Unit labor costs and the current account in the Eurozone

(1)

1 Master Thesis

Universiteit of Amsterdam

Faculty of Economics and Business Study: Economics

Track: International Economics and Globalization Author: Vincent de Vries

E-mail: vdv1990@gmail.com Student number: 6103456

Supervisor: Dr. D.J.M. Veestraeten Second reader: Drs. N.J. Leefmans Date: August 19, 2014

Unit Labor Costs and the Current

Account in the Eurozone

(2)

2

List of Abbreviations

CA Current Account

CAGDP Current Account divided by GDP CAGDPC First difference of CAGDP

EUCAGDP Intra-EU Current Account divided by GDP EUCAGDPC First difference of EUCAGDP

EUEXGDP Intra-EU Exports divided by GDP EUEXGDPC First difference of EUEXGDP EUIMGDP Intra-EU Imports divided by GDP EUIMGDPC First difference of EUIMGDP EXGDP Exports divided by GDP EXGDPC First difference of EXGDP GDP Gross Domestic Product

GDPG Gross Domestic Product growth, relative to the previous year GIIPS Greece, Ireland, Italy, Portugal, and Spain

IMGDP Imports divided by GDP IMGDPC First difference of IMGDP

L1 First Lag

L2 Second Lag

PP Percentage Point

REER Relative Effective Exchange Rate with the 42 main trading partners of the Eurozone, indexed to 100 (2000=100), the higher the level of REER the stronger the Euro REERC First difference of REER, an increase in REERC is an appreciation of the Euro RNULC Relative Nominal Unit Labor Cost, Nominal Unit Labor Cost per Country divided by

the Nominal Unit Labor Cost of the Eurozone (EA12), indexed to 100 (2000=100) RNULCC First difference of RNULC

(4)

4

1. Introduction

Before the financial crisis of 2007, the focus of research on current account imbalances lay mainly on the current account surpluses of Japan, China, and the oil producing Middle Eastern countries on the one hand and the current account deficits of the United States on the other. The Eurozone received little attention in this debate, because the overall current account balance of the Eurozone was close to zero. This changed after the financial crisis of 2007 and the subsequent European sovereign debt crisis that started in 2010. The countries that had difficulties coping with their increasing credit spreads, the GIIPS-countries (Greece, Ireland, Italy, Portugal, and Spain), all had current account deficits. The current account deficits of these countries in 2008 ranged from deficit of 2.4 percent of GDP for Italy up to a deficit of 14.6 percent of GDP for Greece. On the other hand Germany and the Netherlands had current account surpluses of 7.5 and 9.4 percent of GDP respectively.1

It would be wrong however to simply state that the current account deficits were one of the reasons that caused the European sovereign debt crisis. This is because current account deficits do not have to be dangerous per se. A current account deficit implies that capital is flowing into the country2, which is to be expected for countries with a high economic growth potential. It is important however that this capital is used for the right means in order to facilitate the growth of the economy. The lower-income countries were expected to have current account deficits in order for them to increase their productive capabilities. The higher-income countries were expected to have current account surpluses, since they were lending the capital to the lower-income countries. This was because investing in the poorer Eurozone countries would yield a higher expected return on capital than investing in the richer countries would. It was imperative however that productivity levels in the poorer countries would grow in a relatively fast pace, since only then the high economic growth could actually take place.

This did not happen however. Blanchard (2007), for instance, found that economic and

productivity growth in one of the deficit countries, Portugal, stagnated after 2001. The inflow of capital into Portugal in the preceding years had caused an unsustainable economic boom with low

unemployment, which lead to scarcity on the labor market and therefore led to a growth in nominal wage, which was substantially higher than the growth in productivity. This decreased the

competitiveness of Portugal, which led to lower economic growth and higher unemployment. On top of this, the loss of competitiveness put downward pressure on the current account. Portugal was not the

1_{Source: OECD.} 2

(5)

5 only country dealing with this problem. Other GIIPS-countries were dealing with similar issues. This paper focuses on the relationship of the current account and the loss of competitiveness as described in the above by Blanchard (2007).

The relationship investigated in this paper will be between unit labor costs and the current account balances of Eurozone countries. Unit labor costs will be explained more in depth later in the paper, but it can be seen as the ratio between wage costs and the level of productivity. If unit labor costs increase in relative terms compared to another country, one would expect that this would reduce the competitiveness of that country, especially when the exchange rates are irrevocably fixed. An increase of relative unit labor costs is therefore similar to a real exchange rate appreciation and is likely to have a negative effect on the current account balance.

The correlation in Figure 1 shows a negative relationship between the growth of unit labor costs and the change in the current account balance.3 From 1999 to the start of the 2007-financial crisis all GIIPS-countries had relative high unit labor cost growth, while all of their current account balances deteriorated. In Germany and Austria the opposite happened. Relative low unit labor costs growth led to an improvement of the current account balances of those countries.

Many economists advised the GIIPS- countries to improve their competitiveness in order to overcome some of the problems they were and are facing. This improvement in competitiveness should come from a decrease in the unit labor costs relative to other EMU-countries (see for instance Blanchard, 2007, De Grauwe, 2012, or Issing, 2011). This did actually happen from 2007 onward and did have the

3

The correlation coefficient is -0.67. AT BE FI FR DE GR IE IT LU NL PT ES -15 -10 -5 0 5 10 -5 5 15 25 35 45 55 Cu rr e n t A cc o u n t/ GDP C h an ge ( 1999 -2007)

Increase in Unit Labor Cost (1999-2007) in percentage

Fig. 1: Relationship ULC and Current Account (1999-2007)

(6)

6 expected outcome. Figure 2 again shows a negative relationship between unit labor costs and the

current account.4 This time however the GIIPS countries had relatively low unit labor cost growth and a current account balance improvement.

It seems clear from Figure 1 and 2 that there is correlation between the current account and unit labor costs. The question remains however whether there is a causal relationship between the two. This leads to the research question of this paper: Are current account changes in the Eurozone caused by changes in unit labor costs? The main focus of this paper will lay on the period after the introduction of the Euro in 1999, since this was also the moment that the first eleven Eurozone countries irrevocably fixed their exchange rates. This made nominal appreciations and depreciations impossible between Eurozone countries.

The paper will first explain the concept on unit labor costs more elaborately and have a closer look at the development of unit labor costs within the Eurozone. Thereafter, the current account will be investigated, starting with a study of the concept of the current account and followed by a study of the current account balances of the Eurozone countries. Chapter 3 will examine the relationship between unit labor costs and the current account and give arguments both for and against a causal relationship between the two. The theory will then be tested by using econometric analysis in order to investigate the link between unit labor costs and the current account. The last chapter will give a short summary of the findings of this paper.

4

The correlation coefficient is -0.86.

