respectively.

Table 2.

Income as Determinant of Exports: ADL model—OLS (no structural breaks)

Dependent Variable: LX_T

Method: Least Squares

Date: 04/08/16 Time: 18:09

Sample (adjusted): 1951 2014

Included observations: 64 after adjustments

HAC standard errors & covariance (Quadratic-Spectral kernel, Andrews bandwidth = 1.0154)

article image
Sources: WTO (2015) and author’s calculations.

Structural breaks

With a changing relationship over the long time span, there were likely structural breaks in the estimated equation. The Bai and Perron (2003) test of unknown number of structural breaks suggests six structural breaks in the estimated equation (1961, 1975, 1981, 1991, 1997 and 2008). The estimated structural breaks are reported in Table 3. The breakpoints identified by the test appear to be associated with well-defined events. The breaks in 1975 and 1981 reflect the two oil price shocks. 1991 can be associated with the breakup of the former Soviet Union and rise in trade due to the integration of the central/eastern European countries. 1997 reflects the effects of the Asian Crisis. Finally, 2008 picks up the GTC.

Table 3.

Income as Determinant of Exports: Structural breaks

article image
Sources: WTO (2015) and author’s calculations.

The implied short and long run elasticities for the full sample and seven subsamples are collected in Figure 3. The changes in the elasticities are revealing, with ebbs and flows over the subsamples. First, the short and long run elasticities do not rise and fall in tandem, with a large range around the values for the full sample. Second, the oil price shocks of the 1970s dampened the impact of income growth on exports, possibly reflecting supply side shocks. Third, the highest estimated elasticities (both short and long run) are in the first half of the 1990s, and declined thereafter suggesting that the Asian crisis may have dampened trade and was not fully offset by the rising share of China in global trade, the long rise in global income during the 2000s and the GVCs. The rise in the short run elasticity during 2008-14 picks up the oversized decline in global exports (and the subsequent larger recovery) in 2009. At the same time, the decline in the long run elasticity could be suggestive of a slowdown since 2010. These estimates corroborate Irwin (2002) where the long-run elasticity of exports was the highest during 1985-2000. In Figure 3, the highest elasticity is during the first half of the 1990s, a period during which the incorporation of central/eastern European and former Soviet Union countries into global trade led to a reclassification of internal trade into international trade.

Figure 3.
Figure 3.

Elasticity of Exports to Income, 1950-2014

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

Stationarity and cointegration

Unit root tests suggest that the order of integration for yt is I(0) while the null hypothesis that xt is a unit root process is not rejected at the 5 percent level of significance (Table 4).8 The null hypothesis that Δxt is a unit root process is not accepted at the 5 percent level of significance implying that xt is an I(1) process. With this, and to ensure all variables in the equation are stationary, we re-specify and estimate (1) as an error correction specification:

Table 4.

Unit Root Tests for Global Exports and Income

article image
*MacKinnon (1996) one-sided p-values.Sources: WTO (2015) and author’s calculations.
(2)Δxt=β0+β1Δyt+ρx(xt1γxyt1)+ɛxt

where (xt-1 — γ yt-1) is the long run cointegrating relationship between x and y, θs= β1 θL= γx and ρx is the speed of adjustment to the long run co-integrating vector. The Johansen test confirms the existence of a co-integrating vector. The estimated equation for (2) in Table 5 implies, as before, short and long run elasticities of 2.1 and 1.8, respectively. In addition, we have an estimate of ρx=-4.7.

Table 5.

Income as Determinant of Exports: ECM—OLS (no structural breaks)

Dependent Variable: DLX_T

Method: Least Squares (Gauss-Newton / Marquardt steps)

Date: 04/01/16 Time: 16:32

Sample (adjusted): 1951 2014

Included observations: 64 after adjustments

Convergence achieved after 7 iterations

HAC standard errors & covariance using outer product of gradients (Quadratic-Spectral kernel, Andrews bandwidth = 0.9851)

DLX_T=C(1)+C(2)*DLY+C(3)*(LX_T(-1)-C(4)*LY(-1))

article image
Sources: WTO (2015) and author’s calculations.

