## 1. Forces Shaping the Future of Work

## Source details

**Canonical URL:** [1. Forces Shaping the Future of Work](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019288-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019288-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019288-print-pdf.pdf.json)

---

### Key context and motivation
- Public debate dominated by concerns about automation and job losses; commonly cited estimates:
  - "47 percent of US jobs are at high risk of automation" (Frey and Osborne, 2013).
  - "9 percent of jobs in the OECD’s 21 member countries are automatable" (Arnt, Gregory and Zierahn, 2016).
  - "400 to 800 million jobs worldwide could be automated by 2030" (Manyika and others, 2017).
- Growing public pessimism: large majorities believe robots will do much of human work within 50 years; automation cited as having political consequences (e.g., role in the 2016 US election).
- Understanding factors shaping perceptions of automation can inform economic and policy responses.

### Data source and scope
- Survey: Boston Consulting Group’s Henderson Institute (BHI), Future of Work, May 2018.
- Sample:
  - 11,000 workers across 11 countries (US, UK, Germany, France, Spain, Sweden, Japan, India, Indonesia, China, Brazil).
  - Cleaned sample used for main analysis: 7,689 respondents (country counts: lowest Japan 504, highest India 837).
  - Target population deliberately excludes highly-educated workers (top tier bachelor’s degrees, master’s degrees, and higher); focuses on less educated, lower income, and middle-skilled workers.
  - Income composition: "60-to-85 percent of respondents with household income below their respective national averages."

### Framing of automation and main dependent variable
- "Automation" used broadly to refer to new workplace technologies—such as automation and artificial intelligence—as posed in the survey.
- Main dependent variable "automation" constructed from two questions:
  1. "How much of an impact new technology at workplace (e.g., automation, AI) will have on your own personal future?"
  2. "Do you believe the impact to be positive or negative?"
- Variable "automation" values: 1 (negative), 2 (neutral), 3 (positive).
- Dummies for some analyses: automation_pos = 1 if automation = 3; automation_neg = 1 if automation = 1.

### Literature review — public attitudes and education (selected survey findings)
- 2017 Gallup (April-May 2017, 1,100 U.S. workers): 26 percent thought it likely their job would be eliminated by new technologies within the next 20 years; about 13 percent said within the next five years.
- September-October 2017 Northeastern Univ./Gallup: 73 percent of Americans expected AI to destroy more jobs than it would create; employed Americans with less than a bachelor’s degree almost twice as likely to feel at risk.
- 2017 Eurobarometer (May 2017, ~28,000 EU citizens): 74 percent expressed concerns that robots and AI could cause more jobs to disappear than be created; respondents exposed to information about AI in last 12 months were more likely to view AI and robots positively.
- 2018 Pew (September 2018, 1,000 respondents in each of 10 countries): most respondents believed increasing automation would have negative consequences for jobs; pessimism correlated with views on the current state of the economy.
- 2017 PwC (10,000 respondents across China, India, Germany, UK, US): worry by education — 30 percent of respondents with basic education worried about their future vs. 13 percent of university graduates and 11 percent of post-graduates.
- 2018 Fuze (6,600 knowledge workers): 66% not worried about impact of automation on their jobs.

### Data and stylized facts from the BHI/BCG survey
- Respondents rated 15 forces expected to shape the future of work for personal impact (positive/negative).
- Top-ranked forces shaping the future workplace included:
  - Changes in customer needs and business models
  - Increase in the level of skills required
  - Freelancing & labor sharing as a source of income
  - Employees’ expectations for flexible work
  - Increase in the level of formal education required
- New technologies affecting human labor (automation, AI) ranked in the middle and on average slightly positive.
- Aggregate perceptions:
  - 41.3 percent of workers think new technologies (automation, AI) will have a positive impact in the workplace.
  - 25.5 percent of workers think the effect will be negative.
- Country heterogeneity:
  - India and Indonesia: more than 50 percent of respondents have a positive opinion about the effect of automation on the workplace.
  - Germany and Sweden: the percentage with positive views barely surpasses 30 percent.
- Personal characteristic associations (descriptive):
  - Education: More educated people (especially college graduates) have a more positive view.
  - Age: Older respondents tend to be less positive; substantial difference between those below 29 years and those above 50 years.
  - Income: Respondents from higher income quintiles are clearly more positive.
  - Gender: No strong overall effect; share negative slightly higher among women than men.

