and injury leave: results from a difference in difference approach - - PowerPoint PPT Presentation

This project has received funding from the European Union’s Seventh Programme for Research, Technological Development and Demonstration under Grant Agreement No 312691

A European Social Sciences Research Infrastructure Project

Summer School

11ht of May 2016, Noisy le Grand - France

Mohamed Ben Halima (CEE et TEPP), Nathalie Greenan (CEE et TEPP), Joseph Lanfranchi (CEE et LEMMA), Laetitia Otte (DARES)

Company organisational changes and long term sickness absence and injury leave: results from a difference in difference approach

2

Outline of the presentation

• Motivation and aims
• Data description

Coi and Hygie databases Definition of organisational changes Long term sickness absence Sample and descriptive statistics

• Econometric framework

Treatment effects during and after the changes Difference in differences methodology

• Sample average effects of organizational changes on health

Results for the overall population of employees in changing firms Results by gender

• Concluding remarks

3

Motivation

• Standard models assume that employers make adjustments to the production

process to maximise profits, rather than employee wellbeing (Bloom and Van Reenen, 2007; Freeman and Kleiner, 2005).

• However are firms perfectly rational?
• Furthermore, more and more employers do not share the simplistic view of

Friedman saying that “The social responsibility of business is to increase its profit”

• Indeed, from the society point of view, the social consequences of poorly

managed changes at work appear as quite serious,

• The International Labour Organization estimated at 4% of the GDP the economic

losses created by work accidents and occupational diseases.

• In 2009, for each European worker, the European Commission estimates that

respectively 1,3 and 2,1 working days are lost because of work accidents and work related health problems.

• In France, the daily benefits for work accidents and occupational diseases

experience the fastest increase in 2010, the average costs of a work accident are 3000 euros, 24000 euros for cumulative trauma disorders.

4

Organisational changes and health at work (1)

• There is uncertainty about the impact organisational changes are

likely to have on employees’ health.

• On the one hand, if those changes enrich employees’ working lives,

this is likely to improve their mental and physical health.

• On the other hand, if these changes are simply a means of

intensifying worker effort, this may lead to a higher incidence of illness, injury, absence and stress.

• Furthermore, even if organisational changes enhance workers’

control over their job, the process of their introduction can generate uncertainty leading to increased anxiety among workers.

• Also, a supplementary question is the length of this alleged effect of
• rganisational changes on employees’ health. These effects on

long-term sickness absence are unlikely to persist since those worst affected will choose to leave the organisation while the remainder are liable to adapt over time (Kahneman et al., 1999).

5

Organisational changes and health at work (2)

• Different studies show that the introduction of new organisational

practices tends to increase working intensity and consequently deteriorates health: Green (2004), Cottini and Lucifora (2010).

• The process of innovating can also generate anxiety: Bordia et al.

(2004),

• Workplace reorganisations causes different work-related mental

and physical health problems: Pollard (2001), Osthus (2007).

• Social support can help workers cope with workplace innovation,

Bryson, Dale-Olsen and Barth (2014) find supportive evidence for the buffering effect of unionisation in mitigating the negative impact of workplace innovation on job anxiety.

• In France, Euzénat et al, (2013) found that obtaining ISO9001

standard decreases work accidents in firms with more than 200 employees, whereas adopting goods and services labelling increases work accidents.

6

What type of database do we need ?

• Employer-employee linked dataset with:

indicators of innovations in terms of tools or practices defining the technology and organisation of production at large at firm or establishment levels. Precise measures of health indicators at employee level:

• health status
• degrees of physical functioning
• problems with work caused by physical health
• degrees of bodily pain and the extent to which pain interferes with normal work
• work accidents
• occupational disease
• sick leave

7

The COI Survey

• The observed units are private firms in the non-agricultural market

sector with 10 employees or more. Financial sectors and research and development are included, but services to individuals are excluded.

• The data were stratified by industry and company size with a

comprehensive layer beyond 500 employees. This sample contains 13697 units.

• Very precise set of information about the use in 2006 and 2003 of a

large set of tools which diffusion within the population of companies was large enough to justify a question in a national survey:

– like just in time, ISO certification, traceability, enterprise research planning etc,

8

The Hygie database

• Merging two French administrative files : National retirement pension

fund (CNAV) and National Health Insurance Fund (CNAM-TS)

• The CNAV data served as the entry point with a sample of 804,599

beneficiaries in 2005 aged from 22 to 70 with at least one work quarter qualifying for retirement during the course of their lives. The CNAM-TS data concern National Health Insurance beneficiaries for which at least one reimbursement was received in 2004 or 2005.