B. Exports as an “Engine of Growth”

In this subsection we estimate a specification, along the lines of (2), to capture the notion that trade is a determinant of income following studies such as Irwin and Terviö (2002).

Specification and estimation

We specify, as in (2), an ECM model for global income with exports and lagged income as explanatory variables:

(3)Δyt=α0+α1Δxt+ρy(yt1γyxt1)+ɛyt

Using the co-integration results from the previous section, we estimate (3). From the estimated equation (Table 6), the estimated short- and long-run elasticities of income to exports (δs1 and δL=γy, respectively) are 0.3 and 0.5, respectively. The estimated coefficient confirms the positive impact of exports on income, both in the short and long run. The estimated speed of adjustment to the long-run co-integrating vector (ρy) is –4.

Table 6.

Exports as “Engine of Growth”: ECM—OLS (no structural breaks)

Dependent Variable: DLY

Method: Least Squares (Gauss-Newton / Marquardt steps)

Date: 04/01/16 Time: 16:33

Sample (adjusted): 1951 2014

Included observations: 64 after adjustments

Convergence achieved after 5 iterations

HAC standard errors & covariance using outer product of gradients

(Quadratic-Spectral kernel, Andrews bandwidth = 3.0538)

DLY=C(1)+C(2)*DLX_T+C(3)*(LY(-1)-C(4)*LX_T(-1))

article image
Sources: WTO (2015) and author’s calculations.

Structural breaks

Once again, with changes in the relationship likely over such a long time span, we use the Bai and Perron (2003) test of an unknown number of structural breaks (using an ADL specification). The test now suggests eight breaks in the estimated equation at 1957, 1963, 1969, 1975, 1983, 1991, 1997 and 2008 (Table 7). While some of these structural breaks mirror the ones estimated above, there are clearly differences in both number and timing. The estimated short and long run elasticities for the full sample and nine subsamples are shown in Figure 4. As in the earlier section, the effect of exports on income varies, both in the short and long run over the various sample periods. In some ways, these effects, by definition, are inverses of the elasticities of income to exports, although the different structural breaks points provide a point of departure.

Table 7.

Exports as “Engine of Growth”: Structural breaks

article image
Sources: WTO (2015) and author’s calculations.
Figure 4.
Figure 4.

Elasticity of Income to Exports, 1950-2014

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

C. Global Exports and Income: Feedback Loops and Common Shocks

Given that income and exports are related and may indeed be co-determined by each other and other factors, there is a clear possibility of feedback loops. To account for this bi-directional relationship and the possibility of common shocks, we turn to a VAR specification. We first postulate a bivariate structural VAR (SVAR) in global exports and income. We interpret the structural innovations to exports as “globalization” shocks and those to income as “technology” shocks. The “globalization” shocks may be seen as reflecting a variety of developments in global trade, including liberalization and the process of countries joining the WTO. On the negative side, these shocks may come from protectionism, disruptions to financial sector, or wars and other conflicts. We interpret the “technology” shocks partly in a standard macroeconomic fashion as coming from technological change and the resulting increases in total factor productivity, although there could be other plausible interpretations as well including the large balance sheet effects of the type now associated with the GFC.

We first estimate an unstructured VAR (UVAR) from which we retrieve the SVAR by imposing identifying restriction. We estimate the SVAR under the long-run restriction that the cumulative impact of “globalization” shocks on income is zero.9 In the short-run, therefore, both shocks affect global exports and income. We use the SVAR specification to retrieve the impulse response functions and the variance decompositions to examine quantify the impact of shocks on the endogenous variables. To extract short- and long-run elasticities and speeds of adjustment, we estimate a vector error correction (VEC) representation of the SVAR.