### Empirical findings on determinants of perceptions
- Personal and employment characteristics:
  - Negative perceptions more prevalent among workers who are older, poorer, and exposed to job volatility.
  - Positive perceptions associated with higher job satisfaction and higher educational attainment.
- Regional heterogeneity:
  - Respondents from emerging market economies more likely to have a favorable view of automation than respondents from advanced economies.
  - Evidence consistent with technology contributing to labor share declines in advanced economies (Dao et al., 2017).
- Labor market characteristics:
  - Respondents from countries with a higher degree of automation tend to perceive automation more negatively.
  - Respondents from countries with a higher degree of labor protection tend to view automation more positively.

### Reskilling, expectations, and policy preferences
- Reeducation/retraining expectations and responsibilities:
  - Nearly 55 percent of respondents expect new technologies to continuously raise skills requirements.
  - Most workers consider themselves first responsible to prepare for the future of work; exception: United States where more than 70 percent consider themselves first responsible.
  - Sense of personal responsibility higher in emerging market economies than in advanced economies.
  - Government as first responsible prevalent in France; Spain and Sweden near 30 percent. In Germany and Japan, companies seen as having a higher role than governments.
- If not limited by finances or time, most workers would build skills and take more time for themselves:
  - Independent workers: 46 percent would prefer to remain independent but focus on building new skills.
  - Non-contingent workers: 22 percent would reduce employer hours to rebuild skillset.
- Main obstacles preventing action: financial constraints, lack of time, lack of clarity about options.
- Econometric determinants (equation (3); Observations 7,689):
  - Reeducation (formal education):
    - Positive associations: higher education, higher income, job_happiness, automation_pos (automation_pos coefficients reported as 0.803*** and 0.882*** in models including automation_pos).
  - Retraining (on-the-job training):
    - Preexisting education not a significant predictor.
    - Middle-age workers and women more positive toward on-the-job training.
    - Higher income and job_happiness positively associated.
  - Interpretation: Workers with positive perceptions of automation more likely to support reeducation and retraining.

### Government policies: expectations and determinants
- Dependent variables (equation (4); Observations 7,689):
  - gov_protection: support for government protection of existing forms of work.
  - gov_benefits: support for government regulation of new forms of benefits.
- Key determinants:
  - Women and workers who have suffered job volatility more likely to expect government protection and favor new benefits.
    - female coefficients: gov_protection 0.083* and 0.102**; gov_benefits 0.167*** and 0.190*** in specifications with automation_pos included.
    - job_volatility (2.job_volatility) positive and significant for gov_protection and gov_benefits (e.g., 0.171***).
  - Older workers have a more negative view of government protection and new benefits:
    - 4.age_group coefficients: gov_protection -0.223*** and -0.138**; gov_benefits -0.263*** and -0.175**.
  - Workers with positive perceptions of automation (automation_pos) more likely to expect government protection and benefits (automation_pos coefficients ~0.753*** and 0.796***).
- Interpretation:
  - Women and those experiencing job volatility more supportive of government intervention.
  - Older workers may fear reallocating existing benefits toward new programs.

### Conclusions and policy implications
- Survey scope: ~11,000 workers in advanced and emerging economies; main analysis on 7,689 respondents.
- Overall sentiment:
  - Workers generally feel more positive than negative about automation; positivity stronger in emerging markets.
- Key patterns:
  - Negative views concentrated among older, poorer workers, those with recent job volatility, and respondents in countries with high robot penetration.
  - Positive views associated with job satisfaction, higher education (especially in emerging countries), and strong labor protection.
- Reskilling and responsibility:
  - Majority willing to prepare themselves; government seen as only partially responsible; companies also expected to play a role (stronger expectation in advanced economies).
  - Barriers to reskilling: lack of time and financial resources; if removed, a majority would retrain.
- Policy implications highlighted by the authors:
  - Additional fiscal space may be needed to finance major reskilling programs and/or new social benefits.
  - Targeted programs: demand for protection and new benefits more significant among women and workers who suffered job volatility—policymakers could better target these groups within new program design and existing budgets.