• CNAV and CNAM-TS data matching allowed to build the HYGIE panel of

538,870 beneficiaries from 2005 to 2010. It records:

– individual information about socio-demographic characteristics, professional career, medical consumption, sick leaves – information about the identity of employers – complete retrospective career information including data about periods of long term sickness absence and injury leave before 2005.

9

Matching Hygie and COI

• Within HYGIE, we kept all individuals on whom we observe the identity of

the employer between 2003 and 2005, that is the period of observation of employer characteristics in the COI survey.

• We were left with a set of 477 250 individuals employed in the whole

private sector. Then, we matched the 13 697 units surveyed in COI with the employers of this set of individuals.

• We found 12 366 COI firms employing employees in Hygie, that is a match

rate of around 90%.

• The total number of matched employees is 26 499 individuals.
• Our working sample retains those 26 321 employees who have

contributed at least four months to social security.

10

Timing of changes

2000-2001-2002 2003-2004-2005 2006-2007-2008

Experimental Period

Experimental period: 2003 to 2008 Before organisational changes period: 2000 to 2002 During organisational changes period: 2003 to 2005 After organisational changes period: 2006 to 2008 We compare the absence behaviour of employees before, during and after the changes have been implemented by their company

11 11

Measurement of employer Changes ? (1)

• Tools used by the organisation= models of organised action
• Through the adoption or dropping of tools, employers reveal their

intentions of change

• Focus on cumulative adoption: it weighs stronger on evolutions of work

than the adoption of any specific tool because: – There is a strong heterogeneity in the uses by employers of any given tool – Cumulative adoption of tools reveals a strong intention of change / a new orientation in work practices

12 12

• We build two synthetic indicators to measure changes in the uses
• f two families of tools often described as complementary from

the point of view of economic performance – Information and Communication Technologies → equip the information system – Management tools → equip the production system – Over 2003-2006/2007

• We build indicators that are comparable over time

Measurement of employer Changes ? (2)

13 13

Selected tools

Management ICTs

• 1. Quality certification
• 2. Environmental and ethical certification
• 3. Methods of problems solving
• 4. Tools for labelling goods and services
• 5. Satisfaction surveys of customers
• 6. Management of production just in time
• 7. Tools for tracing goods and services
• 8. Contractual commitment to provide a

product or a customer service within a limited time

• 9. Requirement for suppliers to meet tight

deadlines

10.Long term relationships with suppliers 11.Call and contact Centres 12.Teams or autonomous work groups 13.Customer relationship management

• 1. Web site
• 2. Local area network
• 3. Intranet
• 4. Extranet
• 5. Electronic data interchange system
• 6. Database(s) on the management of

human resources

• 7. Database(s) for R&D
• 8. Tools for data analysis
• 9. Tools for interfacing databases

10.Tools for automated data archiving 11.ERP 12.Software or firmware for the

management of human resources

13.Software or firmware for R&D 14.Groupware 15.Workflow software

14

Three treatment groups and a control group

• We performed Multiple Correspondence Analyses on each families of

tools to obtain a continuous scale measuring the extent of the changes in each dimension. The higher the firm value on this scale, the more intense are the organisational changes in the given dimension

• For each type of change, we consider that under a given threshold its

magnitude is marginal and build two change dummies from which we distinguish four change states

• ICT changes only
• Managerial changes only
• Both ICT and managerial changes
• Inertia
• According to the type of change of their employer, employees are

considered as belonging to three treatment groups or to a control group (inertia)

15

Long term absence and injury leave

• The outcome we examine is the occurrence of a long term

absence or injury leave.

• For the National pension fund, a private sector worker will validate

for his pension a period equivalent to a working period when he has experienced 60 consecutive days of absence from work compensated by the National Health Insurance System.

• Every year of the professional career of the individuals from the

Hygie database, we identify if they experienced such a long term absence.

• The National Health Insurance System identifies four different

causes for this long term absence:  severe illness,  work accidents,  occupational disease,  maternity.

16

Selection of the sample of study

• After the merging of the Hygie database with the COI survey, we

were left with 26 321 individuals

• We evaluate the effects of organisational changes between 2003

and 2005 on long term absence.

