Erratum

Previous article FreeErratumMatthieu Chemin and Etienne WasmerMatthieu CheminUniversity of Quebec at Montreal and CIRPEE Search for more articles by this author and Etienne WasmerSciences Po Paris and OFCE Search for more articles by this author University of Quebec at Montreal and CIRPEESciences Po Paris and OFCEOriginal articleUsing Alsace‐Moselle Local Laws to Build a Difference‐in‐Differences Estimation Strategy of the Employment Effects of the 35‐Hour Workweek Regulation in FrancePDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreWe thank Olivier Godechot (Sciences Po) for bringing to our attention two main points from our article titled "Using Alsace-Moselle Local Laws to Build a Difference-in-Differences Estimation Strategy of the Employment Effects of the 35-Hour Workweek Regulation in France" (Journal of Labor Economics 27, no. 4 [2009]: 487–524): a coding mistake and a discussion of border workers. A summary of his discussion can be found on his home page.1 The two main identification strategies of our original paper are not affected by the coding mistake, and they were independently replicated by him. The coding mistake affects only a third identification strategy based on firm size presented in the appendix of our paper. In that test, we incorrectly classified individuals with a missing value for firm size as workers in firms of more than 20 employees. This coding mistake affects columns 1–3 of appendix table A2. In addition, the coefficient of years 1999–2000 in table A2, column 2, cannot be estimated unless firm size is imputed for years that were missing.The firm size variable is missing for 62% of observations. Additional variables in the survey cannot be used to impute the size of the firm. It is therefore impossible to produce a rigorous identification strategy with this variable. Columns 1–3 of appendix table A2 should therefore be dropped from the analysis. The coding mistake does not affect columns 4–6 of appendix table A2 (identification based on sectors affected).It does not affect columns 1–3 of table 1 (identification based on occupations affected). It also does not affect the main results on employment and unemployment (tables 2, 3), since we did not use firm size as an identification strategy in these tables. Therefore, it does not affect the general conclusion of the paper.The "pure control group" in columns 4–6 of table 1 was an intersection of occupations affected, sectors affected, and firms affected. It should be only the intersection of occupations affected and sectors affected. We present here an updated Table 1 of the original paper. Columns 6 and 8 show the same results as in our original paper: employees in Alsace-Moselle worked more than those in the rest of France after 2000. In other words, the 35-hour workweek was implemented with a lower intensity in Alsace-Moselle than in the rest of France. The coefficient is not significantly different from 0.35, as in our original article.Table 1. Replication of Table 1, Dropping the Test Based on the Size of the Firm Samplea Excluding Border Workers Occupations Affected (DD) (1)Occupations Unaffected (DD) (2)Full (DDD: 1–2) (3)Affected Individuals (DD) (4)Unaffected Individuals (DD) (5)Full (DDD: 4–5) (6)Affected Individuals (DD) (7)Full (DDD) (8)Full (DDD: 1–2) (9)Full (DDD: 4–5) (10)Full (DDD: 4–5) (11)(Alsace-Moselle) × (2003)−.26−1.12**−1.11**−.46−.46**−.47**−.57−.49**−1.10**−.42*−.46** (.22)(.43)(.46)(.60)(.19)(.19)(.71)(.20)(.46)(.24)(.25)(Alsace-Moselle) × (2001–2).38*−.43*−.37.72***.12.12.96**.11−.50**−.22*−.25* (.19)(.24)(.28)(.27)(.20)(.20)(.40)(.21)(.24)(.13)(.14)(Alsace-Moselle) × (1999–2000).14−.19−.16.28**.06.06.17.07−.17−.07−.05 (.17)(.65)(.66)(.11)(.28)(.28)(.12)(.28)(.66)(.28)(.28)(Alsace-Moselle) × (2003) × (affected individuals) .87* .04 .04.91*.14.14 (.52) (.61) (.61)(.50)(.58)(.58)(Alsace-Moselle) × (2001–2) × (affected individuals) .75*** .61*** .61***.45**.44**.44** (.22) (.22) (.22)(.20)(.19)(.19)(Alsace-Moselle) × (1999–2000) × (affected individuals) .31 .24 .24.19.36*.35* (.51) (.21) (.21)(.52)(.18)(.18)Border × GDP growth 5.41−1.51 −2.67 (4.98)(3.84) (3.76)Department fixed effects (95)YesYesYesYesYesYesYesYesYesYesYesControl variables (14)YesYesYesYesYesYesYesYesYesYesYesOccupation fixed effects (30)YesYesYesYesYesYesYesYesYesYesYesYear × (affected individuals dummy) fixed effectsNoNoYesNoNoYesNoYesYesYesYesDepartment × (affected individuals dummy) fixed effectsNoNoYesNoNoYesNoYesYesYesYesOccupation × (affected individuals dummy) fixed effectsNoNoYesNoNoYesNoYesYesYesYesObservations370,50788,100458,60765,455393,152458,60753,310458,607453,792453,792453,792R2.166.410.361.181.366.357.204.356.363.359.359p-value of the t-test of coefficient of interest = .35.88 .06.16 .23.13.14.63.63.63Note. Ordinary least squares results. Robust standard errors are in parentheses, clustered at the level of department. The dependent variable in all columns is the number of hours usually worked per week. In col. 1, the sample is restricted to occupations affected by the 35-hour reform. In col. 2, the sample is restricted to occupations unaffected by the 35-hour reform. In col. 3, the coefficient of interest is (Alsace-Moselle) × (2001–2) × (affected individuals), where affected individuals is a dichotomous variable equal to one if the individual has an occupation affected by the 35-hour reform, zero otherwise. In col. 4, the sample is restricted to the affected individuals (individuals working in occupations affected by the 35-hour reform and in sectors affected by the 35-hour reform). In col. 5, the sample is restricted to the unaffected individuals. In all columns, 95 department fixed effects, eight year fixed effects, 14 control variables (five age dummies, sex, size of the household, and seven diploma dummies), and 30 occupation fixed effects are included. Columns 7 and 8 add border × GDP growth. In cols. 3, 6, and 8–11, 95 interaction dummies between departments and the affected individuals dummy, seven interaction dummies between year dummies and the affected individuals dummy, and 30 interaction dummies between occupations and the affected individuals dummy are included. In columns 9–11, we exclude all border workers. The p-value of the t-test coefficient of interest is =.35 is the p-value of the t-test (Alsace-Moselle) × (2001–2) = .35 in cols. 1, 4, and 7 and (Alsace-Moselle) × (2001–2) × (affected individuals) = .35 in cols. 3, 6, and 8–11. The coefficients of interest are highlighted in bold.a In cols. 4–11, "Affected Individuals" designate the intersection of affected occupations and affected sectors.*. Significant at 10%.**. Significant at 5%.***. Significant at 1%.View Table Image: 1 | 2Column 9 presents an additional robustness check. In places near the border, some workers reside in France but work abroad. Most of them are likely not affected by the 35-hour workweek, although one can argue about the exact fraction. We exclude from the analysis all of these "border workers." Columns 9–11 show that the triple differences coefficients are similar—in fact, they are closer to 0.35, the theoretical estimate.Table 4 of our 2009 paper concerning hourly wages is also affected by the coding mistake. We reported that in three specifications (cols. 1–3) hourly wages increased in Alsace-Moselle with marginal significance (at the 10% level), and in the next two (cols. 4, 5) results were not significant. This indicated mixed evidence for hourly wages, leaning toward positive. Given the impossibility of using firm size, columns 1–3 should therefore be dropped. Since then, we discovered that monthly earnings rose significantly in Alsace-Moselle relative to those in France. One explanation is that employers in Alsace-Moselle could not cut wages as much as in the rest of France since workers in Alsace-Moselle have the outside option of working abroad. This may explain why hourly wages did not significantly change over the period of interest. Notes Contact the corresponding author, Matthieu Chemin, at [email protected].1 http://olivier.godechot.free.fr/hoprubrique.php?id_rub=0. Previous article DetailsFiguresReferencesCited by Journal of Labor Economics Volume 35, Number 4October 2017 Published for the Society of Labor Economists, Economics Research Center/ NORC Article DOIhttps://doi.org/10.1086/693983 © 2017 by The University of Chicago. All rights reserved.PDF download Crossref reports the following articles citing this article:Luis Cárdenas, Paloma Villanueva Challenging the working time reduction and wages trade-off: a simulation for the Spanish economy, Cambridge Journal of Economics 45, no.22 (Jan 2021): 333–351.https://doi.org/10.1093/cje/beaa055Related articlesUsing Alsace‐Moselle Local Laws to Build a Difference‐in‐Differences Estimation Strategy of the Employment Effects of the 35‐Hour Workweek Regulation in France18 Jul 2015Journal of Labor Economics

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.