AT BE FI FR DE GR IE IT LU NL PT ES -10 -5 0 5 10 15 -15 -10 -5 0 5 10 15 20 25 30 35 40 Cu rr e n t A cc o u n t/ GDP c h an ge ( 2007 -2012)

Increase in Unit Labor Cost (2007-20012) in percentages

Fig. 2: Relationship ULC and Current Account (2007-2012)

(7)

7

2. Reviewing the definitions and developments of Unit Labor Costs and

the Current Account

Before we can look at the causal relationship between unit labor costs and the current account, we first have to have a closer look at these terms separately. This chapter will first focus on the definition of unit labor costs and then will show the development of unit labor costs in the Eurozone. This chapter will thereafter investigate the current account and trade patterns within the Eurozone.

2.1 Defining Unit Labor Costs

Nominal unit labor costs are the indicator used when looking at the relationship between wage costs per worker, or labor compensation, and the average level of labor productivity. Nominal unit labor cost is used here because Eurozone countries have the same currency. Using real unit labor costs would require using a price deflator and this will distort outcomes. Price deflators are typically constructed per country and this will yield distortions in comparing unit labor costs across countries with the same currency. It is clear that labor costs and the level of productivity are important on a micro-economic level to firms, because firms would like to produce their goods at a minimum cost. For firms these costs will decline if, ceteris paribus, the average level of productivity rises or if labor costs decline. This will lower the marginal costs of producing a certain good and thus will improve competitiveness of the firm. Marginal costs will also decline if the productivity growth is larger than the growth in wage costs. This can all be easily seen when looking at Equation 1.5

(Eq. 1)

When looking at Equation 1 it might be easier to think of a unit in terms of a product, such as for instance pens. Then the unit labor cost becomes average labor cost of producing a single pen. Labor costs are not just the nominal wage the employer pays his worker, but also contains other labor-related costs such as for instance social security payments, the contribution of employers to corporate pension schemes, or severance pay. This implies that governments play an important role in influencing labor costs, because in general the government is involved in deciding on the costs that make the difference between direct compensation for workers and total labor costs.

5_{This is a more micro-economic transformation compared to the definition used in Equation 2 (see for instance} Chen et al (2009)).

(8)

8 Average labor productivity is calculated by looking at total output of the firm, the amount of pens it produces, and dividing it by the amount of workers. This then tells you how many pens are produced per worker on average. The ratio between the total wage costs and the average labor

productivity will then yield unit labor costs, or how much labor cost is required to produce a single pen. This amount is in nominal terms, since the numerator is measured in nominal terms, while the

denominator is measured in quantities.

On the level of an individual firm it is clear that lower unit labor costs will imply a better competitive position, since the firm is more competitive due to lower costs. It is more difficult however when looking at unit labor costs on a macro-economic level. An economy typically produces more than just one single good. This makes the intuition behind unit labor costs harder to grasp. First of all, because every firm in a country within the same industry will not have the same unit labor costs. This is because firms are not identical and do not compete in a perfectly competitive environment. If we however still want to say something about unit labor costs on a national level we have to aggregate unit labor costs. The numerator remains quite similar to Equation 1, all wage costs are aggregated and then divided by the number of workers. This now however happens on a national level , while before it was only on a firm-level.

The denominator however becomes more difficult, since it is clear that a sector or the total economy will produce more than one single good. This makes it more abstract, since now we cannot use as single good anymore, but still have to use define these ‘units’. Eurostat, for instance, shifts from ‘pens to units’, by using Gross Domestic Product as the amount produced6:

(Eq. 2)

Here the alc is the average labor costs, alp is the average labor productivity, and L is the number of workers in the whole economy. The ratio between the total labor costs and real GDP will give the unit labor costs in a sector or in a nation. This, as before, is given in nominal terms, since the numerator is nominal, while the denominator is given in real terms. It is interesting now to compare unit labor costs, or the growth of unit labor costs, with other countries, since competitiveness is generally measured relative to others. The more competitive a country is relative to other countries, the more likely it is to have more demand for its products. This can then lead to an improvement of the current account

6_{Eurostat uses Nominal ULC= (total compensation of employees in national currency / total employees in persons)/} (GDP in market prices in national currency / total employees in persons). (see Eurostat, 2014.)

(9)

9 balance of a country. This theoretical relationship between relative unit labor costs growth and the current account will be investigated in more detail in Chapter 3.

2.2 Development of Unit Labor Costs in the Eurozone

After reviewing the definition of unit labor costs, we will now look at how unit labor costs behaved in the Eurozone after the introduction of the Euro. First we will look at the period from the start of the Euro until the year before the start of the global financial crisis, 2007. Thereafter we will examine more recent data from after the start of the financial crisis that erupted in 2007. Why this divide is made will be clear when looking at the trends.

Figure 3 shows the unit labor cost growth from 2000 to 2007 for the individual Eurozone countries and for the countries combined. It is clear that the countries that are often depicted as the peripheral or GIIPS-countries (Greece, Italy, Ireland, Portugal, and Spain) are on the right side of the spectrum. This implies that labor cost increases in the peripheral countries have been higher than their growth in production. On the other hand Germany has actually had negative growth over this time period, which implies that production grew faster than labor costs.

-5 0 5 10 15 20 25 30 35

Fig. 3: Nominal ULC growth 2000-2007 (in %)

(10)

10 We can also distinguish the two components used to calculate the aggregated unit labor costs, namely total wage costs and GDP. This will give a clearer picture of which of the components is the reason for the growth, this is shown in Figure 4.

Figure 2 shows that Germany had very low wage growth in this time period. De Grauwe (2009) claims that the German D-Mark was overvalued when Germany entered the Eurozone. This

overvaluation had to be corrected and this was done through wage moderation. Other reasons De Grauwe (2009) mentions are labor market reforms and the decline of the power of German labor unions, which gives greater power to the employers in wage negotiations.

When looking at the other end of the spectrum we find that the peripheral countries had increasing unit labor costs for somewhat different reasons. Greece, Spain, and Ireland all had high GDP growth relative to the other countries. They however had the highest wage growth of all Euro area countries. When looking at Portugal and Italy we find that their wage growth has been close to the average. They however accounted for the lowest GDP growth levels of all 12 member countries. This is somewhat unexpected since Portugal and Italy have relatively low per capita incomes. Lower-income countries are generally expected to generate higher growth than relatively richer countries. This catching up would be achieved through a higher growth in GDP (See Blanchard and Giavazzi, 2002). The final country that draws attention is Luxembourg. Luxembourg has the highest per capita income in the European Union, but nevertheless had very high GDP growth between 2000 and 2007. This is most probably due to the relatively large banking sector and small population. This raises the question

0 10 20 30 40 50 60 70 80 90

Fig. 4: Growth in Labor Costs and GDP 2000-2007 (in %)

Total Wage Costs GDP

(11)

11 whether Luxembourg should be included in the sample of countries investigated in the empirical

research in Chapter 4 of this paper. The nature of its economy would simply be too different in comparison to the other Euro area-12 countries.