• In the treatment group are included workers employed in the same

changing firms for the whole three years period : 5 745 workers

• In the control group, we impose a similar condition in inert firms to

prevent any selection effect: 8 875 workers

• We are left with a total of 14 620 individuals in 4030 firms.
• Among these, 10,7% were hired in a firm sampled in the COI survey

at the beginning of the treatment period (2003-2005).

17

Sample and descriptive statistics

Original population : 14620 individuals in 4030 firms TREATED Individuals working for three consecutive years in a single firm with at least one significant

• rganizational change.

5745 individuals (39,30%) 1368 changing firms (33,95%) CONTROL Individuals working for three consecutive years in a single firm with no significant

• rganizational change.

8875 individuals (60.70%) 2662 inert firms (66.05%) Individuals working with the same firm before the 2003-2005 treatment period 13054 individuals (89,29%) Individus newly hired in the firm in 2003 1566 individus (10.71%)

18

Sample and descriptive statistics

Variables All Firms Changing Firms Occurrence of Long term absence in 2005 5,72 % 5,55 % female 32,9 % 33,6 % ≤35 29,5 % 38,2 % [36-45] 33,2 % 38,3 % [46-55] 29,0 % 19,8 % ≥56 8,3 % 3,7 % Entry wage 5950 € 6080 € Hired in 2003 10,71 % 10,6 % Long term disease Before 2003 4,56 % 4,21 % ratio of long term absence before 2003 1 % 0,95 % Managers and professionals 23,6 % 24,8 % Technicians and associate professionals 16,1 % 17,5 % Clerical, services and sales workers 12,8 % 13,2 % Blue collar workers 36,1 % 34,7 % Percentage of firms Size<20 6,2 % 3,5 % [20-50[ 21,0 % 15,3 % [50,249[ 38,5 % 38,4 % [250,499[ 15,6 % 18,6 % >500 18,7 % 24,2% ICT Changes (CC) 20,0 % Management Changes (MC) 10,9 % Both types of Changes (CC&MC) 8,4 %

19

Treatment effects during and after changes

• Observation of long term sickness absence of employees working for

firms sampled in COI during the whole period of organisational changes (2003-2005) and who remained employed with the same employer during the years following the organisational changes (from 2006 until 2008).

• Changes intervene between 2003 and the end of 2005 but may create

instantaneous disorders in the work environment of employees, Unlike treatment due to a change modifying the rules straightaway, our treatment and its effects may be confounded in time.

• However, organisational change may also be a lengthy process (ISO

certification for example) and have positive or negative impacts on the long run.

20

Selection of workers in changing firms

• The selection into firms experiencing organisational changes is

addressed by contrasting their employees with highly comparable employees who display a similar propensity to work for changing

• firms. After conditioning on labour market and health history and

individual characteristics, the choice of working in changing firms is then assumed random.

• Matching of treated and control individuals using the Coarsened

exact matching method proposed by Iacus, King and Porro (2008). Matching involves pruning observations that have no close matches

• n pre-treatment covariates in both the treated and control groups.
• Once the coarsened exact matching algorithm run, we use the output

from it and run estimation of the sample average treatment effect for the treated using classic difference in differences estimator.

• The rate of matching between treated and control individuals is very

high, superior to 90%.

21

Difference in differences estimator (1)

To identify the effect of organisational changes implemented by changing firms, we include  period effects, to capture time-series modification in the absences in the changing firms.  changing firm effects to control for the time invariant characteristics of the treatment group

𝑇𝑗𝑢 = 𝛽 + 𝛾1. 𝑢 = 𝑒𝑣𝑠𝑗𝑜𝑕 𝑝𝑠 𝑢 = 𝑏𝑔𝑢𝑓𝑠 𝑗𝑢 + 𝛿11. 𝑗𝜗𝐷𝐷

𝑘 = 1 ⋂𝑁𝐷 𝑘 = 0 𝑗𝑢

+ 𝛿21. 𝑗𝜗𝐷𝐷

𝑘 = 0 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢 + 𝛿31. 𝑗𝜗𝐷𝐷 𝑘 = 1 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢 + 𝑌𝑗𝑢𝜁

+ 𝑍𝑗𝑢𝜇 + 𝑣𝑗𝑢

Controls:

• Employee: gender, age, occupation, entry wage on the labour market, long term disease

before 2003, ratio of long term absence before 2003

• Employer: size and industry

22

Difference in differences estimator (2)

