Understanding the relationship between trade union membership and collective bargaining across 42 countries ILPC conference, London 7 April 2014 Kea Tijdens Maarten van Klaveren
What do unions do? Bargain about wages & working hours Do workers join a union for their bargaining efforts? TU member but not covered currently not covered but would like to be so reasons beyond CBC (TU provide public goods, ) CB covered but no TU member extension within company/industry, centralised bargaining Research objectives Introduction To what extent are trade union (TU) membership and collective bargaining coverage (CBC) related? Is TU membership a good predictor of CBC? 7-Apr-14 2
Data from WageIndicator web survey WageIndicator web survey on work and wages Web survey is continuously posted on national WageIndicator websites on work and wages These websites are in almost 80 countries, managed by WageIndicator Foundation Attracting millions of web-visitors through branding, search engine optimization and links from media partners Visitors are encouraged to complete survey (prize incentive) Why this data? data is comparable across countries and large numbers of observations survey has questions about TU membership and CBC data elsewhere not available: web survey is the only global data set with these variables 7-Apr-14 3
Disadvantage of the data WageIndicator web survey is a volunteer survey Not randomly sampled findings cannot be generalised The study is therefore an explorative study How biased? workers aged 50+ are underrepresented part-timers 20 hrs or less are underrepresented workers aged 25-45 (M+F) are slightly overrepresented volunteer survey, thus not a representative sample, though the higher Internet access rates, the more likely the Internet population reflects the national population 7-Apr-14 4
The data used survey data from 2006-2012, pooled to obtain sufficient observations for as many countries as possible selection 1: employees (no self-employed, job seekers, etc) selection 2: observations with valid values on dependent and independent variables number of observations per country varies largely, selection 3: only countries with at least 100 observations selected data set has 474,462 observations from 42 countries Models The web survey data binary logistic regression on CBC, affected bytu membership, controlled for age, gender, firm size, wage decile, industry also affected by national bargaining level (ICTWSS database Visser)? 7-Apr-14 5
Guatemala Colombia China Paraguay Kenya El Salvador Chile France Australia United States Peru United Kingdom Angola Argentina Czech Republic Uganda Portugal Senegal Slovakia Sri Lanka Tanzania Zambia Mozambique Mexico Zimbabwe Egypt Ghana Poland Pakistan Germany Brazil Denmark Netherlands Indonesia Spain South Africa Hungary Italy Sweden Finland Belgium Vietnam Trade union membership and collective bargaining coverage 100% 80% 60% 40% 20%.60.50.40.30.20.10.00 -.10 0% -.20 Member of a trade union Covered by agreement correlation Source: WageIndicator non-representative web-survey, pooled 2006-2012 data 7-Apr-14 6
Finland Vietnam Belgium Sweden Denmark Italy Hungary Spain Egypt South Africa Brazil Indonesia Netherlands Argentina Germany Tanzania Senegal Pakistan Mozambique Poland Ghana Zimbabwe Portugal Uganda France Mexico Zambia Slovakia Sri Lanka Chile Angola China Czech Republic United Kingdom Paraguay El Salvador United States Peru Kenya Australia Colombia Guatemala Trade union membership and collective bargaining coverage 2 100% 80% 60% 40% 20% 0% No CB coverage & no TU member CB coverage & no TU member No CB coverage & TU member CB coverage & TU member Source: WageIndicator non-representative web-survey, pooled 2006-2012 data 7-Apr-14 7
Relationship wage deciles - CBC 100% 80% 60% 40% 20% 0% D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 7-Apr-14 8
Explaining CB coverage (N=474,462) Exp(B) sig Exp(B) sig TU member 2.832 *** 2.480 *** Female.960 *** Age (15-64) 1.010 *** Firm size (1=0-20,.., 8=1000+) 1.285 *** Wage decile D1-D10.989 *** NACE2.0 (ref=manufacturing) Trade, transport, hospitality.778 *** Commercial services.606 *** Public sector, health care, education 1.403 *** COUNTRY controls yes *** yes *** -2 Log likelihood 565181.96 565181.96 Nagelkerke R Square 0.22 0.30 7-Apr-14 9
Explaining CB coverage (20 countries) Exp(B) sign TU member 2.484 *** Female.975 ** Age (15-64) 1.010 *** Firm size (1=0-20,.., 8=1000+) 1.301 *** Wagedecile D1-D10.984 *** NACE2.0 *** Trade, transport, hospitality.784 *** Commercial services.591 *** Public sector, health care, education 1.469 *** Coordination (1=company lvl,.., 5=economy wide) 1.301 ns -2 Log likelihood 477653.99 Nagelkerke R Square 0.27 N 426085 7-Apr-14 10
Conclusions Are TU membership and CBC related? yes, correlation pos. and sign. in 36 of 42 countries TU membership good predictor of CBC? yes, odds ratio increases two times for TU members coordination level has no sign. influence on CBC Other variables predicting CBC higher CBC for males, older workers, workers in larger firms, workers in public sector no clear relationship between wage deciles and CBC 7-Apr-14 11
The end... Thank you for your attention Questions, comments? See WageIndicator website www.wageindicator.org m.vanklaveren@uva.nl k.g.tijdens@uva.nl 7-Apr-14 12