### Additional tabulated findings — Gender, Education, Age (selected statistics)
- Gender (by Level) sentiment distributions (Negative; Neutral; Positive):
  - Level 1 — Negative: 28.4; Neutral: 25.6; Positive: 25.1
  - Level 2 — Negative: 24.8; Neutral: 41.6; Positive: 38.3
  - Level 3 — Negative: 34.1; Neutral: 24.8; Positive: 30.0
  - Level 4 — Negative: 36.1; Neutral: 40.7; Positive: 50.4
- Education Level sentiment distributions (Negative; Neutral; Positive) by age bracket headings (labels appear as education in source):
  - Below 29 — Negative: 25.3; Neutral: 25.0; Positive: 26.6
  - 30-39 — Negative: 24.9; Neutral: 26.4; Positive: 28.5
  - 40-49 — Negative: 33.3; Neutral: 42.2; Positive: 48.3
  - Above 50 — Negative: 46.5; Neutral: 40.1; Positive: 32.8
- Age Group sentiment distributions (Q1–Q5) (Negative; Neutral; Positive):
  - Q1 — Negative: 26.0; Neutral: 24.9; Positive: 25.9
  - Q2 — Negative: 24.8; Neutral: 25.6; Positive: 38.4
  - Q3 — Negative: 35.4; Neutral: 32.7; Positive: 30.1
  - Q4 — Negative: 26.8; Neutral: 35.6; Positive: 39.7
  - Q5 — Negative: 41.5; Neutral: 45.1; Positive: 47.6

### Selected econometric results and magnitudes (baseline and labor market characteristics)
- Baseline ordered logit (Observations 7,689). Selected odds ratios (standard errors in parentheses):
  - 2.edu: 1.202** (0.095)
  - 3.edu: 1.311*** (0.091)
  - 4.edu: 1.654*** (0.125)
  - 4.age_group: 0.690*** (0.040)
  - 5.income: 1.311*** (0.090)
  - female: 0.929* (0.040)
  - 3.job_happiness: 1.511*** (0.083)
  - 3.job_volatility: 0.815** (0.076)
- Labor market characteristics (period averages, Ologit; Observations 7,689). Selected coefficients (standard errors in parentheses):
  - 4.edu: 0.420*** (0.077)
  - 3.age_group: -0.248*** (0.066)
  - 4.age_group: -0.363*** (0.062)
  - 5.income: 0.218*** (0.070)
  - 3.job_happiness: 0.388*** (0.055)
  - 3.job_volatility: -0.185** (0.094)
  - rob16: -0.066*** (0.014)
  - lri13: -0.438*** (0.104)

*Source: wpiea2019288-print-pdf*

### 1. Forces Shaping the Future of Work ...................................................................................

### 1. Forces Shaping the Future of Work

### Key context and motivation
- Public debate dominated by concerns about automation and job losses; commonly cited estimates in media include:
  - "47 percent of US jobs are at high risk of automation" (Frey and Osborne, 2013).
  - "9 percent of jobs in the OECD’s 21 member countries are automatable" (Arnt, Gregory and Zierahn, 2016).
  - "400 to 800 million jobs worldwide could be automated by 2030" (Manyika and others, 2017).
- Growing public pessimism: large majorities believe robots will do much of human work within 50 years; political consequences already observed (e.g., automation’s role in the 2016 US election cited).
- Understanding factors shaping perceptions of automation can inform economic and policy responses.

### Data source and scope
- Survey: Boston Consulting Group’s Henderson Institute (BHI), Future of Work, conducted in May 2018.
- Sample: 11,000 workers across 11 countries (US, UK, Germany, France, Spain, Sweden, Japan, India, Indonesia, China, Brazil).
- Target population: deliberately excludes highly-educated workers (top tier bachelor’s degrees, master’s degrees, and higher); focuses on less educated, lower income, and middle-skilled workers.
- Income composition: "60-to-85 percent of respondents with household income below their respective national averages."
- Cleaned sample used for main analysis: 7,689 respondents (country counts: lowest Japan 504, highest India 837).

### Framing of automation
- Term "automation" used broadly to refer to new workplace technologies—such as automation and artificial intelligence—as posed in the survey.
- Main dependent variable "automation" constructed from two survey questions:
  1. "How much of an impact new technology at workplace (e.g., automation, AI) will have on your own personal future?"
  2. "Do you believe the impact to be positive or negative?"
- Variable "automation" takes values 1, 2 and 3 corresponding to negative, neutral, and positive attitudes respectively.
- For some analyses, dummies are created: automation_pos = 1 if automation = 3; automation_neg = 1 if automation = 1.