A Structural VAR (SVAR)

Specification and identification

We propose a bivariate SVAR in first difference in exports and income as follows:

(4)Δxt=βx0+Σj=1nβxxjΔxtj+Σj=0nβxyjΔytj+bxxωxtΔyt=βy0+Σj=0nβyxjΔxtj+Σj=1nβyyjΔytj+byyωyt

where j are lagged values, n is the lag length of the VAR and the vector of error terms [ωxt, ωyt] has an identity variance-covariance matrix Ω, with the interpretation that ωxt and ωyt are exogenous, structural “globalization” and “technology” shocks, respectively. Both shocks are assumed to follow an N(0,1) distribution and are serially and cross uncorrelated. The lag length n for the VAR is chosen using information criteria. The Schwartz criterion suggests a lag length of two for the UVAR.

In matrix form, (4) can be written as follows:

(5)AΔzt=A0+Σj=1nAjΔztj+Bωt

where Δzt = [Δxt, Δyt] is the vector of endogenous variables: The UVAR representation of (5) is:

(6)Δzt=B0+Σj=1nBjΔztj+et

where B0 = A−1A0 Bj = A-1Aj and Aet= t. The identifying restriction that the cumulative impact of “globalization” shocks on income is zero translates into the restriction that the cumulative long-run impact of ωx on Δy is zero. The parameter estimates of the UVAR and estimates of the SVAR restrictions are shown in Table 8 and Table 9, respectively.

Table 8.

Global Exports and Income: VAR Specification

Vector Autoregression Estimates

Date: 04/05/16 Time: 17:09

Sample (adjusted): 1953 2014

Included observations: 62 after adjustments

Standard errors in ( ) & t-statistics in []

article image
Sources: WTO (2015) and author’s calculations.
Table 9.

Global Exports and Income: SVAR Specification

Structural VAR Estimates

Date: 04/07/16 Time: 13:03

Sample (adjusted): 1953 2014

Included observations: 62 after adjustments

Estimation method: method of scoring (analytic derivatives)

Convergence achieved after 6 iterations

Structural VAR is just-identified

article image
Sources: WTO (2015) and author’s calculations.

Innovation Accounting: Variance Decomposition and Impulse Response Functions

As regards variance decompositions (Figure 5), globalization and technology shocks account for 60 and 40 percent, respectively, of the variation in global exports while technology shocks account for the bulk of the variation in global income (90 percent over all horizons). The impulse response functions (IRF) and cumulative IRFs for the SVAR are shown in Figure 6. By construction, the long-run cumulative impact of globalization shocks on income is zero. However, both in the short and long run, the globalization and technology shocks have a measurable impact, especially on exports. For example, a one standard deviation (positive) globalization shock raises global exports by a peak effect of 3½ percent contemporaneously; the cumulative impact declines to around 2¾ percent in year 5. At the same time, globalization shocks do have an impact on income in the short run as well (½ percent), but the impact peters out over subsequent years. Technology shocks have a noticeable impact on global income, with 1¾ percent of the total long run impact of a little under 4 percent coming in the first year. These results complement and enrich the understanding of the relationship between global exports and income that come from elasticity estimates and speed of adjustment to the long run relationship to which we now turn.

Figure 5.
Figure 5.

SVAR Innovation Accounting: Variance Decomposition

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

Figure 6.
Figure 6.

SVAR Innovation Accounting: Impulse Response Functions

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

Source: Authors' computations.Note: 66 percent confidence interval (using Monte Carlo bootstrap procedure).

Vector Error Correction Representation: Elasticities and Speed of Adjustment

The Vector Error Correction Model (VECM) representation of (6) can be written as:

(4)Δxt=γ1Δxt1+γ2Δyt1+ρx(xt1γyt1)+ωxtΔyt=λ1Δyt1+λ2Δxt1+ρy(xt1γyt1)+ωyt

With this alternative specifications, the cointegrating vector (xt-1 — γyt-1) provides the long-run elasticity of exports to income γ and of income to exports (1/γ); the speeds of adjustment of exports and income to long run equilibrium are ρx and ρy, respectively.