Figure 5 shows the period after the financial crisis of 2007 and the following European sovereign debt crisis. What is clear is that Ireland, Portugal, Spain, and Greece are now on the other side of the spectrum. It is interesting to see that those four countries were the only countries within the Eurozone-12 group that received a bailout from the European Union. Figure 6 shows better what the underlying cause was of the divergence in unit labor costs during this time of crisis. Economic growth is very low or even negative, which of course is far lower than the growth rates we saw in Figure 4.

This also reveals one possible flaw in the measurement of aggregate unit labor costs. If we look at for instance Greece, we see that wage costs decrease spectacularly. Unit labor costs however still increased, because GDP shrunk relatively even more. If Equation 2 holds this would imply a significant drop in average labor productivity. It is very questionable however that actual labor productivity also fell by the same amount as production did.

The drop in wage costs in Ireland, Portugal, Spain, and Greece can both be attributed to the reforms the governments had to make in order to receive the bailout and to the high unemployment in those countries. The reforms made sure that all other costs employers had to pay, apart from the nominal wage, decreased. This then decreased the overall wage costs. High unemployment made

-10 -5 0 5 10 15 20 25 30 35

Fig 5: ULC Growth 2007-2012 (in %)

(12)

12 workers accept lower wages, because of the large supply of unemployed labor, thus lowering the

average wage level.

2.3 Current Account Balances and Trade in the Eurozone

In order to assess the research question better we need to have a look at the current account, both as a concept as well as the current account balances within the Eurozone. Within the debate on global current account imbalances the Eurozone has received little attention until the start of the

financial crisis of 2007 and the following sovereign debt crisis. This is because the overall current account balance of the Eurozone has been close to zero before the crisis erupted in 2007, which can be seen in Figure 7. The focus lay more on countries such as China, Japan, and the oil-producing Middle Eastern countries with surpluses on the one hand and the United States on the other hand with persistent deficits. -30 -20 -10 0 10 20 30

Fig. 6: Growth in Labor Costs and GDP 2007-2012 (in %)

Wage Costs GDP

(13)

13 The current account balance of the Eurozone as a whole does not reflect the current account imbalances of the separate member countries within the currency union. Before we look at the current accounts of the Eurozone countries we first have to review the definition of the current account however. This paper will use two ways to investigate the level of the current account and its changes over time, through international trade competitiveness and through the national savings identity.

Explaining changes in the current account through changes in competitiveness focuses on the trade aspect of the current account. If a country becomes relatively more competitive it is expected to improve its current account. Classically the main channel used for a change in competitiveness is the exchange rate. This is related to for instance the Marshall Lerner condition, which says that if the sum of absolute values of the export and import elasticity is larger than one, the current account will improve if the currency depreciates. Within the Eurozone the exchange rates are irrevocably fixed however, since the member countries use the same currency and are therefore not able to devaluate their currency vis-à-vis other member countries. This asks for different determinants of competitiveness than the nominal exchange rate and this will be discussed in Chapter 3.

Another way of explaining the current account is through the national savings identity. It focuses more on the flow of capital and therefore is closer related to the balance of payments. The balance of payments makes sure that the sum of the current account, the capital account and the foreign exchange reserve are in balance and equal to zero. The foreign currency reserve will be left out of the discussion here, since this has little importance within the Eurozone. The exchange rate regime of the Euro is free-floating, which implies that the ECB will not actively use open market operations in order to change the

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 C ur re nt A ccount B al ance (% of GDP )

Fig. 7: Current Account Balance of the Eurozone

(14)

14 nominal exchange rate. This results in very little changes in the foreign exchange reserves, which justifies leaving the change in foreign reserves out of the discussion.

The current account focuses mainly on trade flows, while the capital account consists of short- and long-term capital flows. Since the balance of payments has to remain equal to zero and we dropped the foreign currency reserve we end up with current account = - capital account.

This means that if a country has a positive balance on the current account net capital flows will be negative, meaning that there is a net outflow of capital. The reverse also holds, so a country with a current account deficit borrows capital from abroad in order to pay for its net exports. The capital account can be defined as being net national investments. Because of this the current account can be written as Equation 3.

( ) ( ) (Eq. 3)

Here S and I represent private saving and private investment respectively, which makes it net private savings. T and G represent tax income and government spending, which is the fiscal surplus or net public savings. If the sum of net public and private savings is negative we find a current account deficit and vice versa. Blanchard and Giavazzi (2002) used this approach to set up an intertemporal current account model. This model shows that it is to be expected for countries with a relatively lower income to run current account deficits. Lower-income countries are expected to have higher economic growth, because they are expected to catch up with the more productive wealthy countries.

Firms and citizens of the lower-income countries expect to benefit from this higher economic growth and change their expectations accordingly. This means that they will borrow in order to invest or consume more than their income, since they expect their income to increase. This is related to the permanent income hypothesis. So they will first dissave, or in this case borrow from abroad, and will save when their income is more than the consumption or investment, this means the repayment of the borrowed capital. So this aspect would predict a current account deficit for the lower-income countries at the start of the Eurozone because of an expected capital inflow. Over time this deficit would be expected to decrease, because the lower-income countries would be catching up with the richer countries and thus have achieved economic production closer to higher-income countries.

Another aspect related to the above is the relationship foreign investors have with capital flows. The above dealt with domestic citizens and investors borrowing money from abroad in order to optimize their expenditures over time. This mainly has to do with loans and other portfolio investments within the capital account. Apart from portfolio investments foreign direct investments have to be discussed as well

(15)

15 when reviewing the capital account. Foreign direct investments are investments made by foreigners with the intention of influencing management decisions. Foreign investors and firms expect their returns to be higher in the lower income countries, because of the lower capital-to-labor ratio. This is expected to give a higher return on capital, because capital is relatively scarcer than in richer countries. This is also the reason why investors are willing to make portfolio investments in these countries. On top of that the investors expect that with foreign direct investment they will gain part of the added production from the relatively high economic growth in the poorer countries. This again would yield a positive net capital inflow and thus a current account deficit.

When looking at the Eurozone specifically it should also be noted that firms from outside of the European Union would be willing to invest in the lower wage Eurozone countries in order to produce for the single goods market of the Eurozone and the whole European Union. By doing so firms could jump tariffs, while having access to a very large market without trade restrictions.