---

### Literature review — public attitudes and education
- Public opinion surveys generally show majority concern that robots/AI will cause more jobs to disappear than be created.
- Selected survey findings summarized in the source:
  - 2017 Gallup: April-May 2017, survey of 1,100 U.S. workers found 26 percent thought it likely their job would be eliminated by new technologies within the next 20 years; about 13 percent said within the next five years. College graduates significantly less likely to fear elimination within five years; expectations over 20 years similar across education levels. September-October 2017 Northeastern Univ./Gallup: 73 percent of Americans expected AI to destroy more jobs than it would create; employed Americans with less than a bachelor’s degree were almost twice as likely to feel at risk.
  - 2017 Eurobarometer (May 2017, ~28,000 EU citizens): 74 percent expressed concerns that robots and AI could cause more jobs to disappear than be created; respondents exposed to information about AI in last 12 months were more likely to view AI and robots positively.
  - 2018 Pew (September 2018, 1,000 respondents in each of 10 countries): most respondents believed increasing automation would have negative consequences for jobs; pessimism correlated with views on the current state of the economy; respondents saw a clear role for government in preparing the workforce but also highlighted a role for individuals.
  - 2017 PwC (10,000 respondents across China, India, Germany, UK, US): lower education associated with more worry—30 percent of respondents with basic education worried about their future vs. 13 percent of university graduates and 11 percent of post-graduates.
  - 2018 Fuze (6,600 knowledge workers): 66% not worried about impact of automation on their jobs.

---

### Data and stylized facts from the BHI/BCG survey
- Respondents presented with 15 forces expected to shape the future of work and asked whether each would have a positive or negative impact on their personal situation.
- Top-ranked forces shaping the future workplace included:
  - Changes in customer needs and business models
  - Increase in the level of skills required
  - Freelancing & labor sharing as a source of income
  - Employees’ expectations for flexible work
  - Increase in the level of formal education required
- New technologies affecting human labor (automation, AI) ranked in the middle in terms of perceived impact, and on average were deemed to be slightly positive.
- Aggregate perceptions:
  - 41.3 percent of workers think new technologies (automation, AI) will have a positive impact in the workplace.
  - 25.5 percent of workers think the effect will be negative.
- Country heterogeneity:
  - India and Indonesia: more than 50 percent of respondents have a positive opinion about the effect of automation on the workplace.
  - Germany and Sweden: the percentage with positive views barely surpasses 30 percent.
- Personal characteristic associations (initial, descriptive):
  - Education: More educated people (especially college graduates) have a more positive view than those with lower levels of education.
  - Age: Older respondents tend to be less positive; substantial difference between those below 29 years and those above 50 years.
  - Income: Respondents from higher income quintiles are clearly more positive.
  - Gender: No strong effect overall; share negative slightly higher among women than men.

---

### Empirical findings on determinants of perceptions
- Personal and employment characteristics:
  - Negative perceptions are prevalent among workers who are older, poorer, and exposed to job volatility.
  - Positive perceptions are associated with higher job satisfaction and higher educational attainment.
- Regional heterogeneity:
  - Respondents from emerging market economies are more likely to have a favorable view of automation than respondents from advanced economies.
  - This is consistent with evidence that about half of the total decline in labor shares in advanced economies can be attributed to the impact of technology (Dao et al., 2017).
- Labor market characteristics:
  - Respondents from countries with a higher degree of automation tend to perceive automation more negatively.
  - Respondents from countries with a higher degree of labor protection tend to view automation more positively.

---

### Reskilling, expectations, and policy preferences
- Workers with positive perceptions of automation tend to acknowledge that reeducation and retraining will be needed.
- These workers also expect governments to play a role in shaping the future of work through government protection and new forms of social benefits.
- Policy-relevant heterogeneity:
  - Demand for protection and new benefits is more significant among women and workers that have suffered job volatility.
  - Policy implication suggested by the authors: policymakers could consider better-targeting these groups when designing new programmes to cushion the effects of technological change.

---

### Structure of the chapter (as presented)
- Section III: data and main stylized facts.
- Section IV: factors explaining perceptions of automation, including subsections on regional heterogeneity and labor market characteristics.
- Section V: workers' expectations about responses (reeducation, retraining, government protection, new social benefits).
- Section VI: summary and conclusions.

*Source: IMF staff summary of "1. Forces Shaping the Future of Work" (excerpts from the Boston Consulting Group (2018) survey and related analysis).*

### 1. Gender

### 1. Gender

### Gender (by Level)
- Findings:
  - Sentiment distributions are reported for Level 1 through Level 4 in three categories: Negative, Neutral, Positive.
- Key statistics:
  - Level 1 — Negative: 28.4; Neutral: 25.6; Positive: 25.1
  - Level 2 — Negative: 24.8; Neutral: 41.6; Positive: 38.3
  - Level 3 — Negative: 34.1; Neutral: 24.8; Positive: 30.0
  - Level 4 — Negative: 36.1; Neutral: 40.7; Positive: 50.4

### Education Level
- Findings:
  - Sentiment distributions are reported across education groups in three categories: Negative, Neutral, Positive.
- Key statistics:
  - Below 29 — Negative: 25.3; Neutral: 25.0; Positive: 26.6
  - 30-39 — Negative: 24.9; Neutral: 26.4; Positive: 28.5
  - 40-49 — Negative: 33.3; Neutral: 42.2; Positive: 48.3
  - Above 50 — Negative: 46.5; Neutral: 40.1; Positive: 32.8