The (contemporaneous) short-run elasticities are provided by the IRFs. The IRFs provide, in addition, the lagged impact of shocks on the endogenous variables at other horizons as well. The estimated VECM is shown in Table 10. The estimated long-run elasticity γ is 2.4; ρx and ρy are estimated at 0.03 and 0.02, respectively. While the long-run elasticity is of an order of magnitude similar to the univariate ECMs, the speed of adjustment is significantly smaller suggesting that, following shocks, there could well be extended periods of adjustment to the long run equilibrium of the sort that the global economy experienced after the oil price shock, the break-up of the Soviet Union, the Asian Crisis and most recently the GFC.

Table 10.

Global Exports and Income: VECM Specification

article image
Sources: WTO (2015) and author’s calculations.

IV. The “Great Trade Collapse” and “Back to the Past”?

Using the SVAR to forecast exports and income for various forward looking spans generates an interesting account of expectations that may have been built up during the 1990s and 2000s, until the GFC in 2009. Using the model to generate dynamic stochastic forecasts of exports and income in 1990 (for 1991-2014), 2000 (for 2001-14) and in 2008 (for 2009-14) we arrive at Figure 7.

Figure 7.
Figure 7.

The “Great Trade Collapse” or “Back to the Past”?

Citation: IMF Working Papers 2016, 139; 10.5089/9781475526592.001.A001

Sources: WTO (2015) and author’s calculations.

Using the 1990 forecast, the (actual) growth rate of exports surpassed the expected rate as former central/eastern Europe and Soviet Union countries joined the global trading system. This may, however, have been partly a statistical artifact which translated internal trade into international trade. After the temporary, shallow setback of the 2000-01 dotcom bust, China and the rise of the GVCs (which again may have changed some internal trade into international trade or rerouted existing international trade) contributed to the higher than expected growth of global exports, even by the higher standard of the 1990s, although incorporating the higher base of the 1990s meant that the “overperformace” was smaller (middle panel). All of the remarkable growth of exports appears to have taken place when global income was either underperforming (in the 1990s and in the first half of the 2000s, top panel) or broadly “on track” in the first half of the 2000s (middle panel). Only during the second half of the 2000s do both global exports and income appear to be “overperforming”, a time of easy monetary and financial conditions in advanced economies which culminated in the GFC in 2008Q3. However, even the ostensible “trend” performance of global income in this forecast does not necessarily imply the nonexistence of imbalances across regions and countries. Yet another factor that may have contributed to the slowdown in global trade more recently may be the on-shoring of production processes in China as it has moved up the value chain, displacing imports from relatively more advanced economies, in what seems like an “internalization” process (compared to the “externalization”) at the time of the breakup of the former Soviet Union (IMF, 2016).

The growth of global exports and income appears to be below the expected mean when the 1990-2008 experience, especially the “overperformance” of the second half of the 2000s, is factored into forecast (bottom panel). The step decline in 2009 is, of course, rather evident, and there appears to be a decline in the growth rates as well. Whether this decline in the growth rate is temporary, long-lasting or permanent is be hard to tell, especially given that the VECM suggests that the speed of adjustment to shocks is rather small. The decline in the growth rate may also be reflecting that, with the special factors of the 1990s and 2000s no longer in play, global income and trade may not necessarily be depressed from a long term standpoint, but only back to an old “normal” of the pre-1990s. Furthermore, even if the levels and growth rate are below expected means, they have recovered within a reasonable confidence interval around the (higher) expected mean.

Such an interpretation of the GTC and GFC raises broader questions of what might be needed to restore the “mojo” of both global exports and income, a subject of intense discussion at global policy making institutions and at the individual country level (WEO, 2016). A part of the discussion relates to whether the levers of macroeconomic policy can restore global aggregate demand in the short run into a self-sustaining cycle or deeper structural reform are needed which will have both supply and demand side effects. In the absence of adequate policy action or the next set of positive shocks, would the global economy have to wait longer to return to a higher growth path? In this context, it would be of interest to investigate whether the Trans-Pacific Partnership (TPP) provides the opportunity to kick-start global export growth, with salutary effects on global income.