From the above we can conclude that it was to be expected for the lower income Eurozone countries to have current account deficits. This was also the outcome of the model of Blanchard and Giavazzi (2002). Figure 8 shows that these expectations were in fact right. The three countries with the highest current account deficits were Greece, Portugal, and Spain. On the other hand the rich countries, such as Germany, Luxembourg, and the Netherlands ran current account surpluses. These surpluses and deficits balanced each other and that is why the overall current account balance of the Eurozone

remained close to zero around these years which is shown in Figure 7. -15 -10 -5 0 5 10 15 2000 2001 2002 2003 2004 2005 2006 2007 C ur re nt A ccount B al ance (% of GDP )

Fig. 8: Current Account Balances EZ Countries

Belgium Germany Ireland Greece Spain France Italy Luxembourg Netherlands Austria Portugal Finland Source: Eurostat

(16)

16 From 2007 onwards things started to change, which is captured in Table 1. The GIIPS-countries all improved their current account balance in 2013 in comparison to 2007. They all decreased their imports, but it is interesting to note however that Greece decreased their imports by 13.5 percent, while for instance Spain and Portugal increased their exports. All GIIPS-countries decreased their imports, again with Greece being by far the largest in doing so. The special case of Greece probably has to do with the extra stringent conditions set by the troika, lowering economic activity and thus also trade activity. The better performing countries in the Eurozone remained close to the level of their current account balance of 2007. The overall picture however is that current account balances improved and this can be seen in the increase of the overall Eurozone current account balance in Figure 7.

Table 1: Exports, Imports, and Current Account changes 2007-2013

Countries Export (Change 2007-2013 in %) Import (Change 2007-2013 in %) CA/GDP (Change in PP 2007-2013) CA/GDP 2013 Belgium 9.6% 10.4% -3.5 -1.6 Germany 15.8% 17.9% 0.1 7.5 Ireland 8.5% -8.7% 11.9 6.6 Greece -13.5% -42.9% 15.3 0.7 Spain 14.7% -18.8% 10.8 0.8 France 4.1% 3.9% -0.3 -1.3 Italy -3.0% -13.8% 2.3 1 Luxembourg 3.4% 7.7% -4.9 5.2 Netherlands 14.3% 12.6% 3.7 10.4 Austria 3.8% 1.8% -0.8 2.7 Portugal 14.9% -9.6% 10.6 0.5 Finland -7.5% -1.6% -5.4 -1.1

Source: Calculations by author based data from Eurostat

The final subject of this chapter will investigate trade relationships between the Eurozone countries. The Eurozone countries are assumed to have a high level of trade openness with each other. This is assumed since there is free movement of goods and services and the use of a single currency. It is important however to show that this assumption in fact holds.

Figure 9 shows the export shares of countries to the other Eurozone countries. Finland, Greece, and Ireland are the countries here with the lowest percentage of exports to other Eurozone countries. This is partially because of their geographical location, since their surrounding countries do not have the Euro as their currency. This does not however imply that those countries are not affected by changes within the Eurozone, since most of their trading partners are part of the European Union, and thus have

(17)

17 the same single goods market as the Eurozone countries. Taking this into account for these three

countries it is safe to say that there is a quite substantial integration of markets within the Eurozone. This assumption is needed in order to justify the use of relative nominal unit labor costs in Chapter 4. The large share of exports to the Eurozone shows that the Eurozone countries are important trade partners. It also shows that the markets in the Eurozone are quite integrated. This implies that firms can compete with each other on a Eurozone level without trade barriers. This increases the importance of relative competitiveness for Eurozone countries. Unit labor costs are an important

measure of this competitiveness, especially because competitive nominal exchange rate devaluations are not possible in the Eurozone.

0% 10% 20% 30% 40% 50% 60% 70%

Fig. 9: Exports to members of Eurozone in 2011 (% of all exports)

(18)

18

3. Theoretical Relationship between Unit Labor Costs and the Current

Account

After examining unit labor costs and the current accounts of the Eurozone separately, this chapter will look at the theoretical relationship between the two. The goal of this paper is to investigate whether changes in the current account are caused by changes in unit labor costs. This chapter will discuss arguments for and arguments against this hypothesis that can be found in academic literature. The focus will first lie on arguments in favor of the hypothesis and thereafter on arguments against the hypothesis. It is important to note that the argumentation in this chapter is specifically cast in terms of the Eurozone.

3.1 Arguments supporting a Causal Relationship from Unit Labor Costs to the Current Account

The main rationale in arguing that changes in the current account are caused by changes in unit labor costs comes from an extension of the Ricardian comparative advantage model. The classic

Ricardian model (Ricardo, 1817) relies on unit labor requirements, a. The Ricardian model predicts that a country will produce the good in which it has a comparative advantage, even if it has an absolute

(dis)advantage in all industries. We now extend the unit labor requirements with w, the wage rate per hour. For clarity a 2-country model will be used to explain the model, in which the countries are labelled Home and Foreign and the only production factor is labor. Let w and w* be the hourly wage rate for Home and Foreign, respectively, and a and a* be the unit labor requirement for Home and Foreign. This makes that the cost to produce a single unit will be wa for Home and w*a* for Foreign, or the cost to produce a single good is thus the same as the unit labor cost.

It will be cheaper to produce in Home if wa<w*a*, if we use a theoretical framework with free trade and the absence of transport costs, because the unit labor costs at home are lower than in Foreign. Reversely it is cheaper to produce in Foreign if wa>w*a*. Rewriting this leads to comparatively cheaper production at Home if w/w*<a*/a and comparatively cheaper production in Foreign if w/w*>a*/a. Now assume that the ratio of wage rates, w/w*, is fixed across industries in both countries, but that different industries have different ratios of unit labor requirements, ai*/ai, where subscript i denotes a certain

industry. Production of good i will take place at Home if any industry in which ai*/ai >w/w*, while the

production of goods will take place in Foreign if ai*/ai <w/w*.7 Specializing in certain industries will make

7_{For a more elaborate explanation of comparative advantage, relative wages and specialization patterns, see} Krugman and Obstfeld (2009).

(19)

19 a consumption point outside the production possibly frontier possible for both countries (see Dornbusch et al., 1977). This is the same outcome of the classical Ricardian comparative advantage model and shows that international trade is more efficient than autarky.

If the relative wage or productivity would change, the amount of products the certain country specializes in, and thus exports, will also be affected. If the wage level decreases in Home, ceteris paribus, the model predicts that Home would specialize in more industries, while Foreign specializes in less. Assuming that demand for products remains constant in both countries, this will yield an

improvement of the current account balance of Home. This happens because exports from Home increase and imports decrease, due to the increase of products Home now specializes in. So a wage decrease in Home would lead to a current account balance improvement in Home and a conversely a current account balance deterioration in Foreign. A decrease in relative unit labor costs would thus yield an improvement of the current account balance and an increase of relative unit labor costs would yield a deterioration of the current account balance.