### Age Group
- Findings:
  - Sentiment distributions are reported across quintiles Q1–Q5 in three categories: Negative, Neutral, Positive.
- Key statistics:
  - Q1 — Negative: 26.0; Neutral: 24.9; Positive: 25.9
  - Q2 — Negative: 24.8; Neutral: 25.6; Positive: 38.4
  - Q3 — Negative: 35.4; Neutral: 32.7; Positive: 30.1
  - Q4 — Negative: 26.8; Neutral: 35.6; Positive: 39.7
  - Q5 — Negative: 41.5; Neutral: 45.1; Positive: 47.6

*Source: wpiea2019288-print-pdf - 1. Gender*

### 4. Income

### 4. Income

### Model and key variables
- Dependent variable: automation — perception of how new technologies affect respondents’ own future; values: negative (1), neutral (2), positive (3).
- Baseline ordered logit specification (equation (1)) includes:
  - PPP_i,c: respondent characteristics — education (1–4; 1 = middle school or less; 2 = trade school/vocational training; 3 = high school; 4 = two/three-year college), age group (1–4; 1 = 18-29; 2 = 30-39; 3 = 40-49; 4 = 50-75), income quintile (1–5), female dummy (1 if female, 0 otherwise).
  - EP_i,j,c: employment characteristics — job_happiness (1 unhappy, 2 neutral, 3 happy), job_volatility (1 not at all, 2 once or twice, 3 three or more times in last five years), contingent worker dummy (1 for self-employed, temporary, company owners, or unemployed; 0 otherwise).
- Sample and estimation: Observations 7,689; Estimation Ologit; standard errors reported; significance denoted ***p<0.01, **p<0.05, *p<0.1.

### Baseline results (personal and employment characteristics)
- Education:
  - Higher education associated with more favorable perceptions of technology.
  - Example: increase from high school to college → odds of being more favorable are 1.7 times larger (i.e., a 70 percent increase in the odds), ceteris paribus.
- Income:
  - Higher income associated with more favorable perceptions.
  - Increase from fourth to fifth income quintile → 30 percent increase in the odds of being more positive about automation.
- Age:
  - Older respondents less likely to view automation favorably.
- Gender:
  - Odds of female respondents being more favorable are about 8 percent lower than male respondents.
- Employment:
  - Job_happiness is positively associated with favorable perceptions.
  - Job_volatility is negatively associated with favorable perceptions.
  - Contingent workers have a statistically significant positive view in some specifications.
- Robustness:
  - Results remain robust to inclusion of country, industry, and country-industry fixed effects, though the female dummy, job_volatility, and contingent worker dummy lose significance in some fixed-effects specifications.

### Regional heterogeneity
- Sub-sample regressions conditional on country-industry pair fixed effects separate respondents into Advanced, European, and Emerging economies. Observations 4,607 (Advanced), 3,317 (Europe), 3,082 (Emerging).
- Key observations:
  - Only older respondents in advanced and European economies (not emerging) are likely to have negative perceptions; coefficients for emerging markets are negative but not statistically significant.
  - Positive impact of income on favorable perceptions is restricted to advanced and European countries; coefficients for emerging economies are not statistically significant.
  - Female respondents from advanced and European economies have a statistically significant negative perception of automation’s effects.
- Broader interpretation:
  - Advanced economies: older and female respondents more negative.
  - Emerging markets: respondents more likely to have a favorable view overall.

### Labor market characteristics (country-level)
- Extended specification (equation (2)) adds country-level variables:
  - rob_c: number of robots per thousand workers (International Federation of Robotics) — proxy for degree of automation.
  - lri_c: labor regulation index (CBR Labour Regulation Index) between 0 and 1 — proxy for degree of protection (0 = lowest protection; 1 = highest protection).
- Results (conditional on industry fixed effects; Observations 7,689):
  - rob (latest value): negative and statistically significant.
    - A one-unit increase in robots per thousand workers → 8 percent decline in the odds of being more favorable about automation.
  - lri (latest value): positive and statistically significant.
    - A one unit increase in the labor regulation index → nearly 60 percent increase in the odds of being positive about the impact of automation.
  - Joint estimation of rob and lri: both results hold and are comparable in magnitudes; robustness checks using averages for 2000-2016 (robots) and 2009-2013 (lri) produce similar results, although robot coefficient may lose significance in the joint regression.