Several of the hypotheses noted above and in the literature can be further examined in expanded versions of the model. For one, the paper does not address the issue of exogeneity. It is presumed that there is a bi-directional relationship between global exports and income. Granger causality tests and other tests could be employed to explore this issue. Global growth can be broken down into the contribution of the main blocs to see the relative contribution of advanced and emerging economies (including China) to global exports and growth. Similarly, global exports can be decomposed into the main categories—agricultural, fuels and mining products, and manufacturing to assess relative sectoral contributions, and relatedly the impact of GVCs. From an empirical standpoint, the parameter estimates for the forecasting model estimated in Section III are for the full sample 1950-2014. A part of this could be addressed by re-estimating the SVAR over shorter sample periods, although this may strain the data and could be partly addressed by a shift to Bayesian estimation. It would also be instructive to construct a time-varying SVAR to generate the forecasts, although the main contours of argument would likely remain unchanged.

References

  • Abiad, A., P. Mishra and P. Topalova (2014), How Does Trade Evolve in the Aftermath of Financial Crises? IMF Economic Review, 62: 213247.

    • Search Google Scholar
    • Export Citation
  • Altomonte, C., F. Di Mauro, G. Ottaviano, A. Rungi, and V. Vicard (2012), Global Value Chains During the Great Trade Collapse: A Bullwhip Effect? ECB Working Paper 1412.

    • Search Google Scholar
    • Export Citation
  • Baldwin, R. (ed.), The Great Trade Collapse: Causes, Consequences and Prospects, CEPR for VoxEU.org, London, 2009.

  • Baldwin, R. (2011), Trade and Industrialization after Globalization’s 2nd unbundling: How Building and Joining a Supply Chain Are Different and Why It Matters, NBER Working Paper No. 17716.

    • Search Google Scholar
    • Export Citation
  • Bems, R., R. C. Johnson, and K. Yi (2010), Demand Spillovers and the Collapse of Trade in the Global Recession, IMF Economic Review 58(2): 295326.

    • Search Google Scholar
    • Export Citation
  • Bems, R., R. C. Johnson, and K. Yi (2011), Vertical Linkages and the Collapse of Global Trade, American Economic Review Papers and Proceedings, 101(3): 308312.

    • Search Google Scholar
    • Export Citation
  • Bems, R., R. C. Johnson, and K. Yi (2013), The Great Trade Collapse, Annual Review of Economics, Annual Reviews 5(1): 375400.

  • Bown, C. and M. Crowley (2013), Import Protection, Business Cycles, and Exchange Rates: Evidence from the Great Recession, Journal of International Economics, 90, 1: 5064. 29

    • Search Google Scholar
    • Export Citation
  • Boz, E., M. Bussiere and C. Marsilli (2014), Recent Slowdown in Global Trade: Cyclical or Structural, VoxEU.org.

  • Buono, I. and F. Vergara-Caffarelli (2013), Trade Elasticity and Vertical Specialization, Bank of Italy, Working Paper 924, July 2013.

  • Bussiere, M., G. Callegeri, F. Ghironi, G. Sestieri and N. Yamano (2013), Estimating Trade Elasticities: Demand Composition and the Trade Collapse of 2008-2009, American Economic Journal: Macroeconomics 5(3): 118151.

    • Search Google Scholar
    • Export Citation
  • Constantinescu C., A. Mattoo and M. Ruta (2014), Slow Trade, Finance and Development, International Monetary Fund, Washington D.C., December 2014.

    • Search Google Scholar
    • Export Citation
  • Constantinescu C., A. Mattoo and M. Ruta (2015), The Global Trade Slowdown: Cyclical or Structural, IMF Working Paper 15/6, International Monetary Fund, March 2015.

    • Search Google Scholar
    • Export Citation
  • Constantinescu C., A. Mattoo and M. Ruta (2016), Trade Turbulenece, Finance and Development, International Monetary Fund, March 2016.

