Labor cost, competitiveness, and imbalances within the eurozone

Abstract

This paper examines the impact of the cost of labor on macroeconomic imbalances within the eurozone. For this purpose, we construct a three-country Stock-Flow Consistent (SFC) macroeconomic model in an open economy including the eurozone and the rest of the world. We show that internal devaluation is not effective as an economic policy for getting the eurozone economies to converge. Cutting labor costs cannot kick-start economic activity through a rebound in exports of those countries that do so. Instead, it involves the risk of locking their economies concerned into low-wage production activities. The fall in unemployment it entails is the consequence, then, of a downturn in labor productivity. In contradistinction, we defend the idea that the introduction of a new wage rule making monetary wages dependent on productivity gains and on the target for inflation set by the European Central Bank, together with a budgetary stimulation policy, would be conducive to initiating convergence among European economies. In particular, it would produce a convergence in living conditions by improving labor productivity but it would also bring production structures closer together.

Keywords:

• SFC modeling
• eurozone crisis
• labor cost competitiveness
• wage and budget policies coordination

Disclosure statement

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

Notes

1 The main objectives of the Hartz reforms were to reduce unemployment and increase incentives to look for work. Among other things, these laws extended the scope for temporary work, made it easier to enter into low-wage contracts for a small number of hours worked (mini-Jobs, midi-Jobs, ein euro Jobs), and tightened the conditions for unemployment insurance benefits. For further discussion of the link between the Hartz reforms and Germany's export performance, see Celi et al. (2018, chapter 2), Dadush et al. (2010), Flassbeck and Lapavitsas (2013), Odendahl (2017), Sinn (2014), or Storm and Naastepad (2015).

2 Provided that the commercial and productive structures of eurozone countries do not diverge too much, which is what Flassbeck and Lapavitsas (2013) assume.

3 Fundamental equilibrium exchange rate approach developed by Williamson (1994).

4 The main effects to be expected of an internal devaluation policy brings out the necessity of introducing these variables in the modelling.

5 Godley and Lavoie (2006) assume that the exchange rate between the eurozone and the rest of the world is free floating.

6 If intra-eurozone credit were added to the modelling, it could have an influence if domestic banks ration lending to domestic firms, but banks in the rest of the eurozone do not engage in such rationing behaviour. In this case, investment by domestic firms would be greater if foreign banks granted them credit, compared with the situation where domestic banks rationed credit and, ultimately, investment by domestic firms and foreign banks did not lend to domestic firms. This channel has been studied by Duwicquet and Mazier (2010, 2015).

7 In this section, we only present the most decisive equations in the macroeconomic dynamics of the model and the simulated scenario. The full model is shown in the Appendices for a more detailed view. In addition, the Eviews simulations are available on request.

8 The calibration of this equation is based on Duwicquet (2020). Also, as in Duwicquet (2020), it is assumed that the profit rate can be decomposed as follows: $\frac{\mathit{\text{UP}}}{\mathit{K}}\text{= }\frac{\mathit{\text{UP}}}{{\mathit{Y}}_{\mathit{r}}{\mathit{P}}_{\mathit{P}}}\frac{{\mathit{Y}}_{\mathit{r}}}{{\mathit{K}}_{\mathit{r}}}\frac{{\mathit{P}}_{\mathit{P}}}{\mathit{P}}.$

9 The calibration of this equation is also based on Storm and Naastepad (2012).

10 It is assumed that working hours are constant. Changes in per capita labor productivity gains therefore corresponds to changes in hourly labor productivity gains.

11 The calibration of this equation is based on Duwicquet (2020).

12 It is assumed that they are fixed rate bonds maturing in 10 years.

13 If the Central Banks were to actively pursue quantitative easing policies in the model, rather than behaving passively as residual buyers of government securities, this would have the effect of reducing interest rates on bonds.

14 The model used in the article has two versions: a basic version without a wage rule, and a modified version with a wage rule. The model and simulations in the basic version run for 65 periods. On the other hand, the modified version with a wage rule runs for 28 periods, or even 20 periods depending on the European economy being studied. For this reason, we present the shocks over the shortest time period, i.e., 20 periods, in order to make the results of the simulations comparable.