The extension of the Ricardian model described in the above is obviously very simplified and relies heavily on the assumption of perfect competition. It also predicts that countries would fully specialize in certain products. The real world however is far more complex, markets are typically imperfectly competitive, and international intra-industry trade appears.

Golub and Hsieh (2000) find however that this Ricardian model performs fairly well in predicting the sign of net bilateral trade flows between the US and its trade partners, while looking at different sectors of industry. This does not support the complete specialization of the Ricardian model, but does support the thesis that the difference in unit labor costs are likely to determine the sign of the bilateral trade balance. Countries that are relatively more competitive, in unit labor cost terms, are likely to have a positive trade balance vis-à-vis their trade partner. The Eurozone obviously consists of more than two countries. The rationale of the above still holds however, the more competitive the country is the more likely it is to have a positive current account balance.

Economic theory, such as the Marshall-Lerner condition, often predicts a relationship between the current account balance and the real exchange rate. A real depreciation is likely to have a positive effect on the current account. The depreciation will make the domestic product cheaper for other countries and will thus stimulate export. On the other hand foreign products will become more expensive, which will lead to a demand shift from foreign to domestic goods. Within the Eurozone the nominal exchange rate is irrevocably fixed however, so the only way for a real appreciation or

(20)

20 (2007) one way to measure this is through differences in relative unit labor costs. This means that an increase of relative unit labor costs will have the same effect as a real exchange rate appreciation. This will then worsen the current account balance. They also find that the consumer price index is a good measure for relative price changes, thus implying that the trend of unit labor costs is similar to the trend of the consumer price index.

Arghyrou and Chorteas (2008) find that there is a positive relationship between the current account and the real exchange rate. This implies that a real exchange rate appreciation will lead to a worsening of the current account balance. They use the consumer price index to investigate this relationship between the current Eurozone countries from 1970 up until 2005 and thus find evidence that supports the claim that there is a positive relationship between real depreciation and current account improvements among Eurozone-countries.

Belke and Dreger (2011) find similar results. They also find that the real exchange rate has a positive relationship with respect to the current account balance of the Eurozone countries for the period from 1982-2008. This again means that an appreciation of the real exchange rate is likely to lead to a current account balance deterioration. Apart from price competitiveness they also use real per capita income in order to make sure that relationship between catching up of the lower income countries and current account balances is not omitted from the regression. Remember here the framework of Blanchard and Giavazzi (2002), which was described in the previous chapter. Their model predicted that the lower income countries in the Eurozone would have current account deficits, because of the higher economic growth that was to be expected in those countries.

In the study of Belke and Dreger (2011) the significance of this catching up effect diminishes over time, which implies that price competitiveness becomes more important the closer we get to the

present. This only holds however for the analysis of the complete panel. While looking at the high deficit, peripheral, countries separately they find that price competitiveness is the most relevant indicator, while for surplus and low deficit, core, countries the relative income effect is more important. This can imply that for the surplus countries an effect other than price competitiveness is of more importance. The authors believe that the relative income indicator picked up the effect of non-price competitiveness. The logic here is that the relatively richer countries have a more advanced economy that produces more high-end products. High-end products typically depend more on non-price competitive factors, such as knowledge and patents, and less on price competitive factors, such as labor, which unit labor costs of relatively less importance for the richer countries.

(21)

21 This is related to a point made by Chen et al. (2013), who find that the upcoming economies in Asia and the Middle East started to compete with the deficit countries and thus making price

competitiveness, thus unit labor costs, more important. The surplus countries produce products that are hard to imitate, which gives them more market power and makes their current account balance less influenced by changes in unit labor costs. The production of these advanced products generally also relies less on labor, which again makes the cost of labor less important. The upcoming markets also started to demand more high-end products from the more advanced countries, which increased the demand for export products from the core countries. This had a positive effect on the current account balance of the core countries (see also Zemanek et al. (2010)). The relative appreciation of the Euro vis-à-vis the rest of the world also made it harder for the peripheral countries to compete with the emerging markets (Chen et al. (2013)).

3.2 Arguments against a Causal Relationship from Unit Labor Costs to the Current Account The main argument against a causal relationship from unit labor costs to the current account can be illustrated with the intertemporal current account model of Blanchard and Giavazzi (2002). This model showed that it was expected for the lower-income Eurozone countries to have current account deficits, while higher-income countries were expected to run current account surpluses. This is because higher economic growth was expected in the lower-income countries. This mechanism is explained in more detail in Chapter 2.

Lane and Pels (2011) find that expected economic growth is actually a strong predictor of the current account balance of countries. They established this using old IMF predictions and relating those to the current account balances of the Eurozone countries. This shows that the intertemporal model of Blanchard and Giavazzi could be used with respect to the Eurozone. The introduction of the Euro also led to more financial integration, increasing capital flows between member countries. This contributed to an increase in current account imbalances of the Eurozone countries (Berger and Nitsch, 2010). The

elimination of exchange rate risks and low interest rates made it easy for the peripheral countries to attract capital, while it made core countries more willing to supply that capital. This further divergence of current account balances did not have to a negative effect of the introduction of the Euro, as Schmitz and von Hagen (2011) point out. It could namely just be a sign that integration within the Eurozone was working. The peripheral countries could now attract capital more easily, which in turn would

accommodate faster growth for those countries. They did stress the importance however that these increased capital inflows needed to be handled with prudence.

(22)

22 Giavazzi and Spaventi (2010) claim that the peripheral countries did not manage to do so. Total factor productivity was not growing with the pace investors expected it to. This was because the peripheral countries were not employing the capital in order to achieve the highest sustainable economic growth. Greece and Portugal were claimed to have consumed excessively. Spain and Ireland invested heavily in the non-tradables sectors of the economy, such as construction. Non-tradables are consumed domestically, so the use of foreign funds to invest in non-tradable products is similar to borrowing funds from abroad in order to consume and invest at home. Both countries were thus in effect increasing consumption. Part of this rise in consumption was to be expected (Ca’Zorzi and Rubaszek, 2012), since the prospect of high economic growth increased consumption today. This is because people try to optimize or smooth their consumption pattern over time.8

The implication of this is the expectation that people in the poorer countries will borrow money now in order to smooth their consumption pattern. In order to be able to pay back this borrowed money the expected economic growth has to actually take place.9 To make sure this actually occurs the

peripheral countries have to handle their capital inflow with great care. It is important to make sure private and public consumption is not excessive on the one hand and on the other hand there is the need that investments should be put towards economic activities that will create sustainable growth.

The sustainable economic growth did not occur however, partially because of the excessive consumption and partially because of the variety of foreign investments entering the peripheral

countries. The foreign investment was mainly in the form of portfolio investment and bank loans and not as foreign direct investment. Foreign direct investment is more likely to increase productivity and is thus preferred when trying to achieve sustainable growth (Jaumotte and Sodsriwibbon, 2010). The excessive consumption, the excessive investment in real estate, and the lack of investment aimed at productivity gains led to an unsustainable demand boom and growth.

