Aedín Doris Donal ONeill amp Olive Sweetman Maynooth University 4 th Administrative Data Seminar Dublin Castle April 20 2015 Our Questions How have wages in Ireland changed as a result of the Great RecessionCrisis ID: 1028518
Download Presentation The PPT/PDF document "Using Job Churn Administrative Data for ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
1. Using Job Churn Administrative Data for Labour Market AnalysisAedín Doris, Donal O’Neill & Olive SweetmanMaynooth University4th Administrative Data Seminar,Dublin Castle,April 20, 2015
2. Our QuestionsHow have wages in Ireland changed as a result of the Great Recession/Crisis?What degree of flexibility in wages is there in the Irish labour market?To what extent did firms differ in the pay cuts/rises they gave?
3. Advantages & Disadvantages of ‘Job Churn’ dataFree from reporting errors – usually a major concern in measuring wage changesBroad income measure – ‘gross taxable pay’, includes bonuses and benefit-in-kindBut net of pension contributionsNo information on hours worked so cannot consider hourly wagesNo information on education levels, firm tenure or occupation, so potential for detailed analysis based on individual characteristics is limitedNo labour market status information, so we cannot separately identify the unemployed, the non-participants and the emigrantsVery limited information on firms – but potential for merger with other datasets to overcome this?Firm identifiers allow identification of ‘stayers’ – those in the same job in successive periodsMatching of employers and employees allows analysis of within-firm changes – huge potential for firm and industry level analysis
4. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
5. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
6. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
7. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
8. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
9. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
10. Earnings Dynamics, Job StayersYearMedian ChangeMean ChangeProportion Freezes Proportion Cuts Proportion Increases 2005/06 0.060 0.0700.0250.1720.8042006/07 0.061 0.0700.0250.1760.7992007/08 0.045 0.0490.0280.2290.7422008/09-0.006-0.0200.0330.5270.4402009/10-0.011-0.0130.0440.5520.4032010/11 0.006 0.0130.0680.3930.5392011/12 0.008 0.0160.1030.3420.5552012/13 0.008 0.0200.1000.3430.557
11. Annual Earnings Changes, Job Stayers, 2005/09
12. Annual Earnings Changes, Job Stayers, 2009/13
13. Median Cumulative Pay Changes: By Industrial Sector
14. Measuring Wage RigidityDownward wage rigidity can be the result of contracts being negotiated in advance, or firms’ reluctance to cut wages due to negative morale effectsWe want to measure the proportion of workers who do not receive pay cuts, but who would in the absence of rigidityInternational Wage Flexibility Project (Dickens et al., 2007) covered 16 countries, including Ireland. Proposed measure of rigidity: where C is the proportion of workers affected by pay cuts, F is the proportion of workers affected by pay freezesAssumes that all wage changes that are prevented by wage rigidity are ‘piled up’ into wage freezesDownward rigidity in Ireland reported as being the lowest of all countries included in the study (4% vs. mean of 28%)But note that ECHP data used for Ireland suffers from measurement error
15. The Time Pattern of Wage RigidityYearWage Rigidity2005/060.1812006/070.1732007/080.1642008/090.1012009/100.1272010/110.2132011/120.2752012/130.274
16. The Time Pattern of Wage RigidityYearWage Rigidity2005/060.1812006/070.1732007/080.1642008/090.1012009/100.1272010/110.2132011/120.2752012/130.274
17. The Time Pattern of Wage RigidityYearWage Rigidity2005/060.1812006/070.1732007/080.1642008/090.1012009/100.1272010/110.2132011/120.2752012/130.274
18. The Time Pattern of Wage RigidityYearWage Rigidity2005/060.1812006/070.1732007/080.1642008/090.1012009/100.1272010/110.2132011/120.2752012/130.274
