BUENAS PRÁCTICAS DE MANUFACTURA (CONDICIONES SANITARIAS) PARA ESTABLECIMIENTOS DE ALIMENTACIÓN COLECTIVA
11) Condiciones higiénicas de manipulación a. Recepción o Compra Directa
The e¤ects of the changes in legislation that ease regulations on …xed-term contracts may be di¤erentiated according to the stage of the …rm’s business cycle (Bentolila and Saint-Paul, 1992) and job ‡ows in shrinking sectors may be more negatively a¤ected by EPL than job ‡ows in growing sectors (Messina and Vallanti, 2007). For that reason, we re-estimate the model for narrow sectors with negative and positive employment growth separately (Figures 3.9 and 3.10, respectively).
In expanding sectors, we …nd that an innovation to the index of ‡exibility at the margin does not Granger cause job ‡ows. Conversely, in narrow sectors with negative employment growth, a one standard deviation in the index of ‡exibility at the margin causes an increase in the destruction and a decrease in the creation of …xed-term jobs for a 4-year period.
Figure 3 9: Response of job ‡ows to a shock in ‡exibility at the margin in shrinking sectors
Figure 3 10: Response of job ‡ows to a shock in ‡exibility at the margin in expanding sectors
Looking at the variance decompositions for both samples, we can conclude that the index of ‡exibility at the margin explains a non-negligible higher share of the destruction rate of …xed-term jobs in shrinking sectors (7.4%) than in expanding sectors (5.4%) in the 6 periods ahead (Table D.10 and D.11, respectively). It is also relevant to note that the destruction of …xed-term jobs explains a slightly higher percentage of the variation in job creation of …xed-term jobs in shrinking sectors than in expanding sectors (8.9% and 7.1% one period ahead, respectively), while the destruction of open-ended jobs explains a higher share of the permanent creation rate in expanding than in shrinking sectors (4.7% and 1.8%, respectively, in the following year).
These results show that the promotion of ‡exibility at the margin is especially important for …rms that are reducing jobs, since those …rms adjust the composition of job ‡ows when a shock arises in the previous period. This is quite intuitive and is in line with the …ndings of Varejão and Portugal (2007) that show that the proportion of …xed-term contracts has a larger e¤ect on the probability of establishments adjusting employment downwards rather than upwards. This pattern may signal that …rms use …xed-term contracts longer and avoid converting them when they are in a downturn, which is in fact evidence that …xed-term contracts are also used as bu¤er stocks for open-ended contracts in the Portuguese labour market.
Therefore, the mean e¤ect of ‡exibility at the margin described in subsection 3.5.1 is explained by the behaviour of shrinking rather than expanding sectors.
3.5.4 Robustness Analysis
We rely on the Cholesky decomposition to identify the impulse response functions. Although the ordering of the variables was based on stylised results found in the literature, we test the robustness of our …ndings by changing the ordering. Namely, we reversed the ordering putting the ‡exibility index last (Figure C.4 and C.5). The main conclusions remain about the impact of an innovation to job ‡ows: there is less creation and more destruction of …xed-term jobs. Moreover, a negative (positive) response on the index of ‡exibility at the margin is observed following a shock in the destruction (creation) of …xed-term jobs. Since the results are qualitatively
unchanged, we are con…dent that they are not driven by the ordering of the variables in the system.
The results are also unchanged if we exclude the agriculture sector, which is under sampled in the data source used. The same also occurs if we compute the index of ‡exibility at the margin with the exogenous weights as proposed by Alexandre et al. (2010) (Figure C.6 and C.7).
As discussed by Dolado et al. (2002), the relative wage of permanent and …xed- term contracts may have an impact on the new equilibrium proportion of …xed-term contracts. The introduction and facilitation of the use of …xed-term contracts may indeed increase the bargaining power of workers with permanent contracts, especially in countries with high employment protection levels (Bentolila and Dolado, 1994). As such, we include the relative wage of permanent workers (the ratio between the average log hourly wage21 of workers on permanent contracts and the average log
hourly wage of workers on …xed-term contracts) in the model and conclude that the results remain qualitatively unchanged (Figures C.8, C.9 and C.10). Although the relative wage is not caused by the innovations to job ‡ows, a shock in the relative wage Granger causes higher temporary job destruction, at a 5% signi…cance level. This is in line with Dolado et al.’s (2002) argument, since if workers on permanent contracts can negotiate higher wages, …rms may want to adjust employment by relying more heavily on …xed-term jobs that entail lower dismissal costs so as to tackle the higher wage bill. This e¤ect may be more marked in sectors with higher union power.
Finally, we assess the robustness of our results to the inclusion of employment growth calculated for each narrow sector as a proxy for the business cycle at the sector level, which captures the idiosyncratic behaviour, namely the phase of the cycle and the future prospects of each unit j. The results in Figures C.12 and C.13 are similar to those discussed above. It is also stressed that although a shock in ‡exibility at the margin has a negative e¤ect on employment growth, there is no causal relationship between the index of ‡exibility at the margin and employment growth (Figure C.11).