The economic boom brought about wage increases, while total productivity growth was still lagging behind. The wage increases actually happened first in the non-tradables sector and then spilled over to the tradables sector. This mechanism can be seen as a sort of reversed Balassa-Samuelson effect. Within the Balassa-Samuelson effect the productivity growth in the sector of tradables leads to a wage increase in that sector. This then leads to a wage increase in the sector of non-tradables, since wages in a country are expected to be equal in both sectors. In the peripheral countries the opposite happened and the channel was the reverse of what was expected (Zemanek et al., 2010). Moreover, the excessive

8_{Think of the Permanent Income Hypothesis, as introduced by Friedman (1957).} 9

(23)

23 investments in the non-tradables sector, such as construction, led to price and wage increases in that sector. The increase of prices led to lower purchasing power of the wages in the tradables sector and fueled rising inflation. Workers in the tradables sector then bargained for higher wage compensation to mitigate their loss of purchasing power. This then increased wages in the tradables sector. The

investment in the non-tradables sector thus led to higher wages in the non-tradables sector and as a secondary effect resulted in higher wages in the tradables sector as well. The order of wage increases was thus the opposite of what was expected according to the Balassa-Samuelson theorem. The higher wage increases with low productivity growth led to a relative rise in unit labor costs in the peripheral countries.

This suggests that the current account deficits could cause an increase in unit labor costs instead of the other way around. Gaullier and Vicard (2012) agree with the above and state that the current account imbalances are caused by an asymmetric shock on the demand side, instead of competitiveness problems. They note that export growth has been similar in both the core and peripheral countries. They also state that the increases in aggregate unit labor costs are mainly driven by increases in the growth of the wages in the non-tradables sector. This could happen because these sectors are most sheltered from international competition.

Dieppe et al. (2012) agree that the wage increase was caused by a demand boom in the sector of non-tradables. They however state that the current account imbalances in the Eurozone are caused by lack of price competitiveness, high unit labor costs, in the peripheral countries on the one side and higher non-price competitiveness in the core countries. This point is related to the first part of this chapter, claiming that price competitiveness is relatively more important for the peripheral countries. This is because of the higher international competition in markets of these lower-end goods. The core countries on the other hand rely less on labor and more on non-price competition factors, such as for instance patents. This makes unit labor costs relatively less important for the core countries.

Unit labor costs and the current account are related, as is shown in this chapter. This paper will now show a further investigation into the question whether changes in unit labor costs affect the current account balances of the Eurozone countries. In the following chapter econometric analysis will be used in order to study this research question with a more quantitative approach.

(24)

24

4. Empirical Application

This chapter will empirically investigate whether changes in the unit labor costs affect the current account balances within the Eurozone. Firstly, the sample used in the empirical application will be described, thereafter the methodology will be explained. Finally, the results of the empirical

application will be discussed in order to investigate the statistical relationship between unit labor costs and the current account balance.

4.1 Data description and methodology

The sample used in this paper consists the data of 11 Eurozone countries, who were part of the Euro Area-12 (Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Portugal, and Spain). Luxembourg was left out, because the economic structure of Luxembourg is different in comparison to that of the other countries. This is due to the small size of the country and the predominant role of the financial industry within its economy.10 The time period studied is between 2000, the year after the introduction of the Euro, until 2013. Annual data are used because of greater availability. Using yearly data also makes the data better comparable, since seasonal effects on the current account balance and unit labor costs will not interfere with the analysis when using annual data.

Panel data analysis will be used in order to investigate the relationship between the current account balance and unit labor costs. Next to the analysis of the panel for all 11 countries combined, a separate analysis will be examined using the peripheral and core countries separately in order to see if there are differences between the two groups of countries. The data investigated in this chapter comes from Eurostat or were created on the basis of data from Eurostat.

The first goal of the data analysis is to investigate whether the link between the level of unit labor costs and the current account balance holds as was suggested in the literature.11 To estimate this an autoregressive distributed lag (ADL) model will be used to estimate the effect of unit labor costs on the current account. The autoregressive distributed lag model used will consist of one lagged value of the dependent variable and the contemporaneous and one or more lagged explanatory variables.

By constructing the model in this fashion it will be possible to conduct a Granger causality test. If this test shows a significant relationship between the variables it can be said that the independent variable Granger-causes the dependent variable. It is important to note however that the Granger causality test mainly tests whether the independent variable is a useful predictor for the dependent

10_{This point is also made in Chapter 2.} 11

(25)

25 variable. It would therefore be more appropriate to say that the independent variable Granger-predicts instead of Granger-causes (Stock and Watson, 2012).

In the first part of the analysis the dependent variable will be the current account balance divided by GDP. Both the current account balance with the world and the current account vis-à-vis the EU will be used here to investigate differences between the two. Possible differences in outcomes can show whether trade flows inside and outside of the EU are influenced differently by changes in unit labor costs. Differences, or the lack thereof, can then help explain how changes in unit labor costs affect the relative competitiveness of the Eurozone countries. The expectation is however that both the current account balances with the EU and the world will be negatively influenced by changes in unit labor costs.

The main explanatory variable that will be used in the regressions is relative nominal unit labor cost. Nominal unit labor cost will be used, since real unit labor costs relies on inflation in the separate Eurozone countries. The focus of this paper lies on the price competitiveness, through unit labor costs, within the Eurozone. The definition of nominal unit labor costs given in Chapter 2 was total labor costs divided by real GDP. The nominal unit labor cost is thus the labor cost of producing a single unit of goods denominated in euros. The assumption is that producers in the Eurozone compete on price in nominal terms and thus look at unit labor costs in nominal terms. This in combination with the assumption of high openness of trade in the Eurozone makes that unit labor costs should not be corrected for national inflation. This is why nominal unit labor costs are used instead of real unit labor costs, which are adjusted for inflation.

The nominal unit labor costs are divided by the Eurozone average to yield relative nominal unit labor costs, because of the assumption that the Eurozone countries mainly compete with each other. The higher the relative nominal unit labor costs, the less competitive the respective country is expected to be. High relative nominal unit labor costs are thus expected to worsen the current account balance.

The first part of the analysis will use country-specific fixed effects. These fixed effects are added to mitigate the effects of the omitted variable bias that can appear because of the different economic structure of countries. No time specific dummies are used, since Chapter 2 showed that there is no clear overall year to year dynamic in the current accounts of the Eurozone countries.

