Bubble economics and structural change: the cases of Spain and France compared

Abstract

This paper delves into the recent events that led to the formation of the housing bubble in Spain and the resulting structural change that is arguably needed to put the economy back into the right track. For this purpose, we calibrate a model with different equilibria descriptive of the labor markets in Spain and France, where the unemployment rates went from the same initial spot to very different levels. In addition to this, we run a counterfactual analysis that throws some more light on the performance of the Spanish labor market and the housing bubble. Our results suggest that the unemployment rate in Spain has jumped to much higher levels while switching between equilibria or, what is the same, because of structural change. Moreover, our counterfactuals indicate that, first, there has been an important misdirection of resources into the construction industry mainly fueled by excessively low real interest rates and, second, the Spanish labor reform has fallen short of its own goals.

Keywords:

• structural change
• bubble
• unemployment
• interest rate
• productivity

JEL classification:

• J64
• O57

Acknowledgement

We also thank Romain Espinosa for thoughtful remarks and we remain grateful to the participants at the 18th INFER Conference and the 65th French Economic Society Meeting.

Notes

1. See Figure 1 and Table 1 below, as well as Figures 3 and 4 in the next section.

2. Such preference has been called out from different quarters lately, see among many, the OECD Economic Surveys for Spain (2005, 2014), the OECD Report (2007) on competitiveness and the value chain, the IMF multi-authored Working Paper on competitiveness in the Southern Euro Area (2008), and the McKinsey&Company – Fedea Growth Agenda for Spain (2010).

3. Appendix 1 exhibits all the data sources used in this study.

4. See the Banco de España Annual Report (2014) for a comprehensive study on the change of course in the Spanish economy.

5. See, among others, Akin et al. (2014), Antipa and Lecat (2010), Fernández-Kranz and Hon (2006), Fuentes-Castro (2011), and Garcia, Giannikos, and Guirguis (2007).

6. For early references on the Austrian School in relation to booms and busts cycles and monetary expansion see, most notably, Mises (1912), Hayek (1931), and more recently Huerta de Soto (1997); for other more up-to-date references not necessarily in line with the Austrians see, for example, Bordo and Landon-Lane (2013) and the ones cited in Calvo (2013).

7. Euribor-12 m, for example, is the interest rate most commonly used to calculate mortgage payments in Spain, where mortgages are usually of the variable rate type and are calculated as Euribor + X% with X anything between 0.75 and 2%.

8. Correlation between the interest rates, on the one hand, and construction jobs or mortgages on the other, is strongly negative as clearly inferred from Figure 2.

9. It is true that under the Aznar period (1996–2004) a series of structural reforms were adopted that might have had a lasting positive effect into later years, but these high growth rates are still suspicious when put next to the poor productivity performance. See de Quirós and Ricardo (2006) for an extensive treatment of the Spanish economic model during 1996–2004.

10. Ireland stands a bit apart from the lot of periphery countries as it is fast recovering due in part to its extremely flexible labor market conditions. Doris, O’Neill, and Sweetman (2015), for example, stress the downward nominal wage flexibility in Ireland in the pre-crisis period. In addition, Mario Draghi recently went on record by saying, at the 2014 Jackson Hole Economic Symposium, that Ireland entered the crisis with a relatively flexible labor market, having adopted more labor market reforms under its 2010 European Union/International Monetary Fund (EU/IMF) program.

11. Our database is rather limited and we should remain on the conservative side when reading through our simulations. See Agnese and Sala (2009) for a methodological reference.

12. Different vacancy costs for high- and low-skill firms are introduced to reproduce the situation in Spain.

13. The rate at which jobs arrive to workers m(θ) is defined in (1). Low-skill workers only have a chance to match in low-skill firms, which is measured by ϕ in Equation (2). High-skill workers in CSM equilibrium match with any firm and only with high-skill firms in EPS equilibrium.

14. Bentolila et al. (2012) develop a parallel analysis for France and Spain with quarterly data yet with a focus on the transition between equilibria.

15. Remember that, unlike France, for the Spanish case we propose different combinations of c_L and c_H to mirror the two-tier structure of the labor market there (see Bentolila et al. 2012).

16. There is evidence that the construction boom has affected schooling decisions in Spain (Aparicio-Fenoll 2014). If the schooling decision were endogenous in our model, there would be an even stronger pull towards the EPS equilibrium in the aftermath of the crisis (once we widen the productivity gap). A worsening of the conditions of low-skill workers would make some individuals go for skill upgrading, giving rise to an overall improvement that takes the form of a switch to the EPS equilibrium (see Albrecht, van den Berg, and Vroman 2009 or Agnese and Hromcová 2016).

17. Even when France is not precisely known for being a paramount example of labor flexibility, the vacancy cost to total output ratio stands at 30% – much lower than the 46% of the baseline Spanish calibration.

18. The unemployment rate as derived from the model above is non-linear and quite burdensome for estimation purposes. The formal derivation is available on request but we will stick to the linearized version as this does not pose a serious drawback for our analysis.

19. Introducing py in the unemployment equation answers the necessity of accounting for y_L , yet the former is a measure of aggregate productivity and the other the productivity of low-skill workers alone.

20. The rationale for this was presented with the first counterfactual exercise. Notice that other periphery EU countries have undergone similar bubble episodes but none as significant as in Spain; for example, Ireland’s bubble was short-lived due to its highly flexible labor market, while Greece’s bubble got diluted with its debt problem and the ensuing political instability.

21. Notice that Equations (18) and (19) now imply that high-skill unemployed workers will only take high-skill jobs and low-skill firms will only hire low-skill workers.