### Reskilling: attitudes and determinants
- Survey findings:
  - Nearly 55 percent of respondents expect new technologies to continuously raise skills requirements.
  - Most workers consider themselves first responsible to prepare for the future of work; exception: United States where more than 70 percent consider themselves first responsible.
  - Sense of personal responsibility is higher in emerging market economies than in advanced economies.
  - Government as first responsible is prevalent in France; Spain and Sweden show numbers close to 30 percent. In Germany and Japan, companies are seen as having a higher role than governments.
  - If not limited by finances or time, most workers would build up skills and take more time for themselves.
    - Independent workers: 46 percent would prefer to remain independent but focus on building new skills.
    - Non-contingent workers: 22 percent would reduce employer hours to rebuild skillset.
  - Main obstacles preventing action: financial constraints, lack of time, lack of clarity about options; affordability and time cited in Figure 6.
- Econometric determinants of reeducation and retraining (equation (3); Observations 7,689; Country-Industry Dummies Yes):
  - Reeducation (need for more formal education):
    - Positive associations: higher education, higher income, job_happiness, positive perception of automation (automation_pos).
    - automation_pos dummy: strong positive effect (automation_pos coefficient significant; example coefficients: 0.803*** and 0.882*** in models including automation_pos).
  - Retraining (more on-the-job training):
    - Preexisting education level not a significant predictor.
    - Age and gender matter: middle-age workers and women more positive toward on-the-job training.
    - Higher income and job_happiness also positively associated with retraining attitudes.
  - Interpretation: Workers with a positive perception of automation are more likely to support reeducation and retraining.

### Government policies: expectations and determinants
- Dependent variables (equation (4); Observations 7,689; Country-Industry Dummies Yes):
  - gov_protection: support for government protection of existing forms of work.
  - gov_benefits: support for government regulation of new forms of benefits.
- Key determinants:
  - Women and workers who have suffered job volatility are more likely to expect government protection and to favor new benefits.
    - female coefficients: gov_protection 0.083* and 0.102**; gov_benefits 0.167*** and 0.190*** in specifications with automation_pos included.
    - job_volatility (2.job_volatility) positive and significant for gov_protection and gov_benefits (e.g., 0.171***).
  - Older workers have a more negative view of government protection and new benefits (e.g., 4.age_group coefficients negative and significant: gov_protection -0.223*** and -0.138**; gov_benefits -0.263*** and -0.175**).
  - Workers with positive perceptions of automation (automation_pos) are more likely to expect government protection and benefits (automation_pos coefficients ~0.753*** and 0.796***).
- Interpretation:
  - Women and those experiencing job volatility more supportive of government intervention.
  - Older workers may fear reallocating existing benefits toward new programs.

### Conclusions and policy implications
- Survey scope: ~11,000 workers in advanced and emerging economies.
- Overall sentiment:
  - Workers generally feel more positive than negative about automation; positivity is stronger in emerging markets.
- Key patterns:
  - Negative views concentrated among older, poorer workers, those with recent job volatility, and respondents in countries with high robot penetration.
  - Positive views associated with job satisfaction, higher education (especially in emerging countries), and strong labor protection.
- Reskilling and responsibility:
  - Majority willing to prepare themselves; view government as only partially responsible; companies also expected to play a role (stronger expectation in advanced economies).
  - Barriers to reskilling: lack of time and financial resources; if removed, a majority would retrain.
- Policy implications highlighted by the authors:
  - Additional fiscal space may be needed to finance major reskilling programs and/or new social benefits.
  - Targeted programs: because demand for protection and new benefits is more significant among women and workers who suffered job volatility, policymakers could better target these groups within new program design and within existing budgets.