The second part of the empirical research will investigate the effect of changes in unit labor costs on the exports and imports of the Eurozone countries. The dependent variable will now be the change in exports or imports divided by GDP. Both exports and imports to and from the world and to and from other EU countries will be investigated. In this part of the analysis country fixed effects will be used again in order to account for differences in economic structure between the countries, which can affect the

(26)

26 changes in exports and imports. The exports and imports of most countries have a unit root, this is why first differences are used in the estimations. Most countries have a drop in exports and imports in 2009, this is why a dummy variable will be used for the year 2009. This drop is related to the start of the global recession of 2009, which was related to the preceding global financial crisis. This time dummy was not included in the first part of the analysis, since the effects of a drop in exports over GDP and imports over GDP simultaneously did not result in a substantial change in overall current account balances.

Apart from relative nominal unit labor costs two other explanatory variables are added to the regression in order to explain changes in the current account balance: the real effective exchange rate and GDP growth. The real effective exchange rate is added, because of the assumption that a strong Euro will have a negative effect on the current account balance. This is because Eurozone products will

become relatively more expensive for the rest of the world, and thus less in demand. The addition of the real exchange rate builds on the point made by Chen et al. (2013). They show that the appreciation of the euro affects exports of the peripheral countries negatively, since those countries lose price

competitiveness in comparison to emerging countries. The peripheral countries are affected more by a real exchange rate appreciation than the core, since they produce relatively more low-end goods.12 Price competitiveness is generally more important for these goods, since they are easier to imitate. Core countries are thus expected to be less affected by changes in the real effective exchange rate, since they produce more high-end goods. For the production of these goods non-price competitiveness is relatively more important than for the lower-end goods. The market for low-end goods is also more competitive, because of the emergence of upcoming economies in the Middle East and Asia.13

A strong Euro will result in a high real effective exchange rate in the estimations. The hypothesis thus is that a high real effective exchange rate will worsen the current account balance, because of the relative loss of competitiveness to emerging markets. The real effective exchange rate used is the rate for the Eurozone as a whole. This is done in order to make sure that different rates of inflation of the Eurozone member countries do not distort the outcomes of the analysis. This again builds on the assumption that the high openness of trade within the Eurozone mitigates the need for inclusion of differences in inflation. Using the country-specific real effective exchange rate would include the consumer price index for each country separately.

Ca’Zorzi and Schnatz (2007) show that the consumer price index and unit labor costs follow similar trends. Including country-specific real effective exchange rates would thus introduce a

12_{See Belke and Dreger (2011).} 13

(27)

27 component within the regression that is similar to unit labor costs. The real effective exchange rate is added to the regression in order to investigate how the real exchange rate of the euro affects the price competitiveness relative to countries outside of the Eurozone. Relative nominal unit labor costs are included to study the price competitiveness relative to other countries within the Eurozone.

The growth rate of GDP is added, because it is expected that high GDP growth rates worsen the current account balance. This builds on the expectations of Blanchard and Giavazzi (2002) and the findings of Lane and Pels (2011). High economic growth is likely to attract more capital from abroad, because it will yield higher returns for investors. The extra inflow of capital will increase the net capital inflows. This increases the capital account for the countries with high GDP growth and thus decreases the current account.14

4.2 Empirical results

4.2.1 The Current Account and Relative Nominal Unit Labor Costs

The results of the first part of the research show that there is a significant negative relationship between the current account balance over GDP and relative nominal unit labor costs. This can be seen because of the significantly negative coefficient of relative nominal unit labor costs in the first regression. It is interesting to note however that the fourth regression shows no significantly negative coefficient of relative unit labor costs, thus suggesting that changes in relative nominal unit labor costs do not affect the current account balance with other EU-members. This implies that unit labor costs have a significant influence on the current account balance vis-à-vis the world, but not with the EU. The second and fifth regression in Table 2 shows a negative relationship between the current account level and a change in nominal relative unit labor cost, since the contemporaneous variable is significantly negative and the lagged variable significantly positive. This again is only the case for the model containing the current account balance with the world. The current account balance vis-à-vis the other EU member countries is again not significantly influenced by relative nominal unit labor costs.

14

(28)

28 To establish whether relative nominal unit labor costs Granger-cause the current account

balance, the contemporaneous and lagged value of relative nominal unit labor costs are jointly Wald-tested to be different from zero. This is only the case for the second regression, such that we can only reject the null hypothesis for this regression. 15 This implies that in the second regression previous values of the main explanatory variable, relative nominal unit labor costs, help predict the current level of the

15

P-value of 0.0001 for model 2, while 0.2337 for model 5. This thus rejects the null hypothesis for regression 2 with a significance level of over 99.9 percent, while the fifth model does not reach the 95 percent significance level and the null hypothesis is thus not rejected for the fifth model.

Table 2: Current Account and Relative Nominal Unit Labor Costs

(1) (2) (3) (4) (5) (6)

Dependent variables CAGDP CAGDP CAGDP EUCAGDP EUCAGDP EUCAGDP Independent variables L1.CAGDP 0.726*** 0.726*** 0.624*** (0.0724) (0.0463) (0.0611) L1.EUCAGDP 0.827*** 0.824*** 0.792*** (0.0263) (0.0383) (0.0515) RNULC -0.195** -0.381*** -0.396** -0.00955 -0.169 -0.205 (0.0454) (0.0704) (0.0961) (0.0236) (0.0963) (0.151) L1.RNULC 0.222* 0.154 0.205 0.271 (0.0805) (0.102) (0.112) (0.259) L2.RNULC 0.0317 -0.0341 (0.0391) (0.121) REER -0.0377* -0.0608*** -0.0860** -0.0338** -0.0373* -0.0417 (0.0138) (0.0105) (0.0191) (0.0104) (0.0134) (0.0206) GDPG -0.310** -0.215** -0.202** -0.0926 -0.0278 -0.0288 (0.0808) (0.0656) (0.0593) (0.0641) (0.0658) (0.0661) Constant 11.28*** 10.11*** 11.73*** 2.065 -0.232 1.346 (3.111) (2.782) (3.336) (2.334) (2.759) (3.115) N 152 141 130 152 141 130 adj. R-sq 0.737 0.781 0.776 0.704 0.716 0.691

Note: Robust standard errors in parentheses, * p<0.05 ** p<0.01 *** p<0.001, Country specific fixed effects are used in the estimations

(29)

29 current account balance. Traditionally this is done with solely lagged variables. The contemporaneous variable is also included in the regression however, because of the assumption that the level of relative unit labor costs help explain the current account balance of that same year. The third and sixth model in Table 2 show that adding an extra lagged variable to the model is not preferred, since the second lag is not close to being significant and, albeit slightly, even reduces the adjusted R-squared of the model. It is thus not preferred to add a second lag.