To further identify the effect of organisational changes on the long term absence of newly hired employees in 2003 and tenured employees, we include interaction term between changing firm dummy and newly hired employees dummy. 𝑇𝑗𝑢 = 𝛽 + 𝛾1. 𝑢 = 𝑒𝑣𝑠𝑗𝑜𝑕 𝑝𝑠 𝑢 = 𝑏𝑔𝑢𝑓𝑠 𝑗𝑢 + 𝛿11. 𝑗𝜗𝐷𝐷

𝑘 = 1 ⋂𝑁𝐷 𝑘 = 0 𝑗𝑢

+ 𝛿21. 𝑗𝜗𝐷𝐷

𝑘 = 0 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢 + 𝛿31. 𝑗𝜗𝐷𝐷 𝑘 = 1 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢

+ 𝜀11. 𝑢 = 𝑒𝑣𝑠𝑗𝑜𝑕 𝑝𝑠 𝑢 = 𝑏𝑔𝑢𝑓𝑠 𝑗𝑢. 1. 𝑗𝜗𝐷𝐷

𝑘 = 1 ⋂𝑁𝐷 𝑘 = 0 𝑗𝑢

+ 𝜀21. 𝑢 = 𝑒𝑣𝑠𝑗𝑜𝑕 𝑝𝑠 𝑢 = 𝑏𝑔𝑢𝑓𝑠 𝑗𝑢. 1. 𝑗𝜗𝐷𝐷

𝑘 = 0 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢

+ 𝜀31. 𝑢 = 𝑒𝑣𝑠𝑗𝑜𝑕 𝑝𝑠 𝑢 = 𝑏𝑔𝑢𝑓𝑠 𝑗𝑢. 1. 𝑗𝜗𝐷𝐷

𝑘 = 1 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢

+ 𝜃1. ℎ𝑗𝑠𝑓𝑒 𝑗𝑜 2003 = 1 𝑗𝑢 + 𝜊11. ℎ𝑗𝑠𝑓𝑒 𝑗𝑜 2003 = 1 𝑗𝑢. 1. 𝑗𝜗𝐷𝐷

𝑘 = 1 ⋂𝑁𝐷 𝑘 = 0 𝑗𝑢

+ 𝜊21. ℎ𝑗𝑠𝑓𝑒 𝑗𝑜 2003 = 1 𝑗𝑢. 1. 𝑗𝜗𝐷𝐷

𝑘 = 0 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢

+ 𝜊31. ℎ𝑗𝑠𝑓𝑒 𝑗𝑜 2003 = 1 𝑗𝑢. 1. 𝑗𝜗𝐷𝐷

𝑘 = 1 ⋂𝑁𝐷 𝑘 = 1 𝑗𝑢 + 𝑌𝑗𝑢𝜁 + 𝑍 𝑗𝑢𝜇

+ 𝑣𝑗𝑢 The models are estimated in a linear probability framework for simplicity and ease of interpretation.

23

DD before (2000-2002) vs during (2003-2005) the organisational changes

Model specification Time and treatment dummies Model (1) + individual characteristics Model (2) + firm variables Model (3) Model 3 with coarsened exact matching Model (4) Model 4 with differences between newly hired and tenured employees Working for firms implementing Organisational Changes ICT Changes

• 0,006

(0,004)

• 0,006

(0,004)

• 0,006

(0,004)

• 0,007*

(0,004)

• 0,006*

(0,004) Managerial Changes

• 0,009*

(0,005)

• 0,009*

(0,005)

• 0,009*

(0,005)

• 0,011**

(0,005)

• 0,011**

(0,005) Both Changes 0,020** (0,008) 0,020** (0,008) 0,020** (0,008) 0,026*** (0,008) 0,026*** (0,008) Hired during Organisational Changes Hired in 2003 Hired during ICT Changes

• 0,007

(0,006) Hired during Managerial Changes 0,003 (0,008) Hired during both Changes

• 0,004

(0,013) N / R² 86 918 / 0,001 86 918 / 0,037 86 918 / 0,038 83 914 / 0,036 83 914 / 0,036

Sample average effects of organisational changes on long-term sickness absences

• f treated employees during the period of changes (Coi-Hygie)

24

Sample average effects of organizational changes on long-term sickness absences

• f treated employees after the period of changes (Coi-Hygie)

DD before (2000-2002) vs after (2006-2008) the organisational changes

Model specification Time and treatment dummies Model (1) + individual characteristics Model (2) + firm variables Model (3) Model 3 with coarsened exact matching Model (4) Model 4 with differences between newly hired and tenured employees Working for firms implementing Organisational Changes ICT Changes