*Source: wpiea2019288-print-pdf - 4. Income*

### References

### References

### Bibliographic citations
- Adams, Z., Bishop, L. and Deakin, S. (2016) “CBR Labour Regulation Index (Dataset of 117 Countries)” Cambridge: Centre for Business Research, https://www.repository.cam.ac.uk/bitstream/handle/1810/256566/cbr-lri  -117-countries-codebook-and-methodology.pdf;sequence=1
- Acemoglu, A., and P. Restrepo (2019), “Automation and New Tasks: How Technology Displaces and Reinstates Labor”, Journal of Economic Perspectives, 33 (2):3-30.
- Arntz, M., T. Gregory, and U. Zierahn (2016), “The Risk of Automation for Jobs in OECD Countries: A Comparative Analysis”, OECD Social, Employment and Migration Working Papers No. 189. https://dx.doi.org/10.1787/5jlz9h56dvq7-en
- Autor, D. and A. Salomon (2018), “Is Automation Labor Share-Displacing? Productivity Growth, Employment, and the Labor Share”, Brookings Papers of Economic Activity, Spring: 1-63.
- Boston Consulting Group (2018), “The Future of Work”, BCG Henderson Institute Survey
- Eurobarometer (2017), “Attitudes Towards The Impact Of Digitisation And Automation On Daily Life”. European Commission. https://ec.europa.eu/digital-single-market/en/news/attitudes-towards-impact-digitisation-and-automation-daily-life
- Frey, B. and M.A. Osborne, M.A. (2013), “The Future of Employment: How Susceptible are Jobs to Computerisation?”, University of Oxford Working Paper. https://www.oxfordmartin.ox.ac.uk/downloads/academic/future-of-employment.pdf
- Frey, B., T. Berger and C. Chen (2018), “Political Machinery: Did Robots Swing the 2016 US Presidential Election?”, Oxford Review of Economic Policy, 34(3). https://academic.oup.com/oxrep/article/34/3/418/5047377
- Fuze (2018), “Workforce Futures: The Role of People in the Future of Work” https://www.fuze.com/files/documents/Fuze-WorkforceFutures.pdf
- International Federation of Robotics (2016) World Robotics. https://ifr.org/ifr-press-releases/news/world-robotics-report-2016
- Manski, C.F., and J. Straub (1999), “Worker Perceptions of Job Insecurity in the Mid-1990s: Evidence from the Survey of Economic Expectations”, NBER Working Paper, No. 6908. https://www.nber.org/papers/w6908
- Manyika, J., S. Lund, M. Chui, J. Bughin, J. Woetzel, P. Batra, R. Ko, and S. Sanghvi (2017), “Jobs Lost, Jobs Gained: What the Future of Work will Mean for Jobs, Skills, and Wages.” McKinsey Global Institute. https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-of  -work-will-mean-for-jobs-skills-and-wages
- Newport, F. (201), “One in Four U.S. Workers Say Technology Will Eliminate Jobs”, Gallup Surveys. https://news.gallup.com/poll/210728/one-four-workers-say-technology-eliminate-job.aspx
- Newport, F. (2018), “American Workers Remain Upbeat About Job Security”. Gallup Surveys. https://news.gallup.com/poll/242078/american-workers-remain-upbeat-job-security.aspx
- Price Waterhouse Coopers (2017), “Workforce of the Future: The Competing Forces Shaping 2030”. https://www.pwc.com/gx/en/services/people-organisation/workforce-of  -the-future/workforce-of-the-future-the-competing-forces-shaping-2030-pwc.pdf
- Reinhart, R.J. (2018), “Most U.S. Workers Unafraid of Losing Their Jobs to Robots”, Gallup Surveys. https://news.gallup.com/poll/226841/workers-unafraid-losing-jobs-robots.aspx
- Wike, R. and B. Stokes (2018) “In Advanced and Emerging Economies Alike, Worries About Job Automation”, Pew Research Center. https://www.pewglobal.org/wp-content/uploads/sites/2/2018/09/Pew-Research-Center_In-Advanced-and-Emerging-Economies-Alike-Worries-about-Job-Automation_2018-09-13.pdf
- Winick, E (2018), “Every study we could find on what automation will do to jobs, in one chart”, MIT Technology Review. https://www.technologyreview.com/s/610005/every-study-we-could-find-on-what-automation-will-do-to-jobs-in -one-chart/

