In 2009 it became clear that the Greek economy was facing very serious challenges arising from the country’s ‘twin deficits’ – both the budget of the Greek government and the current account of the country were in deficit by almost 15% of GDP. The two deficits required large amounts of financing that was not forthcoming in the aftermath of the Global Crisis of 2007-09, leading to the eruption of the Greek crisis. Since 2009, however, policymakers and the public have focused almost exclusively on the budget deficit. This column argues that trade adjustment (or lack thereof) constitutes a very important element of the Greek crisis.
To study Greece’s trade performance, we compare developments in Greece with those in Ireland, Portugal and Spain. The comparison is informative because of these countries’ shared predicament – following the introduction of the euro, all four peripheral countries of the Eurozone experienced large net capital inflows and ever expanding current account deficits which were abruptly reversed in the aftermath of the Global Crisis (Figure 1). Furthermore, the adjustment to lower capital inflows was so disruptive that all four countries experienced severe economic crises leading to bailouts regardless of their pre-Crisis fiscal position (see Galenianos 2015 for a detailed account of the Crisis that emphasises the importance of the balance-of-payments dimension).1
Figure 1. Current account balance (% of GDP)
Rebalancing trade to an environment with low net capital inflows is the common challenge faced by all peripheral countries of the Eurozone, one which requires a combination of higher exports and lower imports. While trade balances have indeed improved dramatically and fairly uniformly across the periphery, the pattern of improvement varies considerably across countries. Portugal, Spain and especially Ireland have managed to significantly expand their exports, but Greek exports actually declined between 2007 and 2012, as shown in Figure 2 (the most recent year in our analysis is 2012 for reasons of data availability). The weak performance of exports suggests that the Greek economy has not adjusted to the new international environment.
Figure 2. Change in trade balance between 2007 and 2012 (% of 2007 GDP)
This observation raises important questions. Is there something fundamental about the Greek economy that inhibits exports? If so, then the decline in capital inflows will necessarily lead to a large decline in GDP. If not, then why haven’t exports adjusted? And what is the economic cost of this delayed adjustment?
In Arkolakis et al. (2014), we study the sources of Greece’s export performance. We compare the actual performance of all peripheral countries with a benchmark of the expected response of trade to the reduction in net capital inflows. To derive a plausible benchmark, we turn to trade theory.
We calibrate an equilibrium model of trade, based on the work of Eaton and Kortum (2002) and Dekle et al. (2007). The calibration yields measures of the productivity and trade costs for each country which we use for our counterfactual experiments. We use data from 2007 on GDP, the current account and trade deficits and bilateral trade flows for 32 countries, which account for 80% of global GDP and 98% of EU GDP in 2007, as well as a residual ‘rest of the world’ country.2
We perform two counterfactual experiments. In the first counterfactual, we introduce the current account and trade balances from 2012 to derive the predicted response of variables of interest (exports, imports, wages, prices) to the decline in net capital inflows. This calculation delivers our benchmark for ‘frictionless adjustment’, since prices and wages are fully flexible and workers and productive assets are seamlessly reallocated.
According to this benchmark, in response to the observed decline in external financing Greek exports should increase by approximately 25% rather the mild decline that was actually observed (Figure 3). Therefore exports should account for a large part of the adjustment even after taking into consideration Greece’s relatively low export base in the pre-Crisis years. Furthermore, the export performance of the other countries was much closer to the model’s prediction (even if Portugal and Spain also ‘underperformed’ with respect to the frictionless benchmark). This suggest that Greek exports should have been expected to do better and that the reasons for the underperformance are to be found in Greek-specific rather than global factors.
Figure 3 Percentage change between 2007 and 2012: Data vs. frictionless model
Source: Eurostat, EIU, authors’ calculations
We further explore the reasons for Greece’s export underperformance by examining the adjustment of prices in labour and product markets. Figure 4 plots the predicted wage change from the model and the actual wage change with respect to the Eurozone average between 2007 and 2012.3 The model predicts that wages fall in all four countries, which is consistent with the idea that they need to become more competitive in response to a decline in net capital inflows. Remarkably, Greece is the only country that comes close to achieving the level of wage reductions predicted by the model. Therefore, it does not appear that labour markets are an impediment to adjustment.