• 0,009*

(0,005)

• 0,008

(0,005)

• 0,009*

(0,005)

• 0,008

(0,005)

• 0,009*

(0,005) Managerial Changes

• 0,005

(0,007)

• 0,005

(0,007)

• 0,005

(0,007)

• 0,017**

(0,006)

• 0,016**

(0,007) Both Changes 0,019* (0,011) 0,020* (0,011) 0,020** (0,011) 0,030*** (0,010) 0,030*** (0,010) Hired during Organisational Changes Hired in 2003 Hired during ICT Changes 0,015 (0,012) Hired during Managerial Changes 0,000 (0,015) Hired during both Changes 0,010 (0,025) N / R² 51 195 / 0,001 51 195 / 0,029 51 195 / 0,029 48 690 / 0,028 48 690 / 0,028

25

Discussion of results (1)

• In all regressions we observe the same core result:

– Negative impact of changes in one dimension only on long term absence – Positive impact of joint changes in ICT and management tools

• Possible explanation :

– Cumulative changes are a bigger shock on work organisation and create more disorder that prevents employees from using health preserving strategies because of the related increase in work intensity , there are less hazards associated with a change in one dimension only which is more likely to be mastered by the organisation

• Note 1: the literature on the complementarities between the two families of change

stress the fact the performance return is higher when both families of changes are implemented together

• Note 2: during the observed period, the most frequent configuration of changes is ICT

changes only

• Link with the new stress-desequilibrium theory by Karasek (2008)

26

DD before (2000-2002) vs during (2003-2005) the organisational changes

Model specification Time and treatment dummies Model (1) + individual characteristics Model (2) + firm variables Model (3) Model 3 with coarsened exact matching Model (4) Model 4 with differences between newly hired and tenured employees Working for firms implementing Organisational Changes ICT Changes

• 0,006

(0,004)

• 0,006

(0,004)

• 0,006

(0,004)

• 0,009**

(0,004)

• 0,009**

(0,004) Managerial Changes

• 0,002

(0,005)

• 0,002

(0,005)

• 0,001

(0,005)

• 0,002

(0,005)

• 0,002

(0,005) Both Changes 0,003 (0,008) 0,003 (0,008) 0,003 (0,008) 0,008 (0,008) 0,008 (0,008) Hired during Organisational Changes Hired in 2003 Hired during ICT Changes

• 0,013*

(0,007) Hired during Managerial Changes 0,015* (0,008) Hired during both Changes

• 0,001

(0,013) N / R² 58 418 / 0,001 58 418 / 0,037 58418 / 0,038 56 818 / 0,040 56 818 / 0,040

Sample average effects of organizational changes on long-term sickness absences

• f treated male employees during the period of changes (Coi-Hygie)

27

Sample average effects of organizational changes on long-term sickness absences

• f treated male employees after the period of changes (Coi-Hygie)

DD before (2000-2002) vs after (2006-2008) the organisational changes

Model specification Time and treatment dummies Model (1) + individual characteristics Model (2) + firm variables Model (3) Model 3 with coarsened exact matching Model (4) Model 4 with differences between newly hired and tenured employees Working for firms implementing Organisational Changes ICT Changes

• 0,010*

(0,006)

• 0,010*

(0,006)

• 0,009*

(0,006)

• 0,011**

(0,005)

• 0,013**

(0,005) Managerial Changes

• 0,00

(0,007)

• 0,004

(0,007)

• 0,004

(0,007)

• 0,005

(0,007)

• 0,004

(0,007) Both Changes 0,032*** (0,011) 0,033*** (0,011) 0,032*** (0,011) 0,031*** (0,010) 0,033*** (0,011) Hired during Organisational Changes Hired in 2003 Hired during ICT Changes 0,009 (0,013) Hired during Managerial Changes

• 0,013

(0,015) Hired during both Changes

• 0,024

(0,026) N / R² 33 146 / 0,003 35 146 / 0,030 35146 / 0,030 33 838 / 0,028 33 838 / 0,028

28

DD before (2000-2002) vs during (2003-2005) the organisational changes

Model specification Time and treatment dummies Model (1) + individual characteristics Model (2) + firm variables Model (3) Model 3 with coarsened exact matching Model (4) Model 4 with differences between newly hired and tenured employees Working for firms implementing Organisational Changes ICT Changes

• 0,006

(0,009)