### Annex: Table A.1 – Automation and the Future of Work: Baseline, Odds Ratios
- Dependent variable: automation — captures perception of how survey respondents view the impact of new technologies on their own future; values negative (1), neutral (2), or positive (3).
- Estimation: Ologit across specifications; Observations: 7,689 in each column.
- Selected odds ratios and standard errors (in parentheses):
  - 2.edu: 1.202** (0.095); 1.191** (0.095); 1.284*** (0.107); 1.181** (0.094); 1.273*** (0.106); 1.261*** (0.107)
  - 3.edu: 1.311*** (0.091); 1.312*** (0.092); 1.264*** (0.094); 1.302*** (0.091); 1.256*** (0.094); 1.250*** (0.095)
  - 4.edu: 1.654*** (0.125); 1.619*** (0.123); 1.443*** (0.121); 1.591*** (0.122); 1.421*** (0.120); 1.420*** (0.123)
  - 2.age_group: 0.0946 (0.059); 0.0926 (0.058); 0.949 (0.060); 0.928 (0.059); 0.950 (0.061); 0.952 (0.061)
  - 3.age_group: 0.783*** (0.050); 0.760*** (0.049); 0.822*** (0.055); 0.764*** (0.050); 0.824*** (0.055); 0.830*** (0.056)
  - 4.age_group: 0.690*** (0.040); 0.653*** (0.040); 0.783*** (0.051); 0.663*** (0.041); 0.788*** (0.051); 0.789*** (0.052)
  - 2.income: 1.110 (0.071); 1.090 (0.070); 1.095 (0.071); 1.086 (0.070); 1.090 (0.071); 1.105 (0.072)
  - 3.income: 1.128* (0.074); 1.100 (0.073); 1.106 (0.073); 1.092 (0.072); 1.100 (0.073); 1.109 (0.075)
  - 4.income: 1.263*** (0.084); 1.223*** (0.082); 1.238*** (0.083); 1.210*** (0.082); 1.227*** (0.083); 1.227*** (0.084)
  - 5.income: 1.311*** (0.090); 1.245*** (0.086); 1.269*** (0.088); 1.227*** (0.086); 1.254*** (0.088); 1.247*** (0.089)
  - female: 0.929* (0.040); 0.925* (0.040); 0.948 (0.041); 0.932 (0.041); 0.946 (0.042); 0.951 (0.043)
  - 2.job_happiness: 1.234*** (0.080); 1.202*** (0.079); 1.288*** (0.080); 1.199*** (0.079); 1.195*** (0.080)
  - 3.job_happiness: 1.511*** (0.083); 1.547*** (0.087); 1.508*** (0.083); 1.544*** (0.087); 1.550*** (0.088)
  - 2.job_volatility: 0.923* (0.044); 0.938 (0.045); 0.927 (0.044); 0.943 (0.046); 0.946 (0.046)
  - 3.job_volatility: 0.815** (0.076); 0.838* (0.079); 0.826** (0.078); 0.850* (0.080); 0.858 (0.083)
- Model controls across columns vary: Country Dummies (No/Yes), Industry Dummies (No/Yes), Country-Industry Dummies (No/Yes).
- Significance notation: ***p<0.01, **p<0.05, *p<0.1

### Annex: Table A.2 – Automation and the Future of Work: Labor Market Characteristics, Averages
- Dependent variable: automation — perception of impact of new technologies on respondents' own future; values negative (1), neutral (2), or positive (3).
- Columns 1 and 2 include labor market characteristics (period averages) that proxy degree of automation and degree of protection; Column 3 jointly estimates them.
- Estimation: Ologit across specifications; Observations: 7,689 in each column. Country Dummies: No. Industry Dummies: Yes. Country-Industry Dumm(y): No.
- Selected coefficients and standard errors (in parentheses):
  - 2.edu: 0.196** (0.080); 0.166** (0.080); 0.190** (0.080)
  - 3.edu: 0.266*** (0.070); 0.321*** (0.071); 0.311*** (0.072)
  - 4.edu: 0.420*** (0.077); 0.500*** (0.077); 0.457*** (0.078)
  - 2.age_group: -0.066 (0.063); -0.080 (0.063); -0.070 (0.063)
  - 3.age_group: -0.248*** (0.066); -0.268*** (0.065); -0.251*** (0.066)
  - 4.age_group: -0.363*** (0.062); -0.384*** (0.062); -0.350*** (0.063)
  - 2.income: 0.087 (0.065); 0.088 (0.065); 0.090 (0.065)
  - 3.income: 0.093 (0.066); 0.096 (0.066); 0.098 (0.067)
  - 4.income: 0.195*** (0.067); 0.198*** (0.068); 0.201*** (0.068)
  - 5.income: 0.218*** (0.070); 0.210*** (0.070); 0.220*** (0.070)
  - female: -0.061 (0.044); -0.071 (0.044); -0.063 (0.044)
  - 2.job_happiness: 0.187*** (0.065); 0.175*** (0.065); 0.166** (0.065)
  - 3.job_happiness: 0.388*** (0.055); 0.391*** (0.055); 0.376*** (0.055)
  - 2.job_volatility: -0.077 (0.048); -0.073 (0.048); -0.074 (0.048)
  - 3.job_volatility: -0.185** (0.094); -0.193** (0.094); -0.188** (0.094)
  - rob16: -0.066*** (0.014); -0.054*** (0.015) [appears across columns where included]
  - lri13: -0.438*** (0.104); -0.351*** (0.108) [appears across columns where included]
- Significance notation: ***p<0.01, **p<0.05, *p<0.1

*Content from wpiea2019288-print-pdf - References and Annex tables as provided.*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019288-print-pdf.pdf_