The relationship between the current account balance and real effective exchange rate is negative as was expected. This confirms the expectation that a strong Euro will have a worsening effect on the current account balances of the Eurozone countries. A stronger Euro makes products from the Eurozone relatively more expensive and thus diminishes demand for products from the Eurozone. This then has a worsening effect on the current account balance. It is interesting to note however that the real effective exchange rate also affects trade within the EU in the same manner. Relative nominal unit labor cost does not affect the current account vis-à-vis the EU however. This might point towards the real effective exchange rate picking up part of effect of the changes in the unit labor costs. That is not the case however, since estimating regressions 4, 5, and 6 of Table 2 again without the real effective exchange rate still does not yield significant results for relative nominal unit labor costs.

Table 2 also shows that the current account balance is negatively affected by GDP growth. This could be a confirmation that higher economic growth worsens the current account balance. High economic growth would attract capital, which would stimulate an inflow of more capital and thus a worsening of the current account balance.

The results in Table 2 might be surprising, since the relative nominal unit labor costs seem to be only statistically relevant for the current account balance with the world. The results with respect to the the current account balance with the world were as expected. Higher relative unit labor costs, and thus lower competitiveness, results in a deterioration of the current account balance over GDP. When looking at the current account balance with the other EU countries this significantly negative relationship does not occur. This suggests that the unit labor costs do not affect the current account imbalances within the Eurozone.

Table 3 contains the same variables as regressions 1, 2, 4, and 5 in Table 2, but the regressions are now estimated for the peripheral and core countries separately. Chapter 3 predicted that price competitiveness would be of relatively more importance to the peripheral countries than to the core countries. Separating the two groups of countries allows us to investigate that hypothesis.

(30)

30 Table 3 shows a significantly negative relationship between GDP growth and the current account balance over GDP for the peripheral countries in every regression, thus supporting the hypothesis that high economic growth puts downward pressure on the current account balance for those countries.16 The current account balances of the core countries are unaffected however by GDP growth.

This however should not imply that the peripheral countries should aim for low economic growth in order to improve their current account balance, since this would not be an economically sound policy. It might be important however to aim for prudent use of net capital inflows. This means that the capital inflow will have to be used carefully in order to increase levels of productivity. The capital should be used in the sectors where it will have the greatest productive value. As seen in Chapter 3, this could be done by investing in the tradables sector. This will lead to a better competitive position for the peripheral countries and create economic growth. A way of establishing this is to attract more capital in

16_{As is seen in Lane and Pels (2011), who found that high forecasted economic growth was a strong predictor of} the current account balance of Eurozone countries.

Table 3: Current Account and Relative Nominal Unit Labor Costs separately

(Core1) (Peri1) (Core2) (Peri2) (Core3) (Peri3) (Core4) (Peri4) Dependent variables CAGDP CAGDP CAGDP CAGDP EUCAGDP EUCAGDP EUCAGDP EUCAGDP Independent veriables L1.CAGDP 0.773** 0.734*** 0.684*** 0.775*** (0.140) (0.0401) (0.0931) (0.0378) L1.EUCAGDP 0.846*** 0.835*** 0.845*** 0.770** (0.0248) (0.0710) (0.0448) (0.103) RNULC -0.0903 -0.235** -0.418* -0.234** 0.00786 0.0239 -0.325* -0.0289 (0.159) (0.0462) (0.130) (0.0359) (0.0416) (0.0314) (0.126) (0.0835) L1.RNULC 0.255 0.0364 0.359 0.0648 (0.119) (0.0619) (0.155) (0.125) REER -0.0259 -0.00521 -0.0636 -0.0417 -0.0129 -0.0679 -0.0209 -0.0858* (0.0369) (0.0367) (0.0261) (0.0307) (0.0134) (0.0259) (0.0133) (0.0281) GDPG 0.0187 -0.464** -0.0445 -0.433** 0.134 -0.215* 0.0486 -0.186* (0.0266) (0.0633) (0.0735) (0.0513) (0.0859) (0.0499) (0.0670) (0.0508) Constant 12.22 25.01*** 23.66 24.81** 0.970 4.105 -0.606 4.493 (19.33) (2.129) (13.01) (4.423) (5.437) (2.353) (3.761) (4.779) N 83 69 77 64 83 69 77 64 adj. R-sq 0.707 0.831 0.718 0.846 0.738 0.735 0.745 0.727

Note: Robust standard errors in parentheses, * p<0.05 ** p<0.01 *** p<0.001

Core = Austria, Belgium, Finland, France, Germany, Netherlands. Peri = Greece, Ireland, Italy, Portugal, Spain Country specific fixed effects are used in the estimations

(31)

31 the form of foreign direct investment, which is more likely to increase productivity, instead of portfolio investments and bank loans, which might create bubbles of high, unsustainable, growth.17 Lower, but sustainable, growth would thus be preferred over higher, unsustainable, growth. Aiming for merely sustainable economic growth might then lead to lower current account deficits for the peripheral countries. This would be because of the lower net capital inflow to those countries.

The first regressions of Table 3 furthermore show the relationship between the level of relative nominal unit labor cost and the current account for the peripheral countries when no lags for relative nominal unit labor costs are used. This result is in line with the hypothesis that unit labor costs are of more relevance for peripheral countries with respect to the current account. This hypothesis relies on the assumption that the peripheral countries rely relatively more on price competitiveness when competing with the other countries. This is because the peripheral countries produce relatively more low-end goods, which in general are more labor-intensive to produce and face fiercer competition on the international goods market.

The core countries on the other hand rely more on non-price competiveness. This makes unit labor costs of less importance to them, since they compete more on other aspects than the cost of production. The products the core countries typically produce are also relatively less labor-intensive to produce. This makes unit labor costs less important, since the cost of labor is a relative smaller part of the cost of production. This result is in line with the findings of Belke and Dreger (2011). This is why the coefficient of the first regression of the core countries is not significantly negative, as it is for the peripheral countries.

Granger causality only appears in one regression of Table 3.18 This Granger-causal relationship is between the world current account balance and relative nominal unit labor costs of the peripheral countries. This again supports the hypothesis of the above that unit labor costs are of more importance to the peripheral countries, it is interesting however that again no evidence is found that the current account balance vis-à-vis the other EU members is Granger-caused by relative nominal unit labor costs.

Both Table 2 and 3 thus show no support for the importance of relative nominal unit labor costs in determining the current account balance with the other EU members. This might infer that the countries in the Eurozone and the EU do not compete with each other on the basis of unit labor costs. This finding would be strange however, since there are large trade flows between the countries as was seen in the export shares to Eurozone countries in Figure 10 in Chapter 2. Another interesting feat of

17

See Jaumotte and Sodsriwibbon (2010).

18_{The regressions are tested for Granger causality in the same manner as the regressions in Table 2. See Table A1.1} in Appendix 1 for the values of the outcomes of the Wald-test.