  • Davies, G. (2013), Why World Trade Growth Has Lost Its Mojo, Financial Times Blog, 29 September 2013.

  • Donnan, S. (2014), OECD Warns on Global Trade Slowdown, Financial Times, 27 May 2014.

  • Duval, R., K. Cheng, K. H. Oh, R. Saraf and D. Seneviratne (2014), Trade Integration and Business Cycle Synchronization: A Reappraisal with Focus on Asia, IMF Working Paper, 14/52.

    • Search Google Scholar
    • Export Citation
  • Escaith, H., N. Lindenberg, and S. Miroudot (2010), International Supply Chains and Trade Elasticity in Times of Global CrisisWTO Staff Working Paper ERSD- 2010-08.

    • Search Google Scholar
    • Export Citation
  • Evenett, S. J. (2014), The Global Trade Disorder, The 16th GTA Report, CEPR, November.

  • Ferrantino, M. and D. Taglioni (2014), Global Value Chains in the Current Trade Slowdown, World Bank Economic Premise 137. 30

  • Freund, C. (2009), The Trade Response to Global Downturns: Historical Evidence, Policy Research Working Paper Series 5015, World Bank, Washington, DC.

    • Search Google Scholar
    • Export Citation
  • Grossman, G. M. and E. Rossi-Hansberg (2008), Trading Tasks: A Simple Theory of Offshoring, American Economic Review, 98:5, 19781997

  • Hoekman, B. (2015), The Global Trade Slowdown: A New Normal? A VoXEU Book, CEPR Press.

  • IMF, World Economic Outlook, April 2016.

  • IMF, Asia and Pacific Regional Economic Outlook, April 2016.

  • Kee, H. L. and H. Tang (2014), “Domestic Value Added in Exports: Theory and Firm Evidence from China,mimeo, World Bank.

  • Krugman, P. (2013), “Should Slowing Trade Growth Worry Us?New York Times Blog, 30 September 2013.

  • Krugman, P. (2014), “Flattening Flattens”, New York Times Blog, 3 November 2014.

  • Irwin, D. (2002), “Long-Run Trends in World Trade and Income,World Trade Review, 1:1, 89100.

  • Levchenko, A., L. Lewis, and L. Tesar (2010), “The Collapse of International Trade during the 2008–09 Crisis: In Search of the Smoking GunIMF Economic Review 58(2): 214253.

    • Search Google Scholar
    • Export Citation
  • WTO (2013), “Report on G20 Trade Measures,World Trade Organization, December 2013.

  • WTO (2014a), Trade and Development: Recent Trends and the Role of the WTO, World Trade Report 2014, WTO, Geneva.

  • WTO (2015), International Trade Statistics, WTO, Geneva.

1

The paper uses real GDP to represent real income, as in Irwin (2002).

2

The growth of global export and income during 2001-08 was 5.1 percent and 2.8 percent, respectively.

3

The paper employs an AR(1) single equation model in log levels to compute the elasticities.

4

The share of manufacturing in (real) total global exports surpassed 50 percent in the mid-1970s.

5

Studies in Hoekem (2015) which emphasize the role of cyclical factors in the global trade slowdown include Boz et. al., Bussiere & Marsilli, Ollivaud & Schwellnus, Gangnes et. al., and Veenendall et. al.

6

Constantinescu et. al., Gaulier et. al., Escaith & Miroudot, Crozet et. al., Bark, Ito & Wakasugi, Thorbecke, Chinn, and Pei et. al.

7

In the estimated equations in the tables, LX_T is the log of real exports and DLX_T is the first difference of logs (or the growth rates of exports). Similarly, LY and DLY are log and growth rate of income, respectively.

8

We use the Phillips-Perron unit root tests.

9

We could alternatively achieve identification of the SVAR under the restriction that the “technology’ shocks do not have a long run impact on global exports, which seems rather implausible. We need only one restriction for the SVAR to be just-identified.

6½ Decades of Global Trade and Income: “New Normal” or “Back to Normal” after GTC and GFC?
Author: Sanjay Kalra