19. Variation Between Firms in Wage ChangesAs well as having data on individuals, we have firm identifiers, so we can characterize within-firm wage distributions fullyWe can examine whether firms adopted different approaches to wage changesDid some firms give only pay rises, while others gave only pay cuts?If so, we should find substantial between-firm variation in firms’ mean pay changesTo examine this, we calculate the variance of individual workers’ wage changes and decompose this variance into two parts:The part due to within-firm variance in wage changesThe part due to between-firm variance in mean wage changes
20. Decomposition of Variance of Individual Wage Changes, Job StayersYearTotal Variance of Log Pay ChangeProportion Between FirmsProportion Within Firms2005/060.01870.130.872006/070.01960.120.882007/080.02070.150.852008/090.02390.250.752009/100.02070.180.822010/110.02110.150.852011/120.02040.100.902012/130.01850.100.90
21. Decomposition of Variance of Individual Wage Changes, Job StayersYearTotal Variance of Log Pay ChangeProportion Between FirmsProportion Within Firms2005/060.01870.130.872006/070.01960.120.882007/080.02070.150.852008/090.02390.250.752009/100.02070.180.822010/110.02110.150.852011/120.02040.100.902012/130.01850.100.90
22. Decomposition of Variance of Individual Wage Changes, Job StayersYearTotal Variance of Log Pay ChangeProportion Between FirmsProportion Within Firms2005/060.01870.130.872006/070.01960.120.882007/080.02070.150.852008/090.02390.250.752009/100.02070.180.822010/110.02110.150.852011/120.02040.100.902012/130.01850.100.90
23. Decomposition of Variance of Individual Wage Changes, Job StayersYearTotal Variance of Log Pay ChangeProportion Between FirmsProportion Within Firms2005/060.01870.130.872006/070.01960.120.882007/080.02070.150.852008/090.02390.250.752009/100.02070.180.822010/110.02110.150.852011/120.02040.100.902012/130.01850.100.90
24. Distribution of Pay Cut Proportions Across Firms, Job Stayers aged 25-60, Various Years
25. Cross-Firm Variation in Wage Changes: Additional ResultsWe can examine the characteristics – both within and between firms – that are associated with bigger wage increases (or smaller wage cuts)Between firm analysis: use the ‘Between estimator’ – regresses firm average characteristics on firm average pay changesResults: firms with younger workers, with lower-paid workers, with more female workers and with fewer Irish workers had bigger wage increases or smaller cuts in all yearsWithin firm analysis: use the ‘Fixed Effects’ estimator – regresses within-firm deviations from the firm mean characteristic on the deviation from the firm mean wage change of individual pay changesResults: younger, male, lower-paid workers, had bigger pay increases/smaller cuts. Pre-crisis, non-Irish workers had bigger pay rises, but this switched during the crisis so that in 2010-2013, Irish workers had bigger pay rises/smaller cuts
26. Conclusions IMatched employer-employee administrative data allows us to examine the issue of wage flexibility in far greater detail than has been done previouslyPre-crisis: significant degree of downward wage flexibility in annual wages confirmedCrisis Onset: Aggregate figures suggest a limited wage response to the crisis. However, medians mask considerable heterogeneityProportion of workers receiving pay cuts trebled overall; increased by factor of 2.75 in private sectorProportion of freezes stayed low. Measured wage rigidity fell substantiallyMany workers continued to receive pay rises as others received cutsHowever, the firm the worker was in was not the primary driver of the size of their wage change: within-firm variation in wage changes accounted for the bulk of variation in individual pay changes
27. Conclusions II Post-Crisis Recessionary-Recovery Period: Proportion of workers receiving pay cuts declined relative to crisis peak, but still high by international standardsSpike at zero increased significantly – perhaps due to resistance to pay cuts in successive yearsAs a result, measured wage rigidity rose significantly. Irish level now around average across countries found in Dickens et al. (2007). Must be taken in context of remarkable extent of earnings cutsAs expected, the industrial sectors most involved in the crisis suffered most in terms of pay, with the notable exception of Finance