An Empirical Analysis of Natural and Cyclical Unemployment at the Provincial Level in Spain

Abstract

Differences in regional unemployment rates, as well as their formation mechanism and persistence, have given rise to many papers in recent decades. The present work contributes to this strand of literature from two different perspectives. In the first part of our work, we follow the methodological proposal put forward by Hofler and Murphy (1989) and Aysun et al. (2014). We use a stochastic cost frontier to break down actual Spanish provincial unemployment (NUTS-3) into two different estimation components: the first associated with aggregate supply side factors, and the other more related to aggregate demand side factors. The second part of our research analyses the existence of spatial dependence patterns among Spanish provinces in actual unemployment and in the two above-mentioned components. The decomposition carried out in the first part of our research tells us what margin policymakers have when dealing with unemployment reductions by means of aggregate supply and aggregate demand policies. Finally, spatial analysis of unemployment rates in Spanish provinces may also have significant implications from the standpoint of economic policy since we find common formation patterns or clusters of unemployment.

Appendices

APPENDIX 1: Tables and Figures

Table 3 Description of variables and data sources

Table 4 Mean value and deviation of the variables used in the estimation

Fig. 5

Natural unemployment (${U}_{it}^{NR}$) by province (1984–2012). Source: Authors’ own

Fig. 6

Cyclical unemployment (${U}_{it}^C$) by province (1984–2012). Source: Authors’ own

Fig. 7

Global scatterplot diagrams of Moran’s I (Knn = 5). Source: Author’s own

Fig. 8

Local scatterplot diagrams of Moran’s ii (Knn = 5). Source: Authors’ own

APPENDIX 2: Elements of actual unemployment

Figure 9 illustrates the different components of the actual unemployment rate. The curve ${L}_{it}^S$ is the labour force and displays a positive slope. This is because, as the market wage increases (w_it), more individuals join the labour market given that their “static” reservation wage (or that of the choice model between consumption and leisure) is lower than the market wage.

The curve ${N}_{it}^S$ shows the actual labour supply. The difference between ${L}_{it}^S$ and ${N}_{it}^S$ establishes that not all active workers are immediately available for work. As the market wage increases and rises above the “dynamic” reservation wage (or that of the job search theory), more workers will accept the jobs they find. For this reason, the distance between ${L}_{it}^S$ and ${N}_{it}^S$ is lower for higher wages. This difference between the two previously mentioned curves is frictional unemployment (${U}_{it}^F$).

Figure 9 also shows two situations of the demand for employment, depending on how aggregate production stands. Situation (y₀) represents economic expansion, whereas situation (y₁) reflects an economic recession. Point “A” shows how, in a situation in which production is booming and the market wage is at its equilibrium level, with a ${L}_{i,t}^D\left({y}_0\right)$ demand of work, there are unemployed workers due to the presence of frictional unemployment.

Fig. 9

Frictional, structural and cyclical unemployment. Source: Authors’ own

In addition, the existence of a collective bargaining system impacts on the mechanism for establishing wages by fixing a “collective bargaining” (${w}_{it}^{CB}$) wage which is above the competitive equilibrium wage (${w}_{it}^{PC}$). This is one example of institutional inflexibility which leads to an excess of labour supply, and triggers structural unemployment (${U}_{it}^{ST}$).⁵⁶ As in the previous case, some structural unemployment would also exist even if demand for work were ${L}_{i,t}^D\left({y}_0\right)$, which is associated to an economic boom. Several studies (Bentolila and Jimeno 2003; Simón et al. 2006; Bande et al. 2008) point out that the “inflexibility” of wages prevents these from acting as an equilibrium mechanism in the Spanish labour market.⁵⁷ In line with this, “via price” adjustment does not function “correctly”, thus leading to adjustments in the labour market occurring “via quantity.⁵⁸

Finally, in Fig. 9, cyclical unemployment is reflected in the horizontal distance between curves ${L}_{it}^D\left({y}_0\right)$ and ${L}_{it}^D\left({y}_1\right)$. This kind of unemployment may be corrected in the short term by applying expansive aggregate demand policies.

APPENDIX 3: Model with spatial lags

This additional specification seeks to correct spatial dependence in the estimation. To do this, we incorporate the spatial lag of the dependent and independent variables into the baseline model. In this case, we perform a specification based on the idea of the Spatial Durbin Model (SDM)⁵⁹ as shown in the following expression:

U_it = β₀ + ρWU_it + β₁X_it + β₂Y_it + β₃Z_it + δD2001 + γ₁WX_it + γ₂WY_it + γ₃WZ_it + v_it + μ_i where W is the spatial weight matrix (Knn = 5); WU_it is the spatial lag of the dependent variable; WX_it captures the effect of the industry composition of labour in neighbouring provinces; WY_it are the spatial interactions of the demographic variables and WZ_it is a vector of spatial lags for the human capital of the active population. We compute this new specification so as to test whether spatial dependence in our model is a consequence of endogenous or exogenous interaction effects.

The results of the estimated model are presented in table 5. The spatial lag of the dependent variable shows a positive and statistically significant coefficient as well as the majority of the spatial lags of the independent variables. This implies that the values of the dependent and independent variables in the neighbouring provinces capture part of the variability in the unemployment rate.⁶⁰ In a second step, we carried out the same spatial dependence analysis as in the baseline model within the main text, with results still pointing to the presence of positive spatial correlation. It can therefore be concluded that spatial dependence also seems to be caused by factors different from those variables considered in this paper.

Table 5 Spatial lags model