• 0,006

(0,009)

• 0,006

(0,009)

• 0,004

(0,009)

• 0,004

(0,009) Managerial Changes

• 0,025**

(0,012)

• 0,024**

(0,012)

• 0,025*

(0,012)

• 0,022**

(0,011)

• 0,022**

(0,011) Both Changes 0,053*** (0,018) 0,052*** (0,018) 0,053** (0,018) 0,051*** (0,017) 0,051*** (0,017) Hired during Organisational Changes Hired in 2003 Hired during ICT Changes

• 0,000

(0,013) Hired during Managerial Changes

• 0,000

(0,018) Hired during both Changes

• 0,035

(0,029) N / R² 28 500 / 0,002 28 500 / 0,027 27 369 / 0,028 27 369 / 0,026 27 369/ 0,026

Sample average effects of organizational changes on long-term sickness absences

• f treated female employees during the period of changes (Coi-Hygie)

29

Sample average effects of organizational changes on long-term sickness absences

• f treated female employees after the period of changes (Coi-Hygie)

DD before (2000-2002) vs after (2006-2008) the organisational changes

Model specification Time and treatment dummies Model (1) + individual characteristics Model (2) + firm variables Model (3) Model 3 with coarsened exact matching Model (4) Model 4 with differences between newly hired and tenured employees Working for firms implementing Organisational Changes ICT Changes

• 0,006

(0,011)

• 0,003

(0,011)

• 0,007

(0,011)

• 0,014

(0,011)

• 0,014

(0,011) Managerial Changes

• 0,006

(0,015)

• 0,005

(0,015)

• 0,007

(0,015)

• 0,029**

(0,015)

• 0,032**

(0,015) Both Changes

• 0,0004

(0,024)

• 0,008

(0,023)

• 0,007

(0,023) 0,025 (0,023) 0,022 (0,023) Hired during Organisational Changes Hired in 2003 Hired during ICT Changes 0,026 (0,024) Hired during Managerial Changes 0,059 (0,037) Hired during both Changes 0,031 (0,055) N / R² 16 049 / 0,001 16 049 / 0,026 16 049 / 0,028 15 275 / 0,028 15 275/ 0,028

30

Discussion of results (2)

• There are gendered differences in the timing and strength of impacts

– Women are mainly impacted during the period when changes are implemented and impacts (positive and negative) are stronger – Men are impacted after the period of change

• Possible explanations:

– Men have on average more voice than women in workplaces, they are better able to influence the content of changes and to adapt them to their needs. This effect goes partly through part time work which is negatively associated with voice

• References: Green, 2012; Howel et al., 2015
• Note : we checked that gendered differences are not related to maternity leave as

younger women are not more absent than older ones

– Health behaviour of men and women differ: facing similar health problems, women are more likely to contact earlier their physician

• Reference: Courtenay, 2000

31

Discussion of results (3)

• There are gendered differences in the forms of change that are negatively

associated with long term absence

– Women are positively impacted by managerial changes only. This impact is the

• nly one that seems to persist and to become stronger after the period when

the changes are implemented. – Men are positively impacted by ICT changes only and this impact is the only

• ne that starts to show up during the period when the changes are

implemented

• Possible explanation :

– existence of a digital gender divide, men are in a better position than women to reap the benefits of new technologies.

• Reference: Erickson et al., 2004

– In search of an explanation for the positive impact of management changes…

32

Discussion of results (4)

• Men with more seniority in the company suffer more from ICT

changes than from managerial changes.

– Men that were already in the company before the period of change suffer from more sickness absence than new recruits when ICT changes only are implemented than when managerial changes only are implemented.

• Possible explanation :

– higher skill obsolescence associated with ICT changes for men

• Reference: Erickson et al., 2004

33

Conclusion

• 1. Joint changes in ICT and managerial tools increase long term absence when

changes in one dimension only tend to reduce it

• More intense and complex changes would generate a disequilibrium within

the organisation which increases occupational risks

• 2. There are gendered differences in how changes impact long term absence
• Change impacts are stronger for women than for men and are more likely to
• ccur during the period of change, when for men they are more likely to
• ccur after the period of change
• Managerial changes only reduce women’s long term absence when ICT

changes only reduce men’s long term absence

• However, men with more seniority in the company suffer more from ICT

changes than from managerial changes

• Need to better understand the gendered construction of health behaviours as

well as that of technology and managerial tools in devising occupational safety and health policies in contexts of organisational change