Figure 4 Percentage change between 2007 and 2012: Data vs. model
Source: Eurostat, EIU, authors’ calculations.
Prices, however, followed a very different pattern. Prices in Greece were predicted to have the largest decline (by 8%) while in the data they actually increased (of 2%) in comparison to the Eurozone average, a deviation of almost 10%.4 The other peripheral countries exhibited much smaller differences between predicted and actual declines in relative prices. Therefore, price rigidity, most likely due to limited product market competition, is a leading candidate for why Greek exports underperformed.
In our second counterfactual, we measure the GDP cost of the delayed adjustment. We recalibrate the trade costs in our model to match the exact changes in exports and imports separately, which yields our ‘frictional’ benchmark. In the case of Greece, for instance, this calibration yields an increase in the cost of exporting to rationalize the decline in exports. Figure 5 presents the prediction of this frictional model for the change in GDP in response to the decline in net capital inflows and the observed change in exports and imports separately. Incorporating the slow adjustment of exports fully explains the output drops in Portugal and Spain and captures about one third of the magnitude of Greece’s recession.
Figure 5 Percentage change in real GDP between 2007 and 2012: Data vs. frictional model
Source: Eurostat, EIU, authors’ calculations
Greece’s trade deficit declined by 10% of GDP between 2007 and 2012, removing one of the great economic imbalances of the pre-crisis years. However, this reduction was achieved exclusively through import compression while exports fell over that period, thereby worsening the economic crisis. Our main findings are that:
- Given past performance, Greece’s exports should have increased by 25% rather than falling by 5% between 2007 and 2012;
- Labour markets have adjusted to the new economic environment;
- Product markets have not adjusted, hindering the recovery of competitiveness; and
- Export underperformance is responsible for a third of the decline in Greek GDP since 2007.
Arkolakis C, A Doxiadis and M Galenianos (forthcoming), “The challenge of trade adjustment in Greece”, forthcoming in Reforming the Greek economy, edited by C. Meghir, C. Pissarides, D. Vayanos, N. Vettas, MIT Press.
Dekle, Robert, Jonathan Eaton and Samuel Kortum (2007), “Unbalanced Trade,” American Economic Review, May, 97 (2), pp. 351-355.
Eaton, J and S Kortum (2002), “Technology, Geography and Trade,” Econometrica, September, 70, pp. 1741-1780.
Galenianos, M (2015), “The Greek crisis: origins and implications”, The crisis observatory, research paper 16/2015, January.
1 Greece, Ireland and Portugal received help from the EU and the IMF to finance their governments’ budgets and, in Greece’s and Ireland’s case, to recapitalise their banks. Spain received funds from the EU to recapitalize its banks.
2 The World Bank provides detailed data on the trade of non-oil goods while the United Nations and the Organization of Economic Cooperation and Development provide data on the trade in services. Data on GDP and the current account comes from the Economist Intelligence Unit.
3 We use Eurostat data on the average hourly nominal wage for the whole economy, except for public administration where wages are set under different constraints than the ones under consideration in this study. Nominal wages are deflated for every country using the Eurozone GDP deflator. We do not use each country’s individual GDP deflator because we want to examine the evolution of a country’s wages relative to other Eurozone countries.
4 The increase of VAT taxes is partly responsible for the observed price increases in Greece: VAT taxes are responsible for a 3.2% price increase in Eurozone counties but a 6.6% increase in Greek prices, as calculated by Eurostat. Therefore, had VAT increased in Greece at the same rate as in the rest of Europe, Greece would have experienced a 1.8% drop in relative prices instead of an increase of 1.8% – still much below the 8.6% drop predicted by the model.