15 This notation is adopted so as to simplify the writing. Nevertheless, we are aware of the limits of this modelling. In particular, it does not allow us to represent the European economies under study line by line. It is fairer then, for example, to interpret the modelling for France as that of an economy seeking to approximate the structural characteristics of the French economy. The same applies for the other economies under study. Here we try primarily to compare economies with different structures (degree of openness, geographical or productive specialization) so as to highlight the role played by these structural differences when economic policies are simulated.

16 In order to propose a realistic calibration of the European economies studied in the article, we use national accounts data. In particular, we use the Ameco, Eurostat and EU-Klems databases.

17 Further tests are carried out in the Appendices to assess the robustness of the model and the results obtained.

18 Based on Storm and Naastepad (2015, 2016), it is assumed that strong productive specialization is reflected by high export elasticity to foreign real income and vice versa. Using the classification proposed above, we assign a value to each European economy studied for the elasticity of exports to foreign real income, according to the group to which it belongs. Logically, economies with a high level of productive specialization are assigned a higher elasticity, and vice versa.

19 To shed light on the presentation, we set out in the case of France the macroeconomic dynamics of the various scenarios using graphs. Given the different countries, scenarios and calibrations studied in the model, it seems more relevant to present the results of all the other scenarios in the Appendices to simplify the presentation.

20 We use the following formula to calculate the relative unemployment rate (and likewise for the other variables): unemployment (percentage) scenario under study—unemployment (percentage) reference scenario.

21 From now on the abbreviation “pp” is used for percentage points.

22 In the model, we can observe the evolution of the three balances of the institutional sectors (consolidated government, external and private), as suggested by Godley (1999). In doing so, we show that the policy of internal devaluation leads to unsustainable consequences. Indeed, for the economy that implements it, it leads to an ever-greater increase in the trade balance and the public deficit.

23 To this end, Carnevali, Fontana, and Veronese Passarella (2020) also stress the importance of firms' strategic behaviour during a devaluation policy. Even if the Marshall-Lerner condition is satisfied, it is quite possible that a devaluation of the national currency will not reduce export prices, if firms do not pass on the fall in the nominal exchange rate to their export prices, which results in an increase in margins. In our model, we find the same result for an internal devaluation policy. Indeed, if firms also adopt this mark-up behaviour, the reduction in unit labor costs is not fully reflected in export prices, reducing the effectiveness of the internal devaluation policy.

24 See especially Bofinger (2015), Flassbeck and Lapavitsas (2013), Stockhammer (2011), and Wren-Lewis (2014, 2015, 2016).

25 The recovery policy consists in increasing public spending by an ex-ante amount equivalent to 1% of real GDP in the reference scenario (scenario 5) in period 10.

26 Scenarios 1, 2, 3, and 4 are analysed with respect to the baseline scenario presented in the Appendices. Alteration of the rule for monetary wage formation necessarily entails the development of a new reference scenario. Scenarios 6 and 7 are then used relative to scenario 5, which serves as the new baseline scenario.

27 Like Botta et al. (2023) and Nair (2023), we believe that the effectiveness of scenarios 6 and 7 could be enhanced if the stimulus policy were accompanied by policies for the structural transformation of the economy, such as an industrial policy or trade controls. By substituting imports and/or increasing the elasticity of exports to foreign income, these policies would have three effects in the model: structurally improve the external balance, which stimulates growth and tax revenues and increases the government's fiscal space for manoeuvre; reinforce the effectiveness of the stimulus by reducing the propensity to import, since, as Nair (2023) writes, “the fall in import propensity allows leakages […] to be slow enough that the […] economy has enough time to benefit from the stimulation”; the economy's non-price competitiveness would be strengthened if, as Botta et al. (2023) write, “increased public expenditures will concentrate in public investment feeding productive development and structural change.”

28 In this respect, see the empirical results on the relationship between export growth and relative unit labor cost growth obtained by Danninger and Joutz (2007), Diaz Sanchez and Varoudakis (2013), Di Mauro and Forster (2010), Felipe and Kumar (2011), or Le Bayon et al. (2014).

29 In the baseline and scenario 5, the growth rate of real GDP, unemployment and the growth rate of labor productivity are presented as percentages. The trade balance, public deficit and public debt are presented as a percentage of GDP. Scenarios 1, 2, 3, and 4 are interpreted relative to the baseline scenario. Scenarios 6 and 7 are interpreted in relation to scenario 5 which, as explained above, becomes the new baseline scenario. Scenarios 1, 2, 3, 4, 6, and 7 are then presented in percentage point.